Clinical evaluation of deep learning-based risk profiling in breast cancer histopathology and comparison to an established multigene assay.

PURPOSE: To evaluate the Stratipath Breast tool for image-based risk profiling and compare it with an established prognostic multigene assay for risk profiling in a real-world case series of estrogen receptor (ER)-positive and human epidermal growth factor receptor 2 (HER2)-negative early breast cancer patients categorized as intermediate risk based on classic clinicopathological variables and eligible for chemotherapy. METHODS: In a case series comprising 234 invasive ER-positive/HER2-negative tumors, clinicopathological data including Prosigna results and corresponding HE-stained tissue slides were retrieved. The digitized HE slides were analysed by Stratipath Breast. RESULTS: Our findings showed that the Stratipath Breast analysis identified 49.6% of the clinically intermediate tumors as low risk and 50.4% as high risk. The Prosigna assay classified 32.5%, 47.0% and 20.5% tumors as low, intermediate and high risk, respectively. Among Prosigna intermediate-risk tumors, 47.3% were stratified as Stratipath low risk and 52.7% as high risk. In addition, 89.7% of Stratipath low-risk cases were classified as Prosigna low/intermediate risk. The overall agreement between the two tests for low-risk and high-risk groups (N = 124) was 71.0%, with a Cohen's kappa of 0.42. For both risk profiling tests, grade and Ki67 differed significantly between risk groups. CONCLUSION: The results from this clinical evaluation of image-based risk stratification shows a considerable agreement to an established gene expression assay in routine breast pathology.

Comparing deep learning and pathologist quantification of cell-level PD-L1 expression in non-small cell lung cancer whole-slide images.

Programmed death-ligand 1 (PD-L1) expression is currently used in the clinic to assess eligibility for immune-checkpoint inhibitors via the tumor proportion score (TPS), but its efficacy is limited by high interobserver variability. Multiple papers have presented systems for the automatic quantification of TPS, but none report on the task of determining cell-level PD-L1 expression and often reserve their evaluation to a single PD-L1 monoclonal antibody or clinical center. In this paper, we report on a deep learning algorithm for detecting PD-L1 negative and positive tumor cells at a cellular level and evaluate it on a cell-level reference standard established by six readers on a multi-centric, multi PD-L1 assay dataset. This reference standard also provides for the first time a benchmark for computer vision algorithms. In addition, in line with other papers, we also evaluate our algorithm at slide-level by measuring the agreement between the algorithm and six pathologists on TPS quantification. We find a moderately low interobserver agreement at cell-level level (mean reader-reader F1 score = 0.68) which our algorithm sits slightly under (mean reader-AI F1 score = 0.55), especially for cases from the clinical center not included in the training set. Despite this, we find good AI-pathologist agreement on quantifying TPS compared to the interobserver agreement (mean reader-reader Cohen's kappa = 0.54, 95% CI 0.26-0.81, mean reader-AI kappa = 0.49, 95% CI 0.27-0.72). In conclusion, our deep learning algorithm demonstrates promise in detecting PD-L1 expression at a cellular level and exhibits favorable agreement with pathologists in quantifying the tumor proportion score (TPS). We publicly release our models for use via the Grand-Challenge platform.

Training pathologists to assess stromal tumour-infiltrating lymphocytes in breast cancer synergises efforts in clinical care and scientific research.

A growing body of research supports stromal tumour-infiltrating lymphocyte (TIL) density in breast cancer to be a robust prognostic and predicive biomarker. The gold standard for stromal TIL density quantitation in breast cancer is pathologist visual assessment using haematoxylin and eosin-stained slides. Artificial intelligence/machine-learning algorithms are in development to automate the stromal TIL scoring process, and must be validated against a reference standard such as pathologist visual assessment. Visual TIL assessment may suffer from significant interobserver variability. To improve interobserver agreement, regulatory science experts at the US Food and Drug Administration partnered with academic pathologists internationally to create a freely available online continuing medical education (CME) course to train pathologists in assessing breast cancer stromal TILs using an interactive format with expert commentary. Here we describe and provide a user guide to this CME course, whose content was designed to improve pathologist accuracy in scoring breast cancer TILs. We also suggest subsequent steps to translate knowledge into clinical practice with proficiency testing.

