A radiomics-based interpretable machine learning model to predict the HER2 status in bladder cancer: a multicenter study

Abstract

Objective: To develop a computed tomography (CT) radiomics-based interpretable machine learning (ML) model to preoperatively predict human epidermal growth factor receptor 2 (HER2) status in bladder cancer (BCa) with multicenter validation.

Methods: In this retrospective study, 207 patients with pathologically confirmed BCa were enrolled and divided into the training set (n = 154) and test set (n = 53). Least absolute shrinkage and selection operator (LASSO) regression was used to identify the most discriminative features in the training set. Five radiomics-based ML models, namely logistic regression (LR), support vector machine (SVM), k-nearest neighbors (KNN), eXtreme Gradient Boosting (XGBoost) and random forest (RF), were developed. The predictive performance of established ML models was evaluated by the area under the receiver operating characteristic curve (AUC). The Shapley additive explanation (SHAP) was used to analyze the interpretability of ML models.

Results: A total of 1218 radiomics features were extracted from the nephrographic phase CT images, and 11 features were filtered for constructing ML models. In the test set, the AUCs of LR, SVM, KNN, XGBoost, and RF were 0.803, 0.709, 0.679, 0.794, and 0.815, with corresponding accuracies of 71.7%, 69.8%, 60.4%, 75.5%, and 75.5%, respectively. RF was identified as the optimal classifier. SHAP analysis showed that texture features (gray level size zone matrix and gray level co-occurrence matrix) were significant predictors of HER2 status.

Conclusions: The radiomics-based interpretable ML model provides a noninvasive tool to predict the HER2 status of BCa with satisfactory discriminatory performance.

Critical relevance statement: An interpretable radiomics-based machine learning model can preoperatively predict HER2 status in bladder cancer, potentially aiding in the clinical decision-making process.

Key points: The CT radiomics model could identify HER2 status in bladder cancer. The random forest model showed a more robust and accurate performance. The model demonstrated favorable interpretability through SHAP method.

Keywords: Bladder cancer; Computed tomography; HER2; Machine learning; Radiomics.

Fig. 1

The flowchart of the patient recruitment process

Fig. 2

The overall workflow of this study. CT, computed tomography; ICC, inter- and intra-class correlation coefficient; ROI, region of interest; ROC, receiver operating characteristic; SHAP, Shapley additive explanation; LASSO, the least absolute shrinkage and selection operator

Fig. 3

Performance comparison of ML models for HER2 prediction in bladder cancer patients. a–c Displays the ROC curve, performance radar chart and calibration curve of the ML models in the training set, respectively. d–f Displays the ROC curve, performance radar chart and calibration curve of the ML models in the test set, respectively. AUC, area under the curve; ML, machine learning; ROC, receiver operating characteristic

Fig. 4

The interpretability of the ML radiomics model was assessed using the SHAP method. a The SHAP bar chart shows the importance of each feature based on the mean SHAP values. b The SHAP summary plot shows the impact of each feature on the model predictions. Individual dots symbolize patients, and different colors represent different levels of influence on the model output. ML, machine learning; SHAP, Shapley additive explanation

Fig. 5

Four representative cases correctly predicted as HER2-negative (patients A and B) and HER2-positive (patients C and D) were individually visualized by SHAP method. SHAP, Shapley additive explanation