Facial Anonymization and Privacy Concerns in Total-Body PET/CT

Computer vision-based personal identification using 2D maximum intensity projection CT images

OBJECTIVES

Computer vision (CV) mimics human vision, enabling the automatic comparison of radiological images from recent examinations with a vast image database for unique identification. This method offers significant potential in emergencies involving unknown individuals. This study assesses whether maximum intensity projection (MIP) images from thoracic computed tomography (CT) examinations are suitable for automated CV-based personal identification.

METHODS

The study analyzed 12,465 native CT examinations of the thorax from 8177 individuals, focusing on MIP images to assess their potential for CV-based personal identification in 300 cases. CV automatically identifies and describes features in images, which are then matched to reference images. The number of matching points was used as an indicator of identification accuracy.

RESULTS

The identification rate was 98.67% (296/300) at rank 1 and 99.67% (299/300) at rank 10, among over 8177 potential identities. Matching points were higher for images of the same individual (7.43 ± 5.83%) compared to different individuals (0.16 ± 0.14%), with 100% representing the maximum possible matching points. Reliable matching points were mainly found in the thoracic skeleton, sternum, and spine. Challenges arose when the patient was curved on the table or when medical equipment was present in the image.

CONCLUSION

Unambiguous identification based on MIP images from thoracic CT examinations is highly reliable, even for large CV databases. This method is applicable to various 2D reconstructions, provided anatomical structures are comparably represented. Radiology offers extensive reference images for CV databases, enhancing automated personal identification in emergencies.

KEY POINTS

Question Computer vision-based personal identification holds great potential, but it remains unclear whether maximum intensity projection images from thoracic-CT scans are suitable for this purpose. Findings Maximum intensity projection images of the thorax are highly individual, with computer vision-based identification achieving nearly 100% rank-1 accuracy across a potential 8177 identities. Clinical relevance Radiology holds a vast collection of reference images for a computer vision database, enabling automated personal identification in emergency examinations. This improves patient care and communication with relatives by providing access to medical history.

Total Body PET/CT: Future Aspects

Total-body (TB) positron emission tomography (PET) scanners are classified by their axial field of view (FOV). Long axial field of view (LAFOV) PET scanners can capture images from eyes to thighs in a one-bed position, covering all major organs with an axial FOV of about 100 cm. However, they often miss essential areas like distal lower extremities, limiting their use beyond oncology.TB-PET is reserved for scanners with a FOV of 180 cm or longer, allowing coverage of most of the body. LAFOV PET technology emerged about 40 years ago but gained traction recently due to advancements in data acquisition and cost. Early research highlighted its benefits, leading to the first FDA-cleared TB-PET/CT device in 2019 at UC Davis. Since then, various LAFOV scanners with enhanced capabilities have been developed, improving image quality, reducing acquisition times, and allowing for dynamic imaging. The uEXPLORER, the first LAFOV scanner, has a 194 cm active PET AFOV, far exceeding traditional scanners. The Panorama GS and others have followed suit in optimizing FOVs. Despite slow adoption due to the COVID pandemic and costs, over 50 LAFOV scanners are now in use globally. This review explores the future of LAFOV technology based on recent literature and experiences, covering its clinical applications, implications for radiation oncology, challenges in managing PET data, and expectations for technological advancements.

Analyzing the TotalSegmentator for facial feature removal in head CT scans

BACKGROUND

Facial recognition technology in medical imaging, particularly with head scans, poses privacy risks due to identifiable facial features. This study evaluates the use of facial recognition software in identifying facial features from head CT scans and explores a defacing pipeline using TotalSegmentator to reduce re-identification risks while preserving data integrity for research.