Immune cell infiltrate in ductal carcinoma in situ and the risk of dying from breast cancer: case-control study.

BACKGROUND: Studies identifying risk factors for death from breast cancer after ductal carcinoma in situ (DCIS) are rare. In this retrospective nested case-control study, clinicopathological factors in women treated for DCIS and who died from breast cancer were compared with those of patients with DCIS who were free from metastatic disease. METHODS: The study included patients registered with DCIS without invasive carcinoma in Sweden between 1992 and 2012. This cohort was linked to the National Cause of Death Registry. Of 6964 women with DCIS, 96 were registered with breast cancer as cause of death (cases). For each case, up to four controls (318; women with DCIS, alive and without metastatic breast cancer at the time of death of the corresponding case) were selected randomly by incidence density sampling. Whole slides of tumour tissue were evaluated for DCIS grade, comedo necrosis, and intensity of periductal lymphocytic infiltrate. Composition of the immune cell infiltrate, expression of oestrogen receptor, progesterone receptor, human epidermal growth factor receptor 2, and proliferation marker Ki-67 were scored on tissue microarrays. Clinical information was obtained from medical records. Information on date, site, and histological characteristics of local and distant recurrences was obtained from medical records for both cases and controls. RESULTS: Tumour tissue was analysed from 65 cases and 195 controls. Intense periductal lymphocytic infiltrate around DCIS was associated with an increased risk of later dying from breast cancer (OR 2.21. 95% c.i. 1.01 to 4.84). Tumours with more intense lymphocytic infiltrate had a lower T cell/B cell ratio. None of the other biomarkers correlated with increased risk of breast cancer death. CONCLUSION: The immune response to DCIS may influence the risk of dying from breast cancer.

Image-based multiplex immune profiling of cancer tissues: translational implications. A report of the International Immuno-oncology Biomarker Working Group on Breast Cancer.

Recent advances in the field of immuno-oncology have brought transformative changes in the management of cancer patients. The immune profile of tumours has been found to have key value in predicting disease prognosis and treatment response in various cancers. Multiplex immunohistochemistry and immunofluorescence have emerged as potent tools for the simultaneous detection of multiple protein biomarkers in a single tissue section, thereby expanding opportunities for molecular and immune profiling while preserving tissue samples. By establishing the phenotype of individual tumour cells when distributed within a mixed cell population, the identification of clinically relevant biomarkers with high-throughput multiplex immunophenotyping of tumour samples has great potential to guide appropriate treatment choices. Moreover, the emergence of novel multi-marker imaging approaches can now provide unprecedented insights into the tumour microenvironment, including the potential interplay between various cell types. However, there are significant challenges to widespread integration of these technologies in daily research and clinical practice. This review addresses the challenges and potential solutions within a structured framework of action from a regulatory and clinical trial perspective. New developments within the field of immunophenotyping using multiplexed tissue imaging platforms and associated digital pathology are also described, with a specific focus on translational implications across different subtypes of cancer. © 2024 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Multiresolution Self-Supervised Feature Integration via Attention Multiple Instance Learning for Histopathology Analysis.

Digital histopathology image analysis of tumor tissue sections has seen great research interest for automating standard diagnostic tasks, but also for developing novel prognostic biomarkers. However, research has mainly been focused on developing uniresolution models, capturing either high-resolution cellular features or low-resolution tissue architectural features. In addition, in the patch-based weakly-supervised training of deep learning models, the features which represent the intratumoral heterogeneity are lost. In this study, we propose a multiresolution attention-based multiple instance learning framework that can capture cellular and contextual features from the whole tissue for predicting patient-level outcomes. Several basic mathematical operations were examined for integrating multiresolution features, i.e. addition, mean, multiplication and concatenation. The proposed multiplication-based multiresolution model performed the best (AUC=0.864), while all multiresolution models outperformed the uniresolution baseline models (AUC=0.669, 0.713) for breast-cancer grading. (Implementation: https://github.com/tsikup/multiresolution-clam).

Spatial analyses of immune cell infiltration in cancer: current methods and future directions: A report of the International Immuno-Oncology Biomarker Working Group on Breast Cancer.

