Development of Genome-Derived Tumor Type Prediction to Inform Clinical Cancer Care

Abstract

Diagnosing the site of origin for cancer is a pillar of disease classification that has directed clinical care for more than a century. Even in an era of precision oncologic practice, in which treatment is increasingly informed by the presence or absence of mutant genes responsible for cancer growth and progression, tumor origin remains a critical factor in tumor biologic characteristics and therapeutic sensitivity. OBJECTIVE To evaluate whether data derived from routine clinical DNA sequencing of tumors could complement conventional approaches to enable improved diagnostic accuracy. DESIGN, SETTING, AND PARTICIPANTS A machine learning approach was developed to predict tumor type from targeted panel DNA sequence data obtained at the point of care, incorporating both discrete molecular alterations and inferred features such as mutational signatures. This algorithm was trained on 7791 tumors representing 22 cancer types selected from a prospectively sequenced cohort of patients with advanced cancer. RESULTS The correct tumor type was predicted for 5748 of the 7791 patients (73.8%) in the training set as well as 8623 of 11 644 patients (74.1%) in an independent cohort. Predictions were assigned probabilities that reflected empirical accuracy, with 3388 cases (43.5%) representing high-confidence predictions (>95% probability). Informative molecular features and feature categories varied widely by tumor type. Genomic analysis of plasma cell-free DNA yielded accurate predictions in 45 of 60 cases (75.0%), suggesting that this approach may be applied in diverse clinical settings including as an adjunct to cancer screening. Likely tissues of origin were predicted from targeted tumor sequencing in 95 of 141 patients (67.4%) with cancers of unknown primary site. Applying this method prospectively to patients under active care enabled genome-directed reassessment of diagnosis in 2 patients initially presumed to have metastatic breast cancer, leading to the selection of more appropriate treatments, which elicited clinical responses. CONCLUSIONS AND RELEVANCE These results suggest that the application of artificial intelligence to predict tissue of origin in oncologic practice can act as a useful complement to conventional histologic review to provide integrated pathologic diagnoses, often with important therapeutic implications.

Machine learning analysis using 77,044 genomic and transcriptomic profiles to accurately predict tumor type

“…The accuracy of MI GPSai compares favorably to recent data on the use of DNA NGS panels for tissue of origin identification or guidance of utilization of targeted- and immunotherapies [10] , [28] . However, overall accuracy of these approaches is limited.…”

“…However, overall accuracy of these approaches is limited. For example, predictions made by a Random Forrest Classifier using results from a 468-gene NGS panel as input, resulted in an overall accuracy of 74.1% [10] . Analysis of circulating tumor DNA data from a commercial 70-gene NGS panel revealed potentially targetable mutations.…”

CUP-AI-Dx: A tool for inferring cancer tissue of origin and molecular subtype using RNA gene-expression data and artificial intelligence

“…The path to improving the tumour type classification accuracy may be to consider including other potential features such as somatic point mutations [79] and histopathology images [80] in the model. Mutational profiling of tumours is steadily being incorporated into mainstream work-up of cancer patients and recently several tissue of origin classification methods have been developed based on DNA features alone either from panel [81] whole-exome, and whole-genome sequencing (WGS) [82] . Interestingly, the reported accuracy of these methods especially when using WGS passenger mutational profiles for the tissue of origin classification is similar to using gene-expression profiling and DNA methylation classification.…”

Cancer of Unknown Primary: Challenges and Progress in Clinical Management

“…In their report, Penson et al [ 184 ] applied a machine-learning approach for tumor type prediction using target DNA sequencing data. The algorithm was trained on a cohort of 7791 patients of advanced tumors belonging to 22 different classes, and later tested on an independent set of 11,644 cases.…”