Deep learning predicts HRD and platinum response from histology slides in breast and ovarian cancer
Bergstrom, E. N.; Abbasi, A.; Diaz-Gay, M.; Galland, L.; Lippman, S.; Ladoire, S.; Alexandrov, L. B.
Show abstract
Breast and ovarian cancers harboring homologous recombination deficiencies (HRD) can benefit from platinum-based chemotherapies and PARP inhibitors. Standard diagnostic tests for detecting HRD utilize molecular profiling, which is not universally available especially for medically underserved populations. Here, we trained a deep learning approach for predicting genomically derived HRD scores from routinely sampled hematoxylin and eosin (H&E)-stained histopathological slides. For breast cancer, the approach was externally validated on three independent cohorts and allowed predicting patients response to platinum treatment. Using transfer learning, we demonstrated the methods clinical applicability to H&E-images from high-grade ovarian tumors. Importantly, our deep learning approach outperformed existing genomic HRD biomarkers in predicting response to platinum-based therapies across multiple cohorts, providing a complementary approach for detecting HRD in patients across diverse socioeconomic groups. One-Sentence SummaryA deep learning approach outperforms molecular tests in predicting platinum response of HRD cancers from histological slides.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Myeloid cell-associated resistance to PD-1/PD-L1 blockade in urothelial cancer revealed through bulk and single-cell RNA sequencing 95%
- Circulating tumor DNA analysis in advanced urothelial carcinoma: insights from biological analysis and extended clinical follow-up 94%
- Cell-free urine- and plasma DNA mutational analysis predicts neoadjuvant chemotherapy response and outcome in patients with muscle invasive bladder cancer 94%
Similar papers in this journal
- A single-cell based precision medicine approach using glioblastoma patient-specific models 94%
- Single Cell RNA-sequencing of BCG naive and recurrent non-muscle invasive bladder cancer reveals a CD6/ALCAM mediated immune-suppressive pathway 93%
- Deep learning inference of cell type-specific gene expression from breast tumor histopathology 93%
Similar papers in this journal
- Detection and monitoring of translocation renal cell carcinoma via plasma cell-free epigenomic profiling 93%
- Development and validation of multivariate predictors of primary endocrine resistance to tamoxifen and aromatase inhibitors in luminal breast cancer reveal drug-specific differences 93%
- Targeting EIF4A triggers an interferon response to synergize with chemotherapy and suppress triple-negative breast cancer 93%
Similar papers in this journal
- Cyclin E1 protein is stabilized in BRCA1 mutated breast cancers leading to synergy between CDK2 and PARP inhibitors 94%
- RNA Sequencing-Based Single Sample Predictors of Molecular Subtype and Risk of Recurrence for Clinical Assessment of Early-Stage Breast Cancer 94%
- The breast pre-cancer atlas illustrates the molecular and micro-environmental diversity of ductal carcinoma in situ 92%
Similar papers in this journal
- Integrated radiogenomics models predict response to neoadjuvant chemotherapy in high grade serous ovarian cancer 95%
- Multiplexed RNA-FISH-guided Laser Capture Microdissection RNA Sequencing Improves Breast Cancer Molecular Subtyping, Prognostic Classification, and Predicts Response to Antibody Drug Conjugates 95%
- Deactivation of ligand-receptor interactions enhancing lymphocyte infiltration drives melanoma resistance to Immune Checkpoint Blockade 95%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.