Arduino Robotic Arm

1

Scientific Exploration of Challenging Planetary Analog

Environments With A Team Of Legged Robots

PHILIP ARM1,+,*, GABRIEL WAIBEL1,+, JAN PREISIG1, TURCAN TUNA1, RUYI ZHOU1,2, VALENTIN BICKEL3,4, GABRIELA LIGEZA5, TAKAHIRO MIKI1, FLORIAN KEHL6,7,8, HENDRIK KOLVENBACH1, AND MARCO HUTTER1 1Robotic Systems Lab, ETH Zurich, Zurich, Switzerland 2State Key Laboratory of Robotics and System, Harbin Institute of Technology, Harbin, China 3Laboratory of Hydraulics, Hydrology and Glaciology, ETH Zurich, Zurich, Switzerland +These authors contributed equally to this work.

Compiled July 20, 2023

The interest in exploring planetary bodies for scientific investigation and in-situ resource utilization is ever-rising. Yet, many sites of interest are inaccessible to state-of-the-art planetary exploration robots because of the robots’ inability to traverse steep slopes, unstructured terrain, and loose soil. Additionally, current single-robot approaches only allow a limited exploration speed and a single set of skills. Here, we present a team of legged robots with complementary skills for exploration missions in challenging planetary analog environments. We equipped the robots with an efficient locomotion controller, a mapping pipeline for online and post-mission visualization, instance segmentation to highlight scientific targets, and scientific instruments for remote and in-situ investigation.

Furthermore, we integrated a robotic arm on one of the robots to enable high-precision measurements. Legged robots can swiftly navigate representative terrains, such as granular slopes beyond 25 degrees, loose soil, and unstructured terrain, highlighting their advantages compared to wheeled rover systems. We successfully verified the approach in analog deployments at the BeyondGravity ExoMars rover testbed, in a quarry in Switzerland, and at the Space Resources Challenge in Luxembourg. Our results show that a team of legged robots with advanced locomotion, perception, and measurement skills, as well as task-level autonomy, can conduct successful, effective missions in a short time. Our approach enables the scientific exploration of planetary target sites that are currently out of human and robotic reach.

Introduction

Robotic planetary exploration is invaluable for advancing our under- standing of the solar system and enabling the prospection of potential resources. The recent commitment of national and commercial entities to return to the Moon - targeting a sustainable, long-term human pres- ence - boosted the development of robotic exploration technologies.

Many science-, exploration-, and resource extraction-relevant tar- gets across the lunar surface lie in hard-to-reach areas or areas with substantial potential to host unknown physical surface properties. Ex- amples include pyroclastic vents, volcanic rilles, caves, irregular mare patches, and fresh impact craters [1, 2]. Consequently, developing robotic exploration systems that can efficiently traverse challenging terrain without compromising their explorative, scientific, and resource prospection capabilities remains a top priority.

Notably, several lunar exploration efforts revolve around the Na- tional Aeronautics and Space Administration’s (NASA) Artemis pro- gram, which focuses on robotic and crewed science and exploration at the lunar south pole . One of the first Artemis program missions, the Volatiles Investigating Polar Exploration Rover (VIPER) , will venture into several permanently shadowed regions - cold, volatile-rich topographic depressions that have not been illuminated for millions of years . Ultimately, Artemis 3 will lead humans to the south pole in 2025 . All of those missions will need to navigate the challenging south-polar terrain, including steep slopes, impact ejecta and boulder fields, and potentially anomalous physical regolith properties [4, 9, 10].

One promising way to foster the development of lunar exploration and prospection technologies is challenge-driven innovation. The European Space Agency (ESA) and the European Space Resources

2

Innovation Centre (ESRIC) established the Space Resources Challenge (SRC) in 2021 to evaluate and advance the state of the art of robotic lu- nar prospecting technologies. The main technical goal of the challenge was prospecting a lunar analog environment for resource-enriched ar- eas (REAs), meaning areas that contain minerals suitable for In Situ Resource Utilization (ISRU), such as ilmenite, rutile, and titanium dioxide. The two rounds of the challenge took place in 2021 and 2022, with the final competition in a lunar analog terrain in Luxembourg. The competition involved adverse conditions found at the lunar south pole, including a previously unknown terrain, loose granular soil, high solar incidence angle illumination creating long, high-contrast shadows, and network communications with high latency 5.0 s round trip time (RTT)) and intermittent complete loss of signal. The SRC was an important inspiration in this work and was one of the two major field deployments of our robotic exploration team.

In this article, we propose a team of legged robots for quick, ef- ficient, and safe exploration and prospection of challenging plane- tary analog environments. The team approach allows us to cover a larger area, deploy a wider variety of scientific payloads, investigate more scientific targets, and gain more in-depth knowledge per target than possible with a single-robot approach or non-teamed multi-robot approaches. Additionally, the increased redundancy allows mission completion even if multiple robots fail. To control the robotic team, the operators send high-level navigation, remote measurement, and in-situ measurement tasks to the robots. The robots execute these tasks autonomously using their state-of-the-art mobility and navigation sys- tems, as well as a complementary and redundant set of payloads. The level of autonomy allows continued scientific data collection, even if communication becomes unreliable or during a complete loss of signal (LoS). Simultaneously, the scientists in the operations team can select and prioritize scientific targets during the mission.

Until now, most planetary exploration robots relied on wheeled locomotion. Their locomotion system did not fundamentally change since the first rover - Lunokhod 1 - touched down on the surface of the Moon in 1970 . Other prominent examples are the Lunar Roving Vehicle (LRV), Lunokhod 2, and Yutu 2 , or martian rovers such as the Sojourner Rover , Spirit, Opportunity , Curiosity , and Perseverance . Although these systems can build upon well-tested heritage technology and provide robustness in relatively flat terrain, wheeled rovers reach their limitations on steep slopes, on loose granular terrain, and in unstructured environments. On Mars, the Spirit rover was lost in anomalously loose soil and Opportunity got temporarily stuck in a dune . On the Moon, the Apollo 15 LRV was trapped in loose regolith and had to be manually retrieved by the astronauts . Similarly, Lunokhod 2 encountered excessive wheel sinkage (>20 cm) near Le Monnier crater . The Yutu-2 team reported that entering craters would be of great scientific interest.

However, they do not target craters because of the increased probability of locomotion failure . This locomotion limitation prevents current missions from investigating high-priority targets [1, 2, 21–23].

Meanwhile, terrestrial legged robots have reached a high level of robustness in exploring unknown environments. Their robust locomo- tion system allows them to traverse unstructured, challenging natural terrain, including mud, gravel, snow, vegetation, and sand [24, 25].

Several researchers have developed legged robots with the intent to use them in space in the past . Until now, we have focused on the usage of dynamically walking legged robots on steep, planetary soil analogs and low-gravity environments , showcasing the potential of the technology. However, for these robots to be useful in real-world scenarios, they need to be advanced beyond locomotion tasks. They have to interact with their environment in realistic analog missions, for example by deploying scientific instruments or taking samples. We advanced in this direction at the first field trial of the SRC, where we deployed a legged robot with base-mounted instruments .

Heterogeneous robotic teams have been used as a viable solution in terrestrial real-world missions. Notably, all top-ranking teams in the DARPA Subterranean Challenge 2021 used heterogeneous robotic teams with diverse skills . To succeed in the challenge, the teams developed robust solutions for locomotion, localization, multi- robot mapping, local planning, and exploration planning. In this work, we build upon these advances - specifically on the systems of our team CERBERUS - and address the unique challenges presented by analog space missions, including instrument deployment, efficient and robust locomotion with robotic arms, redundancy to component or system failures, and validation in realistic missions with high latency communication. Additionally, although the robotic teams in the Subter- ranean Challenge were diverse in their locomotion skills, we consider diversity in scientific investigation capabilities in this work.

Several robotic teams for planetary exploration have been devel- oped and tested in analog environments. For example, the German Aerospace Center (DLR) deployed a drone and two wheeled robots autonomously to set up a distributed radio telescope and perform ge- ological exploration on Mt. Etna . Although they showed a high level of autonomy, the wheeled rovers were limited in their locomotion capabilities. The German Research Centre for Artificial Intelligence (DFKI) developed a heterogeneous robotic team of a wheeled and a legged robot for sample collection in a lunar analog environment .

In , the authors demonstrate a teleoperated sample return mission using a robotic team in a martian analog environment. The operations team operated the robots via waypoints and tested capabilities such as sampling in isolated tests instead of the full mission deployment. More recently, NASA JPL built upon their NEBULA solution to explore ana- log martian caves with multiple Spot robots in the NASA BRAILLE project . One of the robots was equipped with a robotic arm to take close-up images and swab samples. However, the details of this work are not yet published. Lastly, the first heterogeneous robotic team for planetary exploration is currently operating on Mars: The Ingenuity helicopter supports the Mars 2020 mission by scouting potential tar- gets for the Perseverance rover and inspecting targets the rover cannot access . This actual Mars mission is a remarkable example of a heterogeneous robotic team for planetary exploration.

We present our teamed exploration approach with dynamically walking robots for planetary environments (Movie 1). We designed a team of legged robots with a diverse set of scientific investigation skills and redundancy measures and validated our system in three challenging analog environments: The ExoMars locomotion test facility, a quarry site, and the competition site of the SRC. We report our results and lessons learned from these deployments and identify opportunities for future developments.

To achieve effective analog mission deployments, we developed, improved and validated critical subsystems: We validated the baseline locomotion policy on planetary analog terrain and developed a locomotion policy with a focus on efficiency. Our policy includes arm observations, making it robust for legged robots with robotic arms for scientific investigation in challenging environments. Additionally, we built a two-pronged mapping approach for lightweight real-time online mapping and high-resolution, realistic post-mission visualization. Our instance segmentation pipeline highlights potential scientific targets to support online mission planning in previously unknown environ- ments. By distributing a balanced set of remote and close-up scientific instruments, we achieve effective and safe exploration missions.

We showcase the capabilities of our legged robots in martian and lunar analog environments, demonstrating that our technology can enable robots to investigate scientifically transformative targets on the Moon and Mars that are unreachable at present using wheeled rover systems.

3

Fig. 1. System architecture of our team of legged robots. (A) Robotic and scientific payloads on the Scout, Scientist, and Hybrid. Robotic payloads and scientific payloads are labeled in orange and white, respectively. (B) High-level overview of the software architecture of our system. With a balanced combination of shared and specialized modules per robot, we designed a safe yet efficient multirobot system.

4

Fig. 2. Example of task allocations in our system. Task allocations are shown as solid lines. Alternative allocation paths are shown as dashed lines. When a task allocation becomes invalid, for example, because a payload or robot malfunctions (red lines), tasks can be reallocated accord- ing to our redundancy concept.

A Team Of Legged Robots For Planetary Exploration

The core of our system comprises a team of three four-legged robots. ANYmal serves as the base platform for all robots. Each of the three robots has a dedicated role and a unique set of payloads for exploration and scientific data collection. Fig. 1-A presents our team of legged robots: the Scout, the Hybrid, and the Scientist. The Scout’s primary task is rapidly exploring the environment using its additional LiDARs and RGB cameras. It provides the operations team, consisting of the robot operators and planetary scientists, with an overview of the previously unknown area and allows the team to identify potential scientific targets. Its secondary task is to capture images of potential scientific targets in various spectral bands using a pan-tilt context imager augmented with a custom-built filter wheel (CTX-FW) (Fig. S1).

The Hybrid’s main task is collecting scientific data of numerous targets using a pan-tilt context imager with an additional thermal camera (CTX-TH). Moreover, it is equipped with a base-mounted Metrohm Instant Raman Analyzer XTR DS (MIRA) with a zoom lens to acquire Raman spectra of targets of interest. The Scientist performs an in-depth scientific analysis of previously identified targets. It features a custom 6-degrees-of-freedom (DoF) robotic arm with a MIRA on the forearm and a custom microscopic imager (MICRO) on the wrist.

