From:  A systematic review on cardiovascular risk stratification in the artificial intelligence paradigm using radiogenomics

 Applications of radiogenomics in oncology.

AuthorYearCancer typeImaging modalityAIResultsAim
Yan et al. [63]2021GliomaMRIBayesian-regularization neural networksAUC(IDH): 0.884
AUC(1p19q): 0.815
AUC(TERT): 0.669		To predict the molecular groups (IDH, 1p19q, and TERT) in gliomas and assess their prognosis
Verduin et al. [76]2021GlioblastomaMRIMultivariable Cox-regression modelAUC(EGFR amplification): 0.707
AUC(MGMT methylation): 0.667		To predict overall survival and key molecular markers in glioblastoma patients
Zeng et al. [67]2021Clear cell renal cell carcinomaCTRFAUC: 0.971To predict molecular characteristics and overall survival in ccRCC
Lu et al. [64]2023Colorectal cancerCTYolov7AUC: 0.9591To determine the colorectal tumor location and predict the stage, and RAS gene mutation
Lee et al. [65]2023Colorectal cancer (stage IV)18F-FDGPETML algorithms with the best performance achieved by kNNAUC: 0.791To predict tumor mutational burden (TMB) and prognosis in patients with stage IV colorectal cancer
Zhou et al. [66]2024Triple-negative breast cancerDynamic contrast enhanced-MRILRAUC: 0.93To predict pCR from radiogenomic models
Ogbonnaya et al. [77]2024Prostate cancerMRIPearson’s correlationAUC: 0.95To create a radiogenomics map and predict prostate cancer
Buzdugan et al. [28]2025IDH-wild-type glioblastomaMultiparametric MRIRF, XGBoost, LightGBM, DNNCI: 0.86To refine survival prediction in glioblastoma patients
Return to article view

AUC: area under the curve; CI: concordance index; ccRCC: clear cell renal cell carcinoma; DNN: deep neural network; EGFR: epidermal growth factor receptor; FDGPET: fluorodeoxyglucose positron emission tomography; IDH: isocitrate dehydrogenase; kNN: k-nearest neighbors; MGMT: O6-methylguanine–DNA methyltransferase; pCR: pathological complete response; RAS: rat sarcoma; RF: random forest; TERT: telomerase reverse transcriptase.