METHODS

1404 high-quality renderings from the UCLH EIT Stroke dataset, both with and without defacing were analysed. The performance of defacing with the face mask created by TotalSegmentator was compared to a state-of-the-art CT defacing algorithm. Face detection was performed using deep learning models. The cosine similarity between facial embeddings for intra- and inter-patient images was compared. A Support Vector Machine was trained on cosine similarity values to assess defacing performance, determining if two renderings came from the same patient. This analysis was conducted on defaced and non-defaced images using 5-fold cross-validation.

RESULTS

Faces were detected in 76.5 % of non-defaced images. Intra-patient images exhibited a median cosine similarity of 0.65 (IQR: 0.47-0.80), compared to 0.50 (IQR: 0.39-0.62) for inter-patient images. A binary classifier performed moderately on non-defaced images, achieving a ROC-AUC of 0.69 (SD = 0.01) and an accuracy of 0.65 (SD = 0.01) in distinguishing whether a scan belonged to the same or a different individual. Following defacing, performance declined markedly. Defacing with the TotalSegmentator decreased the ROC-AUC to 0.55 (SD = 0.02) and the accuracy to 0.56 (SD = 0.01), whereas the CTA-DEFACE algorithm brought the performance down to a ROC-AUC of 0.60 (SD = 0.02) and an accuracy of 0.59 (SD = 0.01). These results demonstrate the effectiveness of defacing algorithms in mitigating re-identification risks, with the TotalSegmentator providing slightly superior privacy protection.

CONCLUSION

Facial recognition software can identify facial features from partial and complete head CT scan renderings. However, using the TotalSegmentator to deface images reduces re-identification risks to a near-chance level. We offer code to implement this privacy-preserving pipeline.

IMPLICATIONS FOR PRACTICE

Utilizing the TotalSegmentator framework, the proposed pipeline efficiently removes facial features from CT images, making it ideal for multi-site research and data sharing. It is a useful tool for radiographers and radiologists who must comply with medico-legal requirements necessitating the removal of facial features.

Deep learning-based defacing tool for CT angiography: CTA-DEFACE

The growing use of artificial neural network (ANN) tools for computed tomography angiography (CTA) data analysis underscores the necessity for elevated data protection measures. We aimed to establish an automated defacing pipeline for CTA data. In this retrospective study, CTA data from multi-institutional cohorts were utilized to annotate facemasks (n = 100) and train an ANN model, subsequently tested on an external institution's dataset (n = 50) and compared to a publicly available defacing algorithm. Face detection (MTCNN) and verification (FaceNet) networks were applied to measure the similarity between the original and defaced CTA images. Dice similarity coefficient (DSC), face detection probability, and face similarity measures were calculated to evaluate model performance. The CTA-DEFACE model effectively segmented soft face tissue in CTA data achieving a DSC of 0.94 ± 0.02 (mean ± standard deviation) on the test set. Our model was benchmarked against a publicly available defacing algorithm. After applying face detection and verification networks, our model showed substantially reduced face detection probability (p < 0.001) and similarity to the original CTA image (p < 0.001). The CTA-DEFACE model enabled robust and precise defacing of CTA data. The trained network is publicly accessible at www.github.com/neuroAI-HD/CTA-DEFACE . RELEVANCE STATEMENT: The ANN model CTA-DEFACE, developed for automatic defacing of CT angiography images, achieves significantly lower face detection probabilities and greater dissimilarity from the original images compared to a publicly available model. The algorithm has been externally validated and is publicly accessible. KEY POINTS: The developed ANN model (CTA-DEFACE) automatically generates facemasks for CT angiography images. CTA-DEFACE offers superior deidentification capabilities compared to a publicly available model. By means of graphics processing unit optimization, our model ensures rapid processing of medical images. Our model underwent external validation, underscoring its reliability for real-world application.