Fig. 1-B shows a system overview. Two operators on two mission control stations send high-level navigation, remote measurement, and in-situ measurement goals to the robots. "Remote measurement tasks" describe scientific tasks that the robot can conduct remotely using the CTX-FW and the CTX-TH payloads. "In-situ measurement tasks" describe close-up investigations, namely MIRA and MICRO measure- ments. All data packets between mission control and the robots are delayed with a round-trip-time (RTT) of 5.0 s to simulate lunar opera- tion. Navigation goals are handled by the same module on each robot.

3D remote measurement goals are only used by the CTX imagers on the Scout and the Hybrid. The 6D in-situ measurement targets are processed by the Hybrid and the Scientist. The robots send feedback about their state, navigation images, a sparse map representation, and scientific data of targets of interest to the mission control stations.

Although all robots have their designated role, they share many exploration capabilities and payloads, enabling a high redundancy level. If a robot fails, the operations team can reallocate tasks between the robots. Fig. 2 visualizes how tasks can be reallocated for examples of exploration and measurement tasks thanks to the payload redundancy concept. Table S1 summarizes which scientific measurement tasks can be executed under which failure conditions.

Experimental Setup On Analog Sites

We conducted end-to-end missions in two lunar analog environments and a locomotion validation test in a martian analog testbed. In both end-to-end analog missions, the operations team consisted of five peo- ple: A team of one robot operator and one planetary scientist each operated one of the two mission control stations. The robot operators sent tasks to the robots, and the planetary scientists selected and priori- tized targets based on the data received from the robots. Additionally, one supervisor overlooked the mission and ensured communication between the two control teams. Both mission control stations could be used to interact with any of the robots. However, the team ensured that each robot only received commands from one control team at any given time to prevent conflicting commands, where the newer com- mand would supersede the previous one. In the presented deployments, we used one mission control station to control the Scout and the Hybrid

5

and the second control station to control the Scientist.

Space Resources Challenge, Luxembourg

One field campaign occurred at the ESA/ESRIC Space Resources Challenge in Esch-sur-Alzette, Luxemburg (September 2022). The competition area measured 1800 m2, and the terrain was unknown before the challenge. The ground was covered in coarse, granular basalt with a substantial fine fraction (clay to silt). The light conditions closely resembled those at the lunar south pole because of a powerful illumination source at a high incidence angle in the corner of the area (Fig. 3-A). The analog scenario contained several locations of interest, such as resource-enriched areas, boulders, craters, and a lunar habitat prototype. The mission control room was separated from the competition area with a 5.0s RTT to communicate to all systems in the competition area. We summarize the most important objectives and rules of the SRC in Table S2.

Quarry Site, Switzerland

The quarry site is an active gravel quarry operated by KIBAG, situ- ated in Neuheim, Switzerland. The quarry consists of poorly sorted, fine and coarse sediments, including meter-sized boulders, leading to locomotion challenges such as sinkage and slippage, especially on steep inclines. The site includes a headwall with a maximum slope of about 20◦. We simulated realistic lighting conditions in a lunar south pole scenario. To this end, we conducted the test at night to minimize the influence of naturally occurring light and illuminated the test site with a 180 W LED lamp (Aputure LS 120D II) at a high illumination angle of roughly 87◦. As shown in Fig. 4-A, the illumination leads to characteristic long, high-contrast shadows as expected in the vicinity of the lunar south pole. Furthermore, all communication between mission control and the robots passed through a delay simulator, which created a round-trip-time of 5.0 s.

We selected distinct boulders on the site to simulate scientific targets of interest. Furthermore, we spread patches of basalt, ilmenite, rutile, and titanium dioxide in different mass fractions on the terrain to create realistic resource-enriched areas for a lunar prospecting mission.

Locomotion Validation Testbed, Beyond Gravity

We used the planetary soil testbed at Beyond Gravity (Fig. S2) to validate the Scout’s locomotion controller on steep, granular soil analogs. The testbed was initially designed to test the locomotion subsystem of the ExoMars rover Rosalind Franklin. It features a 6 m x 6 m tiltable container that can be filled with various analog soil simulants. We used ESA’s ES-4 martian soil simulant and a row of Jurassic limestone plates as martian bedrock analogs for our tests.

The testbed is tiltable up to 25◦at a 0.1◦resolution.

Analog Mission Results

This section provides an overview of our analog deployments: the end-to-end deployments at the SRC and in the quarry, as well as the locomotion validation tests.

Space Resources Challenge Mission Overview

During the SRC, the challenge’s core objectives provided by the or- ganizers were mapping the competition area, locating boulders and resource-enriched areas (REAs), and characterizing them. Based on that, we derived goals for the robotic system: These comprised of mapping the entire competition area, locating REA candidates and all boulders, and providing scientific data of the boulders and potential REAs to enable trained geologists to characterize them.

Fig.3-F shows the mission overview of the SRC deployment. We first deployed the Scout to map the area and use the navigation cameras and CTX-FW to help the operations team prioritize the targets of interest. After the first LoS, we decided that we had enough targets of interest to deploy the Scientist. The Scientist’s task was mainly to focus on potential REAs. After 76 min, we deployed the Hybrid to support the boulder characterization with the thermal imager and the

Mira.

We mapped 95% of the competition area, located seven out of eight boulders, and identified 18 potential REAs. We prioritized five boulders for closer investigation, taking CTX-FW images of all of them and MIRA measurements of three boulders. Additionally, we investigated six of the 18 potential REAs with MICRO and MIRA, taking several measurements per target when the data quality was insufficient.

The SRC deployment illustrates several advantages of the teamed exploration approach. Three robots operated simultaneously for 20 min, thanks to the two mission control stations that can be used inter- changeably to interact with all robots. During this time, the Scout was completing the map in yet unexplored areas, and the Hybrid and the Scientist were collecting scientific data to characterize the targets of interest. Our autonomy features and intuitive GUI further supported deploying three robots with two operator stations. Additionally, be- cause the environment was already known after the Scout deployment, the Scientist could deploy MICRO and MIRA every 3-5 min, which would be unfeasible with an approach where the same robot has to map and identify targets.

Quarry Mission Overview

To test all payloads on a number of targets with variability, we de- fined a minimum of five targets to investigate with each instrument. Furthermore, we intended to map a large area to test our localization and mapping systems on a relevant scale. We accordingly defined the following mission goals for the deployment at the quarry: Exploring and mapping an area of at least 1000 m2, and identifying at least five targets of interest, such as boulders or terrain patches. To test our instruments, we set the goal to acquire measurements of at least five targets with each instrument (CTX, filter wheel, thermal imager, MIRA, and MICRO).

Fig. 4-E shows the mission overview of the quarry deployment. We did not use the CTX-FW on the Scout and limited the use of the MIRA on the Scientist because the hardware malfunctioned. Thus, according to our redundancy concept, we deployed the Hybrid early to take over the CTX tasks of the Scout and the MIRA tasks of the Scientist. The Scientist deployed once the operations team had identified and priori- tized enough targets of interest to maximize the payload utilization on the Scientist.

In a total mission time of 68 min, we mapped an area of 1375 m2 and identified twelve targets of interest (ten boulders and two area patches). We collected 16 CTX and thermal images of seven rocks, MIRA spectra of six rocks, and MICRO images of three rocks and two area patches (two MICRO datasets of the same area patch). Using our two control stations, we could efficiently control the Hybrid and the Scientist in parallel during one-third of the mission. Thanks to the redundancy concept in our robotic team approach, we fulfilled all mission objectives except the acquisition of filter wheel images, despite the inactive payloads. A single-robot system or a system without a redundancy concept could not have achieved the objectives related to the CTX and the MIRA measurements under these circumstances.

Locomotion Results

In the validation test campaign at Beyond Gravity, the Scout could climb and descend slopes of up to 25◦- the maximum of the testbed - on ES-4 and bedrock using our existing locomotion control policy . On both grounds, we conducted three tests, both in ascent and descent, without a locomotion failure, which we define as either a fall or a failure to continue the traverse, for example due to slippage. Even on the maximum slope of the testbed, the robot reached a top speed

6

Fig. 3. Performance of our team of legged robots during the SRC. (A) Overview of the competition area. (B) Postprocessed high-resolution map of the exploration area and the associated path of the Scout. The region in the blue dotted box corresponds to (C). (C) Online mesh map of the Scout for target identification. (D) Postprocessed high-resolution height map and scientific acquisition paths of the Hybrid and the Scientist.

(a and b) Thermal images acquired during the SRC. (c) Example of the rock instance segmentation of a Navcam image. (d) Example of the rock instance segmentation of a CTX image. (E) Example of a panorama image with rock instance segmentation acquired during the SRC. (F) Mission summary of the SRC.

7

Fig. 4. Performance of our team of legged robots during the analog mission on the Neuheim quarry site. (A) Experimental test yard with a variety of scientific targets. (B) Postprocessed and shaded high-resolution map of the Neuheim quarry site with the path of the Scout. (C) Rock segmentation results under low-illumination conditions. (D) Postprocessed high-resolution elevation map of the Neuheim quarry test site. (E) Mission summary of the end-to-end analog deployments at the quarry. The Scout, the Hybrid, and the Scientist were all deployed sequentially with an overlap of the Hybrid and the Scientist. The whole mission duration was 68 min.

8

Fig. 5. Locomotion capability of our customized ANYmal over different challenging planetary analog terrains. (A) ANYmal climbs a steep sandy slope at around 20◦on the Neuheim quarry site. (B) ANYmal escapes from the analog lunar impact crater rim during the SRC. C) ANYmal escapes from deep foot sinkage in the locomotion validation test bed. D) ANYmal walks over a high bedrock step.

of 0.7 m/s, outperforming state-of-the-art systems that can climb such inclinations (Table S3). We conducted additional tests at a 10◦slope on ES-4 with steps of bedrock and very loose hills in all directions. The robot could negotiate the hills despite the high sinkage (Fig. 5-C) and seamlessly transition high steps between ES-4 and bedrock (Fig. 5-D).

Our developed controller for the Hybrid and Scientist enables simi- lar robustness as the existing baseline controller , but with added heavy payloads such as the robotic arm. The moving robotic arm did not impede the Scientist while walking on flat terrain (Fig. S3, Movie S1). Additionally, with a static arm, the robot consumed 15% less power in a mock mission on flat ground when using our controller compared to the baseline controller (see Supplementary Results).

During the mission deployments in the quarry and the SRC, the robots traversed steep granular slopes up to 20◦(Fig. 5-A), and a crater rim (Fig. 5-B). At the SRC, the robots covered a total distance of 358m of granular terrain (Fig. 3-B and Fig. 3-D). The Scientist, carrying the robotic arm, showed the same level of robustness as the Scout and the Hybrid. During all pre-tests and mission deployments, no single locomotion failure occurred on any of the robots.

Mapping And Target Identification

We used a lightweight mesh representation for online operations and a dense mapping framework for mission post-processing in both analog missions (Fig. 6). Furthermore, we ran an instance segmentation pipeline on our navigation cameras and CTX images to identify and highlight boulders as potential targets of interest for the operations team, substantially reducing the operational overhead. Fig. 3-C-E and Fig. 4-C-E show the mapping and perception results in the quarry and at the SRC, respectively.

The robot sent a downsampled point cloud in a nine-meter radius around the robot to mission control. The mission control PC generated the mesh and fused single mesh instances automatically to gradually build a mesh map of the covered area. Offloading the meshing operation to the mission control PC allowed us to transmit a small point cloud instead of a heavy map. Fig. 3-C shows a fraction of the mesh map of the SRC built with three mesh instances. Boulders 4 and 5 and impact craters 4 and 5 are distinguishable in the mesh. Together with the navigation cameras, the resolution of the mesh map allowed the operations team to identify and mark these targets of interest for further investigation. Therefore, the resolution was high enough to understand the robot’s environment, select targets on the mesh map, and make mission scheduling decisions while still allowing the map to be transferred over the network.

