Far-UVC- and UVB-induced DNA damage depending on skin type.

Far-UVC radiation sources of wavelengths 222 nm and 233 nm represent an interesting potential alternative for the antiseptic treatment of the skin due to their high skin compatibility. Nevertheless, no studies on far-UVC-induced DNA damage in different skin types have been published to date, which this study aims for. After irradiating the skin with far-UVC of the wavelengths 222 and 233 nm as well as broadband UVB, the tissue was screened for cyclobutane pyrimidine dimer-positive (CPD+ ) cells using immunohistochemistry. The epidermal DNA damage was lower in dark skin types than in fair skin types after irradiation at 233 nm. Contrary to this, irradiation at 222 nm caused no skin type-dependent differences, which can be attributed to the decreased penetration depth of radiation. UVB showed the relatively strongest differences between light and dark skin types when using a suberythemal dose of 3 mJ/cm2 . As melanin is known for its photoprotective effect, we evaluated the ratio of melanin content in the stratum basale and stratum granulosum in samples of different skin types using two-photon excited fluorescence lifetime imaging (TPE-FLIM) finding a higher ratio up to skin type IV-V. As far-UVC is known to penetrate only into the upper layers of the viable skin, the aforementioned melanin ratio could explain the less pronounced differences between skin types after irradiation with far-UVC compared to UVB.

Semantic modeling of cell damage prediction: a machine learning approach at human-level performance in dermatology.

Machine learning is transforming the field of histopathology. Especially in classification related tasks, there have been many successful applications of deep learning already. Yet, in tasks that rely on regression and many niche applications, the domain lacks cohesive procedures that are adapted to the learning processes of neural networks. In this work, we investigate cell damage in whole slide images of the epidermis. A common way for pathologists to annotate a score, characterizing the degree of damage for these samples, is the ratio between healthy and unhealthy nuclei. The annotation procedure of these scores, however, is expensive and prone to be noisy among pathologists. We propose a new measure of damage, that is the total area of damage, relative to the total area of the epidermis. In this work, we present results of regression and segmentation models, predicting both scores on a curated and public dataset. We have acquired the dataset in collaborative efforts with medical professionals. Our study resulted in a comprehensive evaluation of the proposed damage metrics in the epidermis, with recommendations, emphasizing practical relevance for real world applications.

From modern CNNs to vision transformers: Assessing the performance, robustness, and classification strategies of deep learning models in histopathology.

While machine learning is currently transforming the field of histopathology, the domain lacks a comprehensive evaluation of state-of-the-art models based on essential but complementary quality requirements beyond a mere classification accuracy. In order to fill this gap, we developed a new methodology to extensively evaluate a wide range of classification models, including recent vision transformers, and convolutional neural networks such as: ConvNeXt, ResNet (BiT), Inception, ViT and Swin transformer, with and without supervised or self-supervised pretraining. We thoroughly tested the models on five widely used histopathology datasets containing whole slide images of breast, gastric, and colorectal cancer and developed a novel approach using an image-to-image translation model to assess the robustness of a cancer classification model against stain variations. Further, we extended existing interpretability methods to previously unstudied models and systematically reveal insights of the models' classification strategies that allow for plausibility checks and systematic comparisons. The study resulted in specific model recommendations for practitioners as well as putting forward a general methodology to quantify a model's quality according to complementary requirements that can be transferred to future model architectures.

A deep-learning algorithm to classify skin lesions from mpox virus infection.