Modern histologic imaging platforms coupled with machine learning methods have provided new opportunities to map the spatial distribution of immune cells in the tumor microenvironment. However, there exists no standardized method for describing or analyzing spatial immune cell data, and most reported spatial analyses are rudimentary. In this review, we provide an overview of two approaches for reporting and analyzing spatial data (raster versus vector-based). We then provide a compendium of spatial immune cell metrics that have been reported in the literature, summarizing prognostic associations in the context of a variety of cancers. We conclude by discussing two well-described clinical biomarkers, the breast cancer stromal tumor infiltrating lymphocytes score and the colon cancer Immunoscore, and describe investigative opportunities to improve clinical utility of these spatial biomarkers. © 2023 The Pathological Society of Great Britain and Ireland.

Pitfalls in machine learning-based assessment of tumor-infiltrating lymphocytes in breast cancer: A report of the International Immuno-Oncology Biomarker Working Group on Breast Cancer

The clinical significance of the tumor-immune interaction in breast cancer is now established, and tumor-infiltrating lymphocytes (TILs) have emerged as predictive and prognostic biomarkers for patients with triple-negative (estrogen receptor, progesterone receptor, and HER2-negative) breast cancer and HER2-positive breast cancer. How computational assessments of TILs might complement manual TIL assessment in trial and daily practices is currently debated. Recent efforts to use machine learning (ML) to automatically evaluate TILs have shown promising results. We review state-of-the-art approaches and identify pitfalls and challenges of automated TIL evaluation by studying the root cause of ML discordances in comparison to manual TIL quantification. We categorize our findings into four main topics: (1) technical slide issues, (2) ML and image analysis aspects, (3) data challenges, and (4) validation issues. The main reason for discordant assessments is the inclusion of false-positive areas or cells identified by performance on certain tissue patterns or design choices in the computational implementation. To aid the adoption of ML for TIL assessment, we provide an in-depth discussion of ML and image analysis, including validation issues that need to be considered before reliable computational reporting of TILs can be incorporated into the trial and routine clinical management of patients with triple-negative breast cancer. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Survival Outcomes, Digital TILs, and On-treatment PET/CT During Neoadjuvant Therapy for HER2-positive Breast Cancer: Results from the Randomized PREDIX HER2 Trial.

PURPOSE: PREDIX HER2 is a randomized Phase II trial that compared neoadjuvant docetaxel, trastuzumab, and pertuzumab (THP) with trastuzumab emtansine (T-DM1) for HER2-positive breast cancer. Rates of pathologic complete response (pCR) did not differ between the two groups. Here, we present the survival outcomes from PREDIX HER2 and investigate metabolic response and tumor-infiltrating lymphocytes (TIL) as prognostic factors. PATIENTS AND METHODS: In total, 202 patients with HER2-positive breast cancer were enrolled and 197 patients received six cycles of either THP or T-DM1. Secondary endpoints included event-free survival (EFS), recurrence-free survival (RFS), and overall survival (OS). Assessment with PET/CT was performed at baseline, after two and six treatment cycles. TILs were assessed manually at baseline biopsies, while image-based evaluation of TILs [digital TILs (DTIL)] was performed in digitized full-face sections. RESULTS: After a median follow-up of 5.21 years, there was no difference between the two treatment groups in terms of EFS [HR = 1.26; 95% confidence interval (CI), 0.54-2.91], RFS (HR = 0.69; 95% CI, 0.24-1.93), or OS (HR = 0.52; 95% CI, 0.09-2.82). Higher SUVmax at cycle 2 (C2) predicted lower pCR (ORadj = 0.65; 95% CI, 0.48-0.87; P = 0.005) and worse EFS (HRadj = 1.27; 95% CI, 1.12-1.41; P < 0.001). Baseline TILs and DTILs provided additional prognostic information to clinical parameters and C2 SUVmax. CONCLUSIONS: Long-term outcomes following neoadjuvant T-DM1 were similar to neoadjuvant THP. SUVmax after two cycles of neoadjuvant therapy for HER2-positive breast cancer may be an independent predictor of both short- and long-term outcomes. Combined assessment with TILs may facilitate early selection of poor responders for alternative treatment strategies.

External Quality Assessment 2.0: The Importance of a Standardized Implementation of TILs for Daily and Trial Practices.

Increasing data suggests that an intact immune system is required for improvedoutcomes in patients with Human Epidermal Growth Factor Receptor 2 (HER2+) and Triple Negative Breast Cancer (TNBC) [...].