Fig. 3-D-c,d and Fig. 4-C show the instance segmentation output on the navigation cameras in the SRC and quarry missions, respec- tively. The images show that, even under difficult lighting conditions, the pipeline can identify and highlight boulders in the image, support- ing the operations team in the target identification. Additionally, we segmented selected panoramic images (Fig. 3-E).

Aside from the online mesh to guide operations, all robots main- tained a high-resolution point cloud map in both missions for post- mission visualization. Fig. 3-B shows the post-processed and shaded dense map of the complete SRC competition environment, with the path of the Scout overlaid. We explored 95% of the area with the Scout within 96 min, where most of the environment and the scientific targets were already identified after 50 min. Similarly, Fig. 4-B shows the dense map created during the quarry mission. Additionally, we colored the dense maps by elevation (Fig. 3-D and Fig. 4-D) to create a topographic overview of the environment. For example, the headwall in the quarry is visible in the top left part of the map. Furthermore, we created colored maps (Fig. S4 and Fig. S5) by projecting the navigation camera colors onto the point cloud.

9

Fig. 6. Mission support mapping and localization modules. The blue area represents the postmission operations, whereas the other modules are active during the mission.

Scientific Data Of Targets Of Interest

Once the operations team identified and prioritized targets of interest, they could send remote measurement and in-situ measurement tasks to all robots (Fig. 7). Fig. 8 shows an example of the scientific data we could gather during the SRC. We acquired images in five visual spectral ranges at two zoom levels using CTX-FW (Fig. 8-A). The low-zoom image shows the target’s size, shape, and geomorphic context. The high-zoom image provides information on the target’s surface texture, including the presence of millimeter to centimeter-scale vesicles, the rock’s lithology, and the distribution of minerals. Both low-zoom and high-zoom images indicate that all boulders in the SRC were porous basalts with aphanitic (fine-grained) textures. The high-zoom images with the filter wheel enabled us to study the relative reflectance of each target and compute five-point spectra. We color-corrected selected RBG images with images from a robot-mounted color calibration card.

Fig. 8-B shows in-situ measurement results of a potentially resource- enriched area during the SRC. We collected microscopic images in six different spectral bands (white, red, green, blue, UV, IR), allowing us to compute the target’s five-point spectrum. The images show the coarse granular basalt regolith, which covered the whole competition area at the SRC. MICRO images enabled us to investigate the grain size distribution and presence of potential resources. The MIRA Raman spectrum of the basaltic regolith in Fig. 8-B shows, for example, a prominent peak at 952 cm−1.

Discussion

Using a team of legged robots with powerful locomotion capabilities, a mapping pipeline for online and offline visualization, and segmentation tools for target identification allowed us to collect a substantial amount of scientific data in planetary analog missions with limited mission time.

Compared to the single-robot approach we deployed at the first Space Resources Challenge field trial , we could drastically increase the quality and quantity of scientific data products. The specialization of our robots allowed a high payload utilization, as proven at the SRC, where the Scientist robot deployed the MIRA and MICRO instruments every 3-5 min. In a multi-robot approach without specialization, for example without a dedicated scout robot, every robot would invest a substantial part of the mission time in mapping and target identification, limiting the high-return instruments’ payload utilization.

The quarry deployment showed the importance of the redundancy concept. We could still fulfill six out of seven mission goals despite two malfunctioning payloads. Although we did not experience a robot failure in the deployments, our redundancy concept would allow us to still accomplish most mission goals.

Our payload selection features a balanced mix of remote and in- situ science, but it was subject to budget and time constraints. Other instruments, for example X-ray fluorescence spectrometers (XRF), as used by other teams at the SRC, performed better at identifying REAs. Further expanding our scientific instrument suite, for example, with an XRF or laser-induced breakdown spectrometer (LIBS), would increase the quality and diversity of the mission’s science outputs.

Additionally, adapting our system to allow sample return to a lander with more capable scientific instruments would be a valuable expansion to increase the mission’s scientific output.

One learning from our field deployments is that a round-trip-time of 5s in the communication between mission control and the robot notably affects operations. Standard, reliable protocols such as TCP are not suitable in this setting. When using UDP, we had to reduce the data size of products, such as point clouds and images, to decrease the probability of a single packet loss occurring within the transfer of a data product. Furthermore, every operator interaction costs valuable mission time. We tackled this issue by providing lightweight, expressive data to the operators, such as the mesh map, and employing task-level autonomy. Hence, every task execution only requires one operator interaction. However, a more in-depth analysis of suitable protocols and compression technologies could further alleviate this problem.

A key learning in our deployments was the importance of the human aspect: The robots could execute single tasks autonomously, but the human operators still had to prioritize and allocate tasks. This decision- making can be time-consuming and difficult, especially since the oper- ators must balance the time required for the decision-making process with the mission time. An algorithmic approach to decision-making and task allocation could improve the mission return. The algorithm could either aid the operations team or, ideally, be implemented in an autonomy module on the robots to allow longer-term or even full mission autonomy. However, besides semantic scene understanding, this requires automatic interpretation of scientific data products such as images and spectra, which is an open challenge. Furthermore, more robot-to-robot communication is needed to build collaborative maps, allocate tasks without mission control in the loop and allow for more involved interaction and collaboration between multiple robots.

A higher level of autonomy will additionally improve the system’s scalability to applications with even more challenging communication, such as Mars exploration. Moreover, it will allow scaling the approach to a higher number of robots without increasing the workload on the operations team. Required modules to increase the level of autonomy are a safe, multi-goal planner to visit identified targets, automatic target identification and prioritization as stated above, and automatic task allocation to the robots based on the robot-specific skillset.

In this work, we used legged robots with different scientific in- vestigation skills but almost identical locomotion skills. In future deployments, a more heterogeneous approach could be advantageous.

Drones would enable fast terrain mapping, and wheeled robots, with limited mobility but higher battery life, could investigate easy-to-reach targets while the legged robots focus on hard-to-access areas.

Lastly, a relevant research stream is the scalability of such proto- types from analog deployments to actual space missions . Chal- lenges specific to legged robots mainly concern the power concept and thermal management. Preliminary results show that these as- pects are feasible [45, 46]. However, there are currently no powerful space-graded processors to run the mapping, navigation, and loco- motion algorithms in real-time. Additionally, our localization and mapping pipelines heavily rely on LiDARs, which are currently not space-qualified. However, solid-state LiDARs could pave the way for LiDAR-based localization and mapping in space applications.

Although there remains substantial room for future research and several open challenges in scaling the presented system to actual space missions, this work is an important step towards teams of legged robots for planetary exploration that provide value in both scientific and commercial missions.

Hardware Description

All described robots in this report are based on ANYmal by ANYbotics. The Scout is an ANYmal C (2019), whereas the Hybrid and the Scien- tist are based on ANYmal D (2021). Additionally, the Scientist carries a DynaArm, a custom 6DoF robotic arm. ANYmal weighs 50 kg, has a payload capacity of 15 kg, and a nominal operation time of 90 min while continuously walking. We show an overview of the on-board computers in Table S4.

Robotic Payloads

In the factory configuration, the ANYmals are equipped with a VLP16 Puck LITE LiDAR by Velodyne for mapping and localization, four

11

Fig. 7. In situ measurement task workflow. (A) The operator selects the desired target in the depth camera interface to define the instrument pose. The pose can be adjusted using a 6-DoF interactive marker. The robot receives the 6-DoF goal and uses it to deploy the scientific instru- ments at the desired location. (B) Boulder measurement using the MIRA Raman spectrometer and associated data products. (C) Boulder mea- surement using MICRO and associated data products. (D) Ground patch measurement using MICRO and associated data products.

12

Fig. 8. Examples of scientific data products at the SRC. (A) Multispectral visual data acquired by the CTX-FW of boulder 5. The images show a vesicular basaltic boulder. (B) MICRO and Raman data products of a REA candidate. The images and spectra show the basaltic soil of the competition area. The relative reflectance has an arbitrary scale.

13

(ANYmal C) or six (ANYmal D) RealSense D435 active stereo sensors by Intel (front, left, right, rear) for elevation mapping (Sec. Elevation Mapping), and two FLIR Blackfly wide-angle cameras (front and rear) to provide RGB image streams. We customized all robots with two high-power, air-cooled LEDs at the front and rear to illuminate dark environments. Since the Scout is the desired mapping robot, we exchanged the four RealSense D435 with two Robosense RS-Bpearl LiDARs (front and rear) for more accurate elevation mapping (Sec.

Elevation Mapping) and dense point cloud mapping (Sec. Dense Point Cloud Mapping). Additionally, we added Sevensense’s Alphasense Core visual-inertial sensor as a high-performance navigation camera array on the Scout. It features three monochrome and four color cameras (front, left, right, and top). Each camera has 0.4 MP and a field of view of 126◦x 92.4◦(HxV). In addition, four high-power, air- cooled LEDs are encapsulated in the same housing. The Alphasense Core allows the operator to quickly gain a first overview in an unknown environment and enables RGB-colored mapping (Sec.

Science Payloads

The Scout carries an ANYbotics inspection payload, comprising a 10x optical zoom camera and a spotlight on a pan-tilt unit. We augmented the zoom camera with a spectral filter wheel. The filter wheel contains five narrow band-pass filters (390 nm, 470 nm, 530 nm, 620 nm, and 940 nm) and one position with no filter. In this configuration, we refer to the payload as CTX-FW. We selected CTX-FW to obtain high-resolution images of the targets and their surroundings as well as multi-spectral information since lunar minerals are known to exhibit distinct reflective features at different wavelengths. On the back of the robot, we mounted a color calibration target with 24 color squares to color-calibrate CTX images.

The Hybrid carries a newer ANYbotics inspection payload for the ANYmal D. It consists of a 20x optical zoom camera, a thermal camera, and a spotlight on a pan-tilt unit. We refer to this payload as CTX-TH. Aside from high-resolution images, CTX-TH enables the thermophysical analysis of targets. Rocks and regolith have different thermal signatures depending on their composition. Additionally, we equipped the Hybrid with a Metrohm Instant Raman Analyzer XTR DS (MIRA), a ruggedized, portable Raman spectrometer . MIRA is equipped with an auto-focus attachment that enables measurements of targets from a distance of up to two meters. The MIRA allows compositional analysis of the target’s mineralogy, which provides a more in-depth investigation than the CTX-FW and CTX-TH imaging techniques.

The Scientist uses the DynaArm to deploy the instruments for in-situ measurements. The DynaArm has a total weight of 10 kg, including the 2.3 kg of the scientific payloads. It has a reach of 0.9 m without any tools attached and a nominal payload capacity of 8 kg at half of the reach with a maximum tool speed of up to 13.5 m/s. The forearm of the DynaArm carries a MIRA. Additionally, we mounted our custom- built UV-VIS-NIR microscopic imager (MICRO) as the end-effector (Fig. S1-B). It contains a USB microscope (Dinolite AD4113T-I2V) mounted on a linear actuator mechanism, a ring of 48 RGB LEDs, a time-of-flight sensor, and control electronics. The front of the casing is equipped with a foam ring, which ensures that no stray light enters the case when MICRO is in contact with a target. The internal UV and NIR LEDs of the Dino-Lite microscope and the RGB LED ring allow for acquiring multi-spectral microscopic images in the entire spectral range from UV to NIR at 395 nm, 470 nm, 525 nm, 620 nm, and 940 nm. The time-of-flight sensor determines the distance between the microscope and the target. The distance feedback allows the micro- scope, in combination with the linear actuator and an autofocus routine, to acquire sharp images of the samples, independent of the precise placement of the instrument. Using MICRO’s multi-spectral images, trained geologists can perform petrographic assessments which are more detailed than the analyses based on CTX-FW.

