End-to-end reproducible AI pipelines in radiology using the cloud.

Artificial intelligence (AI) algorithms hold the potential to revolutionize radiology. However, a significant portion of the published literature lacks transparency and reproducibility, which hampers sustained progress toward clinical translation. Although several reporting guidelines have been proposed, identifying practical means to address these issues remains challenging. Here, we show the potential of cloud-based infrastructure for implementing and sharing transparent and reproducible AI-based radiology pipelines. We demonstrate end-to-end reproducibility from retrieving cloud-hosted data, through data pre-processing, deep learning inference, and post-processing, to the analysis and reporting of the final results. We successfully implement two distinct use cases, starting from recent literature on AI-based biomarkers for cancer imaging. Using cloud-hosted data and computing, we confirm the findings of these studies and extend the validation to previously unseen data for one of the use cases. Furthermore, we provide the community with transparent and easy-to-extend examples of pipelines impactful for the broader oncology field. Our approach demonstrates the potential of cloud resources for implementing, sharing, and using reproducible and transparent AI pipelines, which can accelerate the translation into clinical solutions.

Generating synthetic computed tomography for radiotherapy: SynthRAD2023 challenge report.

Radiation therapy plays a crucial role in cancer treatment, necessitating precise delivery of radiation to tumors while sparing healthy tissues over multiple days. Computed tomography (CT) is integral for treatment planning, offering electron density data crucial for accurate dose calculations. However, accurately representing patient anatomy is challenging, especially in adaptive radiotherapy, where CT is not acquired daily. Magnetic resonance imaging (MRI) provides superior soft-tissue contrast. Still, it lacks electron density information, while cone beam CT (CBCT) lacks direct electron density calibration and is mainly used for patient positioning. Adopting MRI-only or CBCT-based adaptive radiotherapy eliminates the need for CT planning but presents challenges. Synthetic CT (sCT) generation techniques aim to address these challenges by using image synthesis to bridge the gap between MRI, CBCT, and CT. The SynthRAD2023 challenge was organized to compare synthetic CT generation methods using multi-center ground truth data from 1080 patients, divided into two tasks: (1) MRI-to-CT and (2) CBCT-to-CT. The evaluation included image similarity and dose-based metrics from proton and photon plans. The challenge attracted significant participation, with 617 registrations and 22/17 valid submissions for tasks 1/2. Top-performing teams achieved high structural similarity indices (≥0.87/0.90) and gamma pass rates for photon (≥98.1%/99.0%) and proton (≥97.3%/97.0%) plans. However, no significant correlation was found between image similarity metrics and dose accuracy, emphasizing the need for dose evaluation when assessing the clinical applicability of sCT. SynthRAD2023 facilitated the investigation and benchmarking of sCT generation techniques, providing insights for developing MRI-only and CBCT-based adaptive radiotherapy. It showcased the growing capacity of deep learning to produce high-quality sCT, reducing reliance on conventional CT for treatment planning.

Foundation model for cancer imaging biomarkers.

Foundation models in deep learning are characterized by a single large-scale model trained on vast amounts of data serving as the foundation for various downstream tasks. Foundation models are generally trained using self-supervised learning and excel in reducing the demand for training samples in downstream applications. This is especially important in medicine, where large labelled datasets are often scarce. Here, we developed a foundation model for cancer imaging biomarker discovery by training a convolutional encoder through self-supervised learning using a comprehensive dataset of 11,467 radiographic lesions. The foundation model was evaluated in distinct and clinically relevant applications of cancer imaging-based biomarkers. We found that it facilitated better and more efficient learning of imaging biomarkers and yielded task-specific models that significantly outperformed conventional supervised and other state-of-the-art pretrained implementations on downstream tasks, especially when training dataset sizes were very limited. Furthermore, the foundation model was more stable to input variations and showed strong associations with underlying biology. Our results demonstrate the tremendous potential of foundation models in discovering new imaging biomarkers that may extend to other clinical use cases and can accelerate the widespread translation of imaging biomarkers into clinical settings.

