Face Recognition Robot

Abstract— Recent advances in generative modeling have enabled the generation of high-quality synthetic data that is applicable in a variety of domains, including face recognition.

Here, state-of-the-art generative models typically rely on condi- tioning and fine-tuning of powerful pretrained diffusion models to facilitate the synthesis of realistic images of a desired identity.

Yet, these models often do not consider the identity of subjects during training, leading to poor consistency between generated and intended identities. In contrast, methods that employ identity-based training objectives tend to overfit on various aspects of the identity, and in turn, lower the diversity of images that can be generated. To address these issues, we present in this paper a novel generative diffusion-based framework, called ID-Booth. ID-Booth consists of a denoising network responsible for data generation, a variational auto-encoder for mapping images to and from a lower-dimensional latent space and a text encoder that allows for prompt-based control over the genera- tion procedure. The framework utilizes a novel triplet identity training objective and enables identity-consistent image gener- ation while retaining the synthesis capabilities of pretrained diffusion models. Experiments with a state-of-the-art latent diffusion model and diverse prompts reveal that our method facilitates better intra-identity consistency and inter-identity separability than competing methods, while achieving higher image diversity. In turn, the produced data allows for effective augmentation of small-scale datasets and training of better- performing recognition models in a privacy-preserving manner.

The source code for the ID-Booth framework is publicly available at https://github.com/dariant/ID-Booth.

Ntroduction

Deep learning models are nowadays utilized as backbones in a variety of recognition systems . These models typi- cally require sufficiently large and diverse training datasets to achieve competitive performance. However, obtaining suit- able datasets can be difficult in the field of biometrics, due to copyright, consent, and privacy issues , . With recent advancements in generative models, researchers are increasingly exploring the use of synthetic data to address the data needs of contemporary deep learning models. This is especially true in face recognition, where synthetic data may be used to train models or augment existing datasets by enriching the variation present in real-world data .

State-of-the-art generative models are currently dominated by diffusion-based techniques, which offer unparalleled syn- thesis capabilities in terms of quality and diversity of the Supported in parts by the Slovenian Research and Innovation Agency (ARIS) through Research Programmes P2-0250 (B) ”Metrology and Bio- metric Systems” and P2–0214 (A) “Computer Vision”, the ARIS Project J2-50065 ”DeepFake DAD”, and the ARIS Young Researcher Programme.

-Booth Samples

Fig. 1. Samples generated with the proposed ID-Booth framework. The framework enables fine-tuning of pretrained diffusion models for generating diverse identity-consistent face images based on images gathered in a constrained setting with the consent of subjects.

generated data, while enabling synthesis guided by text prompts . Recently, diffusion models have also been utilized to produce datasets suitable for face recognition tasks, i.e., containing images of multiple identities with multiple samples each. To this end, approaches rely on

Identity-Conditioning , , And Fine-Tuning Of

pretrained diffusion models. Nevertheless, most solutions fo- cus mainly on image reconstruction during training, resulting in poor consistency between the desired and the generated identities. To address this issue, PortraitBooth proposed an additional identity-based training objective, which can be used to extend the fine-tuning DreamBooth method.

However, the solution only considers the identity similarity of input samples and the generated samples during training. In turn, it tends to overfit on input identity features, including undesired characteristics, e.g., the pose, age, hair, acces- sories, thereby reducing the diversity of generated images.

In this paper, we present a solution for the outlined issues, in the form of a new generative framework, called ID-Booth. The proposed framework entails three main components, in- cluding (i) a denoising network that produces data based on input noise, (ii) a Variational Auto-Encoder (VAE) that maps images to and from a more efficient latent space on which the denoising network operates, and (iii) a text encoder that enables prompt-based conditioning of the denoising network. The proposed framework utilizes a novel triplet identity objective, which considers both positive and negative identity samples during training, to facilitate the generation