Legged Locomotion In Planetary Environments

We use a reinforcement learning (RL) approach to design our locomo- tion policy because RL approaches have shown robust performance in challenging environments [24, 25, 48]. We base our work on the perceptive locomotion pipeline originally published in , which has already been successfully used in other field deployments such as the DARPA SubT challenge . In this pipeline, the control policy is trained in two stages: First, we train a teacher policy with all ground truth information. Then, we train a student policy to output the same action as the teacher from noisy and limited information.

We use the existing controller from on the Scout. On the Hy- brid and the Scientist, however, the scientific payloads and the robotic arm can lead to big disturbances for the controller, resulting in reduced robustness and inefficient motion. Therefore, we developed an updated locomotion policy with explicit payload measurements: We add pay- load mass randomization during the training and explicitly give the mass information to the teacher policy. Secondly, we add arm position and velocity observations to allow the policy to actively counteract the wrench on the base caused by the arm’s motion. Additionally, we increased the joint torque reward to further improve energy efficiency.

During training, we gave a random arm initial position and random arm target position and controlled each arm joint via PD control to simulate the arm movements. We present more detail on our locomotion setup in the Supplementary Methods.

Localization, Mapping, And Perception

In this section, we describe our localization and mapping system, designed to globally localize the robots and generate an accurate rep- resentation of the environment during the mission. Furthermore, we describe the instance segmentation algorithm that helps the operations team understand the environment.

Robust Lidar-Based Localization

We rely on LiDAR-based SLAM because of the long range and large field of view of LiDARs. A core challenge of LiDAR-based SLAM in planetary analog missions is LiDAR degeneracy. To cope with degen- erate environments, we use a modified version of CompSLAM , which has shown to perform well under degeneracy through numer- ous deployments . CompSLAM is a complementary multi-modal SLAM system based on the Iterative Closest Point algorithm .

CompSLAM can use visual, thermal, or inertial data as a robust prior of the LiDAR mapping module. In this work, we rely only on the inertial data as a prior. If the environment degenerates, CompSLAM identifies degenerate directions and directly uses the prior to integrate the point cloud scan into the map.

Dense Point Cloud Mapping

CompSLAM maintains only a sparse map to keep the SLAM optimiza- tion problem tractable. We therefore require a separate module for dense mapping to create a high-resolution environment representation.

To this end, we maintain an octree-based map with a voxel size of 30 mm on the robots. Firstly, we merge and filter the point clouds of all LiDAR sensors on the robot. After registering the point cloud into the dense map using the pose estimate of the SLAM system, we clip the map at a two-meter height to remove unnecessary data. Furthermore, we apply an outlier filter by removing points that do not have at least five neighbors within a 200 mm radius. Lastly, we apply a multi-robot crop filter to reject points that lie on the other robots. The mapping pipeline additionally

14

maintains a colorized map by projecting the RGB information from the navigation cameras onto a copy of the point cloud. After the mission, we fetch the high-resolution point cloud map from the robot, simulating long-term data transmission. First, a GPU- based pre-processor calculates the surface normals at each point using the nearest 40 points. Subsequently, we use a statistical noise filter and a radius filter to filter the map. The statistical noise filter considers the nearest 10 points and filters out-of-distribution points (standard deviation > 2σ). After the pre-processing, we reconstruct a triangle mesh with a tree depth of 12 using the Poisson surface reconstruction method included in the open3D library to generate the high- resolution, continuous mesh representation. This process takes up to five minutes on a computer with an Intel i9 12700K CPU depending on the terrain complexity. Later, the generated high-resolution mesh is visualized with an eye dome lighting shader .

Lightweight Mesh Representation

Since an unreliable, high-latency network cannot transport the dense map in real-time, we periodically send a downsampled version of the dense map (voxel size of 150 mm) in a 9 m radius around the robot to mission control. Before transmission, we compress the point cloud using the Draco point cloud compression library . The mission control PC uses this downsampled and compressed point cloud to create a lightweight mesh representation (Fig. S6). With downsampling and compression, we reduce the size of the point cloud that is transported over the high-latency network from 3 MB to 250 kB. Furthermore, the robot only sends the point cloud if it contains substantial new information, for example when the robot moves, or when the user requests a mesh of a certain area in the user interface. The meshing operation on the mission control PC is limited to a maximum of 0.25 Hz.

The mission control PC calculates a triangle mesh using the Poisson surface reconstruction method with a depth parameter of eight. We recalculate the surface normals of the mesh to correct the orientation and smoothness of the mesh surfaces. Additionally, we filter out vertices and triangles with less than ten support points. This mesh is then merged into the mesh map by fusing the non-overlapped vertices and triangles and averaging the overlapping areas.

Elevation Mapping

The elevation map is a local robot-centric 2.5D grid representation where each cell indicates the terrain height. We use it as a terrain representation with a high update rate for the local planning module (Sec. Autonomous Local Navigation). The elevation map is 8 m x 8 m with a resolution of 40 mm. We use an elevation mapping pipeline run- ning on a GPU to integrate the point cloud data into the elevation map in parallel, based on the robot’s odometry.

Instance Segmentation

We apply a boulder instance segmentation network to autonomously identify each boulder instance in the RGB images to contextualize the scientific data. We built our approach upon Mask R-CNN . The network predicts each boulder instance’s bounding box, outline, and confidence. We finetuned the original model on a custom-built dataset containing hundreds of images collected in the first field trial of the SRC with instance labels. In this custom-built dataset, ANYmal ac- quired the images in a dark analog lunar terrain, similare to the setting in the final challenge. We deployed our rock instance segmentation network on the mission control PC to segment navigation and CTX im- ages. Fig. S7 shows several prediction results, indicating the network’s robust performance under variable light conditions and different input cameras.

Scientific Payload Integration And Deployment

The operator can send 3D remote measurement targets to request images from CTX-FW or CTX-TH. For CTX-FW, the operator can additionally specify which spectral filters to use. The operator marks a target on the map or uses an already-marked target to send a remote measurement task. When requesting RGB images, the operator can select between a high and a low zoom level, with an image width of 0.3 m and 3 m in the focal plane, respectively. The robot sends the images to mission control, which automatically saves them by the target identifier. Additionally, the operator can request panoramic images. The robot then takes nine images at a fixed zoom level, which are stitched in post-processing.

The MIRA communicates to the onboard PC of the robot through a USB interface. It is fully integrated into the software stack through a custom ROS wrapper that provides a ROS action interface to trigger a measurement. The measurement procedure is then fully automated: the auto-focus attachment measures the distance to the target, focuses on this distance, and then the MIRA initiates a measurement. The MIRA matches the received spectroscopic response against the custom-built library of 355 different spectra of relevant lunar minerals (for example oxides, olivines, pyroxenes, and feldspars) (Table S5). The robot sends the spectroscopic response and the corresponding match to mission control for the operations team to evaluate.

The MICRO has a USB interface and communicates to the onboard PC of the robot through rosserial. As for the MIRA, its measurement sequence is fully automated. On a measurement request, the device focuses on the target automatically, records an image sequence at all available spectral bands (UV, R, G, B, NIR), and returns a complete measurement data package.

Payload Deployment With The Robotic Arm

To deploy the MIRA, the arm aligns the Raman laser with the target at the desired distance of 0.7 m. The deployment of MICRO happens in two stages: Firstly, the arm aligns the MICRO with the target pose at a predefined offset of 100 mm in the inverse heading direction of the target pose. In the second stage, a PID controller moves MICRO along the heading axis of the target pose to bring the front foam piece (Fig. S1) in contact with the target surface using the ToF distance measurement as feedback. The two-stage approach allows MICRO to accurately make contact with the target surface, even if the original 6D in-situ measurement target is not precisely on the surface. We provide information on the arm controller in the Supplementary Methods.

Autonomy

We use behavior trees (BT) to handle the autonomous task execu- tion on the robots. The modularity of BTs allows us to reuse modules across the different robots. The BT on the robot processes all operator interactions and starts, monitors, and stops the requested task. In this way, the operator sets high-level objectives while retaining the ability to stop or change the task easily. We show more details of our behavior trees in the Supplementary Methods.

Autonomous Local Navigation

The operators guide the robots via high-level navigation goals. We specifically avoided a fully autonomous approach because we preferred the operations team to select and prioritize scientific targets during the mission in unknown environments.

To allow the operator to safely operate the robots via waypoints, we use a sampling-based local planner that operates on the local elevation map. Unlike most state-of-the-art navigation planners, it does not assign traversability values to discrete terrain patches. Instead, the robot morphology is approximated by reachability volumes for the feet and a collision volume for the torso, which makes the planner more

15

suitable for legged robots. Using a PID controller, a path-tracking mod- ule converts the traversable path to twist commands for the locomotion controller.

Autonomous In-Situ Measurement Acquisition

Once the Scientist is close to a target, the operator can select the in- situ measurement point on the grey-scale infrared image stream of the Realsense camera. We use the grey-scale infrared image from the Realsense camera because it does not provide an RGB image stream and the infrared image provides good output in bad lighting conditions.

The user interface spawns a 6-DoF interactive marker, which can be adjusted if necessary. The operator can specify the exact measurement task, that is, a MIRA measurement, a MICRO measurement, or both.

Once the robot receives the measurement task, it approaches the target and aligns the base at a predefined offset of 1 m to the target. Then, the robot deploys the instruments and triggers a measurement automatically. Depending on the task definition, the robot takes a MIRA measurement, a MICRO measurement, or both. If an error occurs during the measurement process, the arm will move back to the default position, and the robot will notify the operator.

Loss Of Signal Operation

If a LoS occurs, the robots finish their currently allocated task and then switch to a dedicated LoS behavior. The Scout and the Hybrid perform an autonomous measurement routine. First, they record a panoramic image of their environment. Second, they image their own footprints, which provide information about soil mechanics. Third, the robots acquire images of each other and the Scientist to assess the status of the hardware. Fourth, the Scout takes an image of the color calibration card. This image is used in post-processing to perform color correction. The robots send the acquired data to mission control once communication is restored. The Scientist remains in a nominal standing configuration since the system currently does not autonomously detect new measurement targets.

Resilient Communication For High-Latency Networks

Resilient communication in high-latency networks is crucial to ensure data transmission between mission control and the robots. We use commercial off-the-shelf radio devices by Rajant to create a re- liable mesh network. The mesh setup includes the base station and the robots, each acting as a mesh node. The robots are equipped with a BreadCrumb DX2 and the base station consists of a BreadCrumb ES1 with a panel antenna. All radios operate at 5.8 GHz. The base station and the mission control PCs are connected by wire via a delay simulator that ensures a 5.0 s RTT delay.

We use the Robot Operating System (ROS1) on all robots and mission control stations to manage onboard communications. Each robot and mission control station had a separate rosmaster. However, ROS1 is not designed to communicate over high latency networks, as a TCP handshake is required to establish a connection, even if the data is transmitted over UDP. Therefore, we use Nimbro Network to estab- lish communication between different rosmasters [62, 63]. Nimbro Network was specifically designed to work in unreliable high-latency networks and used in the SpaceBot cup, a robot competition for lunar exploration .

Using the predominant Internet protocol TCP is problematic be- cause of the RTT delay and high bandwidth. Consequently, we transmit all data via UDP. UDP does not provide congestion control, and there is no guarantee that data will arrive. Therefore, mission control and the robots only exchange essential data. We show the network usage of mission control 1 at the SRC in Fig. S8. The robot-to-robot com- munication, however, runs on TCP because the delay is in the low millisecond range.

