Machine learning-based prediction of microsatellite instability and high tumor mutation burden from contrast-enhanced computed tomography in endometrial cancers

Harini Veeraraghavan, Claire F Friedman, Deborah F DeLair, Josip Ninčević, Yuki Himoto, Silvio G Bruni, Giovanni Cappello, Iva Petkovska, Stephanie Nougaret, Ines Nikolovski, Ahmet Zehir, Nadeem R Abu-Rustum, Carol Aghajanian, Dmitriy Zamarin, Karen A Cadoo, Luis A Diaz, Mario M Leitao, Vicky Makker, Robert A Soslow, Jennifer J MuellerBritta Weigelt, Yulia Lakhman

Research output: Contribution to journalArticlepeer-review

Abstract

To evaluate whether radiomic features from contrast-enhanced computed tomography (CE-CT) can identify DNA mismatch repair deficient (MMR-D) and/or tumor mutational burden-high (TMB-H) endometrial cancers (ECs). Patients who underwent targeted massively parallel sequencing of primary ECs between 2014 and 2018 and preoperative CE-CT were included (n = 150). Molecular subtypes of EC were assigned using DNA polymerase epsilon (POLE) hotspot mutations and immunohistochemistry-based p53 and MMR protein expression. TMB was derived from sequencing, with > 15.5 mutations-per-megabase as a cut-point to define TMB-H tumors. After radiomic feature extraction and selection, radiomic features and clinical variables were processed with the recursive feature elimination random forest classifier. Classification models constructed using the training dataset (n = 105) were then validated on the holdout test dataset (n = 45). Integrated radiomic-clinical classification distinguished MMR-D from copy number (CN)-low-like and CN-high-like ECs with an area under the receiver operating characteristic curve (AUROC) of 0.78 (95% CI 0.58-0.91). The model further differentiated TMB-H from TMB-low (TMB-L) tumors with an AUROC of 0.87 (95% CI 0.73-0.95). Peritumoral-rim radiomic features were most relevant to both classifications (p ≤ 0.044). Radiomic analysis achieved moderate accuracy in identifying MMR-D and TMB-H ECs directly from CE-CT. Radiomics may provide an adjunct tool to molecular profiling, especially given its potential advantage in the setting of intratumor heterogeneity.

Original languageEnglish
Pages (from-to)17769
JournalSci. Rep.
Volume10
Issue number1
DOIs
Publication statusPublished - Oct 20 2020
Externally publishedYes

Keywords

  • Aged
  • Carcinoma, Endometrioid/diagnosis
  • Cohort Studies
  • Computer Simulation
  • DNA Mismatch Repair/genetics
  • DNA Polymerase II/genetics
  • Endometrial Neoplasms/diagnosis
  • Female
  • Humans
  • Machine Learning
  • Microsatellite Instability
  • Middle Aged
  • Mutation/genetics
  • Neoplasm Staging
  • Poly-ADP-Ribose Binding Proteins/genetics
  • Prognosis
  • Tomography, X-Ray Computed/methods
  • Tumor Suppressor Protein p53/metabolism
  • Uterus/diagnostic imaging

Fingerprint Dive into the research topics of 'Machine learning-based prediction of microsatellite instability and high tumor mutation burden from contrast-enhanced computed tomography in endometrial cancers'. Together they form a unique fingerprint.

Cite this