Evaluation of large language models as a diagnostic aid for complex medical cases.

Background: The use of large language models (LLM) has recently gained popularity in diverse areas, including answering questions posted by patients as well as medical professionals. Objective: To evaluate the performance and limitations of LLMs in providing the correct diagnosis for a complex clinical case. Design: Seventy-five consecutive clinical cases were selected from the Massachusetts General Hospital Case Records, and differential diagnoses were generated by OpenAI's GPT3.5 and 4 models. Results: The mean number of diagnoses provided by the Massachusetts General Hospital case discussants was 16.77, by GPT3.5 30 and by GPT4 15.45 (p < 0.0001). GPT4 was more frequently able to list the correct diagnosis as first (22% versus 20% with GPT3.5, p = 0.86), provide the correct diagnosis among the top three generated diagnoses (42% versus 24%, p = 0.075). GPT4 was better at providing the correct diagnosis, when the different diagnoses were classified into groups according to the medical specialty and include the correct diagnosis at any point in the differential list (68% versus 48%, p = 0.0063). GPT4 provided a differential list that was more similar to the list provided by the case discussants than GPT3.5 (Jaccard Similarity Index 0.22 versus 0.12, p = 0.001). Inclusion of the correct diagnosis in the generated differential was correlated with PubMed articles matching the diagnosis (OR 1.40, 95% CI 1.25-1.56 for GPT3.5, OR 1.25, 95% CI 1.13-1.40 for GPT4), but not with disease incidence. Conclusions and relevance: The GPT4 model was able to generate a differential diagnosis list with the correct diagnosis in approximately two thirds of cases, but the most likely diagnosis was often incorrect for both models. In its current state, this tool can at most be used as an aid to expand on potential diagnostic considerations for a case, and future LLMs should be trained which account for the discrepancy between disease incidence and availability in the literature.

Characterizing the Increase in Artificial Intelligence Content Detection in Oncology Scientific Abstracts From 2021 to 2023.

PURPOSE: Artificial intelligence (AI) models can generate scientific abstracts that are difficult to distinguish from the work of human authors. The use of AI in scientific writing and performance of AI detection tools are poorly characterized. METHODS: We extracted text from published scientific abstracts from the ASCO 2021-2023 Annual Meetings. Likelihood of AI content was evaluated by three detectors: GPTZero, Originality.ai, and Sapling. Optimal thresholds for AI content detection were selected using 100 abstracts from before 2020 as negative controls, and 100 produced by OpenAI's GPT-3 and GPT-4 models as positive controls. Logistic regression was used to evaluate the association of predicted AI content with submission year and abstract characteristics, and adjusted odds ratios (aORs) were computed. RESULTS: Fifteen thousand five hundred and fifty-three abstracts met inclusion criteria. Across detectors, abstracts submitted in 2023 were significantly more likely to contain AI content than those in 2021 (aOR range from 1.79 with Originality to 2.37 with Sapling). Online-only publication and lack of clinical trial number were consistently associated with AI content. With optimal thresholds, 99.5%, 96%, and 97% of GPT-3/4-generated abstracts were identified by GPTZero, Originality, and Sapling respectively, and no sampled abstracts from before 2020 were classified as AI generated by the GPTZero and Originality detectors. Correlation between detectors was low to moderate, with Spearman correlation coefficient ranging from 0.14 for Originality and Sapling to 0.47 for Sapling and GPTZero. CONCLUSION: There is an increasing signal of AI content in ASCO abstracts, coinciding with the growing popularity of generative AI models.

Artificial intelligence-based morphologic classification and molecular characterization of neuroblastic tumors from digital histopathology.

A deep learning model using attention-based multiple instance learning (aMIL) and self-supervised learning (SSL) was developed to perform pathologic classification of neuroblastic tumors and assess MYCN-amplification status using H&E-stained whole slide digital images. The model demonstrated strong performance in identifying diagnostic category, grade, mitosis-karyorrhexis index (MKI), and MYCN-amplification on an external test dataset. This AI-based approach establishes a valuable tool for automating diagnosis and precise classification of neuroblastoma tumors.