Acknowledgments: We would like to thank the implementation (HSLU), ANYbotics AG, and the maxon SpaceLab. A special thanks to Samuel Tenisch, Alex Brandes, Gerhard Székely, and Nico Steinert for developing the filter wheel, Fabio Mast, Lars Horvath, Hugo Um- bers, and Marco Trentini for supporting the MICRO development, and Daria Larcher and Julian Bernasconi for supporting the scientific data analysis. We would like to thank Metrohm Schweiz AG for providing the Raman spectrometer. We acknowledge the ETH Earth Sciences rock and mineral samples for our experiments. We thank KIBAG Kies Neuheim AG for making the quarry available. This work has been conducted as part of ANYmal Research, a community to advance legged robotics. Funding: This research was supported by the Swiss National Science Foundation (SNF) through the National Centre of Competence in Research Robotics (NCCR Robotics), through the Na- tional Centre of Competence in Digital Fabrication (NCCR dfab), and received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme grant agreements No 852044 and No 101016970. This project has and 4000135310/21/NL/PA/pt. Author contributions All authors con- tributed to the system design and wrote the paper. PA, GW, JP, TT, RZ, VB, GL, FK and HK participated in the field deployments and evaluated the respective data. PA, GW, JP and TM conducted the additional locomotion tests. TT and RZ developed the mapping and perception pipelines. JP developed the arm controller. TM developed the locomotion controller. VB, GL and FK selected and designed the scientific payloads and evaluated the respective data. Competing inter- ests: The authors declare no competing interest. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Material. Other material

Authors:

Peder EZ Larson 1, 2,* , Jenna ML Bernard1, James A Bankson 3, Nikolaj Bøgh 4, Robert A Bok1, Albert P. Chen 5, Charles H Cunningham 6,7, Jeremy Gordon1, Jan-Bernd Hövener 8, Christoffer Laustsen 4, Dirk Mayer 9,10, Mary A McLean11 12, Franz Schilling13, James Slater1, Jean-Luc Vanderheyden5, 14, Cornelius von Morze 15, Daniel B Vigneron1, 2, Duan Xu1, 2, and the HP 13C

94143, Usa.

Denmark. 5 GE Healthcare, Menlo Park, California, USA. 6 Physical Sciences, Sunnybrook Research Institute, Toronto, Ontario, Canada.

8 Section Biomedical Imaging, Molecular Imaging North Competence Center (MOIN CC), Medicine, Baltimore, MD, USA. Cambridge, United Kingdom.

14Jlvmi Consulting Llc, Dousman, Wi, Usa

#See Acknowledgements for a list of all HP 13C MRI Consensus Group Members This work was supported by the ISMRM Hyperpolarized Media MR Study Group, the ISMRM Hyperpolarization Methods & Equipment Study Group, and the Hyperpolarized MRI Technology Resource Center (NIH/NIBIB grant P41EB013598).

Abstract

MRI with hyperpolarized (HP) 13C agents, also known as HP 13C MRI, can measure processes such as localized metabolism that is altered in numerous cancers, liver, heart, kidney diseases, and more. It has been translated into human studies during the past 10 years, with recent rapid growth in studies largely based on increasing availability of hyperpolarized agent preparation methods suitable for use in humans. This paper aims to capture the current successful practices for HP MRI human studies with [1-13C]pyruvate - by far the most commonly used agent, which sits at a key metabolic junction in glycolysis. The paper is divided into four major topic areas: (1) HP 13C-pyruvate preparation, (2) MRI system setup and calibrations, (3) data acquisition and image reconstruction, and (4) data analysis and quantification. In each area, we identified the key components for a successful study, summarized both published studies and current practices, and discuss evidence gaps, strengths, and limitations. This paper is the output of the “HP 13C MRI Consensus Group” as well as the ISMRM Hyperpolarized Media MR and Hyperpolarized Methods & Equipment study groups. It further aims to provide a comprehensive reference for future consensus building as the field continues to advance human studies with this metabolic imaging modality.

Keywords: Hyperpolarized MRI, metabolic imaging, carbon-13, pyruvate, dissolution dynamic

Introduction

MRI with hyperpolarized 13C agents, also known as hyperpolarized (HP) 13C MRI, has shown great potential as a novel imaging modality, particularly for its ability to probe metabolic processes in real time. The first human studies with HP [1-13C]pyruvate were performed in 2011 in prostate cancer patients (1).

Since then, there have been over 60 papers published with imaging results of human subjects from 13 different sites, with applications including prostate cancer, brain tumors, breast cancer, kidney cancer, pancreatic cancer, metastatic disease, liver disease, ischemic heart disease, diabetes and cardiomyopathies. The vast majority of these studies used [1-13C]pyruvate (1–63), where [2-13C]pyruvate (64) and 13C-urea (56) have been demonstrated too.

As clinical HP 13C MRI advances, there is a growing need to build consensus for best practices, which are critical for comparing data across sites, performing multi-site trials,deploying methods to new sites, partnering with vendors, and potentially for obtaining broader regulatory approvals.

In March 2022, we initiated an effort to build consensus within the HP 13C MRI community with this opportunity in mind, and it was greeted with strong enthusiasm. The “HP 13C MRI Consensus Group”, containing over 55 members from 27 sites, identified the area of greatest need and opportunity for consensus building to be HP [1-13C]pyruvate human

●

Pyruvate is the most mature and widely used HP agent and has the most significant translational evidence emphasizing the potential clinical impact.

●

Clinical trials, particularly multi-site trials, have the strongest need for consensus methods to ensure that data can be combined across sites. This work is a Position Paper for which the goal is to describe current successful practices and study methods for HP [1-13C]pyruvate human studies along with justification to support those practices. This is divided into four major topic areas: (1) HP 13C-pyruvate preparation, (2) MRI system setup and calibrations, (3) data acquisition and image reconstruction, and (4) data analysis and quantification (Fig. 1). The current successful practices and study methods include a literature review of published peer-reviewed journal papers showing human HP [1-13C]pyruvate study data, up to September 2022 (1–63), as well as new unpublished information from surveys of HP 13C study sites. Based on this information, we also highlight the evidence gaps, strengths, and limitations of current practices which are summarized at the end of each section.

Figure 1: Illustration of the HP 13C MRI human study process, including the 4 major areas covered in this paper: Hyperpolarized 13C-pyruvate preparation, MRI system setup and calibration, Acquisition and Reconstruction, and Data Analysis and Quantification.

Figure 2: Anatomical targets of HP [1-13C]pyruvate MRI human studies published up to September 2022.

Hyperpolarized 13C-Pyruvate Preparation

This section covers the processes for creating the HP agent, 13C pyruvate, and will include many aspects and considerations that are needed to safely and effectively prepare doses for metabolic imaging studies in human subjects. These include material, personnel, equipment and facility, fluid path preparation, quality control, and release.

It is helpful to understand that the specifications of a dose of 13C pyruvate suitable for in vivo MR HP metabolic imaging were shaped in part by early preclinical studies performed by GE HealthCare summarized in Ref. (65). In short, the safety of the two novel drug components, 13C pyruvate and the electron paramagnetic agent (EPA) AH111501, were demonstrated in those studies. The more precise formulation of the dose suitable for human use was then determined from clinical studies (66) that included two Phase 1 clinical trials in young and elderly healthy volunteers without hyperpolarization of the 13C nuclei and another Phase 1/2a dose escalation and imaging feasibility study with HP 13C pyruvate in 31 prostate cancer patients at the With the exception of the first HP 13C imaging clinical trial, which utilized a prototype device in a cleanroom (1), all HP 13C studies performed in humans to date have utilized the SPINlab polarizer (manufactured by GE HealthCare). Consequently all doses of the HP 13C pyruvate delivered by SPINlab have been produced using the “SPINlab Pharmacy Kit” that serves as the container-closure system for the various drug components (13C pyruvic acid and EPA mixture, dissolution medium, and neutralization and dilution medium) during sample polarization, dissolution and quality control (QC) processes. Thus many aspects of the HP sample preparation considerations discussed below are related to the SPINlab instrument and the consumables designed to be used with it (67).

General Considerations

While more than 860 patients or healthy subjects having been injected with HP 13C pyruvate as of January 2022 without reports of any serious adverse events (68), HP 13C pyruvate injection remains an investigational MR contrast agent and can only be administered by those with Investigational New Drug (IND) exemption from the Food and Drug Administration (FDA) in the USA, a Clinical Trial Application (CTA) in Canada, approval from National Research Ethics Committee Services in the UK, or approval from the relevant local regulatory body. Thus, methods and processes involved to produce a dose should have patient safety as the first priority. Since utilizing dissolution dynamic nuclear polarization (dissolution-DNP) for human use is still a relatively new development, there are no existing published regulatory guidelines specifically for this method.

There are two major production styles that determine how various sites approach the agent preparation. In the US, the most common approach is to rely on a sterilizing filter (“Terminal Sterilization”) to ensure sterility of the final product, akin to PET tracer production, where a starting molecule with a radioisotope is processed using various other ingredients to make the final, desired and injectable contrast agent within a necessarily short amount of time (69). For these sites, sterilization of the components and accessories upstream of this filter are not required, although many of them were manufactured and tested following Good Manufacturing Practice (GMP) or Good Laboratory Practice (GLP) requirements. The filling process is usually performed under an ISO 5 laminar flow hood, but a clean room or an isolator is not required.

This approach is typically accompanied by testing the integrity of the sterilizing filter prior to release of the dose for injection. Typically, post release endotoxin and sterility tests are performed using an aliquot reserved from each released dose.

In the UK and EU, the most common approach is to more-closely follow sterile pharmaceutical compounding guidelines (70), where all components and ingredients are required to be sterile or manufactured under GMP guidelines and are assembled and filled within a clean room environment or an isolator system (“Sterile Preparation”). Typically a batch of Pharmacy Kits for HP 13C pyruvate injection are prepared together. The sterility of the final dose is also ensured by batch validation testing, in addition to the sterility of the ingredients and the sterile compounding process. The endotoxin and sterility testing are performed for the process validation but are not performed for each injected dose.

Some institutions fill and assemble the Pharmacy Kit required for a specific study on the same day or the day prior to polarization, dissolution, and patient administration, but others have also demonstrated the feasibility of preparing a batch of kits, keeping them in a -20ºC freezer and using them over a period of a few months.

Beyond the obvious requirements that the process and the facility has to ultimately produce a dose that is safe to inject into a human, regulatory authorities will also focus on the question “Are you in control of your processes?”. To be in control of your process requires an in-depth and broad understanding of all processes involved in pre, post, and during the production process.

Personnel

It is typical and may be required to have licensed personnel involved in the production process depending on local regulations.Typically a pharmacist, radiopharmacist or other similarly qualified person (QP), in charge of the facility where the Pharmacy Kit filling and preparation is taking place, is responsible for the overall process and the release of the injectable dose.

Qualified cleanroom technicians are often involved in the Pharmacy Kit filling under the supervision of the pharmacist or QP. As is required for pharmaceutical compounding or PET tracer production, training requirements and training records for all personnel need to be maintained and available for audit by the FDA or equivalent.

Equipment And Facility

The facility and all equipment need to have standard operating procedures (SOPs) that describe how equipment is used, maintained, and calibrated to comply with relevant legislation. Currently, almost all the filling of the Pharmacy Kit takes place within a compounding laminar flow hood or isolator (typically ISO 5). At some sites, the filling is conducted within a cleanroom, while at others, it is conducted in a dedicated non-cleanroom space, reflecting differences in cleanroom approach and specifications between regulators worldwide (71). Some equipment or facilities, such as the compounding hood or cleanroom, may require external certified laboratories for testing.

Material Handling

Material handling guidelines (69,70) require SOPs detailing a system to track all of the materials involved in the HP production process for a particular patient dose, similar to current good manufacturing practice (cGMP) requirements for material handling for drug compounding. This includes acceptance standards, storage conditions, amount used in the patient dose for each ingredient and materials used in the assembly of the fluid path and Pharmacy Kit. Currently some users choose to open and inspect and sometimes modify the Pharmacy Kits upon arrival, but some users keep them in the sealed packaging until they are required for dose preparation.