Arxiv:2504.07392V6 [Cs.Cv] 13 Oct 2025

of identity-consistent images while retaining the synthesis capabilities of pretrained models. Throughout the experi- ments, we explore the suitability of ID-Booth for addressing privacy concerns by generating diverse synthetic in-the-wild images of identities from the Tufts Face Database , which contains images gathered in a constrained setting with subject consent, as shown in Figure 1. We perform fine- tuning of a state-of-the-art diffusion model conditioned on diverse prompts and compare synthesis results with Dream- Booth and a PortraitBooth-based version of it in terms of image quality, fidelity and diversity as well as intra- identity consistency and inter-identity separability. Further- more, we investigate the real-world utility of the produced synthetic samples for augmenting existing datasets to train modern face recognition models in a privacy-preserving man- ner. We demonstrate that our fine-tuning framework enables the generation of more diverse synthetic samples with better intra-identity consistency and inter-identity separability. As showcased by improved recognition performance, across five real-world verification benchmarks, this makes our approach more suitable for augmenting small-scale training datasets than existing solutions , . Overall, the paper makes

For

generating highly-diverse identity-consistent privacy- preserving face images. • We propose a novel triplet identity learning objective for fine-tuning that improves identity consistency while retaining better image diversity.

• We demonstrate the suitability of the produced data for augmenting existing small-scale datasets and show that training with the mixed images leads to better performing face recognition models.

Related Work

Image generation. The field of image synthesis has under- gone rapid development since the introduction of deep gener- ative models. Generative Adversarial Networks (GANs) were the initial models to achieve the synthesis of convincing images, with a generator and a discriminator network. Exten- sive improvements followed, namely StyleGAN facili- tated higher image quality and better control over the gener- ation process. However, the synthesis capabilities of GANs have nowadays been surpassed by recent diffusion mod- els , which generate images by gradually removing noise from initial noisy samples. This denoising process is learned with a convolutional encoder-decoder by predicting the noise that is added to training samples at different scales .

Recently, Latent Diffusion Models (LDMs) achieved improved efficiency and efficacy by moving the denoising process from the pixel space to a lower-dimensionality latent space of a pretrained variational autoencoder. Their remarkable synthesis capabilities and conditioning on text prompts via a pretrained text encoder have led to their broad adoption, namely of the open-source Stable Diffusion

(Sd) Model . Image Resolution Of These Models Has

been further improved by utilizing a larger U-Net backbone along with two text encoders and additional conditioning schemes . Recent approaches have also enhanced control over the generation process, e.g., ControlNet conditions the model on segmentation masks or depth maps via an auxiliary trainable copy of the model, while IP-Adapter utilizes image features as a condition through a decoupled cross-attention mechanism. Fine-tuning approaches have also been developed to incorporate new concepts into pretrained diffusion models by training on a minimal set of images .

Generating synthetic face recognition data. Generative models and synthetic data hold considerable potential in face recognition by enabling the creation of large-scale (training and test) datasets with predefined characteristics, facilitating augmentation in data-scarce application scenarios, and bal- ancing data across different demographics . To enable con- trol over various characteristics of generated faces, Deng et al. conditioned StyleGAN on input 3D face priors.

However, recognition models trained on the generated data achieved worse performance than those trained on real-world data. To tackle this issue, Qiu et al. introduced identity and domain mixup of synthetic and real data during training.

Boutros et al. proposed to condition StyleGAN2 on one-hot encoded identity labels. This improved intra-identity diversity at the cost of lowered inter-identity separability and a limited amount of possible identities. To address this, Tomaˇsevi´c et al. instead utilized identity features from a pretrained face recognition model as the condition, in addition to enabling the generation of multispectral data.

Recently, Boutros et al. achieved the generation of identity-specific images with latent diffusion models by con- ditioning the denoising network on face recognition features.

The proposed contextual partial dropout also prevented over- fitting on identities and enabled control over inter-identity separability and intra-identity diversity. Differently, more recent approaches relied on pretrained diffusion models rather than training the models from scratch. Ruiz et al.

presented the DreamBooth method that can associate a new identity to a rare text token through fine-tuning on images of the identity. During training, face images generated by the pretrained model are also used to preserve prior synthesis capabilities. Arc2Face instead replaces the identity token with recognition features and fine-tunes the model on a large- scale dataset. The textual-part of the prompt is frozen, so that control is tied primarily to the identity features, thus enabling more consistent generation of input identities. However, this comes at the cost of losing powerful prompt-based control. The recent IP-Adapter has also been modified to use identity features as the condition, while retaining control of text prompts through decoupled cross-attention.

InstantID extends these capabilities by incorporating spatial control with an auxiliary ControlNet-based mod- ule conditioned on facial landmarks and features. Despite advancements, identity consistency remained problematic, as the identity aspect was not considered in training objectives.

