Objective To explore the value of TCGA/TCIA-CT radiomics for the noninvasive prediction of the status of classical oncogenic signaling pathways in bladder cancer(BLCA).Methods Sixty-nine BLCA patients(55 in the train-ing cohort and 14 in the validation cohort)screened from the TCGA/TCIA public database were retrospectively included in this study and used to analyze the relationship of radiomics with P13K,RTK RAS,NOTCH,and TP53 classical oncogenic signaling pathway status.1130 radiomics features were extracted based on enhanced CT arterial phase images.Least absolute shrinkage and selection operator algorithm was used to develop radiomics scores.The performance,calibration of radiomics scores were confirmed and validated.Results All 4 oncogenic signaling pathways exhibited a positive rate of 40％or more.CT radiomics scores were developed based on 7,4,5 and 7 features and showed moderate or higher predictive per-formance(AUC＞0.70)in both training and validation cohorts with good calibration agreement.There was a positive cor-relation between NOTCH and RTK RAS radiomic scores(r=0.41,P＜0.001)and a negative correlation with P13K(r=-0.60,P＜0.001),reflecting possible synergistic or mutually exclusive effects between pathways.Conclusion Ra-diomics has the value of predicting the status of BLCA classical oncogenic signaling pathways,providing a potential non-in-vasive means of monitoring the effectiveness of targeted tumor therapy.
NETL
NSTL
万方数据