Pharmacy Kit Filling And Assembling

As required by an IND or its equivalent, the preparation of the doses of HP 13C agent are detailed in the Chemistry, Manufacturing, and Control (CMC) section of an applicable regulatory submission; an example of this has been made available (72). It describes the processes of filling the Pharmacy Kit with the different components that make up the final drug product, and of assembling the final kit for either storage or immediate use in the polarizer. Special attention should be given to the laser welding process in order to satisfy installation qualification (IQ) and operational qualification (OQ). Typically, the final developed process is validated by process qualification (PQ) runs, during which 3 or more Pharmacy Kits are filled and used and the final HP 13C products are tested for endotoxin and sterility and to confirm that they meet the dose specifications for injections (usually including pyruvate concentration, residual EPA concentration, pH, liquid state polarization level and dose temperature). The data from 3 consecutive PQ runs are submitted as part of the IND submission (or its equivalent), and are often also reviewed by the Institutional Review Board (IRB) where the studies are conducted.

Quality Control And Dose Release

The quality control (QC) and dose release can be separated into two aspects: one is the QC and release of the filled Pharmacy Kit, and second is the QC and release of the HP 13C agent for injection, after polarization and dissolution. For institutions filling a batch of kits and storing them to use over a period of time, typically the batch can be released based on initial validation, environmental monitoring data from the day of kit production, and if filters are used during preparation of any of the components, filter integrity testing. But in some cases one or more kits are used for validation before the batch of kits are released for future use. For institutions that fill only the kits required for specific studies shortly before the experiment, the filled kits often do not go through separate release tests before they are used.

The quality control of the HP 13C pyruvate solution post dissolution is primarily performed to ensure that the agent meets the dose specifications (Table 1) before it is administered to the subject. These specifications target both safety (pH, residual EPA, temperature) and efficacy (pyruvate concentration, polarization, volume). Typically, the pyruvate concentration, residual EPA concentration, pH, dose temperature, dose volume, and liquid state polarization are measured by the QC accessory associated with the SPINlab polarizer. Some users perform a secondary measurement for one of the parameters, such as pH, using a different instrument or pH paper. For sites that do not go through a separate release testing process for batch filled kits, the integrity of the sterilization assurance filter, a part of the Pharmacy Kit, is typically tested as a part of the dose release. It is also common for these users to preserve an aliquot of the final HP 13C pyruvate solution for post-release endotoxin and sterility testing. This testing cannot be completed fast enough to test an individual dose prior to injection, but this is why other processes such as PQ runs and validation testing are done to minimize the chance a subject could be injected with a contaminated dose.

The Final Dose Release And Injection

should be done under the supervision of a licensed professional, based on local regulations.

Some Key Challenges

Many of the challenges associated with HP 13C pyruvate preparation can be attributed to the conditions required for the dissolution-DNP method of high magnetic field (~3-7 T) and very low temperature (~1 K) during polarization, with pressurized and superheated water necessary for the rapid dissolution event. These extreme conditions are quite challenging for the design of the container-closure and fluid path system. In particular, the cryogenic temperature in the polarizer requires special attention to any moisture or ambient (moist) air introduced into that portion of the fluid path, which can form an ice block at ~1 K. This ice can lead to flow restriction during the dissolution event and reduce the strength of the laser welded bond between the cryovial and its cap. This can ultimately produce failures in the dissolution step, including variations in final pyruvate concentration and pH that may fail to meet QC release criteria as well as fluid path ruptures that provide no available dose and result in polarizer down-time.

The polarization of the HP 13C pyruvate sample decays quickly over the span of a few minutes after dissolution, and thus the process of dissolution, QC for release, and injection should be completed as fast as possible to preserve the high polarization level achieved. Any delays in the preparation process, such as transportation time or equipment malfunction, can significantly reduce the final polarization and result in lower quality imaging data.

Current Practices

A summary of data collected from all sites performing clinical trials with HP 13C-pyruvate is shown in Fig. 3 and Table 1, including the specification of the final dose and how the quality control and release of the final dose are performed. There is a split in the Production Style, described in the General Considerations section above, with 8/13 sites using Sterile Preparation versus 5/13 using Terminal Sterilization. While many of the dose specifications show notable differences in acceptable ranges, all of these variations listed in tables have been successfully and safely been used to perform HP 13C pyruvate studies in humans. Their differences depend on the institutions’ preferences, resources and their particular regulatory situation. There is high similarity in pyruvate ranges, temperature ranges, EPA limits, and volume limits. There is modest variability in pH ranges and large variability in the endotoxin test limit. There is a 3-fold difference in acceptable polarization levels, which are measured to ensure a futile dose is not injected since the polarization is directly proportional to SNR. This reflects the decision by several sites to believe that useful data can be still be obtained with suboptimal polarizations.

Figure 3: Hyperpolarized agent preparation methods reported by sites currently performing HP

In House

Table 1: HP 13C-pyruvate preparation parameters, methods, and dose specifications used for quality control testing and release as well as validation. These were obtained from a survey of all sites performing clinical trials with HP [1-13C]pyruvate. The parameters used for product release are noted in bold text, otherwise these parameters are measured for batch validation or other QC measurements. The endotoxin and sterility testing are performed during process validation of the batch and/or post-injection, and largely depends on the agent production approach.

Summary

The overall safety record of HP 13C-pyruvate has been very strong, and the SPINlab hyperpolarizer has proven to provide high polarizations at human sized doses while meeting numerous QC and release criteria. A weakness remains the failure modes of the SPINlab Phamacy Kits (e.g. ice blocks, path ruptures), which are placed under extreme requirements particularly during dissolution. The preparation process still requires a high degree of expertise.

Therefore, there is a significant need to improve the reliability, robustness, and ease of operation for generating HP 13C-pyruvate doses for human studies. Furthermore, there is a divide between manufacturing and sterile compounding style preparation as well as other site-specific practices, resulting in variations in SOPs and justification required to relevant regulatory bodies. There have also been no comparisons between these approaches. It is also unclear what release criteria and QC parameters are truly required to ensure patient safety.

However, all of the reported methods are acceptable and approved by the appropriate regulatory authorities, and have led to the rapid expansion of successful human studies in recent years.

Mri System Setup And Calibrations

This section covers the MRI system setup, including the imaging system, RF coils, phantoms, and prescan calibration methods.

Imaging System

The main prerequisite for a given MRI scanner to be capable of supporting studies with HP 13C is its “broadband” capability to transmit and receive radiofrequency (RF) signal at the frequency of 13C, which is around 4 times lower than 1H. This does not come as a default on clinical MR devices. The transmit power of the broadband amplifier should also be sufficient to support the intended flip angle and RF pulse shape with the employed transmission RF coil(s) for 13C. Most studies to date use relatively low flip angles (< 90 degrees) for HP 13C in order to preserve polarization for time-resolved imaging. The capability to receive 13C signal on multiple channels is also desirable to increase SNR, as discussed further in the “RF coils” section.

The choice of magnetic field strength is primarily dependent on the metabolites’ frequency separation due to chemical shift dispersion and 1H imaging. High field strengths do not enhance hyperpolarized 13C signal as they do for 1H because the signal strength in a HP experiment relies on manipulating the population of quantum energy states outside of the MRI scanner.

However, the injected HP 13C-pyruvate and its metabolic products have greater frequency separation at higher fields, and it may thus be easier to separate and quantify these resonances at higher fields. This comes at the cost of a reduction in the achievable T2* and often reduced T1. As the initial polarization is independent of the imaging field strength it has been proposed that the increased T2* at 1.5T can potentially be exploited to increase SNR by adapting the acquisition bandwidth or reduce off-resonance imaging effects in cases when the decay of the transverse magnetization is dominated by T2* (73). In practice, 3T has been used in all published human 13C-pyruvate studies surveyed (Supporting Table S1), and comprises the majority of scanners currently in use for human studies (Table 3). A field strength of 3T is well-suited for 1H MRI anatomical reference and correlative imaging.

Stronger and more rapidly slewing magnetic field gradients support more rapid spatial encoding, particularly for metabolite-specific single-shot imaging using echo-planar imaging (EPI) or spiral imaging (See “Acquisition and Reconstruction”). Although the spatial resolution acquired for HP 13C imaging is typically much coarser than for 1H MRI, the factor of ~4 in gyromagnetic ratio leads to the same reduction factor in performance of the gradient system, so 13C experiments are potentially more limited by gradient hardware performance. To date, all human studies have used the commercially-available integrated gradient systems provided in clinical MRI scanners.

Optimization of scanner design has understandably focused on minimization of artifacts in 1H MRI, where devices such as room lights, the gradient amplifiers, and the motors driving the patient bed are checked to ensure that they do not produce RF interference at the 1H frequency, but artifacts may arise at other frequencies. Eddy current compensation is also not always appropriately adjusted for nuclei at other frequencies (74). In order to optimize for 13C, many sites have performed checks on phantoms for RF interference, gradient artifacts, and eddy currents (74), including the use of post-hoc gradient impulse response function characterisation and correction, and some vendors have fixed these issues as well.

Rf Coils

For HP 13C imaging studies in humans, RF coils for both 1H and 13C nuclei are needed, with 1H MRI providing an anatomical reference for registration and optional additional multiparametric MRI readouts. At the Larmor frequency of 13C nuclei, the relative contributions from coil noise compared to sample noise increase compared to 1H (73,75), although sample noise still is likely the dominant contributor for human-sized coils at 32.1MHz - the resonance frequency of 13C nuclei at 3T.

The key requirement for human 13C-pyruvate RF coils are that the coil geometry and sensitive volume must cover the volume of interest in the subject. Table 2 and Figure 4 shows coil configurations that have been used and optimized for applications in different anatomic regions.

Volume resonators are most commonly used for transmit, as they surround the subject to

Provide B1 Transmit Across The Fov (B1

+). While 1H relies on a large birdcage (“body”) coil built into the scanner, 13C transmit coils must be placed inside the bore. This takes up valuable space within the magnet, and also has led to the use of designs with relatively inhomogeneous

B1

+. Many human studies have used Helmholz pair resonators for transmit, including the “clamshell coil”, which has a notably inhomogeneous B1

+ Profile But Has Been Used Because Of

relatively easy integration into the scanner bore. B1

+ Variation Results In Variations In The Flip

angles that control the use of the hyperpolarized magnetization and creates errors in common HP metrics (9,76). The exception are head coils, where birdcage designs with highly

Homogeneous B1

+ can be placed around the head while easily fitting inside the bore. As with 1H MRI, higher SNR can typically be achieved by smaller receive coil elements, such as surface coils or phased arrays, and the majority of 13C receive coils used have layouts similar to 1H phased arrays.

RF coil quality control is important to ensure proper functioning of the coils to provide consistent imaging quality, especially with limited natural abundance 13C signal in vivo. It typically involves 1) a physical integrity check of the coil cables and connectors and 2) phantom SNR tests to check the coil’s performance and to monitor it over time (see Phantoms below). An useful reference for RF coil quality control is outlined in the MRI accreditation program of the American College of Radiology (77) and can be adapted for 13C coils.

Notably, configurations for brain and prostate studies used dual-tuned 1H/13C coil designs, which greatly simplify workflow and registration of 1H and 13C images, as no switching of coils is needed.

(1)

Table 2: RF coil configurations reported for human HP [1-13C]pyruvate studies.

Tx = Transmit

coil, RX = receive coil. The commonly used “clamshell” TX coil is a Helmholz pair design. For 1H RF configurations, all used the Body coil for TX unless otherwise noted, and “repositioned” indicates the 13C coil was removed for 1H imaging. One representative reference is listed for each configuration. The RF coil configurations reported in the reviewed papers are shown in Supporting Table S1.

