Ability of artificial intelligence to identify self-reported race in chest x-ray using pixel intensity counts.

Purpose: Prior studies show convolutional neural networks predicting self-reported race using x-rays of chest, hand and spine, chest computed tomography, and mammogram. We seek an understanding of the mechanism that reveals race within x-ray images, investigating the possibility that race is not predicted using the physical structure in x-ray images but is embedded in the grayscale pixel intensities. Approach: Retrospective full year 2021, 298,827 AP/PA chest x-ray images from 3 academic health centers across the United States and MIMIC-CXR, labeled by self-reported race, were used in this study. The image structure is removed by summing the number of each grayscale value and scaling to percent per image (PPI). The resulting data are tested using multivariate analysis of variance (MANOVA) with Bonferroni multiple-comparison adjustment and class-balanced MANOVA. Machine learning (ML) feed-forward networks (FFN) and decision trees were built to predict race (binary Black or White and binary Black or other) using only grayscale value counts. Stratified analysis by body mass index, age, sex, gender, patient type, make/model of scanner, exposure, and kilovoltage peak setting was run to study the impact of these factors on race prediction following the same methodology. Results: MANOVA rejects the null hypothesis that classes are the same with 95% confidence (F 7.38, P<0.0001) and balanced MANOVA (F 2.02, P<0.0001). The best FFN performance is limited [area under the receiver operating characteristic (AUROC) of 69.18%]. Gradient boosted trees predict self-reported race using grayscale PPI (AUROC 77.24%). Conclusions: Within chest x-rays, pixel intensity value counts alone are statistically significant indicators and enough for ML classification tasks of patient self-reported race.

Opportunistic detection of type 2 diabetes using deep learning from frontal chest radiographs.

Deep learning (DL) models can harness electronic health records (EHRs) to predict diseases and extract radiologic findings for diagnosis. With ambulatory chest radiographs (CXRs) frequently ordered, we investigated detecting type 2 diabetes (T2D) by combining radiographic and EHR data using a DL model. Our model, developed from 271,065 CXRs and 160,244 patients, was tested on a prospective dataset of 9,943 CXRs. Here we show the model effectively detected T2D with a ROC AUC of 0.84 and a 16% prevalence. The algorithm flagged 1,381 cases (14%) as suspicious for T2D. External validation at a distinct institution yielded a ROC AUC of 0.77, with 5% of patients subsequently diagnosed with T2D. Explainable AI techniques revealed correlations between specific adiposity measures and high predictivity, suggesting CXRs' potential for enhanced T2D screening.

An integrated model for predicting KRAS dependency.

The clinical approvals of KRAS G12C inhibitors have been a revolutionary advance in precision oncology, but response rates are often modest. To improve patient selection, we developed an integrated model to predict KRAS dependency. By integrating molecular profiles of a large panel of cell lines from the DEMETER2 dataset, we built a binary classifier to predict a tumor's KRAS dependency. Monte Carlo cross validation via ElasticNet within the training set was used to compare model performance and to tune parameters α and λ. The final model was then applied to the validation set. We validated the model with genetic depletion assays and an external dataset of lung cancer cells treated with a G12C inhibitor. We then applied the model to several Cancer Genome Atlas (TCGA) datasets. The final "K20" model contains 20 features, including expression of 19 genes and KRAS mutation status. In the validation cohort, K20 had an AUC of 0.94 and accurately predicted KRAS dependency in both mutant and KRAS wild-type cell lines following genetic depletion. It was also highly predictive across an external dataset of lung cancer lines treated with KRAS G12C inhibition. When applied to TCGA datasets, specific subpopulations such as the invasive subtype in colorectal cancer and copy number high pancreatic adenocarcinoma were predicted to have higher KRAS dependency. The K20 model has simple yet robust predictive capabilities that may provide a useful tool to select patients with KRAS mutant tumors that are most likely to respond to direct KRAS inhibitors.

Translating transcriptomic findings from cancer model systems to humans through joint dimension reduction.

Model systems are an essential resource in cancer research. They simulate effects that we can infer into humans, but come at a risk of inaccurately representing human biology. This inaccuracy can lead to inconclusive experiments or misleading results, urging the need for an improved process for translating model system findings into human-relevant data. We present a process for applying joint dimension reduction (jDR) to horizontally integrate gene expression data across model systems and human tumor cohorts. We then use this approach to combine human TCGA gene expression data with data from human cancer cell lines and mouse model tumors. By identifying the aspects of genomic variation joint-acting across cohorts, we demonstrate how predictive modeling and clinical biomarkers from model systems can be improved.

AI recognition of patient race in medical imaging: a modelling study.