Objective assessment of tumor infiltrating lymphocytes as a prognostic marker in melanoma using machine learning algorithms.

BACKGROUND: The prognostic value of tumor-infiltrating lymphocytes (TILs) assessed by machine learning algorithms in melanoma patients has been previously demonstrated but has not been widely adopted in the clinic. We evaluated the prognostic value of objective automated electronic TILs (eTILs) quantification to define a subset of melanoma patients with a low risk of relapse after surgical treatment. METHODS: We analyzed data for 785 patients from 5 independent cohorts from multiple institutions to validate our previous finding that automated TIL score is prognostic in clinically-localized primary melanoma patients. Using serial tissue sections of the Yale TMA-76 melanoma cohort, both immunofluorescence and Hematoxylin-and-Eosin (H&E) staining were performed to understand the molecular characteristics of each TIL phenotype and their associations with survival outcomes. FINDINGS: Five previously-described TIL variables were each significantly associated with overall survival (p<0.0001). Assessing the receiver operating characteristic (ROC) curves by comparing the clinical impact of two models suggests that etTILs (electronic total TILs) (AUC: 0.793, specificity: 0.627, sensitivity: 0.938) outperformed eTILs (AUC: 0.77, specificity: 0.51, sensitivity: 0.938). We also found that the specific molecular subtype of cells representing TILs includes predominantly cells that are CD3+ and CD8+ or CD4+ T cells. INTERPRETATION: eTIL% and etTILs scores are robust prognostic markers in patients with primary melanoma and may identify a subgroup of stage II patients at high risk of recurrence who may benefit from adjuvant therapy. We also show the molecular correlates behind these scores. Our data support the need for prospective testing of this algorithm in a clinical trial. FUNDING: This work was also supported by a sponsored research agreements from Navigate Biopharma and NextCure and by grants from the NIH including the Yale SPORE in in Skin Cancer, P50 CA121974, the Yale SPORE in Lung Cancer, P50 CA196530, NYU SPORE in Skin Cancer P50CA225450 and the Yale Cancer Center Support Grant, P30CA016359.

Real World Evaluation of the Prosigna/PAM50 Test in a Node-Negative Postmenopausal Swedish Population: A Multicenter Study.

Molecular signatures to guide decisions for adjuvant chemotherapy are recommended in early ER-positive, HER2-negative breast cancer. The objective of this study was to assess what impact gene-expression-based risk testing has had following its recommendation by Swedish national guidelines. Postmenopausal women with ER-positive, HER2-negative and node negative breast cancer at intermediate clinical risk and eligible for chemotherapy were identified retrospectively from five Swedish hospitals. Tumor characteristics, results from Prosigna® test and final treatment decision were available for all patients. Treatment recommendations were compared with the last version of regional guidelines before the introduction of routine risk signature testing. Among the 360 included patients, 41% (n = 148) had a change in decision for adjuvant treatment based on Prosigna® test result. Out of the patients with clinical indication for adjuvant chemotherapy, 52% (n = 118) could avoid treatment based on results from Prosigna® test. On the contrary, 23% (n = 30) of the patients with no indication were escalated to receive adjuvant chemotherapy after testing. Ki67 could not distinguish between the Prosigna® risk groups or intrinsic subtypes and did not significantly differ between patients in which decision for adjuvant therapy was changed based on the test results. In conclusion, we report the first real-world data from implementation of gene-expression-based risk assessment in a Swedish context, which may facilitate the optimization of future versions of the national guidelines.

Systematically higher Ki67 scores on core biopsy samples compared to corresponding resection specimen in breast cancer: a multi-operator and multi-institutional study.