Figure 4: Examples of RF coil configurations used for human HP [1-13C]pyruvate brain studies. (A,B) 13C Clamshell TX (Helmholz pair) and 2× 4-channel paddle RX arrays. (C) 13C Birdcage volume TX and 32-channel RX array (RX array slides into TX coil). (D) 13C Birdcage volume TX and 24-channel RX array, combined with a 1H 8-channel RX array. Image reproduced with permission from Ref (16).

Phantoms

Since hyperpolarized magnetization is non-renewable, phantoms containing 13C nuclei are important to: 1) test the multi-nuclear capabilities of the imaging system, including all parts of the signal excitation and receive chain; 2) perform calibration measurements before a scan with hyperpolarized nuclei; and 3) perform necessary pre-scan adjustments (see “Prescan Calibration” section). The phantoms currently in use are listed in Table 3. Their composition must provide sufficient 13C signal, with additional considerations of conductivity, stability, chemical shift(s) present, potential for dynamic imaging, and cost. The phantom geometries are typically either compact, in order to be used alongside the subject during a HP scan, or large enough to mimic the inner volume of a RF coil for system testing.

One popular compact design contains enriched 13C-urea at high concentration, typically 8 M, which provides a single resonance, placed inside a small container ~1 mL. The most common recipe mixes 13C-urea in a 90% water/10% glycerol solution, with glycerol used to increase the urea solubility and doping with a Gd-based contrast agent to shorten T1 which increases the potential SNR per unit time. For example, when Dotarem is added at a 3:1000 volume ratio the 13C-urea T1 is around 500 ms and T2 is around 100 ms. However, when testing pulse sequences influenced by T1 and T2, doping should be used carefully. This phantom is suitable for frequency calibration, transmit gain calibration, sequence testing, and as a fiducial marker when placed next to a patient. However, enriched 13C-urea has a relatively high cost compared to natural abundance compounds.

For larger volumes (>100 ml), the phantoms most often used contain undiluted ethylene glycol, glycerol, or dimethyl silicone. These compounds have sufficiently high carbon concentrations to provide sufficient 13C signal even with the 1.1% natural abundance of 13C. These larger phantoms matching the inner volume of an RF coil are useful for coil testing, including transmit

+) And Receive (B1

-) coil profile mapping, as well as to mimic acquisitions using in vivo FOV requirements. In this case, size and conductivity should match the expected subject size in order to mimic coil loading and get a realistic estimation of B1+. Large-volume natural abundance urea phantoms have also been used by some sites, but suffer from higher conductivity compared to biological tissues. Typically, it is easier to increase the conductivity and hence coil loading of the non-conductive phantom by adding NaCl to match physiological loading (16,78).

Dynamic phantoms that aim to mimic metabolite kinetics have also been developed (79–81), and have the potential to more closely mimic the HP experiment, but so far these are not widely used.

Prescan Calibration

Prior to performing an MRI acquisition, the so-called prescan procedure is used to set the shim parameters to maximize B0 homogeneity over the field of view (FOV) or a specific region of interest (ROI), the scanner center frequency (CF), the RF transmit gain, and the receiver gain.

While this calibration procedure is usually automated for 1H, the lack of sufficient natural abundance 13C signal prevents use of automated methods. (Although natural abundance 13C lipid signal has been detected, there are so far no reports on using this signal for prescan.) Table 3 shows current practices across sites.

Maximizing B0 homogeneity is independent of the nucleus and is therefore performed prior to 13C imaging using the 1H water signal and existing shimming tools, such as by a standard automated process (“Auto Shimming”) or using high order shimming routines. Similarly, the 13C CF can be calculated from the 1H CF using a predetermined scaling factor that depends on the target chemical shift (82). Another common approach used is to have a small, high-concentration 13C phantom, e.g. 8M 13C-urea, integrated in the RF coil or placed next to the scan subject (1). The reference frequency can also be based on real-time measurements after the HP injection but prior to imaging (83). Both the CF and B0 shimming are critical when using spectrally-selective RF pulses, as inmetabolite-specific imaging methods, where the desired excitation bandwidths are typically very narrow and frequency offsets can lead to a failure mode that is only apparent after injection.

The calibration of the RF transmit power is typically performed on a small, high-concentration 13C phantom placed near the region of interest during the scan or on a large 13C phantom of similar size and coil loading as the subject, prior to the subject scan. Reference power is often done by sweeping the power in a pulse-acquire sequence (53,62), or the Bloch-Siegert method (52,84). When using a small phantom, the location of the phantom, B1

+ Inhomogeneity As Well

as any shielding effects, e.g., when the phantom is integrated into a coil (1), may degrade the accuracy. Other methods include real-time Bloch-Siegert method measurements after the HP injection (83), and using the stronger natural abundance 23Na signal that is close enough to the 13C resonance frequency to be detected by 13C coils (82).

The receiver gain is predetermined, either systematically based on independent phantom measurements and assuming the dose and polarization of the HP compound is known prior to injection, or based on past HP imaging studies.

Power [Kw]

Phantom(s) - during study Phantom(s) - before study 13C Frequency

8

13C-bicarbonate doped with dimethyl silicone, various

Maximum Values

Table 3: Summary of the imaging systems, phantoms, and prescan procedures used at sites currently performing HP 13C-pyruvate human studies. These were obtained from a survey of all sites performing clinical trials with HP [1-13C]pyruvate. *Previously performed studies with a Siemens 3T Tim Trio. The imaging systems, phantoms, and prescan procedures reported in the reviewed papers are shown in Supporting Table S1.

Summary

Commercially available 3T MRI systems are by far the most commonly used for human HP 13C-pyruvate studies, although a systematic investigation of the impact of B0 has only recently been investigated (73). The multi-nuclear RF transmit and receive chain has proven sufficient for current acquisition strategies, although many sites have observed artifacts due to RF interference, gradient interference, and residual eddy currents when operating at the 13C frequency. A variety of 13C RF coils, tailored for numerous anatomical targets, have been successfully demonstrated, with the main limitation that most transmit coils take up a lot of additional space inside the bore and provide relatively inhomogeneous B1

+ Profiles. The

phantoms used have converged into generally 2 categories - small phantoms containing 13C-enriched compounds that can be used during the study and human-sized phantoms containing compounds with high carbon concentrations but without 13C enrichment that are used to test and calibrate the coils. There are no standardized compositions or geometry, and dynamic phantoms that recapitulate in vivo kinetics would be desirable but are still an emerging area. Prescan calibration procedures were not well defined in most publications, so we surveyed individual sites to determine current practices. Calibration procedures for the B0 field (13C CF and shimming) for most sites take advantage of 1H signal and methods, while methods

For Calibration Of B1

+ is more variable across sites, likely a reflection of remaining challenges in how to perform this calibration. Standardization of both phantoms and calibration procedures would synergistically improve the robustness and reproducibility of HP 13C studies.

Acquisition And Reconstruction

Data acquisition strategies in human HP [1-13C]pyruvate MRI studies must account for multiple chemical shifts, efficiently utilize the non-renewable HP magnetization, and acquire data quickly relative to metabolism and relaxation decay processes. These studies require spectral encoding to separate metabolites, necessitating pulse sequences that efficiently encode up to 5D data (3 spatial + 1 spectral + 1 temporal dimension). RF pulses must efficiently sample without immediately saturating the non-renewable HP magnetization, and sequences must acquire data quickly and be robust to both experimental and physiologic variation (e.g. B1

+ Inhomogeneity,

variation in perfusion) to ensure reproducibility and minimize scan-to-scan variability. This section covers current successful practices for data acquisition in human [1-13C]pyruvate studies, and accompanying 1H imaging, from different anatomic regions, including scan parameters and image reconstruction.

Acquisition And Reconstruction Methods

The acquisition methods used in human [1-13C]pyruvate studies can be classified into 3 categories: 1) MR spectroscopy or MR spectroscopic imaging (“MRS/I”), 2) chemical shift encoding methods, and 3) metabolite-specific imaging (Fig. 5).

Mrs/I Methods Specifically

resolve a spectrum that can be analyzed to extract expected as well as unexpected resonances, making this approach very robust. It was used in many initial studies (1).

Chemical Shift

encoding methods, most commonly the Iterative Decomposition of water and fat with Echo Asymmetry and Least-squares estimation (IDEAL) method, use imaging sequences acquired with multiple TEs and rely on a model-based separation of expected chemical shifts (85).

Metabolite-specific imaging methods use specialized RF pulses that are spatially and spectrally selective to excite individual metabolites which are then typically imaged with fast k-space trajectories such as echo planar imaging (EPI) or spirals (86).

Their Application To Different

organ systems is described below. The image reconstruction methods used in human [1-13C]pyruvate studies have typically been conventional methods (e.g. FFT, non-uniform FFT, or equivalent). The incorporation of accelerated imaging and advanced reconstruction methods including parallel imaging (4,57,87) and compressed sensing (7) has also been applied in human studies for improved spatial resolution, temporal resolution and coverage, but have the potential for additional artifacts as well as SNR losses due to ill-conditioning of the reconstruction (e.g. g-factor).

The Majority Of

published studies do not use accelerated imaging indicating the resolution and coverage achievable without acceleration is currently adequate for successful data collection. Performing coil combination, even with fully sampled data has also been shown to have specific challenges for HP human images: using naive sum-of-squares methods suffer from high noise amplification in the relatively low SNR regime of HP [1-13C]pyruvate (compared to 1H), motivating several HP 13C-specific methods that include data-driven coil sensitivity estimation which have shown obvious improvements over sum-of-squares (11).

More recently denoising techniques have been applied as post-processing of human HP data(41,42,44). The techniques applied are based on spatial-temporal singular value decomposition for unsupervised estimation of signal and noise components. They have shown improvements in apparent SNR in the brain and liver, while care must be taken to choose parameters such as the rank threshold to avoid oversmoothing and overfitting to the estimated signal components.

Prostate Studies

Prostate cancer was the first human application of HP [1-13C]pyruvate (1), and data was acquired with MRS/I methods: 1D dynamic MRS, single-slice 2D dynamic echo-planar spectroscopic imaging (EPSI), and single time point 3D EPSI. Advances in imaging strategies led to the development and application of new acquisition schemes, including undersampled 3D EPSI with compressed-sensing (7), model-based chemical shift encoding methods that use a priori information (47,59), and metabolite-specific EPI (10), all of which can provide volumetric whole-organ coverage and dynamic acquisitions.

The pyruvate bolus arrival in the prostate can vary by ± 10 s between patients, necessitating dynamic imaging to reliably and consistently capture the pyruvate bolus (18). For this reason, all currently ongoing studies acquire dynamic data. While MRS/I, chemical shift encoding, and metabolite-specific imaging can all achieve dynamic imaging, chemical shift encoding and metabolite-specific imaging provide greater dynamic and volumetric coverage (85). For scan prescriptions, the FOV is designed to provide full prostate coverage and typically to match the orientation of the anatomic imaging used for registration. Flip angles used in current studies are constant through time, as quantification with a variable-through-time flip scheme is highly sensitive to bolus timing (8) and errors in the RF transmit (B1 +) field (76).

Heart Studies

Data acquisition methods for 13C imaging in the heart must be designed to meet the demands of significant cardiac motion and blood flow. To cope with the periodic cardiac motion, most human heart studies to date used gating to the diastolic window, the longest cardiac cycle interval, which has reduced motion (2,22,28,30,35,36,38,45,52). The duration of the diastolic window limits the available data sampling time, making cardiac acquisitions the most time-constrained of the HP 13C MRI applications. The most common acquisition approach is metabolite-specific imaging with spiral k-space trajectories (2). Their single-shot imaging capability makes these methods particularly robust to motion effects. Furthermore, spiral k-space trajectories provide rapid k-space coverage and relatively benign flow and motion artifacts. The majority of studies have used 2D multi-slice acquisitions, but 3D encoding has also been used successfully (35).

