Abstract:Background: The area of computational pathology has made huge progresses due to advances in artificial intelligence (AI) and machine learning technologies. It has been applied to many research and translational tasks which provide great improvement on medical diagnosis and treatment. Cancer stem-like cells (CSC) have been consistently reported for its key role in Triple negative breast cancer (TNBC) . Given the large amount of existing H&E stained histological slides of TNBC, digital identification of CSC could benefit the evaluation of tumor status and prediction of patients' response to chemotherapy. Here we proposed an AI framework based on Convolutional Neural Network (CNN) to predict CSC from the histological images of TNBC patient. And our preliminary work suggested that chromosome 17p loss, a common genetic variations in breast cancer, is linked to cancer stemness. Methods: A modified GoogleNet model was adopted as our CNN classifier. Consecutive breast cancer tissue microarrays (TMA), which stained with H&E ,SOX2, OCT4 and NANOG antibodies respectively by IHC, were used as training dataset for the CNN model. Gene expression data from the TCGA and METABRIC datasets were used to identify gene signatures associated with CSC. The connectivity map (CMAP) and Cancer Cell Line Encyclopedia (CCLE) were used for screening compounds that target stemness in cancer cell lines with chromosomal 17p loss. HS578T and EO771 cells with or without heterozygous 17p loss (11b in EO771) were used for in vitro experiments. Female immunodeficient nude (Nu/J) mice were used for animal studies. Results: The well trained GoogleNet model was applied to TNBC patient diagnosis images in TCGA BRCA dataset. By analyzing patient genomic alteration on chromosomal level, we found that loss of chromosome 17p associate with high cancer stemness in TNBC. Flow cytometry assays also demonstrated higher ALDH1 activity and higher CD44+/CD24−/low cell population in HS578T cells with 17p loss. RNA-seq of HS578T cells revealed that most CSC marker genes were located in the unregulated differentially expressed genes (DEGs) of 17p loss cells. We next compared the cytotoxicity of chemotherapy drugs including doxorubicin, paclitaxel, docetaxel and cisplatin on 17p loss and 17p intact HS578T cells, 11b loss and 11b intact EO771 cells, in terms of IC50 value. The IC50 value of indicated drug on 17p loss HS578T cells with were 3-6 fold higher than their IC50 on 17p intact HS578T cells. Similar result was observed in EO771 cells. Next, 17p loss and 17p intact HS578T cells were orthotopically implanted into the Nu/J mice. Under the doxorubicin treatment, mice bearing 17p loss HS578T derived tumors had larger and heavier tumors in compare to mice bearing 17p intact tumors. Next, we did a computational drug screening to identify compounds that can target the cancer stemness in 17p loss cells. Screened out compounds were tested for their cytotoxicity on 17p loss and 17p intact HS578T cells and FK866 showed the most pronounced efficacy on inhibiting the viability of 17p loss cells, compared to 17p intact cells. Followup experiments demonstrated that FK866 can decrease the CSC features induced by doxorubicin, both in vitro and in vivo. FK866 also potentiates the effect of doxorubicin on treating TNBC cells with 17p loss, which provide a drug combination potential for TNBC patient with 17p loss. Conclusions: A CNN based model was developed to identify CSC from TNBC histopathology images. The images analysis combined with patient genomic data revealed that chromosome 17p loss associated with cancer stemness in TNBC. This result was confirmed using assays on TNBC cells with or without 17p loss. A computational drug screening was performed to identify candidates that targeting stemness in 17p loss cells. FK866 was identified and it potentiates the anti-tumor effect of doxorubicin on treating TNBC cells with 17p loss. Our study provides a novel strategy on applying AI to precision treatment for cancers. Citation Format: tianhan dong, jiannan liu, yuanzhang Fang, Ziyu Liu, Xiongbin Lu, kun huang. Machine learning based histopathology images analysis reveals cancer stemness in TNBC patient with 17p loss [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl) nr P5-14-08.

Abstract 3532: Transforming genomic data into images for enhanced deep learning in precision oncology

Abstract LB392: Transformative pathomics in oncology: Harnessing FFPE tissue slide imagery for clinically relevant gene expression prediction in invasive breast carcinoma

Abstract 888: Uncovering clinically relevant omics signatures from pan-cancer imaging and multi-omics data integration

Abstract 4298: Predicting tumor evolution from digital histology using AI

PathoGen-X: A Cross-Modal Genomic Feature Trans-Align Network for Enhanced Survival Prediction from Histopathology Images

Abstract 5954: Big data-driven discovery of novel oncogenic fusion genes for anticancer therapy

Abstract 5354: Prediction of OncotypeDX high risk group for chemotherapy benefit in breast cancer by deep learning analysis of hematoxylin and eosin-stained whole slide images

Abstract 2059: Machine learning integration of transcriptome-wide spatial sequencing data and ultra-high plex spatial proteomic data enables the prioritization of cancer drug targets

Abstract 2310: Integrating real-world histopathological and clinicogenomic data from 1799 lung cancer patients by applying unsupervised deep learning

Abstract 2034: Discovery of novel therapeutic targets using 3D chromatin conformation and patient-specific gene regulatory networks

Abstract 5440: Deep-learning model for tumor type classification enables enhanced clinical decision support in cancer diagnosis

Abstract 464: AI-powered segmentation and analysis of nuclei morphology predicts genomic and clinical markers in multiple cancer types

Abstract 4938: Prediction of gene mutation from colorectal adenocarcinoma whole slide images via integrated deep learning pipeline

Abstract 2315: AI-enabled precision oncology era: Advanced and interactive interpretation of next-gneneration sequencing (NGS) reports

Abstract 5361: Deep learning models capture high dimensional features for cell morphology analysis

Abstract 188: Deep learning enables label-free profiling of the tumor microenvironment and enrichment of rare cancer cells

Abstract A007: Picturing progression: A new role of whole slide imaging in comprehensive clinical genomic datasets

Abstract 2313: Multi-modal deep learning to predict cancer outcomes by integrating radiology and pathology images

Abstract P5-14-08: Machine learning based histopathology images analysis reveals cancer stemness in TNBC patient with 17p loss

Abstract 6186: Histopathology-derived molecular portrait of serous ovarian cancer

Learning Genomic Representations to Predict Clinical Outcomes in Cancer