Decoding biological age from face photographs using deep learning.

Because humans age at different rates, a person's physical appearance may yield insights into their biological age and physiological health more reliably than their chronological age. In medicine, however, appearance is incorporated into medical judgments in a subjective and non-standardized fashion. In this study, we developed and validated FaceAge, a deep learning system to estimate biological age from easily obtainable and low-cost face photographs. FaceAge was trained on data from 58,851 healthy individuals, and clinical utility was evaluated on data from 6,196 patients with cancer diagnoses from two institutions in the United States and The Netherlands. To assess the prognostic relevance of FaceAge estimation, we performed Kaplan Meier survival analysis. To test a relevant clinical application of FaceAge, we assessed the performance of FaceAge in end-of-life patients with metastatic cancer who received palliative treatment by incorporating FaceAge into clinical prediction models. We found that, on average, cancer patients look older than their chronological age, and looking older is correlated with worse overall survival. FaceAge demonstrated significant independent prognostic performance in a range of cancer types and stages. We found that FaceAge can improve physicians' survival predictions in incurable patients receiving palliative treatments, highlighting the clinical utility of the algorithm to support end-of-life decision-making. FaceAge was also significantly associated with molecular mechanisms of senescence through gene analysis, while age was not. These findings may extend to diseases beyond cancer, motivating using deep learning algorithms to translate a patient's visual appearance into objective, quantitative, and clinically useful measures.

Foundation Models for Quantitative Biomarker Discovery in Cancer Imaging.

Foundation models represent a recent paradigm shift in deep learning, where a single large-scale model trained on vast amounts of data can serve as the foundation for various downstream tasks. Foundation models are generally trained using self-supervised learning and excel in reducing the demand for training samples in downstream applications. This is especially important in medicine, where large labeled datasets are often scarce. Here, we developed a foundation model for imaging biomarker discovery by training a convolutional encoder through self-supervised learning using a comprehensive dataset of 11,467 radiographic lesions. The foundation model was evaluated in distinct and clinically relevant applications of imaging-based biomarkers. We found that they facilitated better and more efficient learning of imaging biomarkers and yielded task-specific models that significantly outperformed their conventional supervised counterparts on downstream tasks. The performance gain was most prominent when training dataset sizes were very limited. Furthermore, foundation models were more stable to input and inter-reader variations and showed stronger associations with underlying biology. Our results demonstrate the tremendous potential of foundation models in discovering novel imaging biomarkers that may extend to other clinical use cases and can accelerate the widespread translation of imaging biomarkers into clinical settings.

Quality of YouTube Videos Related to Colorectal Cancer Screening.

Introduction YouTube, an unregulated video-sharing website, is the second most visited website on the internet. As more patients turn to the internet for information about colon cancer screening, it is important to understand what they are consuming online. Our goal was to evaluate YouTube videos about colon cancer screening to better understand the information patients are accessing. Methods We searched YouTube on October 28, 2020, using the following search terms sorted by relevance and view count: colonoscopy, colon cancer screening, virtual colonoscopy, colonoscopy alternatives, and cologuard. Videos longer than 10 minutes, not in English, and duplicates were excluded. Three evaluators graded each video using the DISCERN criteria. Numerical data were averaged into a composite score. Two-sided t-tests and one-way ANOVA tests were used to compare mean ratings between groups. Results Fifty videos were analyzed, with a total of 23,148,938 views, averaging 462,979 views per video. The average overall rating was 3.16/5. There was no difference between search methods, search terms, or presence of a physician. The average ratings for videos with gastroenterologists (3.08), other physicians (3.35), and non-physicians (3.09) were not significantly different. Videos without physicians had more views on average (1,148,677) compared to videos with gastroenterologists (157,846, p=0.013) or other physicians (35,730, p=0.013). Conclusion YouTube videos related to colon cancer screening were of good quality regardless of search terms, search methods, or presence of a physician. However, videos without physicians were viewed more frequently. Physicians should continue making videos that address deficits while increasing viewership.