BACKGROUND: Previous studies in medical imaging have shown disparate abilities of artificial intelligence (AI) to detect a person's race, yet there is no known correlation for race on medical imaging that would be obvious to human experts when interpreting the images. We aimed to conduct a comprehensive evaluation of the ability of AI to recognise a patient's racial identity from medical images. METHODS: Using private (Emory CXR, Emory Chest CT, Emory Cervical Spine, and Emory Mammogram) and public (MIMIC-CXR, CheXpert, National Lung Cancer Screening Trial, RSNA Pulmonary Embolism CT, and Digital Hand Atlas) datasets, we evaluated, first, performance quantification of deep learning models in detecting race from medical images, including the ability of these models to generalise to external environments and across multiple imaging modalities. Second, we assessed possible confounding of anatomic and phenotypic population features by assessing the ability of these hypothesised confounders to detect race in isolation using regression models, and by re-evaluating the deep learning models by testing them on datasets stratified by these hypothesised confounding variables. Last, by exploring the effect of image corruptions on model performance, we investigated the underlying mechanism by which AI models can recognise race. FINDINGS: In our study, we show that standard AI deep learning models can be trained to predict race from medical images with high performance across multiple imaging modalities, which was sustained under external validation conditions (x-ray imaging [area under the receiver operating characteristics curve (AUC) range 0·91-0·99], CT chest imaging [0·87-0·96], and mammography [0·81]). We also showed that this detection is not due to proxies or imaging-related surrogate covariates for race (eg, performance of possible confounders: body-mass index [AUC 0·55], disease distribution [0·61], and breast density [0·61]). Finally, we provide evidence to show that the ability of AI deep learning models persisted over all anatomical regions and frequency spectrums of the images, suggesting the efforts to control this behaviour when it is undesirable will be challenging and demand further study. INTERPRETATION: The results from our study emphasise that the ability of AI deep learning models to predict self-reported race is itself not the issue of importance. However, our finding that AI can accurately predict self-reported race, even from corrupted, cropped, and noised medical images, often when clinical experts cannot, creates an enormous risk for all model deployments in medical imaging. FUNDING: National Institute of Biomedical Imaging and Bioengineering, MIDRC grant of National Institutes of Health, US National Science Foundation, National Library of Medicine of the National Institutes of Health, and Taiwan Ministry of Science and Technology.

Rad18 mediates specific mutational signatures and shapes the genomic landscape of carcinogen-induced tumors in vivo.

The E3 ubiquitin ligase Rad18 promotes a damage-tolerant and error-prone mode of DNA replication termed trans-lesion synthesis that is pathologically activated in cancer. However, the impact of vertebrate Rad18 on cancer genomes is not known. To determine how Rad18 affects mutagenesis in vivo, we have developed and implemented a novel computational pipeline to analyze genomes of carcinogen (7, 12-Dimethylbenz[a]anthracene, DMBA)-induced skin tumors from Rad18+/+ and Rad18- / - mice. We show that Rad18 mediates specific mutational signatures characterized by high levels of A(T)>T(A) single nucleotide variations (SNVs). In Rad18- /- tumors, an alternative mutation pattern arises, which is characterized by increased numbers of deletions >4 bp. Comparison with annotated human mutational signatures shows that COSMIC signature 22 predominates in Rad18+/+ tumors whereas Rad18- / - tumors are characterized by increased contribution of COSMIC signature 3 (a hallmark of BRCA-mutant tumors). Analysis of The Cancer Genome Atlas shows that RAD18 expression is strongly associated with high SNV burdens, suggesting RAD18 also promotes mutagenesis in human cancers. Taken together, our results show Rad18 promotes mutagenesis in vivo, modulates DNA repair pathway choice in neoplastic cells, and mediates specific mutational signatures that are present in human tumors.

Genetic determinants of cellular addiction to DNA polymerase theta.

Polymerase theta (Pol θ, gene name Polq) is a widely conserved DNA polymerase that mediates a microhomology-mediated, error-prone, double strand break (DSB) repair pathway, referred to as Theta Mediated End Joining (TMEJ). Cells with homologous recombination deficiency are reliant on TMEJ for DSB repair. It is unknown whether deficiencies in other components of the DNA damage response (DDR) also result in Pol θ addiction. Here we use a CRISPR genetic screen to uncover 140 Polq synthetic lethal (PolqSL) genes, the majority of which were previously unknown. Functional analyses indicate that Pol θ/TMEJ addiction is associated with increased levels of replication-associated DSBs, regardless of the initial source of damage. We further demonstrate that approximately 30% of TCGA breast cancers have genetic alterations in PolqSL genes and exhibit genomic scars of Pol θ/TMEJ hyperactivity, thereby substantially expanding the subset of human cancers for which Pol θ inhibition represents a promising therapeutic strategy.

Sensitivity of magnetic resonance imaging for detection of patellofemoral articular cartilage defects.