Neoadjuvant Nivolumab Plus Chemotherapy Followed By Response-Adaptive Therapy for HPV+ Oropharyngeal Cancer: OPTIMA II Phase 2 Open-Label Nonrandomized Clinical Trial.

Importance: Immune checkpoint inhibitors improve survival in recurrent and/or metastatic head and neck cancer, yet their role in curative human papillomavirus-positive oropharyngeal cancer (HPV+ OPC) remains undefined. Neoadjuvant nivolumab and chemotherapy followed by response-adaptive treatment in HPV+ OPC may increase efficacy while reducing toxicity. Objective: To determine the deep response rate and tolerability of the addition of neoadjuvant nivolumab to chemotherapy followed by response-adapted locoregional therapy (LRT) in patients with HPV+ OPC. Design, Setting, and Participants: This phase 2 nonrandomized clinical trial conducted at a single academic center enrolled 77 patients with locoregionally advanced HPV+ OPC from 2017 to 2020. Data analyses were performed from February 10, 2021, to January 9, 2023. Interventions: Addition of nivolumab to neoadjuvant nab-paclitaxel and carboplatin (studied in the first OPTIMA trial) followed by response-adapted LRT in patients with HPV+ OPC stages III to IV. Main Outcomes and Measures: Primary outcome was deep response rate to neoadjuvant nivolumab plus chemotherapy, defined as the proportion of tumors with 50% or greater shrinkage per the Response Evaluation Criteria in Solid Tumors 1.1. Secondary outcomes were progression-free survival (PFS) and overall survival (OS). Swallowing function, quality of life, and tissue- and blood-based biomarkers, including programmed death-ligand 1 (PD-L1) expression and circulating tumor HPV-DNA (ctHPV-DNA), were also evaluated. Results: The 73 eligible patients (median [range] age, 61 [37-82] years; 6 [8.2%] female; 67 [91.8%] male) started neoadjuvant nivolumab and chemotherapy. Deep responses were observed in 51 patients (70.8%; 95% CI, 0.59-0.81). Subsequent risk- and response-adaptive therapy was assigned as follows: group A, single-modality radiotherapy alone or transoral robotic surgery (28 patients); group B, intermediate-dose chemoradiotherapy of 45 to 50 Gray (34 patients); and group C, regular-dose chemoradiotherapy of 70 to 75 Gray (10 patients). Two-year PFS and OS were 90.0% (95% CI, 0.80-0.95) and 91.4% (95% CI, 0.82-0.96), respectively. By response-adapted group, 2-year PFS and OS for group A were 96.4% and 96.4%, and group B, 88.0% and 91.0%, respectively. Lower enteral feeding rates and changes in weight, as well as improved swallowing, were observed among patients who received response-adapted LRT. Pathologic complete response rate among patients who underwent transoral robotic surgery was 67.0%. PD-L1 expression was nonsignificantly higher for deeper responses and improved PFS, and ctHPV-DNA clearance was significantly associated with improved PFS. Conclusions and Relevance: This phase 2 nonrandomized clinical trial found that neoadjuvant nivolumab and chemotherapy followed by response-adapted LRT is feasible and has favorable tolerability, excellent OS, and improved functional outcomes in HPV+ OPC, including among patients with high-risk disease. Moreover, addition of nivolumab may benefit high PD-L1 expressors, and sensitive dynamic biomarkers (eg, ctHPV-DNA) are useful for patient selection. Trial Registration: ClinicalTrials.gov Identifier: NCT03107182.

Artificial intelligence-based epigenomic, transcriptomic and histologic signatures of tobacco use in oral squamous cell carcinoma.