To address this, Peng et al. introduced PortraitBooth, a method that incorporates an identity-based objective into the training process, which can also be applied to the fine- tuning of DreamBooth . However, the solution only relies on the identity similarity of training images and generated noisy images, despite the success of more refined objectives on face recognition tasks . As a result, the approach can overfit even on undesired characteristics of training identities, e.g., their pose, age or face accessories, which lowers the diversity of produced images. In contrast, our proposed ID- Booth framework utilizes a triplet objective that relies on the identity similarity between generated images and both training images (i.e., positive samples) and prior images produced by the initial model (i.e., negative samples). This enables better identity consistency, while better retaining synthesis capabilities of pretrained latent diffusion models.

Ethodology

In this section we present the inner workings of ID-Booth, a framework for generating diverse high-fidelity identity- consistent facial images suitable for augmenting small-scale datasets captured with the consent of subjects.

A. The Id-Booth Framework

The proposed diffusion-based ID-Booth framework con- sists of three primary components, as depicted in Figure 2. This includes (i) the denoising network, responsible for enabling data generation through diffusion, (ii) the Varia- tional Auto-Encoder (VAE) that maps images to and from a more efficient latent space, and (iii) the text encoder, which enables prompt-based control of the generation process.

Fine-tuning of pretrained diffusion models is then achieved with three training objectives, (i) the conventional recon- struction loss on a small set of input samples, (ii) the prior preservation loss, focused on combating overfitting via the reconstruction of images generated before fine-tuning, and (iii) the triplet identity loss, which utilizes a pretrained face recognition model to guide the diffusion model toward better similarity between generated and target identities rather than random identities from prior images. Details of each component and objective are provided below.

Denoising network. At the core of the diffusion model lies the denoising network, which is trained to reverse a noising process q that gradually degrades training images by adding noise at different scales. This entails the corruption of a sample x0 from the real data distribution p(x0) into its noised versions x1, . . , xT through a Markov chain of length T, as:

(1)

for timesteps t = 1, . . , T, where α1, . , αT represent a fixed variance schedule. However, any step of the noised sample can also be efficiently produced using a closed-form expression directly from the input x0 as follows:

S=1 Αs, Which Enables Uniform Sampling Of

t. Through training, the denoising network (i.e., typically a U-Net network ), learns to estimate the real data dis- tribution from a noise-filled standard Gaussian distribution.

This entails gradually denoising a noisy image xT ∼N(0, I) to less noisy samples xt until a denoised data sample x0 is reached. To this end, the denoising network ϵθ(xt, t) predicts the noise ϵ that is added at step t with Equation (2).

Variational Auto-Encoder (VAE). To greatly improve ef- ficiency, the noising and denoising processes are carried out in the latent space of a pretrained Variational Auto- Encoder (VAE) instead of the pixel space . This is achieved by first mapping the input sample x0 to the latent input z0 through the encoder model E of the VAE. Noising with Equation (2) is then performed to obtain noised samples zt on which the denoising network ϵθ is trained. During in- ference, synthetic images can then be generated by randomly sampling a noisy sample zT in the latent space, denoising it with the predictor ϵθ, and then mapping the denoised sample z0 back to the pixel space with the VAE decoder D.

Text encoder. To enable control over the generation process, the denoising network is also conditioned on input text prompts . The text prompt is first tokenized and mapped to corresponding token embeddings, which are then en- coded through a pretrained CLIP text encoder . Encoded prompts c are then passed as conditions to the denoising network through the cross-attention mechanism .

B. Training Objectives Of Id-Booth

Pretrained diffusion models provide unparalleled text- guided synthesis capabilities, owing to training on various datasets of unprecedented scale . However, their knowl- edge of very specific concepts and styles remains limited.

This is also true for their ability to create images of a desired identity as prompting for a specific non-celebrity identity can be difficult or even impossible.

To facilitate the generation of identity-specific images, we propose to fine-tune a pretrained diffusion model on a small set of input images of a desired identity. Our proposed ID-Booth framework utilizes three separate training objectives, to improve identity consistency while retaining the synthesis capabilities of pretrained models. This includes the reconstruction loss LREC, the prior preservation loss LP R and a triplet-identity loss LT ID, which are combined

(3)

as illustrated in Figure 2. Here, the balancing weight λP R is

T )2 To Reduce

the influence of identity supervision at higher timesteps, as image blurriness increases. The training objectives are described in detail below.