Ki67 has potential clinical importance in breast cancer but has yet to see broad acceptance due to inter-laboratory variability. Here we tested an open source and calibrated automated digital image analysis (DIA) platform to: (i) investigate the comparability of Ki67 measurement across corresponding core biopsy and resection specimen cases, and (ii) assess section to section differences in Ki67 scoring. Two sets of 60 previously stained slides containing 30 core-cut biopsy and 30 corresponding resection specimens from 30 estrogen receptor-positive breast cancer patients were sent to 17 participating labs for automated assessment of average Ki67 expression. The blocks were centrally cut and immunohistochemically (IHC) stained for Ki67 (MIB-1 antibody). The QuPath platform was used to evaluate tumoral Ki67 expression. Calibration of the DIA method was performed as in published studies. A guideline for building an automated Ki67 scoring algorithm was sent to participating labs. Very high correlation and no systematic error (p = 0.08) was found between consecutive Ki67 IHC sections. Ki67 scores were higher for core biopsy slides compared to paired whole sections from resections (p ≤ 0.001; median difference: 5.31%). The systematic discrepancy between core biopsy and corresponding whole sections was likely due to pre-analytical factors (tissue handling, fixation). Therefore, Ki67 IHC should be tested on core biopsy samples to best reflect the biological status of the tumor.

Deconstructing Immune Cell Infiltration in Human Colorectal Cancer: A Systematic Spatiotemporal Evaluation.

Cancer-related immunity has been identified as playing a key role in the outcome of colorectal cancer (CRC); however, the exact mechanisms are only partially understood. In this study, we evaluated a total of 242 surgical specimen of CRC patients using tissue microarrays and immunohistochemistry to evaluate tumor infiltrating immune cells (CD3, CD4, CD8, CD20, CD23, CD45 and CD56) and immune checkpoint markers (CTLA-4, PD-L1, PD-1) in systematically selected tumor regions and their corresponding lymph nodes, as well as in liver metastases. Additionally, an immune panel gene expression assay was performed on 12 primary tumors and 12 consecutive liver metastases. A higher number of natural killer cells and more mature B cells along with PD-1+ expressing cells were observed in the main tumor area as compared to metastases. A higher number of metastatic lymph nodes were associated with significantly lower B cell counts. With more advanced lymph node metastatic status, higher leukocyte-particularly T cell numbers-were observed. Eleven differentially expressed immune-related genes were found between primary tumors and liver metastases. Also, alterations of the innate immune response and the tumor necrosis factor superfamily pathways had been identified.

Interplay between copy number alterations and immune profiles in the early breast cancer Scandinavian Breast Group 2004-1 randomized phase II trial: results from a feasibility study.

Emerging data indicate that genomic alterations can shape immune cell composition in early breast cancer. However, there is a need for complementary imaging and sequencing methods for the quantitative assessment of combined somatic copy number alteration (SCNA) and immune profiling in pathological samples. Here, we tested the feasibility of three approaches-CUTseq, for high-throughput low-input SCNA profiling, multiplexed fluorescent immunohistochemistry (mfIHC) and digital-image analysis (DIA) for quantitative immuno-profiling- in archival formalin-fixed paraffin-embedded (FFPE) tissue samples from patients enrolled in the randomized SBG-2004-1 phase II trial. CUTseq was able to reproducibly identify amplification and deletion events with a resolution of 100 kb using only 6 ng of DNA extracted from FFPE tissue and pooling together 77 samples into the same sequencing library. In the same samples, mfIHC revealed that CD4 + T-cells and CD68 + macrophages were the most abundant immune cells and they mostly expressed PD-L1 and PD-1. Combined analysis showed that the SCNA burden was inversely associated with lymphocytic infiltration. Our results set the basis for further applications of CUTseq, mfIHC and DIA to larger cohorts of early breast cancer patients.

Independent Clinical Validation of the Automated Ki67 Scoring Guideline from the International Ki67 in Breast Cancer Working Group.

Ki67 is an important biomarker with prognostic and potential predictive value in breast cancer. However, the lack of standardization hinders its clinical applicability. In this study, we aimed to investigate the reproducibility among pathologists following the guidelines of the International Ki67 in Breast Cancer Working Group (IKWG) for Ki67 scoring and to evaluate the prognostic potential of this platform in an independent cohort. Four algorithms were independently built by four pathologists based on our study cohort using an open-source digital image analysis (DIA) platform (QuPath) following the detailed guideline of the IKWG. The algorithms were applied on an ER+ breast cancer study cohort of 157 patients with 15 years of follow-up. The reference Ki67 score was obtained by a DIA algorithm trained on a subset of the study cohort. Intraclass correlation coefficient (ICC) was used to measure reproducibility. High interobserver reliability was reached with an ICC of 0.938 (CI: 0.920-0.952) among the algorithms and the reference standard. Comparing each machine-read score against relapse-free survival, the hazard ratios were similar (2.593-4.165) and showed independent prognostic potential (p ≤ 0.018, for all comparisons). In conclusion, we demonstrate high reproducibility and independent prognostic potential using the IKWG DIA instructions to score Ki67 in breast cancer. A prospective study is needed to assess the clinical utility of the IKWG DIA Ki67 instructions.