Oral squamous cell carcinoma (OSCC) biomarker studies rarely employ multi-omic biomarker strategies and pertinent clinicopathologic characteristics to predict mortality. In this study we determine for the first time a combined epigenetic, gene expression, and histology signature that differentiates between patients with different tobacco use history (heavy tobacco use with ≥10 pack years vs. no tobacco use). Using The Cancer Genome Atlas (TCGA) cohort (n = 257) and an internal cohort (n = 40), we identify 3 epigenetic markers (GPR15, GNG12, GDNF) and 13 expression markers (IGHA2, SCG5, RPL3L, NTRK1, CD96, BMP6, TFPI2, EFEMP2, RYR3, DMTN, GPD2, BAALC, and FMO3), which are dysregulated in OSCC patients who were never smokers vs. those who have a ≥ 10 pack year history. While mortality risk prediction based on smoking status and clinicopathologic covariates alone is inaccurate (c-statistic = 0.57), the combined epigenetic/expression and histologic signature has a c-statistic = 0.9409 in predicting 5-year mortality in OSCC patients.

Development and validation of a clinical breast cancer tool for accurate prediction of recurrence.

Given high costs of Oncotype DX (ODX) testing, widely used in recurrence risk assessment for early-stage breast cancer, studies have predicted ODX using quantitative clinicopathologic variables. However, such models have incorporated only small cohorts. Using a cohort of patients from the National Cancer Database (NCDB, n = 53,346), we trained machine learning models to predict low-risk (0-25) or high-risk (26-100) ODX using quantitative estrogen receptor (ER)/progesterone receptor (PR)/Ki-67 status, quantitative ER/PR status alone, and no quantitative features. Models were externally validated on a diverse cohort of 970 patients (median follow-up 55 months) for accuracy in ODX prediction and recurrence. Comparing the area under the receiver operating characteristic curve (AUROC) in a held-out set from NCDB, models incorporating quantitative ER/PR (AUROC 0.78, 95% CI 0.77-0.80) and ER/PR/Ki-67 (AUROC 0.81, 95% CI 0.80-0.83) outperformed the non-quantitative model (AUROC 0.70, 95% CI 0.68-0.72). These results were preserved in the validation cohort, where the ER/PR/Ki-67 model (AUROC 0.87, 95% CI 0.81-0.93, p = 0.009) and the ER/PR model (AUROC 0.86, 95% CI 0.80-0.92, p = 0.031) significantly outperformed the non-quantitative model (AUROC 0.80, 95% CI 0.73-0.87). Using a high-sensitivity rule-out threshold, the non-quantitative, quantitative ER/PR and ER/PR/Ki-67 models identified 35%, 30% and 43% of patients as low-risk in the validation cohort. Of these low-risk patients, fewer than 3% had a recurrence at 5 years. These models may help identify patients who can forgo genomic testing and initiate endocrine therapy alone. An online calculator is provided for further study.

Acquired resistance to immunotherapy and chemoradiation in MYC amplified head and neck cancer.

The proto-oncogene MYC encodes a nuclear transcription factor that has an important role in a variety of cellular processes, such as cell cycle progression, proliferation, metabolism, adhesion, apoptosis, and therapeutic resistance. MYC amplification is consistently observed in aggressive forms of several solid malignancies and correlates with poor prognosis and distant metastases. While the tumorigenic effects of MYC in patients with head and neck squamous cell carcinoma (HNSCC) are well known, the molecular mechanisms by which the amplification of this gene may confer treatment resistance, especially to immune checkpoint inhibitors, remains under-investigated. Here we present a unique case of a patient with recurrent/metastatic (R/M) HNSCC who, despite initial response to nivolumab-based treatment, developed rapidly progressive metastatic disease after the acquisition of MYC amplification. We conducted comparative transcriptomic analysis of this patient's tumor at baseline and upon progression to interrogate potential molecular processes through which MYC may confer resistance to immunotherapy and/or chemoradiation and used TCGA-HNSC dataset and an institutional cohort to further explore clinicopathologic features and key molecular networks associated with MYC amplification in HNSCC. This study highlights MYC amplification as a potential mechanism of immune checkpoint inhibitor resistance and suggest its use as a predictive biomarker and potential therapeutic target in R/M HNSCC.