To further retain the capabilities of pretrained models, while still enabling fine-tuning on new identities, our ID- Booth framework also relies on the use of the Low-Rank

Adaptation Method (Lora) . Thus, Instead Of Fine-

tuning the entire diffusion model, all existing weights remain frozen while new low-rank trainable layers are introduced in the denoising network. This allows for better retention of synthesis capabilities, while enabling faster training and more efficient storage of fine-tuned model weights.

Sample

Fig. 2. Overview of the ID-Booth framework. The framework utilizes three training objectives to fine-tune a pretrained diffusion model. LREC and LP R are aimed at the reconstruction of training and prior images. Differently, the proposed triplet identity objective LT ID focuses on the identity similarity between generated samples and both training and prior samples to improve identity consistency without impacting the capabilities of the pretrained model.

Reconstruction loss. The first training objective of our ID- Booth framework is aimed at image reconstruction and is based on the reweighted optimization objective convention- ally used for training diffusion models . Since denoising is performed in the latent space of a pretrained VAE, the loss is based on the noise ϵ that is added to sample z0 at timestep t and the noise that is estimated by the denoising network ϵθ considering the noisy latent sample zt, the timestep t, and the text prompt condition c. Formally, this reconstruction loss

(4)

Prior preservation loss. Fine-tuning a diffusion model on a small set of images with only the reconstruction objective LREC often leads to overfitting on input data and the loss of prior knowledge, e.g., the concept of what a person is. To address this, our ID-Booth framework utilizes an additional training objective aimed at the preservation of prior concepts . To this end, a set of prior images xpr,0 are generated by the initial pretrained model prior to training, with prompts related to the novel concept to be introduced.

Following the initial reconstruction objective, these prior samples are used to form the prior preservation loss LP R

(5)

where the pr notation represents factors related to prior images generated with the initial model. Triplet identity loss. Despite the suitability of LREC and LP R for fine-tuning, both objectives are focused solely on image reconstruction and do not target the consistency of generated identities. This is the case for both consistency with desired input identities and consistency among gener- ated samples. To address this, we propose to incorporate the identity aspect into the training process through the similarity of identity features extracted from images with a pretrained face recognition model. However, to enable the inspection of generated identities during training, suitable face images must be produced at each training step. To this end, we use the predicted noise ϵθ(zt, t, c) and the latent noisy sample zt

(6)

Afterward, we can decode the estimated denoised latent ˆz0 to the estimated input image with ˆx0 = D(ˆz0). Next, we extract the facial region with a face detection model for both the estimated and the input training image, denoted as ˆxf

And Xf

respectively. If the facial region exist, we obtain the identity feature representations of each image with a pretrained face recognition model φ, otherwise the objective is skipped.

To guide the generative model toward better identity consistency, we propose to form a triplet identity objective. The objective utilizes identity features of the reconstructed sample ˆx0 as the anchor, the input image x0 as a positive example of an identity and prior images xpr,0 as a negative example. Formally, our proposed triplet identity objective LT ID can be defined using cosine similarity cos as follows:

(7)

where the notations introduced before apply. In addition, m represents a non-negative margin, i.e., the minimum differ- ence between positive and negative similarities that is re- quired for the loss to be zero. The proposed triplet-objectives also addresses the risk of overfitting on unintentional char- acteristics of training samples, e.g., the pose, age, hair or accessories, which might leak into the identity embeddings.

This is achieved through negative identity examples, which often share similar characteristics with positive examples.

Experiments And Results

Dataset preparation. To fine-tune ID-Booth, we utilize the

Tufts Face Database (Tfd) , Which Contains Images

of subjects. In total, the dataset includes over 10, 000 images of 113 human subjects captured across various light spectra. We focus on images captured with four visible field cameras under constant diffused light in a semi-circle around the subjects. During preprocessing, we remove heavily blurred images and extreme side-profile images lacking key facial features (e.g., two eyes), then crop them to focus on the face region, resulting in 2299 images of 107 subjects. Next, we use eye landmarks, detected with the Multi-Task Cascaded Convolutional Neural Network (MTCNN) to align the faces through an affine transform, and then resize the images to 512×512. For evaluation we rely on the Flickr Faces High- Quality (FFHQ) dataset of 70, 000 diverse in-the-wild unlabeled face images, which we also resize to 512 × 512.