PURPOSE: Chondral defects within the patellofemoral compartment are common and lack the ability to heal on their own. Early detection of these lesions with a noninvasive modality would be beneficial in delaying or preventing their possible progression to osteoarthritis. We hypothesized that magnetic resonance imaging (MRI) is a sensitive, specific, and accurate imaging modality for the detection of patellofemoral chondral defects with substantial interobserver reliability and that MRI has a higher sensitivity, specificity, and accuracy for detecting patellar defects than trochlear defects. METHODS: A systematic review of multiple medical databases was performed by use of the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocol. Analysis of studies that reported diagnostic performance of MRI in the assessment of patellofemoral chondral defects (patella and trochlea), using arthroscopy as the reference gold standard, was performed. Sensitivity, specificity, accuracy, and interobserver reliability were reported. Significant heterogeneity across studies precluded meta-analysis. RESULTS: MRI was more sensitive in detection of patellar (87%) versus trochlear (72%) defects. MRI was similarly specific for patellar (86%) and trochlear (89%) defects. MRI was similarly accurate for patellar (84%) and trochlear (83%) defects. Interobserver agreement was substantial to almost perfect for both patellar and trochlear defects. CONCLUSIONS: MRI is a highly sensitive, specific, and accurate noninvasive diagnostic modality for the detection of chondral defects in the patellofemoral compartment of the knee, using arthroscopy as the reference gold standard. Although there was wide variability in the statistical parameters assessed, MRI was more sensitive for detection of patellar versus trochlear defects and similarly specific and accurate for patellar and trochlear defects. Interobserver reliability is substantial to near perfect in the assessment of these lesions, without a significant difference between patellar and trochlear defects. CLINICAL RELEVANCE: Use of MRI may allow early detection of chondral defects within the patellofemoral compartment, enabling clinicians to adopt strategies to delay or prevent progression to osteoarthritis. LEVEL OF EVIDENCE: Level III, systematic review of Level I, II, and III studies.

Association of apolipoprotein E genotype and Alzheimer disease in African Americans.

BACKGROUND: Alzheimer disease (AD) is the most frequent cause of dementia. Even though the incidence of AD in the African American population is similar to or higher than that in persons of European descent, AD in African Americans is understudied. Identification of genetic risk factors in African Americans is essential for understanding the etiology of AD. OBJECTIVE: To determine the effect of apolipoprotein E (APOE) genotype on the risk of AD in elderly African Americans. DESIGN: Population-based longitudinal study of AD. SETTING: Indianapolis, Ind. PARTICIPANTS: African Americans 65 years and older. MAIN OUTCOME MEASURES: APOE genotype and diagnosis of AD. RESULTS: The APOE genotype was determined in 1822 samples. Of these, 690 were clinically evaluated: 318 were normal, and 162 had a diagnosis of AD. The presence of APOE epsilon4 was significantly associated with increased risk of AD (epsilon3/epsilon4: OR, 2.32; 95% confidence interval [CI], 1.41-3.82; and epsilon4/epsilon4: OR, 7.19; 95% CI, 3.00-17.29, compared with the epsilon3/epsilon3 genotype). There was also a significant protective effect with APOE epsilon2 (epsilon2/epsilon2 and epsilon2/epsilon3: OR, 0.42; 95% CI, 0.20-0.89). CONCLUSIONS: These findings are in marked contrast to the lack of association between APOE and AD in the Ibadan, Nigeria, sample of this project. These results suggest that other genetic factors and different environmental influences may play a role in the risk for AD in individuals of African ancestry.

APOE epsilon4 is not associated with Alzheimer's disease in elderly Nigerians.

Since 1992, research teams from Indiana University and the University of Ibadan have been collecting and comparing data from two diverse, elderly populations to identify risk factors for dementia and Alzheimer's disease. Apolipoprotein E (APOE) was genotyped in 2,245 Nigerian samples. Of these, 830 had a diagnosis: 459 were normal, and 140 had dementia including 123 diagnosed with Alzheimer's disease. In contrast with other populations, the APOE epsilon4 allele was not significantly associated with Alzheimer's disease or dementia. This lack of association in the Yoruba might reflect genetic variation, environmental factors, as well as genetic/environmental interactions.

Conference on plant-made pharmaceuticals. 16-19 March 2003, Québec City, Québec, Canada.

The Conference on Plant-Made Pharmaceuticals was an inaugural event organized by Molecular Farming Association Inc, a non-profit organization created in 2000 to support the emergence of plant-made biopharmaceuticals and plant-factory companies. The meeting was sponsored by the Government of Canada, the Gouvermement du Québec and the Société Générale de Financement, along with 20 companies involved in plant-made pharmaceuticals and related activities. Although there was very little discussion of new biopharmaceuticals under development with these systems, this was the first meeting where participants could survey the entire breadth of technologies and approaches that are being taken to produce biopharmaceuticals in transgenic plants. Participants also heard about the technical, manufacturing and regulatory issues confronting transgenic plant expression systems, as well as from 'new technology' companies that hav not previously presented in a public forum. There was a clear sense emerging from the meeting that the hurdles currently facing the industry will be overcome and that transgenic plant systems will eventually move into the mainstream of biopharmaceutical manufacturing technologies.