Generative Adversarial Networks Accurately Reconstruct Pan-Cancer Histology from Pathologic, Genomic, and Radiographic Latent Features.

Artificial intelligence models have been increasingly used in the analysis of tumor histology to perform tasks ranging from routine classification to identification of novel molecular features. These approaches distill cancer histologic images into high-level features which are used in predictions, but understanding the biologic meaning of such features remains challenging. We present and validate a custom generative adversarial network - HistoXGAN - capable of reconstructing representative histology using feature vectors produced by common feature extractors. We evaluate HistoXGAN across 29 cancer subtypes and demonstrate that reconstructed images retain information regarding tumor grade, histologic subtype, and gene expression patterns. We leverage HistoXGAN to illustrate the underlying histologic features for deep learning models for actionable mutations, identify model reliance on histologic batch effect in predictions, and demonstrate accurate reconstruction of tumor histology from radiographic imaging for a 'virtual biopsy'.

Using histopathology latent diffusion models as privacy-preserving dataset augmenters improves downstream classification performance.

Latent diffusion models (LDMs) have emerged as a state-of-the-art image generation method, outperforming previous Generative Adversarial Networks (GANs) in terms of training stability and image quality. In computational pathology, generative models are valuable for data sharing and data augmentation. However, the impact of LDM-generated images on histopathology tasks compared to traditional GANs has not been systematically studied. We trained three LDMs and a styleGAN2 model on histology tiles from nine colorectal cancer (CRC) tissue classes. The LDMs include 1) a fine-tuned version of stable diffusion v1.4, 2) a Kullback-Leibler (KL)-autoencoder (KLF8-DM), and 3) a vector quantized (VQ)-autoencoder deploying LDM (VQF8-DM). We assessed image quality through expert ratings, dimensional reduction methods, distribution similarity measures, and their impact on training a multiclass tissue classifier. Additionally, we investigated image memorization in the KLF8-DM and styleGAN2 models. All models provided a high image quality, with the KLF8-DM achieving the best Frechet Inception Distance (FID) and expert rating scores for complex tissue classes. For simpler classes, the VQF8-DM and styleGAN2 models performed better. Image memorization was negligible for both styleGAN2 and KLF8-DM models. Classifiers trained on a mix of KLF8-DM generated and real images achieved a 4% improvement in overall classification accuracy, highlighting the usefulness of these images for dataset augmentation. Our systematic study of generative methods showed that KLF8-DM produces the highest quality images with negligible image memorization. The higher classifier performance in the generatively augmented dataset suggests that this augmentation technique can be employed to enhance histopathology classifiers for various tasks.

Dysregulated FGFR3 signaling alters the immune landscape in bladder cancer and presents therapeutic possibilities in an agent-based model.

Bladder cancer is an increasingly prevalent global disease that continues to cause morbidity and mortality despite recent advances in treatment. Immune checkpoint inhibitors (ICI) and fibroblast growth factor receptor (FGFR)-targeted therapeutics have had modest success in bladder cancer when used as monotherapy. Emerging data suggests that the combination of these two therapies could lead to improved clinical outcomes, but the optimal strategy for combining these agents remains uncertain. Mathematical models, specifically agent-based models (ABMs), have shown recent successes in uncovering the multiscale dynamics that shape the trajectory of cancer. They have enabled the optimization of treatment methods and the identification of novel therapeutic strategies. To assess the combined effects of anti-PD-1 and anti-FGFR3 small molecule inhibitors (SMI) on tumor growth and the immune response, we built an ABM that captures key facets of tumor heterogeneity and CD8+ T cell phenotypes, their spatial interactions, and their response to therapeutic pressures. Our model quantifies how tumor antigenicity and FGFR3 activating mutations impact disease trajectory and response to anti-PD-1 antibodies and anti-FGFR3 SMI. We find that even a small population of weakly antigenic tumor cells bearing an FGFR3 mutation can render the tumor resistant to combination therapy. However, highly antigenic tumors can overcome therapeutic resistance mediated by FGFR3 mutation. The optimal therapy depends on the strength of the FGFR3 signaling pathway. Under certain conditions, ICI alone is optimal; in others, ICI followed by anti-FGFR3 therapy is best. These results indicate the need to quantify FGFR3 signaling and the fitness advantage conferred on bladder cancer cells harboring this mutation. This ABM approach may enable rationally designed treatment plans to improve clinical outcomes.