Implementation details. We evaluate the suitability of our framework on the state-of-the-art diffusion model Stable Diffusion 2.1 (SD-2.1) , which is capable of generating high-quality and diverse 512×512 images through 1000 de- noising timesteps, specified by the discrete denoising sched- uler with βstart = 8.5 × 10−4 and βend = 0.012 . We fine-tune the SD-2.1 model on images of each identity in the Tufts Face Database (TFD) . To this end, we utilize the training objectives specified by either DreamBooth , fo- cused primarily on image reconstruction, PortraitBooth , which includes a simple two-point identity objective, or by our proposed ID-Booth framework, that balances identity consistency and image diversity. The identity objectives utilize features extracted with a pretrained ArcFace-based ResNet-100 recognition model from face regions detected

With Mtcnn . The Detection Of Faces Also Acts As

the decision factor for when identity-based objectives are applied. To minimize the effect on the synthesis capabilities of the pretrained model, we utilize the Low-Rank Adaptation (LoRA) method, which freezes the diffusion model but introduces new trainable layers instead. Specifically, we add two linear layers of rank 4 to each cross-attention block, initialized with a Gaussian distribution. We also generate 200 images with the initial SD-2.1 model and the prompt photo of a person, which are used for preservation of prior concepts through LP R . We then perform fine- tuning with images of a desired identity and the prompt photo of [ID] person, where [ID] represents a rare text token that will be tied to the new identity, in our case sks . We utilize an initial learning rate of 10−4 and

The Adamw Optimizer With Β1 = 0.9, Β2 = 0.999,

ϵ = 10−8 and a weight decay of 0.01, along with the half- precision floating point format to lower VRAM usage. Fine- tuning is stopped after 32 epochs (i.e., 6400 steps), based on our initial observations and existing works , .

Data generation. Each fine-tuned SD model is used to generate two synthetic datasets, one with 21 images per identity, as is the case in TFD , and one with 100 images per identity to investigate the scalability of our approach. Data generation is performed with a guidance scale of 5.0 and 30 inference denoising steps with the same discrete denoising scheduler as during training. The goal is to generate diverse synthetic images of desired identities under various scenarios. To produce images that resemble real-world in-the-wild datasets , we utilize a prompt that defines a face image of a specific identity as well as the

Background

Here [ID] represents the identity token, while [G] defines the gender of the person, i.e., female or male. To generate diverse images we also select the environment through [B],

Club

To also produce a variety of poses we randomly select whether the image is a portrait or a side-portrait, represented by [P]. In addition, we rely on the following

Landscape

An ablation study of the main prompt components is avail- able in the supplementary material. Evaluation methodology. We evaluate the suitability of the proposed ID-Booth framework by comparing its synthesis capabilities to those of DreamBooth and a version of it extended with the PortraitBooth identity objective. Other diffusion-based frameworks that produce identity-specific

Images, E.G., Arc2Face And Instantid , Are Not

considered as they are trained on large-scale web-scraped face recognition datasets, without the consent of subjects. Meanwhile, our experiments entail fine-tuning on a limited

Amount Of Images From Tfd Gathered With Suitable

consent. To evaluate the produced images, we compare them to the diverse real-world images of FFHQ . Here, we consider either entire images or only the face regions of a resolution 112 × 112, aligned and cropped based on face landmarks detected by MTCNN . The quality of images is then determined with Fr´echet Distance and Kernel Distance , measured on features extracted with the pre- trained DINOv2-ViT-L/14 model rather than the typical Inception-v3 model, which has been shown to be un- suitable, due to poor correlation with human evaluators and the limitations of the ImageNet dataset . We also evaluate the fidelity and diversity of images separately, with the use of Density and Coverage , measured on features

Of Dinov2-Vit-L/14 . To Compute These Scores We

utilize the generated datasets with 100 samples per identity

And Compare Them To 10.000 Samples From Ffhq . In

addition, we analyze the intra-identity diversity of samples via the extracted features with the Vendi score , which differently from previous measures does not require a refer- ence dataset. Similarly, we rely on the Certainty Ratio Face Image Quality Assessment (CR-FIQA) to evaluate the quality of each face image through relative classifiability with a pretrained ResNet-101 . We also analyze intra-identity diversity by evaluating the pitch, yaw and roll of faces in the images with the 6DRepNet head pose estimator.

