Synapse - Deep-learning model for tumor-type prediction using targeted clinical genomic sequencing data

Deep-learning model for tumor-type prediction using targeted clinical genomic sequencing data Journal Article

Authors:	Darmofal, M.; Suman, S.; Atwal, G.; Toomey, M.; Chen, J. F.; Chang, J. C.; Vakiani, E.; Varghese, A. M.; Balakrishnan Rema, A.; Syed, A.; Schultz, N.; Berger, M. F.; Morris, Q.
Article Title:	Deep-learning model for tumor-type prediction using targeted clinical genomic sequencing data
Abstract:	Tumor type guides clinical treatment decisions in cancer, but histology-based diagnosis remains challenging. Genomic alterations are highly diagnostic of tumor type, and tumor-type classifiers trained on genomic features have been explored, but the most accurate methods are not clinically feasible, relying on features derived from whole-genome sequencing (WGS), or predicting across limited cancer types. We use genomic features from a data set of 39,787 solid tumors sequenced using a clinically targeted cancer gene panel to develop Genome-Derived-Diagnosis Ensemble (GDD-ENS): a hyperparameter ensemble for classifying tumor type using deep neural net-works. GDD-ENS achieves 93% accuracy for high-confidence predictions across 38 cancer types, rivaling the performance of WGS-based methods. GDD-ENS can also guide diagnoses of rare type and cancers of unknown primary and incorporate patient-specific clinical information for improved predic-tions. Overall, integrating GDD-ENS into prospective clinical sequencing workflows could provide clinically relevant tumor-type predictions to guide treatment decisions in real time. SIGNIFICANCE: We describe a highly accurate tumor-type prediction model, designed specifically for clinical implementation. Our model relies only on widely used cancer gene panel sequencing data, predicts across 38 distinct cancer types, and supports integration of patient-specific nongenomic information for enhanced decision support in challenging diagnostic situations. © 2024 The Authors; Published by the American Association for Cancer Research.
Keywords:	osteosarcoma; gene mutation; sequence analysis; somatic mutation; gene deletion; genetics; missense mutation; squamous cell carcinoma; pancreas cancer; prospective study; neoplasm; neoplasms; colorectal cancer; accuracy; pathology; bladder cancer; prediction; histology; lung metastasis; training; algorithm; microsatellite instability; uterine cervix cancer; gene fusion; genomics; clinical decision making; predictive value; linear regression analysis; artificial neural network; decision support system; non small cell lung cancer; head and neck squamous cell carcinoma; procedures; learning algorithm; granulosa cell; workflow; dabrafenib; trametinib; measurement accuracy; humans; human; article; whole genome sequencing; pembrolizumab; entrectinib; deep learning; larotrectinib; deep neural network; selpercatinib; neural networks, computer
Journal Title:	Cancer Discovery
Volume:	14
Issue:	6
ISSN:	2159-8274
Publisher:	American Association for Cancer Research
Date Published:	2024-06-01
Start Page:	1064
End Page:	1081
Language:	English
DOI:	10.1158/2159-8290.Cd-23-0996
PUBMED:	38416134
PROVIDER:	scopus
PMCID:	PMC11145170
DOI/URL:	https://www.scopus.com/inward/record.uri?eid=2-s2.0-85195227372&doi=10.1158%2f2159-8290.CD-23-0996&partnerID=40&md5=ccbb5677a4aa8537b6ee2a603422c0ef
Notes:	The MSK Cancer Center Support Grant (P30 CA008748) is acknowledged in the PDF -- Corresponding authors are MSK authors: Michael F. Berger and Quaid Morris -- Source: Scopus