Slideflow: deep learning for digital histopathology with real-time whole-slide visualization.

Deep learning methods have emerged as powerful tools for analyzing histopathological images, but current methods are often specialized for specific domains and software environments, and few open-source options exist for deploying models in an interactive interface. Experimenting with different deep learning approaches typically requires switching software libraries and reprocessing data, reducing the feasibility and practicality of experimenting with new architectures. We developed a flexible deep learning library for histopathology called Slideflow, a package which supports a broad array of deep learning methods for digital pathology and includes a fast whole-slide interface for deploying trained models. Slideflow includes unique tools for whole-slide image data processing, efficient stain normalization and augmentation, weakly-supervised whole-slide classification, uncertainty quantification, feature generation, feature space analysis, and explainability. Whole-slide image processing is highly optimized, enabling whole-slide tile extraction at 40x magnification in 2.5 s per slide. The framework-agnostic data processing pipeline enables rapid experimentation with new methods built with either Tensorflow or PyTorch, and the graphical user interface supports real-time visualization of slides, predictions, heatmaps, and feature space characteristics on a variety of hardware devices, including ARM-based devices such as the Raspberry Pi.

Evolutionary dynamics of tipifarnib in HRAS mutated head and neck squamous cell carcinoma.

Head and neck squamous cell carcinoma (HNSCC) is a highly prevalent malignancy worldwide, with a significant proportion of patients developing recurrent and/or metastatic (R/M) disease. Despite recent advances in therapy, the prognosis for patients with advanced HNSCC remains poor. Here, we present the case of a patient with recurrent metastatic HNSCC harboring an HRAS G12S mutation who achieved a durable response to treatment with tipifarnib, a selective inhibitor of farnesyltransferase. The patient was a 48-year-old woman who had previously received multiple lines of therapy with no significant clinical response. However, treatment with tipifarnib resulted in a durable partial response that lasted 8 months. Serial genomic and transcriptomic analyses demonstrated upregulation of YAP1 and AXL in metastatic lesions compared with the primary tumor, the evolution of the tumor microenvironment from an immune-enriched to a fibrotic subtype with increased angiogenesis, and activation of the PI3K/AKT/mTOR pathway in tipifarnib treatment. Lastly, in HRAS-mutated PDXs and in the syngeneic HRAS model, we demonstrated that tipifarnib efficacy is limited by activation of the AKT pathway, and dual treatment with tipifarnib and the PI3K inhibitor, BYL719, resulted in enhanced anti-tumor efficacy. Our case study highlights the potential of targeting HRAS mutations with tipifarnib in R/M HNSCC and identifies potential mechanisms of acquired resistance to tipifarnib, along with immuno-, chemo-, and radiation therapy. Preclinical results provide a firm foundation for further investigation of drug combinations of HRAS-and PI3K -targeting therapeutics in R/M HRAS-driven HNSCC.

Validation of the RSClin risk calculator in the National Cancer Data Base.