Recognition experiment details. As part of our exper- iments, we also investigate intra-identity consistency and inter-identity separability based on genuine and imposter distributions. These are formed using the cosine similarity of identity features of synthetic samples and either samples of the corresponding identity (genuine pair) or a different identity (imposter pair), from either TFD or the syn- thetic dataset. The identity features are extracted with a Resnet-101 recognition model trained with the ArcFace

Of The Images Detected With Mtcnn . To Allow For

a fair comparison with samples of TFD , we form the distributions with synthetic datasets that also consist of 21 samples per identity. For each dataset combination, we form all possible genuine pairs along with an equal amount of

Table I

QUANTITATIVE EVALUATION OF QUALITY, FIDELITY AND DIVERSITY OF SYNTHETIC IMAGES. QUALITY IS ASSESSED WITH FR´ECHET DISTANCE AND KERNEL DISTANCE , WHILE FIDELITY AND DIVERSITY ARE MEASURED THROUGH DENSITY AND COVERAGE . RESULTS ARE COMPUTED BY COMPARING DISTRIBUTIONS OF FEATURES EXTRACTED WITH DINOV2-VIT-L/14 FROM SYNTHETIC IMAGES AND REAL-WORLD IMAGES OF FFHQ , CONSIDERING EITHER ENTIRE IMAGES OR ONLY THE FACE REGION. VENDI SCORE IS USED TO EVALUATE INTRA-IDENTITY DIVERSITY, WHILE CR-FIQA MEASURES FACE IMAGE QUALITY OF EACH SAMPLE, BOTH WITHOUT A REFERENCE DATASET.

± 0.181

(↓/ ↑) – Lower / Higher is better; (Bold) – Best result; (Underline) – Second best result

± 2.920

(↓/ ↑) – Lower / Higher is better; (Bold) – Best result; (Underline) – Second best result randomly sampled imposter pairs. We report the mean and standard deviation of distributions along with established metrics, including Equal Error Rate (EER), False Match Rate at a False Non-Match Rate of 1.0% (FMR100) or 0.01% (FMR1000), False Non-Match Rate at a False Match Rate of 1.0% (FNMR100) or 0.01% (FNMR1000), and the Fisher Discriminant Ratio (FDR) . Lastly, we use the produced data to augment the TFD dataset, which is then used to train a ResNet-50 recognition model with the AdaFace loss . For training we utilize a batch size of 128 and the Stochastic Gradient Descent (SGD) optimizer with 0.9 momentum, a weight decay of 5×10−4, and a dropout ratio of 0.4. The learning rate is initially set to 0.1 and is lowered by a factor of 10 after the 22nd, the 30th, and the 35th epoch. Training is stopped once no improvement in 5 epochs is observed on the LFW benchmark. The performance of the trained model is then evaluated on five state-of-the- art verification benchmarks, including Labeled Faces in the Wild (LFW) , its Cross-Age and Cross-Pose subsets CA-

Fw And Cp-Lfw , Celebrities In Frontal-Profile

in the Wild (CFP-FP) and AgeDB-30 . Experimental hardware. The experiments were conducted on a cluster of 4 Nvidia A100 SXM4 40GB GPUs and a Desktop PC with an Nvidia RTX 4090 GPU.

A. Evaluation Of Generated Images

Image quality. We begin our evaluation by assessing the overall quality of images produced by either the proposed ID-Booth framework or its two competitors, the base Dream- Booth and a version of it extended with the Portrait-

-Booth

Fig. 3. Comparison of generated image samples. ID-Booth facilitates better identity consistency than DreamBooth and better image diversity than when utilizing the PortraitBooth identity objective, which can limit the variety of facial features and poses.

Booth identity objective (denoted as PortraitBooth for brevity). To this end, we utilize the Fr´echet Distance and Kernel Distance computed between features extracted

With Dinov2-Vit-L/14 From Synthetic Images And

features extracted from real-world images of the FFHQ dataset. From results reported in Table I and samples in Figures 1 and 3, we can discern that with the SD-2.1 model and our defined prompts we can generate images that better match the quality of in-the-wild FFHQ images than the real-world images of TFD , which were gathered in a constrained environment. Comparing results of the different fine-tuning methods, we see that our ID-Booth framework achieves the best quality results, scoring closest to the non fine-tuned model, while enabling identity-specific generation.

This is the case both when evaluating entire images or only the cropped face regions. Interestingly, both identity-based training objectives of either PortraitBooth or ID-Booth improve the image quality of the base DreamBooth .