Undetected infection and delayed isolation of infected individuals are key factors driving the monkeypox virus (now termed mpox virus or MPXV) outbreak. To enable earlier detection of MPXV infection, we developed an image-based deep convolutional neural network (named MPXV-CNN) for the identification of the characteristic skin lesions caused by MPXV. We assembled a dataset of 139,198 skin lesion images, split into training/validation and testing cohorts, comprising non-MPXV images (n = 138,522) from eight dermatological repositories and MPXV images (n = 676) from the scientific literature, news articles, social media and a prospective cohort of the Stanford University Medical Center (n = 63 images from 12 patients, all male). In the validation and testing cohorts, the sensitivity of the MPXV-CNN was 0.83 and 0.91, the specificity was 0.965 and 0.898 and the area under the curve was 0.967 and 0.966, respectively. In the prospective cohort, the sensitivity was 0.89. The classification performance of the MPXV-CNN was robust across various skin tones and body regions. To facilitate the usage of the algorithm, we developed a web-based app by which the MPXV-CNN can be accessed for patient guidance. The capability of the MPXV-CNN for identifying MPXV lesions has the potential to aid in MPXV outbreak mitigation.

To pretrain or not? A systematic analysis of the benefits of pretraining in diabetic retinopathy.

There is an increasing number of medical use cases where classification algorithms based on deep neural networks reach performance levels that are competitive with human medical experts. To alleviate the challenges of small dataset sizes, these systems often rely on pretraining. In this work, we aim to assess the broader implications of these approaches in order to better understand what type of pretraining works reliably (with respect to performance, robustness, learned representation etc.) in practice and what type of pretraining dataset is best suited to achieve good performance in small target dataset size scenarios. Considering diabetic retinopathy grading as an exemplary use case, we compare the impact of different training procedures including recently established self-supervised pretraining methods based on contrastive learning. To this end, we investigate different aspects such as quantitative performance, statistics of the learned feature representations, interpretability and robustness to image distortions. Our results indicate that models initialized from ImageNet pretraining report a significant increase in performance, generalization and robustness to image distortions. In particular, self-supervised models show further benefits to supervised models. Self-supervised models with initialization from ImageNet pretraining not only report higher performance, they also reduce overfitting to large lesions along with improvements in taking into account minute lesions indicative of the progression of the disease. Understanding the effects of pretraining in a broader sense that goes beyond simple performance comparisons is of crucial importance for the broader medical imaging community beyond the use case considered in this work.

Active multitask learning with uncertainty-weighted loss for coronary calcium scoring.

PURPOSE: The coronary artery calcification (CAC) score is an independent marker for the risk of cardiovascular events. Automatic methods for quantifying CAC could reduce workload and assist radiologists in clinical decision-making. However, large annotated datasets are needed for training to achieve very good model performance, which is an expensive process and requires expert knowledge. The number of training data required can be reduced in an active learning scenario, which requires only the most informative samples to be labeled. Multitask learning techniques can improve model performance by joint learning of multiple related tasks and extraction of shared informative features. METHODS: We propose an uncertainty-weighted multitask learning model for coronary calcium scoring in electrocardiogram-gated (ECG-gated), noncontrast-enhanced cardiac calcium scoring CT. The model was trained to solve the two tasks of coronary artery region segmentation (weak labels) and coronary artery calcification segmentation (strong labels) simultaneously in an active learning scenario to improve model performance and reduce the number of samples needed for training. We compared our model with a single-task U-Net and a sequential-task model as well as other state-of-the-art methods. The model was evaluated on 1275 individual patients in three different datasets (DISCHARGE, CADMAN, orCaScore), and the relationship between model performance and various influencing factors (image noise, metal artifacts, motion artifacts, image quality) was analyzed. RESULTS: Joint learning of multiclass coronary artery region segmentation and binary coronary calcium segmentation improved calcium scoring performance. Since shared information can be learned from both tasks for complementary purposes, the model reached optimal performance with only 12% of the training data and one-third of the labeling time in an active learning scenario. We identified image noise as one of the most important factors influencing model performance along with anatomical abnormalities and metal artifacts. CONCLUSIONS: Our multitask learning approach with uncertainty-weighted loss improves calcium scoring performance by joint learning of shared features and reduces labeling costs when trained in an active learning scenario.

Multiparametric MRI for Characterization of the Basal Ganglia and the Midbrain.

