AUTHOR=Yang Linsha , Du Dan , Zheng Tao , Liu Lanxiang , Wang Zhanqiu , Du Juan , Yi Huiling , Cui Yujie , Liu Defeng , Fang Yuan 

TITLE=Deep learning and radiomics to predict the mitotic index of gastrointestinal stromal tumors based on multiparametric MRI

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.948557

DOI=10.3389/fonc.2022.948557

ISSN=2234-943X

ABSTRACT=Background: Preoperative evaluation of the mitotic index (MI) of gastrointestinal stromal tumors (GIST) represents the basis of individualized treatment of patients. However, the accuracy of conventional preoperative imaging methods is limited. Purpose: The aim of this study was to develop a predictive model based on multi-parameter MRI for preoperative MI prediction. Methods: A total of 112 patients who were pathologically diagnosed with GIST were enrolled in this study. The dataset was subdivided into the development (n = 81) and test (n = 31) sets based on the time of diagnosis. Using T2-weighted imaging (T2WI) and apparent diffusion coefficient (ADC) map, a CNN-based classifier was developed for MI prediction, which used a hybrid approach based on 2D tumor images and radiomic features from 3D tumor shape. The trained model was tested on an internal test set. Then, we comprehensively tested the hybrid model and compared it with the conventional ResNet, shape radiomic classifier and age plus diameter classifier. Results: The hybrid model showed good MI prediction ability at the image level, the AUROC, AUPRC and accuracy in the test set were 0.947 (95% Confidence Interval [CI]: 0.927-0.968), 0.964 (95% CI: 0.930-0.978) and 90.8 (95% CI: 88.0-93.0), respectively. With the average probabilities from multiple samples per patient, good performance was also achieved at the patient level, with AUROC, AUPRC and accuracy of 0.930 (95% CI: 0.828-1.000), 0.941 (95% CI: 0.792-1.000) and 93.6% (95% CI: 79.3-98.2) in the test set, respectively. Conclusion: The deep learning-based hybrid model demonstrated the potential to be a good tool for the operative and noninvasive prediction of MI in GIST patients.