BACKGROUND: Guidelines recommend the use of genomic assays such as OncotypeDx to aid in decisions regarding the use of chemotherapy for hormone receptor-positive, HER2-negative (HR+/HER2-) breast cancer. The RSClin prognostic tool integrates OncotypeDx and clinicopathologic features to predict distant recurrence and chemotherapy benefit, but further validation is needed before broad clinical adoption. METHODS: This study included patients from the National Cancer Data Base (NCDB) who were diagnosed with stage I-III HR+/HER2- breast cancer from 2010 to 2020 and received adjuvant endocrine therapy with or without chemotherapy. RSClin-predicted chemotherapy benefit was stratified into low (<3% reduction in distant recurrence), intermediate (3%-5%), and high (>5%). Cox models were used to model mortality adjusted for age, comorbidity index, insurance, and race/ethnicity. RESULTS: A total of 285,441 patients were identified for inclusion from the NCDB, with an average age of 60 years and a median follow-up of 58 months. Chemotherapy was associated with improved overall survival only for those predicted to have intermediate (adjusted hazard ratio [aHR], 0.68; 95% confidence interval [CI], 0.60-0.79) and high benefit per RSClin (aHR, 0.66; 95% CI, 0.61-0.72). Consistent benefit was seen in the subset with a low OncotypeDx score (<26) and intermediate (aHR, 0.66; 95% CI, 0.53-0.82) or high (aHR, 0.71; 95% CI, 0.58-0.86) RSClin-predicted benefit. No survival benefit with chemotherapy was seen in patients with a high OncotypeDx score (≥26) and low benefit per RSClin (aHR, 1.70; 95% CI, 0.41-6.99). CONCLUSIONS: RSClin may identify high-risk patients who benefit from treatment intensification more accurately than OncotypeDx, and further prospective study is needed.

External Evaluation of a Mammography-based Deep Learning Model for Predicting Breast Cancer in an Ethnically Diverse Population.

Purpose: To externally evaluate a mammography-based deep learning (DL) model (Mirai) in a high-risk racially diverse population and compare its performance with other mammographic measures. Materials and Methods: A total of 6435 screening mammograms in 2096 female patients (median age, 56.4 years ± 11.2 [SD]) enrolled in a hospital-based case-control study from 2006 to 2020 were retrospectively evaluated. Pathologically confirmed breast cancer was the primary outcome. Mirai scores were the primary predictors. Breast density and Breast Imaging Reporting and Data System (BI-RADS) assessment categories were comparative predictors. Performance was evaluated using area under the receiver operating characteristic curve (AUC) and concordance index analyses. Results: Mirai achieved 1- and 5-year AUCs of 0.71 (95% CI: 0.68, 0.74) and 0.65 (95% CI: 0.64, 0.67), respectively. One-year AUCs for nondense versus dense breasts were 0.72 versus 0.58 (P = .10). There was no evidence of a difference in near-term discrimination performance between BI-RADS and Mirai (1-year AUC, 0.73 vs 0.68; P = .34). For longer-term prediction (2-5 years), Mirai outperformed BI-RADS assessment (5-year AUC, 0.63 vs 0.54; P < .001). Using only images of the unaffected breast reduced the discriminatory performance of the DL model (P < .001 at all time points), suggesting that its predictions are likely dependent on the detection of ipsilateral premalignant patterns. Conclusion: A mammography DL model showed good performance in a high-risk external dataset enriched for African American patients, benign breast disease, and BRCA mutation carriers, and study findings suggest that the model performance is likely driven by the detection of precancerous changes.Keywords: Breast, Cancer, Computer Applications, Convolutional Neural Network, Deep Learning Algorithms, Informatics, Epidemiology, Machine Learning, Mammography, Oncology, Radiomics Supplemental material is available for this article. © RSNA, 2023See also commentary by Kontos and Kalpathy-Cramer in this issue.

Mathematical model predicts tumor control patterns induced by fast and slow cytotoxic T lymphocyte killing mechanisms.