Objectives To characterize subcortical nuclei by multi-parametric quantitative magnetic resonance imaging. Materials and Methods: The following quantitative multiparametric MR data of five healthy volunteers were acquired on a 7T MRI system: 3D gradient echo (GRE) data for the calculation of quantitative susceptibility maps (QSM), GRE sequences with and without off-resonant magnetic transfer pulse for magnetization transfer ratio (MTR) calculation, a magnetization-prepared 2 rapid acquisition gradient echo sequence for T1 mapping, and (after a coil change) a density-adapted 3D radial pulse sequence for 23Na imaging. First, all data were co-registered to the GRE data, volumes of interest (VOIs) for 21 subcortical structures were drawn manually for each volunteer, and a combined voxel-wise analysis of the four MR contrasts (QSM, MTR, T1, 23Na) in each structure was conducted to assess the quantitative, MR value-based differentiability of structures. Second, a machine learning algorithm based on random forests was trained to automatically classify the groups of multi-parametric voxel values from each VOI according to their association to one of the 21 subcortical structures. Results The analysis of the integrated multimodal visualization of quantitative MR values in each structure yielded a successful classification among nuclei of the ascending reticular activation system (ARAS), the limbic system and the extrapyramidal system, while classification among (epi-)thalamic nuclei was less successful. The machine learning-based approach facilitated quantitative MR value-based structure classification especially in the group of extrapyramidal nuclei and reached an overall accuracy of 85% regarding all selected nuclei. Conclusion Multimodal quantitative MR enabled excellent differentiation of a wide spectrum of subcortical nuclei with reasonable accuracy and may thus enable sensitive detection of disease and nucleus-specific MR-based contrast alterations in the future.

Risk estimation of SARS-CoV-2 transmission from bluetooth low energy measurements.

Digital contact tracing approaches based on Bluetooth low energy (BLE) have the potential to efficiently contain and delay outbreaks of infectious diseases such as the ongoing SARS-CoV-2 pandemic. In this work we propose a machine learning based approach to reliably detect subjects that have spent enough time in close proximity to be at risk of being infected. Our study is an important proof of concept that will aid the battery of epidemiological policies aiming to slow down the rapid spread of COVID-19.

Diagnostic, clinical, and personality correlates of food anxiety during a food exposure in patients diagnosed with an eating disorder.

BACKGROUND: Eating disorders are characterized by high levels of anxiety, especially while eating. However, little is known about anxiety experienced during meals and specifically what other variables may impact such anxiety. OBJECTIVE: We sought to further quantify and understand the relationship between food anxiety, eating disorders, and related correlates (e.g., comorbid diagnoses, personality). METHODS: In the current study [N = 42 participants diagnosed with an eating disorder (n = 36 participants with anorexia nervosa)], we quantified anxiety before, during, and after a meal using data from a food exposure session in a partial hospital eating disorder center. We examined diagnostic, personality, and clinical factors as correlates of food anxiety. RESULTS: Participants were more likely to experience higher food anxiety if they had a current diagnosis of major depression, obsessive-compulsive disorder, or post-traumatic stress disorder (PTSD). Concern over mistakes was the strongest and most consistent correlate of food anxiety regardless of time during the meal that anxiety was assessed. Other significant correlates were fear of positive evaluation, social appearance anxiety, BMI, and trust. CONCLUSIONS: These findings show how diagnoses, perfectionism (concern over mistakes), and other correlates relate to anxiety during meals. Food exposure interventions may benefit from personalizations that address these factors. LEVEL OF EVIDENCE: IV Evidence from a randomized control trial, but from the first session before effects of the design would be present.

Shearlet-based compressed sensing for fast 3D cardiac MR imaging using iterative reweighting.