Exploring de-anonymization risks in PET imaging: Insights from a comprehensive analysis of 853 patient scans

Due to their high resolution, anonymized CT scans can be reidentified using face recognition tools. However, little is known regarding PET deanonymization because of its lower resolution. In this study, we analysed PET/CT scans of 853 patients from a TCIA-restricted dataset (AutoPET). First, we built denoised 2D morphological reconstructions of both PET and CT scans, and then we determined how frequently a PET reconstruction could be matched to the correct CT reconstruction with no other metadata. Using the CT morphological reconstructions as ground truth allows us to frame the problem as a face recognition problem and to quantify our performance using traditional metrics (top k accuracies) without any use of patient pictures. Using our denoised PET 2D reconstructions, we achieved 72% top 10 accuracy after the realignment of all CTs in the same reference frame, and 71% top 10 accuracy after realignment and mixing within a larger face dataset of 10, 168 pictures. This highlights the need to consider face identification issues when dealing with PET imaging data.

Cloud-based serverless computing enables accelerated monte carlo simulations for nuclear medicine imaging

Objective.This study investigates the potential of cloud-based serverless computing to accelerate Monte Carlo (MC) simulations for nuclear medicine imaging tasks. MC simulations can pose a high computational burden-even when executed on modern multi-core computing servers. Cloud computing allows simulation tasks to be highly parallelized and considerably accelerated.Approach.We investigate the computational performance of a cloud-based serverless MC simulation of radioactive decays for positron emission tomography imaging using Amazon Web Service (AWS) Lambda serverless computing platform for the first time in scientific literature. We provide a comparison of the computational performance of AWS to a modern on-premises multi-thread reconstruction server by measuring the execution times of the processes using between105and2·1010simulated decays. We deployed two popular MC simulation frameworks-SimSET and GATE-within the AWS computing environment. Containerized application images were used as a basis for an AWS Lambda function, and local (non-cloud) scripts were used to orchestrate the deployment of simulations. The task was broken down into smaller parallel runs, and launched on concurrently running AWS Lambda instances, and the results were postprocessed and downloaded via the Simple Storage Service.Main results.Our implementation of cloud-based MC simulations with SimSET outperforms local server-based computations by more than an order of magnitude. However, the GATE implementation creates more and larger output file sizes and reveals that the internet connection speed can become the primary bottleneck for data transfers. Simulating 109decays using SimSET is possible within 5 min and accrues computation costs of about $10 on AWS, whereas GATE would have to run in batches for more than 100 min at considerably higher costs.Significance.Adopting cloud-based serverless computing architecture in medical imaging research facilities can considerably improve processing times and overall workflow efficiency, with future research exploring additional enhancements through optimized configurations and computational methods.

Performance and application of the total-body PET/CT scanner: a literature review

BACKGROUND

The total-body positron emission tomography/computed tomography (PET/CT) system, with a long axial field of view, represents the state-of-the-art PET imaging technique. Recently, the total-body PET/CT system has been commercially available. The total-body PET/CT system enables high-resolution whole-body imaging, even under extreme conditions such as ultra-low dose, extremely fast imaging speed, delayed imaging more than 10 h after tracer injection, and total-body dynamic scan. The total-body PET/CT system provides a real-time picture of the tracers of all organs across the body, which not only helps to explain normal human physiological process, but also facilitates the comprehensive assessment of systemic diseases. In addition, the total-body PET/CT system may play critical roles in other medical fields, including cancer imaging, drug development and immunology.

MAIN BODY

Therefore, it is of significance to summarize the existing studies of the total-body PET/CT systems and point out its future direction. This review collected research literatures from the PubMed database since the advent of commercially available total-body PET/CT systems to the present, and was divided into the following sections: Firstly, a brief introduction to the total-body PET/CT system was presented, followed by a summary of the literature on the performance evaluation of the total-body PET/CT. Then, the research and clinical applications of the total-body PET/CT were discussed. Fourthly, deep learning studies based on total-body PET imaging was reviewed. At last, the shortcomings of existing research and future directions for the total-body PET/CT were discussed.

CONCLUSION

Due to its technical advantages, the total-body PET/CT system is bound to play a greater role in clinical practice in the future.