Brain Studies

For HP 13C MRI of the human brain, the majority of studies have also used 2D (slice selective) acquisitions (10–12,14,16,28,33,40,41,44,51,53,60), with a trend toward volumetric coverage using 2D multi-slice metabolite-specific imaging. 3D metabolite-specific imaging of the whole brain, with phase encoding of the slice direction (34,57), has been shown to provide similar SNR efficiency (88) compared with multislice imaging. A number of studies have employed MRS/I (5,6,29,31–33,50,55) resulting in a spectrum from each voxel, which has the advantage of not requiring a priori information about which peaks to encode. This was important in early brain studies when it was not known which peaks would be detectable. Chemical shift encoding, using a set of images with different echo times and an iterative reconstruction of the individual resonances (i.e. the IDEAL approach (85)), has also been used (12,49,54), with the drawback that coverage in the slice direction was limited due to the time required to acquire multiple echo time images.

Abdomen And Breast Studies

The fundamental approaches to data acquisition and reconstruction in the abdomen and breast are largely similar to the aforementioned applications, but demand attention to particular challenges associated with these anatomic regions, especially relating to respiratory motion.

Although it has been shown that a basic 2D MRSI approach based on phase encoding and FID readout can be successfully applied for HP 13C imaging in breast (15) and kidney (13), major advantages in terms of spatiotemporal resolution and coverage have been realized using tailored approaches based on metabolite-specific imaging (43,62) and chemical shift encoding (43), which have facilitated multi-slice or 3D dynamic acquisitions over large FOVs in the abdomen (4,37,46).

The significant respiratory motion encountered in these regions can directly blur 13C images, and has further favored these rapid acquisition strategies. Motion also degrades B0 homogeneity, which can shift frequency-selective excitation profiles and introduce artifacts into rapid imaging readouts. This makes accurate determination of the acquisition center frequency and shimming essential in these regions which often cover large FOVs. (See “Prescan Calibration” section for more information). In some studies, breath-holding was used to minimize motion effects and enforce frame-to-frame data consistency (42). A pragmatic and reasonably effective approach for dealing with respiratory motion during 13C data acquisition is an initial breath-hold (as long as can be tolerated), followed by free-breathing (46,62).

1H Imaging

Collection of 1H imaging data is essential both for prescribing the 13C acquisition and for interpretation of the resulting 13C data. Multi-planar 1H scouts are acquired prior to 13C acquisition to enable graphical prescription of the 13C imaging region. All human HP 13C-pyruvate imaging studies acquire conventional MRI scans (e.g. T1- and T2-weighted volumes) for anatomic reference, aiming to cover at least the full 13C FOV. Acquiring these anatomic scans as close as possible to the time of 13C imaging (immediately before or after) minimizes potential misregistration between the data sets. Depending on the application, other advanced 1H sequences are also acquired (e.g. diffusion-weighted imaging for cancer imaging).

When contrast-enhanced data is acquired, it is done after 13C imaging, as paramagnetic contrast agents will accelerate 13C relaxation.

Reported Study Parameters

Figures 5 and 6, and Supporting Table S2 shows the reported acquisition study parameters for human HP [1-13C]pyruvate studies published as of September 2022. Figure 5 shows a mixture of MRS/I, metabolite-specific imaging, and chemical shift encoding methods have been successfully used, where spectroscopy-based methods have become less prevalent in recent studies. Figure 6 shows the acquisition timing, including the important start time and interval/temporal resolution, is quite variable across studies.

Figure 5: Acquisition methods used in published HP [1-13C]pyruvate human studies published up to September 2022, classified into: MR spectroscopy and spectroscopy imaging (MRS/I); chemical shift encoding methods, such as IDEAL, that use multiple TEs and model-based reconstructions; and metabolite-specific imaging methods that use spectrally-selective excitation to image a single resonance at a time.

Figure 6: Temporal acquisition characteristics reported in HP [1-13C]pyruvate human studies published up to September 2022. (a) Reported referencing of acquisition start times.

(B)

Acquisition start times reported when using dynamic imaging and when timing was reported relative to the end of the injection. (c) Temporal resolutions. “Not Applicable” indicates dynamic imaging was not used.

Summary

Three general categories of acquisition strategies have been used successfully for human HP 13C-pyruvate studies: MRS/I, model-based chemical shift encoding (e.g. IDEAL) methods, and metabolite-specific imaging methods. These have enabled successful studies in the prostate, heart, brain, abdomen, and breast. Recent studies increasingly have used the imaging-based strategies of metabolite-specific imaging and chemical shift encoding which are the fastest methods, although a heads-to–head comparison between techniques has not been performed.

Metabolite-specific imaging is quite popular because of its speed and compatibility with single-shot imaging, but is sensitive to B0 field variations and thus requires careful calibrations. Nearly all studies surveyed acquired data dynamically, allowing measurement of the bolus and metabolite kinetics. The exact timings and associated flip angles vary quite widely across reported studies, with no consensus yet as to how to choose these parameters. Image reconstruction is typically done directly using Fourier Transform methods, and accelerated imaging strategies are uncommon.

Data Analysis And Quantification

This section covers the analysis of data from human HP [1-13C]pyruvate studies, including modeling and metrics, visualization, as well as considerations for how to store data and metadata. Depending on study design, the analysis may need to give quantitative or semi-quantitative output reflecting a biological process or may just reflect a contrast between different regions of interest for quantitative evaluation.

Metrics

Figure 7: HP [1-13C]pyruvate raw data (A) have typically been quantified using four categories of metrics depending on the acquisition. Data acquired as a single time point are often quantified using normalized metabolite images or metabolite ratios (B). Dynamic data can be quantified using normalized metabolite images or metabolite ratios (B), or with metabolite timings such as time-to-peak (TTP) or pharmacokinetic (PK) models (C). The latter two require the data to be time-resolved. [1-13C]alanine and 13C-bicarbonate are analyzed similarly to [1-13C]lactate but omitted here for display.

Metabolite images are commonly used as summary metrics for HP MRI data, often including some form of normalization as well as summed over time as an area under the time curve (AUC) (17). These are analogous to the visual evaluation that is most used for routine clinical work (89,90). In these metabolite images, we expect that the [1-13C]pyruvate AUC signal is predominantly weighted towards perfusion and uptake, while [1-13C]lactate, [1-13C]alanine and 13C-bicarbonate AUCs represent metabolic conversion. The strength of this approach lies in its simplicity and relatively few underlying assumptions. Limitations to the use of single-metabolite images or AUCs include sensitivity to inhomogeneous coil profiles (57,87,91), the acquisition strategy and acquisition parameters, pyruvate polarization and concentration level, and signal relaxation rates (92). Further, the reader must be careful to interpret all the images in conjunction to better understand the underlying biology; for example, increased [1-13C]lactate in the presence of decreased [1-13C]pyruvate delivery can have a very different meaning compared to increased [1-13C]lactate with increased [1-13C]pyruvate delivery.

In an attempt to address variations in coil sensitivity, polarization level, and pyruvate delivery, AUC images are often computed by normalizing to a specified parameter, such as the maximum pyruvate or average lactate signals, or presented as a ratio such as lactate/pyruvate or divided by “total Carbon” - the sum total of HP 13C signal observed across all metabolites. The AUC ratios between metabolites and pyruvate are proportional to the corresponding forward kinetic rates (81,93), but are not directly comparable to rate constants when magnetization loss rates (e.g. relaxation and losses due to signal excitation) differ between studies. Similarly, the ratios between the produced metabolites (e.g. bicarbonate/lactate) can reflect the balance between downstream metabolic pathways (12,55). Care must be taken to consider how AUC images are calculated and normalized before comparing values between studies.

To further quantify the interpretation, pharmacokinetic (PK) modeling approaches were developed to compute the apparent kinetics of pyruvate-to-metabolite exchange (92,94–99). These yield semi-quantitative to quantitative apparent rate constants, given in s-1. Some models require a vascular input function, while others avoid this requirement (95). PK models can explicitly account for acquisition-specific details such as excitation angle and repetition time, and thus may reduce the effects of these details on quantification. An input-less model, provided in the Hyperpolarized-MRI-Toolbox (https://github.com/LarsonLab/hyperpolarized-mri-toolbox) (100) and thus frequently employed for human data, has been shown to fit well and robustly to prostate and brain data (8,20). PK models are quantitative in nature, arguably provide more relevant biological information (8,20), and appear to be reproducible across sites (51). However, rate constants derived from PK models are still apparent rates, and likely do not reflect a single biological characteristic.

Some additional considerations include whether complex or magnitude data is used, as the noise behaviors will impact the analysis differently. Additionally, cut-off thresholds or other criteria may be used to identify and avoid voxels with insufficient SNR before analysis to improve robustness (20,41).

Regardless of the analysis approach, the underlying biology is not always clearly represented by the data; instead, the metrics may be influenced by perfusion, barrier permeability, intercellular shuttles, enzyme activities, co-substrate concentrations, or combinations thereof, depending on the organ and disease of interest (19,43,94,101–103). This may be addressed by incorporating complementary information. As an example, HP 13C pyruvate data is influenced by perfusion, and thus addition of perfusion MRI could be important for interpretation (98,104,105).

All the methods outlined above have been explored in clinical studies, described in Supporting Table 3 and summarized in Figure 8. As of September 2022, approximately 52% of studies involving human subjects report rate constants derived from a PK model with a few different models reported. A nearly equal fraction (51%) of the studies report AUC ratio values.

Approximately 66% of these studies report metabolite-specific images or AUC values. About 40% report SNR values; this metric is particularly frequent in manuscripts that describe technical developments for clinical HP MRI. Approximately 16% of these studies summarize model-free metrics, and 10% report measurements from a single timepoint. Most studies report a combination of quantities.

Figure 8: Reported metrics used for analysis in HP [1-13C]pyruvate human studies published up to September 2022.

Visualization

A wide variety of approaches have been used for visualizing data from human HP 13C-MRI studies. The challenges and practical considerations are: 1) choosing the appropriate metrics to display, 2) how to encode the parameters (e.g. the colormap), and 3) choosing how to provide anatomical context and other multi-parametric data. The choice of visualization also depends on the goal which could be for diagnostic interpretation, but also quality control, reproducibility among readers and publication.

Metrics

The choice of HP 13C metrics is described in detail above. At this stage in HP 13C development where there is no standardized metric, often a combination of metabolite images and ratios or PK model parameters are shown.

Parameter Encoding

The mapping function chosen should provide an adequate, often quantitative, impression of the parameter mapped. There is a consensus in the visualization field that perceptually uniform maps are best suited to visualize continuous parameters, like the greyscale typically used by radiologists as well as other monochrome (black to blue) and color ranges (fire-type, rainbow-type) (106,107). Multi-color heatmaps have been the most frequently employed method for HP 13C data, while greyscale has infrequently been used but it ensures there is no coloring-based bias as well as facilitating later reuse (Fig. 9a). Among the color schemes employed in the clinical HP 13C literature, fire-type scheme seems to be the most common [similar to “Plasma” or “Inferno” in matplotlib.org]. Next most commonly employed is the rainbow-type scheme [similar to “Rainbow” in matplotlib.org].

Anatomical Context

HP MRI faces the challenge that it does not necessarily depict the anatomical features, similar to PET, and thus requires an anatomical reference. Most often, a grayscale anatomical image is overlaid with a HP colormap (Fig. 9c,d). This approach is very intuitive, but can skew perception as the grey-scale anatomical reference may affect the brightness of the HP data (e.g. signal in the skull). This bias does not occur when showing adjacent maps (Fig. 9a, b). Here, anatomical outlines may help to provide reference (Fig. 9b).

Related Journal Articles & DOI Links

Selected peer-reviewed publications relevant to 12 Lead ECG Acquisition. Click the DOI to access the full paper (may require institutional access).

Why Choose Us?

Bangalore guidance for robotics, Spectre and autonomous systems projects.