High-resolution three-dimensional (3D) cardiovascular magnetic resonance (CMR) is a valuable medical imaging technique, but its widespread application in clinical practice is hampered by long acquisition times. Here we present a novel compressed sensing (CS) reconstruction approach using shearlets as a sparsifying transform allowing for fast 3D CMR (3DShearCS) using 3D radial phase encoding (RPE). An iterative reweighting scheme was applied during image reconstruction to ensure fast convergence and high image quality. Shearlets are mathematically optimal for a simplified model of natural images and have been proven to be more efficient than classical systems such as wavelets. 3DShearCS was compared to three other commonly used reconstruction approaches. Image quality was assessed quantitatively using general image quality metrics and using clinical diagnostic scores from expert reviewers. The proposed technique had lower relative errors, higher structural similarity and higher diagnostic scores compared to the other reconstruction techniques especially for high undersampling factors, i.e. short scan times. 3DShearCS provided ensured accurate depiction of cardiac anatomy for fast imaging and could help to promote 3D high-resolution CMR in clinical practice.

Understanding Singaporean medical students' stress and coping.

Medical school is intrinsically stressful, and high levels of stress have untoward effects. Although surveys have revealed some sources of stress among medical students, little is known about the qualitative aspects of these stressors and their associated coping behaviours, particularly among medical students in Singapore. Our exploratory pilot study found that relationship issues and examinations were the major sources of stress for medical students. The respondents described multiple context-sensitive coping styles, as well as reported 'avoidance' or 'wishful thinking' coping strategies as ineffective. Their stress-and-coping process suggests the influence of Asian culture and medical school culture. Our findings thus indicate the need for further research, potentially using new methodologies such as the critical incident analysis technique, and thoughtful consideration of culture when implementing programmes in Singapore medical schools to improve the students' stress-and-coping responses.

Screening for panic-related anxiety in emergency department patients with cardiopulmonary complaints: A comparison of two self-report instruments.

The present study examined 1) the accuracy of two self-report measures for detecting panic-related anxiety in emergency department (ED) patients with cardiopulmonary complaints; and 2) whether modified scoring resulted in improved performance. English-speaking adults presenting to the ED of a large public hospital with palpitations, chest pain, dizziness, or difficulty breathing were evaluated for the presence of panic-related anxiety with the Structured Clinical Interview for DSM-IV (SCID) over a one-year period. Patients completed the panic disorder modules of the Patient Health Questionnaire (PHQ-PD) and Psychiatric Diagnostic Screening Questionnaire (PDSQ-PD). Sensitivity, specificity, area under the curve (AUC), and predictive values were compared for various cut-offs and scoring algorithms using SCID diagnosis of panic attacks (in the absence of panic disorder) or panic disorder as the reference standard. In this sample of 200 participants, the majority had a chief complaint of chest pain and 46.5% met SCID criteria for panic-related anxiety. The PDSQ-PD demonstrated only fair operating characteristics for panic attacks (AUC = 0.57) and good operating characteristics for panic disorder (AUC = 0.79). The PHQ-PD achieved adequate operating characteristics (AUC = 0.66) for panic attacks and good operating characteristics for panic disorder (AUC = 0.76) using a modified scoring algorithm or a single screening question (AUC = 0.72).

Fear of food prospectively predicts drive for thinness in an eating disorder sample recently discharged from intensive treatment.

Fears of food are common in individuals with eating disorders and contribute to the high relapse rates. However, it is unknown how fears of food contribute to eating disorder symptoms across time, potentially contributing to an increased likelihood of relapse. Participants diagnosed with an eating disorder (N=168) who had recently completed intensive treatment were assessed after discharge and one month later regarding fear of food, eating disorder symptoms, anxiety sensitivity, and negative affect. Cross lagged path analysis was utilized to determine if fear of food predicted subsequent eating disorder symptoms one month later. Fear of food-specifically, anxiety about eating and feared concerns about eating-predicted drive for thinness, a core symptom domain of eating disorders. These relationships held while accounting for anxiety sensitivity and negative affect. There is a specific, direct relationship between anxiety about eating and feared concerns about eating and drive for thinness. Future research should test if interventions designed to target fear of food can decrease drive for thinness and thereby prevent relapse.