Immunotherapy has dramatically transformed the cancer treatment landscape largely due to the efficacy of immune checkpoint inhibitors (ICIs). Although ICIs have shown promising results for many patients, the low response rates in many cancers highlight the ongoing challenges in cancer treatment. Cytotoxic T lymphocytes (CTLs) execute their cell-killing function via two distinct mechanisms: a fast-acting, perforin-mediated process and a slower, Fas ligand (FasL)-driven pathway. Evidence also suggests that the preferred killing mechanism of CTLs depends on the antigenicity of tumor cells. To determine the critical factors affecting responses to ICIs, we construct an ordinary differential equation model describing in vivo tumor-immune dynamics in the presence of active or blocked PD-1/PD-L1 immune checkpoint. Specifically, we identify important aspects of the tumor-immune landscape that affect tumor size and composition in the short and long term. We also generate a virtual cohort of mice with diverse tumor and immune attributes to simulate the outcomes of immune checkpoint blockade in a heterogeneous population. By identifying key tumor and immune characteristics associated with tumor elimination, dormancy, and escape, we predict which fraction of a population potentially responds well to ICIs and ways to enhance therapeutic outcomes with combination therapy.

The use of deep learning state-of-the-art architectures for oral epithelial dysplasia grading: A comparative appraisal.

BACKGROUND: Dysplasia grading systems for oral epithelial dysplasia are a source of disagreement among pathologists. Therefore, machine learning approaches are being developed to mitigate this issue. METHODS: This cross-sectional study included a cohort of 82 patients with oral potentially malignant disorders and correspondent 98 hematoxylin and eosin-stained whole slide images with biopsied-proven dysplasia. All whole-slide images were manually annotated based on the binary system for oral epithelial dysplasia. The annotated regions of interest were segmented and fragmented into small patches and non-randomly sampled into training/validation and test subsets. The training/validation data were color augmented, resulting in a total of 81,786 patches for training. The held-out independent test set enrolled a total of 4,486 patches. Seven state-of-the-art convolutional neural networks were trained, validated, and tested with the same dataset. RESULTS: The models presented a high learning rate, yet very low generalization potential. At the model development, VGG16 performed the best, but with massive overfitting. In the test set, VGG16 presented the best accuracy, sensitivity, specificity, and area under the curve (62%, 62%, 66%, and 65%, respectively), associated with the higher loss among all Convolutional Neural Networks (CNNs) tested. EfficientB0 has comparable metrics and the lowest loss among all convolutional neural networks, being a great candidate for further studies. CONCLUSION: The models were not able to generalize enough to be applied in real-life datasets due to an overlapping of features between the two classes (i.e., high risk and low risk of malignization).

Longitudinal Patient Outcomes in Chronic Dizziness: A Scoping Review.

BACKGROUND: Chronic dizziness can cause significant functional impairment. Outcome measures used in this patient population have not been examined systematically. Consequently, providers lack consensus on the ideal outcome measures to assess the impact of their interventions. OBJECTIVE AND METHODS: We conducted a scoping review to summarize existing literature on outcomes in chronic dizziness (with a minimum of 6 mo of patient follow-up). Among other details, we extracted and analyzed patient demographics, medical condition(s), and the specific outcome measures of each study. RESULTS: Of 19,426 articles meeting the original search terms, 416 met final exclusion after title/abstract and full-text review. Most studies focused on Ménière's disease (75%) and recurrent benign paroxysmal positional vertigo (21%). The most common outcome measures were hearing (62%) and number of attacks by American Academy of Otolaryngology-Head & Neck Surgery criteria (60%). A minority (35%) looked formally at quality-of-life metrics (Dizziness Handicap Index or other). CONCLUSIONS: Ménière's disease and benign paroxysmal positional vertigo are overrepresented in literature on outcome assessment in chronic dizziness. Objective clinical measures are used more frequently than quality-of-life metrics. Future work is needed to identify the optimal outcome measures that reflect new knowledge about the most common causes of chronic dizziness (including persistent postural-perceptual dizziness and vestibular migraine) and consider what is most important to patients.

Deep learning-based subtyping of gastric cancer histology predicts clinical outcome: a multi-institutional retrospective study.

