Synapse - A genomics-driven artificial intelligence–based model classifies breast invasive lobular carcinoma and discovers CDH1 inactivating mechanisms

A genomics-driven artificial intelligence–based model classifies breast invasive lobular carcinoma and discovers CDH1 inactivating mechanisms Journal Article

Authors:	Pareja, F.; Dopeso, H.; Wang, Y. K.; Gazzo, A. M.; Brown, D. N.; Banerjee, M.; Selenica, P.; Bernhard, J. H.; Derakhshan, F.; da Silva, E. M.; Colon-Cartagena, L.; Basili, T.; Marra, A.; Sue, J.; Ye, Q.; Da Cruz Paula, A.; Yeni Yildirim, S.; Pei, X.; Safonov, A.; Green, H.; Gill, K. Y.; Zhu, Y.; Lee, M. C. H.; Godrich, R. A.; Casson, A.; Weigelt, B.; Riaz, N.; Wen, H. Y.; Brogi, E.; Mandelker, D. L.; Hanna, M. G.; Kunz, J. D.; Rothrock, B.; Chandarlapaty, S.; Kanan, C.; Oakley, J.; Klimstra, D. S.; Fuchs, T. J.; Reis-Filho, J. S.
Article Title:	A genomics-driven artificial intelligence–based model classifies breast invasive lobular carcinoma and discovers CDH1 inactivating mechanisms
Abstract:	Artificial intelligence (AI) systems can improve cancer diagnosis, yet their development often relies on subjective histologic features as ground truth for training. Herein, we developed an AI model applied to histologic whole-slide images using CDH1 biallelic mutations, pathognomonic for invasive lobular carcinoma (ILC) in breast neoplasms, as ground truth. The model accurately predicted CDH1 biallelic mutations (accuracy 1⁄4 0.95) and diagnosed ILC (accuracy 1⁄4 0.96). A total of 74% of samples classified by the AI model as having CDH1 biallelic mutations but lacking these alterations displayed alternative CDH1 inactivating mechanisms, including a deleterious CDH1 fusion gene and noncoding CDH1 genetic alterations. Analysis of internal and external validation cohorts demonstrated 0.95 and 0.89 accuracy for ILC diagnosis, respectively. The latent features of the AI model correlated with human-explainable histopathologic features. Taken together, this study reports the construction of an AI algorithm trained using a genetic rather than histologic ground truth that can robustly classify ILCs and uncover CDH1 inactivating mechanisms, providing the basis for orthogonal ground truth utilization for development of diagnostic AI models applied to whole-slide image. Significance: Genetic alterations linked to strong genotypic–phenotypic correlations can be utilized to develop AI systems applied to pathology that facilitate cancer diagnosis and biologic discoveries. ©2024 American Association for Cancer Research.
Keywords:	adult; controlled study; gene mutation; major clinical study; genetics; mutation; histopathology; area under the curve; validation process; diagnostic accuracy; sensitivity and specificity; accuracy; phenotype; allele; breast cancer; basal cell carcinoma; cohort analysis; genotype; pathology; breast neoplasms; algorithms; uvomorulin; histology; neuroendocrine tumor; training; algorithm; artificial intelligence; breast tumor; breast carcinoma; fusion gene; genomics; antigens, cd; cancer classification; bioinformatics; cadherin; cadherins; leukocyte antigen; predictive value; receiver operating characteristic; carcinoma, lobular; copy number variation; gray matter; diagnostic test accuracy study; feature extraction; lobular carcinoma; procedures; machine learning; cdh1 protein, human; estrogen receptor positive breast cancer; humans; human; female; article; whole genome sequencing; droplet digital polymerase chain reaction; whole exome sequencing; cross validation; invasive lobular breast carcinoma; learning rate; optimizer; weight decay
Journal Title:	Cancer Research
Volume:	84
Issue:	20
ISSN:	0008-5472
Publisher:	American Association for Cancer Research
Date Published:	2024-10-15
Start Page:	3478
End Page:	3489
Language:	English
DOI:	10.1158/0008-5472.Can-24-1322
PUBMED:	39106449
PROVIDER:	scopus
PMCID:	PMC11479818
DOI/URL:	https://www.scopus.com/inward/record.uri?eid=2-s2.0-85206398311&doi=10.1158%2f0008-5472.CAN-24-1322&partnerID=40&md5=70910b6af637728c989603a1451aebcf
Notes:	Article -- MSK Cancer Center Support Grant (P30 CA008748) acknowledged in PubMed and PDF -- MSK corresponding author is Fresia Pareja -- Source: Scopus