Frequency-Domain-Based Structure Losses for CycleGAN-Based Cone-Beam Computed Tomography Translation.

Research exploring CycleGAN-based synthetic image generation has recently accelerated in the medical community due to its ability to leverage unpaired images effectively. However, a commonly established drawback of the CycleGAN, the introduction of artifacts in generated images, makes it unreliable for medical imaging use cases. In an attempt to address this, we explore the effect of structure losses on the CycleGAN and propose a generalized frequency-based loss that aims at preserving the content in the frequency domain. We apply this loss to the use-case of cone-beam computed tomography (CBCT) translation to computed tomography (CT)-like quality. Synthetic CT (sCT) images generated from our methods are compared against baseline CycleGAN along with other existing structure losses proposed in the literature. Our methods (MAE: 85.5, MSE: 20433, NMSE: 0.026, PSNR: 30.02, SSIM: 0.935) quantitatively and qualitatively improve over the baseline CycleGAN (MAE: 88.8, MSE: 24244, NMSE: 0.03, PSNR: 29.37, SSIM: 0.935) across all investigated metrics and are more robust than existing methods. Furthermore, no observable artifacts or loss in image quality were observed. Finally, we demonstrated that sCTs generated using our methods have superior performance compared to the original CBCT images on selected downstream tasks.

Artificial Intelligence in Radiation Therapy.

Artificial intelligence (AI) has great potential to transform the clinical workflow of radiotherapy. Since the introduction of deep neural networks, many AI-based methods have been proposed to address challenges in different aspects of radiotherapy. Commercial vendors have started to release AI-based tools that can be readily integrated to the established clinical workflow. To show the recent progress in AI-aided radiotherapy, we have reviewed AI-based studies in five major aspects of radiotherapy including image reconstruction, image registration, image segmentation, image synthesis, and automatic treatment planning. In each section, we summarized and categorized the recently published methods, followed by a discussion of the challenges, concerns, and future development. Given the rapid development of AI-aided radiotherapy, the efficiency and effectiveness of radiotherapy in the future could be substantially improved through intelligent automation of various aspects of radiotherapy.

Head and neck tumor segmentation in PET/CT: The HECKTOR challenge.

This paper relates the post-analysis of the first edition of the HEad and neCK TumOR (HECKTOR) challenge. This challenge was held as a satellite event of the 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2020, and was the first of its kind focusing on lesion segmentation in combined FDG-PET and CT image modalities. The challenge's task is the automatic segmentation of the Gross Tumor Volume (GTV) of Head and Neck (H&N) oropharyngeal primary tumors in FDG-PET/CT images. To this end, the participants were given a training set of 201 cases from four different centers and their methods were tested on a held-out set of 53 cases from a fifth center. The methods were ranked according to the Dice Score Coefficient (DSC) averaged across all test cases. An additional inter-observer agreement study was organized to assess the difficulty of the task from a human perspective. 64 teams registered to the challenge, among which 10 provided a paper detailing their approach. The best method obtained an average DSC of 0.7591, showing a large improvement over our proposed baseline method and the inter-observer agreement, associated with DSCs of 0.6610 and 0.61, respectively. The automatic methods proved to successfully leverage the wealth of metabolic and structural properties of combined PET and CT modalities, significantly outperforming human inter-observer agreement level, semi-automatic thresholding based on PET images as well as other single modality-based methods. This promising performance is one step forward towards large-scale radiomics studies in H&N cancer, obviating the need for error-prone and time-consuming manual delineation of GTVs.

Factors that Influence End-Of-Life Decision Making Amongst Attending Physicians.