In addition, we evaluate the quality of each face region with CR-FIQA . Here, however, high face quality is not necessarily as desired as having a mix of high and low quality images, which can lead to better performing recognition models. This difference can also be observed on real-world datasets, where in-the-wild images of FFHQ achieve a lower mean but higher standard deviation than constrained images of TFD . Similarly, compared to

Reambooth And The Portraitbooth-Based Version,

our ID-Booth framework achieves a higher standard devia- tion but lower mean of CR-FIQA scores, that are closer to those of diverse samples of the non fine-tuned models.

Image fidelity and diversity. Next, we analyze the produced images in terms of fidelity, i.e., the degree to which they resemble real samples, and diversity, i.e., how well they cover the variability of real samples . To this end, we rely on Density and Coverage , respectively, reported in Table I, in addition to qualitative samples in Figures 3 and 4. Compar- ing different datasets with FFHQ , we can observe that images of TFD lack the fidelity and diversity expected of in-the-wild images. Synthetic samples produced by the non-finetuned SD-2.1 model with diverse prompts offer a notable improvement in these areas. In comparison, fine- tuning approaches achieve a higher fidelity of entire images and face regions, but crucially result in lower diversity of face regions. Among the approaches, DreamBooth scores the highest in terms of density (i.e., fidelity), while ID-Booth achieves the highest coverage (i.e., diversity) on both entire images and detected face regions. This can also be observed in Figures 3 and 4, where ID-Booth offers more consistent identities, while enabling a larger variety of poses, ages, accessories and other facial features than DreamBooth or the PortraitBooth-based version of it.

Intra-identity diversity. To further investigate the produced images, we also analyze the intra-diversity of samples with the per-class Vendi score . As reported in Table I, synthetic images generated by all fine-tuning methods offer more intra-identity diversity than the constrained samples of TFD . Our ID-Booth achieves the largest intra-identity diversity among the fine-tuning approaches, both of entire images and only face regions, while ensuring better identity consistency. This can be seen in Figure 4, where ID-Booth samples of the same identity contain a larger variety of poses and face accessories. To obtain deeper insight, we also analyze the pitch, yaw and roll of faces with the 6DRepNet head pose estimator. In Table II we report the mean and standard deviation of standard deviation values obtained from pose distributions of each identity. Results reveal that ID-Booth generates samples with the largest variety of poses per identity, especially in terms of pitch and roll, which represents an important aspect of overall intra- identity diversity. In comparison, DreamBooth and the PortraitBooth-based version often default to more front- facing poses, as seen in Figures 3 and 4.

B. Recognition-Based Experiments

Identity consistency and separability. To determine the suitability of generated images for augmenting recognition datasets we must also examine the consistency and sepa- rability of identities in the images. To this end, we form genuine and imposter distributions either only among syn- thetic identities or between synthetic and real-world iden- tities, based on the similarity of features extracted with the pretrained ArcFace-based recognition model . From

Samples Of An Identity

Fig. 4. Comparison of identity consistency. ID-Booth achieves better identity consistency than DreamBooth , while retaining more diverse synthesis capabilities and ensuring better intra-identity diversity than Por- traitBooth . Reported is the cosine similarity of synthetic and real iden- tity features extracted with the pretrained ArcFace recognition model .

verification scores in Table III that describe these distribu- tions, we can discern that the use of identity-based training objectives from either ID-Booth or PortraitBooth re- sults in better identity consistency and separability than the base DreamBooth . This is the case both either among synthetic or between synthetic and real identities across all verification measures. The only exception is the FNMR1000 score in the latter scenario, which can likely be attributed to a handful of outliers, especially when considering the lower FNMR100 scores and improved FDR values. Compared to the PortraitBooth-based approach, ID-Booth achieves lower FNMR scores among synthetic identities along with lower FMR and FNMR scores between synthetic and real identities, indicating fewer outliers. In combination with a higher FDR score in the second scenario, this signifies better intra-identity consistency and inter-identity separability. ID- Booth samples in Figure 4 further support these observa- tions with better consistency between generated and real identities or among different samples of the same identity.

Overall, these results highlight crucial characteristics of ID- Booth, demonstrating its suitability for augmenting exist- ing datasets in a privacy-preserving manner by producing identity-consistent in-the-wild images of real-world identities from the training dataset collected with subject consent.

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

Power [Kw]

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

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.