INTRODUCTION: The Laurén classification is widely used for Gastric Cancer (GC) histology subtyping. However, this classification is prone to interobserver variability and its prognostic value remains controversial. Deep Learning (DL)-based assessment of hematoxylin and eosin (H&E) stained slides is a potentially useful tool to provide an additional layer of clinically relevant information, but has not been systematically assessed in GC. OBJECTIVE: We aimed to train, test and externally validate a deep learning-based classifier for GC histology subtyping using routine H&E stained tissue sections from gastric adenocarcinomas and to assess its potential prognostic utility. METHODS: We trained a binary classifier on intestinal and diffuse type GC whole slide images for a subset of the TCGA cohort (N = 166) using attention-based multiple instance learning. The ground truth of 166 GC was obtained by two expert pathologists. We deployed the model on two external GC patient cohorts, one from Europe (N = 322) and one from Japan (N = 243). We assessed classification performance using the Area Under the Receiver Operating Characteristic Curve (AUROC) and prognostic value (overall, cancer specific and disease free survival) of the DL-based classifier with uni- and multivariate Cox proportional hazard models and Kaplan-Meier curves with log-rank test statistics. RESULTS: Internal validation using the TCGA GC cohort using five-fold cross-validation achieved a mean AUROC of 0.93 ± 0.07. External validation showed that the DL-based classifier can better stratify GC patients' 5-year survival compared to pathologist-based Laurén classification for all survival endpoints, despite frequently divergent model-pathologist classifications. Univariate overall survival Hazard Ratios (HRs) of pathologist-based Laurén classification (diffuse type versus intestinal type) were 1.14 (95% Confidence Interval (CI) 0.66-1.44, p-value = 0.51) and 1.23 (95% CI 0.96-1.43, p-value = 0.09) in the Japanese and European cohorts, respectively. DL-based histology classification resulted in HR of 1.46 (95% CI 1.18-1.65, p-value < 0.005) and 1.41 (95% CI 1.20-1.57, p-value < 0.005), in the Japanese and European cohorts, respectively. In diffuse type GC (as defined by the pathologist), classifying patients using the DL diffuse and intestinal classifications provided a superior survival stratification, and demonstrated statistically significant survival stratification when combined with pathologist classification for both the Asian (overall survival log-rank test p-value < 0.005, HR 1.43 (95% CI 1.05-1.66, p-value = 0.03) and European cohorts (overall survival log-rank test p-value < 0.005, HR 1.56 (95% CI 1.16-1.76, p-value < 0.005)). CONCLUSION: Our study shows that gastric adenocarcinoma subtyping using pathologist's Laurén classification as ground truth can be performed using current state of the art DL techniques. Patient survival stratification seems to be better by DL-based histology typing compared with expert pathologist histology typing. DL-based GC histology typing has potential as an aid in subtyping. Further investigations are warranted to fully understand the underlying biological mechanisms for the improved survival stratification despite apparent imperfect classification by the DL algorithm.

Artificial Intelligence in Clinical Oncology: From Data to Digital Pathology and Treatment.

Recently, a wide spectrum of artificial intelligence (AI)-based applications in the broader categories of digital pathology, biomarker development, and treatment have been explored. In the domain of digital pathology, these have included novel analytical strategies for realizing new information derived from standard histology to guide treatment selection and biomarker development to predict treatment selection and response. In therapeutics, these have included AI-driven drug target discovery, drug design and repurposing, combination regimen optimization, modulated dosing, and beyond. Given the continued advances that are emerging, it is important to develop workflows that seamlessly combine the various segments of AI innovation to comprehensively augment the diagnostic and interventional arsenal of the clinical oncology community. To overcome challenges that remain with regard to the ideation, validation, and deployment of AI in clinical oncology, recommendations toward bringing this workflow to fruition are also provided from clinical, engineering, implementation, and health care economics considerations. Ultimately, this work proposes frameworks that can potentially integrate these domains toward the sustainable adoption of practice-changing AI by the clinical oncology community to drive improved patient outcomes.