Background: Medical advances prolong life and treat illness but many patients have chronically debilitating conditions that prevent them from making end-of-life (EOL) decisions for themselves. These situations are difficult to navigate for both patient and physician. This study investigates physicians' feelings and approach toward EOL care, physician-assisted suicide (PAS), and euthanasia. Methods: An anonymous, self-administered online survey was distributed through the New Jersey Medical School servers and American College of Surgeons forums. The survey presented clinical EOL vignettes and subjective questions regarding PAS and euthanasia. Results: We obtained 142 responses from attending physicians. Respondents were typically male (61%), married (85%), identified as Christian (54%), had more than 20 years of experience (55%), and worked at a university hospital (57%). Religious beliefs and years of work experience seemed to be significant contributors in EOL decision making, whereas gender and medical specialty were not significantly influential. Conclusion: Factors such as years of work experience and religious belief may influence medical professionals' opinions about PAS and euthanasia and their subsequent actions regarding EOL care. In many cases, the boundaries are blurred and require further study before concrete conclusions can be made.

SULT1A1 (Minoxidil Sulfotransferase) enzyme booster significantly improves response to topical minoxidil for hair regrowth.

BACKGROUND: Minoxidil is a widely used over-the-counter topical treatment for hair loss. The response rate for topical minoxidil is relatively low. Minoxidil is a pro-drug, converted to its active form, minoxidil sulfate, by SULT1A1 enzymes located in the scalp. Recently, a novel topical formula that increases the activity of SULT1A1 in hair follicles was reported. AIMS: To evaluate any benefit of applying the SULT1A1 enzyme booster prior to daily 5% minoxidil treatment. METHODS: Male androgenic alopecia patients were recruited to a randomized blinded placebo-controlled study. Patients were randomized to receive 5% topical minoxidil plus the novel formula or minoxidil plus a sham adjuvant. Patient's hair growth was monitored using global photography over 60 days. RESULTS: Twenty-four males with androgenic alopecia (Norwood scale average 4.4, range 2-6) were randomized and completed the trial: 12 in the active arm and 12 in placebo. 75% of the subjects who used the SULT1A1 adjuvant with their daily minoxidil treatments for 60 days regrew hair versus 33% of those using the placebo adjuvant (p = 0.023). CONCLUSIONS: In a small cohort of androgenetic alopecia men, adding the SULT1A1 adjuvant to their daily minoxidil treatment regimen improved hair regrowth.

Association between aspartate aminotransferase-to-alanine aminotransferase ratio and insulin resistance among US adults.

OBJECTIVE: To determine whether a low aspartate aminotransferase (AST) to alanine aminotransferase (ALT) ratio (AST/ALT ratio) is associated with insulin resistance among those without liver dysfunction. METHODS: In this cross-sectional study of the National Health and Nutrition Examination Survey (NHANES) 2011-2016, we included 2747 (1434 male and 1313 nonpregnant female) adults ≥20 years without evidence of liver dysfunction (ALT<30 in male and <19 in female, negative viral serologies, no excess alcohol consumption, no elevated transferrin saturation, AST/ALT <2). Serum AST/ALT ratio was categorized into sex-specific quartiles (female: <1.12, 1.12-1.29, 1.29-1.47, ≥1.47 and male: <0.93, 0.93-1.09, 1.09-1.26, ≥1.26). The primary outcome was insulin resistance, as determined by Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) index score ≥3. Covariate-adjusted odds ratios (ORs) were estimated. Study analysis completed from 13 March 2020 to 21 April 2021. RESULTS: Among the 2747 individuals, 33% had insulin resistance. Those in the lowest quartile (Q1) of AST/ALT had 75% higher adjusted odds of insulin resistance compared to the highest quartile (Q4) [aOR (95% confidence interval (CI), 1.75 (1.20-2.57)]. This association was more pronounced in those with elevated BMI [Q1 vs. Q4; BMI ≥ 25: 2.29 (1.58-3.33), BMI < 25: 0.66 (0.26-1.69); NAFLD per Fatty Liver Index ≥ 60: 2.04 (1.21-3.44), No NAFLD: 1.68 (0.94-3.01)]. CONCLUSION: Lower AST/ALT ratio is associated with increased insulin resistance among those with healthy-range ALT, especially in those with BMI greater than or equal to 25 kg/m2.