Interobserver Agreement of PD-L1/SP142 Immunohistochemistry and Tumor-Infiltrating Lymphocytes (TILs) in Distant Metastases of Triple-Negative Breast Cancer: A Proof-of-Concept Study. A Report on Behalf of the International Immuno-Oncology Biomarker Working Group.

Patients with advanced triple-negative breast cancer (TNBC) benefit from treatment with atezolizumab, provided that the tumor contains ≥1% of PD-L1/SP142-positive immune cells. Numbers of tumor-infiltrating lymphocytes (TILs) vary strongly according to the anatomic localization of TNBC metastases. We investigated inter-pathologist agreement in the assessment of PD-L1/SP142 immunohistochemistry and TILs. Ten pathologists evaluated PD-L1/SP142 expression in a proficiency test comprising 28 primary TNBCs, as well as PD-L1/SP142 expression and levels of TILs in 49 distant TNBC metastases with various localizations. Interobserver agreement for PD-L1 status (positive vs. negative) was high in the proficiency test: the corresponding scores as percentages showed good agreement with the consensus diagnosis. In TNBC metastases, there was substantial variability in PD-L1 status at the individual patient level. For one in five patients, the chance of treatment was essentially random, with half of the pathologists designating them as positive and half negative. Assessment of PD-L1/SP142 and TILs as percentages in TNBC metastases showed poor and moderate agreement, respectively. Additional training for metastatic TNBC is required to enhance interobserver agreement. Such training, focusing on metastatic specimens, seems worthwhile, since the same pathologists obtained high percentages of concordance (ranging from 93% to 100%) on the PD-L1 status of primary TNBCs.

In Situ Hybridization of PRRSV-1 Combined with Digital Image Analysis in Lung Tissues of Pigs Challenged with PRRSV-1.

Betaarterivirus suid 1 and 2 are the causative agents of porcine reproductive and respiratory syndrome (PRRS), which is one of the most significant diseases of the swine industry, causing significant economic losses in the main pig producing countries. Here, we report the development of a novel, RNA-based in situ hybridization technique (RNAscope) to detect PRRS virus (PRRSV) RNA in lung tissues of experimentally infected animals. The technique was applied to lung tissues of 20 piglets, which had been inoculated with a wild-type, highly pathogenic PRRSV-1 strain. To determine the RNAscope's applicability as a semi-quantitative method, we analysed the association between the proportion of the virus-infected cells measured with an image analysis software (QuPath) and the outcome of the real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) tests performed in parallel. The results of the quantitative approach of these two molecular biological methods show significant association (pseudo R2 = 0.3894, p = 0.004). This is the first time RNAscope assay has been implemented for the detection of PRRSV-1 in experimental animals.

Predicting Molecular Phenotypes from Histopathology Images: A Transcriptome-Wide Expression-Morphology Analysis in Breast Cancer.

Molecular profiling is central in cancer precision medicine but remains costly and is based on tumor average profiles. Morphologic patterns observable in histopathology sections from tumors are determined by the underlying molecular phenotype and therefore have the potential to be exploited for prediction of molecular phenotypes. We report here the first transcriptome-wide expression-morphology (EMO) analysis in breast cancer, where individual deep convolutional neural networks were optimized and validated for prediction of mRNA expression in 17,695 genes from hematoxylin and eosin-stained whole slide images. Predicted expressions in 9,334 (52.75%) genes were significantly associated with RNA sequencing estimates. We also demonstrated successful prediction of an mRNA-based proliferation score with established clinical value. The results were validated in independent internal and external test datasets. Predicted spatial intratumor variabilities in expression were validated through spatial transcriptomics profiling. These results suggest that EMO provides a cost-efficient and scalable approach to predict both tumor average and intratumor spatial expression from histopathology images. SIGNIFICANCE: Transcriptome-wide expression morphology deep learning analysis enables prediction of mRNA expression and proliferation markers from routine histopathology whole slide images in breast cancer.