Radiomics integration into a picture archiving and communication system.

Radiomics is referred to as quantitative imaging of biomarkers used for clinical outcome prognosis or tumor characterization. In order to bridge radiomics and its clinical application, we aimed to build an integrated solution of radiomics extraction with an open-source Picture Archiving and Communication System (PACS). The integrated SQLite4Radiomics software was tested in three different imaging modalities and its performance was benchmarked in lung cancer open datasets RIDER and MMD with median extraction time of 10.7 (percentiles 25-75: 8.9-18.7) seconds per ROI in three different configurations.

Clinical outcomes and resource utilization analysis in patients with rheumatoid arthritis undergoing endoscopic retrograde cholangiopancreatography.

BACKGROUND AND AIM: The literature is lacking on associations of endoscopic retrograde cholangiopancreatography (ERCP) related outcomes in rheumatoid arthritis (RA) patients. The aim of this study is to evaluate the effects of RA on clinical outcomes and hospital resource utilization in patients undergoing ERCP. METHODS: The National Inpatient Sample database was used to identify hospitalized patients who had underwent an ERCP study from 2012 to 2014 using International Classification of Diseases-Ninth Edition (ICD-9) codes. Primary outcomes were mortality, hospital charges, and length of stay. Secondary outcomes were ERCP-related complications. Chi-squared tests for categorical data and independent t-test for continuous data were utilized. Multivariate analysis was performed to assess the primary outcomes. RESULTS: There was 83 890 ERCP procedures performed, of which 970 patients had RA. In patients with RA, 74.2% were female, and the average age was 65.7 years. RA primary outcomes of mortality rate and hospital cost were lower and statistically significant. There was no statistically significant difference in secondary outcomes except for lower cholecystectomy rates in RA patients. CONCLUSION: With a high inflammatory state, it was hypothesized that RA would be associated with worse outcomes after ERCP. Yet, the primary outcomes of mortality and hospital cost were found to be lower than controls, with no difference in secondary outcomes. We posit that immunosuppressants used to treat RA provides a protective effect to overall complications with ERCP.

Association between alanine aminotransferase within the normal range and all-cause and cause-specific mortality: A nationwide cohort study.

BACKGROUND AND AIM: We sought to determine the association between alanine aminotransferase (ALT) in the normal range and mortality in the absence of liver dysfunction to better understand ALT's clinical significance beyond liver injury and inflammation. METHODS: A cohort of 2,708 male and 3,461 female adults aged 20-75 years without liver dysfunction (ALT<30 in males & <19 in females, negative viral serologies, negative ultrasound-based steatosis, no excess alcohol consumption) from the National Health and Nutrition Examination Survey (NHANES)-III (1988-1994) were linked to the National Death Index through December 31, 2015. Serum ALT levels were categorized into sex-specific quartiles (Females: <9, 9-11, 11-14, ≥14 IU/L, Male: <12, 12-15, 15-20, ≥20 U/L). The primary outcome was all-cause mortality. Hazard ratios (HRs) were estimated, adjusting for covariates and accounting for the complex survey design. RESULTS: Relative to males in the lowest quartile (Q1), males in the highest quartile (Q4) had 44% decreased risk of all-cause mortality (aHR [95% CI]: 0.56 [0.42, 0.74]). Females in Q4 had 45% decreased risk of all-cause mortality (aHR [95% CI]: 0.55 [0.40, 0.77]). Males with BMI <25 kg/m2 in Q4 had significantly lower risk of all-cause mortality than Q1; however, this association did not exist in males with BMI ≥25 (BMI<25: 0.36 [0.20, 0.64], BMI≥25: 0.77 [0.49, 1.22]). Risk of all-cause mortality was lower in males ≥50 years than in males<50 (age≥50: 0.55 [0.39, 0.77], age<50: 0.81 [0.39, 1.69]). These age- and BMI-related differences were not seen in females. CONCLUSION: ALT within the normal range was inversely associated with all-cause mortality in U.S. adults.