Individualized Prediction of Colorectal Cancer Metastasis Using a Radiogenomics Approach

Abstract

Objectives: To evaluate whether incorporating the radiomics, genomics, and clinical features allows prediction of metastasis in colorectal cancer (CRC) and to develop a preoperative nomogram for predicting metastasis. Methods: We retrospectively analyzed radiomics features of computed tomography (CT) images in 134 patients (62 in the primary cohort, 28 in the validation cohort, and 44 in the independent-test cohort) clinicopathologically diagnosed with CRC at Dazhou Central Hospital from February 2018 to October 2019. Tumor tissues were collected from all patients for RNA sequencing, and clinical data were obtained from medical records. A total of 854 radiomics features were extracted from enhanced venous-phase CT of CRC. Least absolute shrinkage and selection operator regression analysis was utilized for data dimension reduction, feature screen, and radiomics signature development. Multivariable logistic regression analysis was performed to build a multiscale predicting model incorporating the radiomics, genomics, and clinical features. The receiver operating characteristic curve, calibration curve, and decision curve were conducted to evaluate the performance of the nomogram. Results: The radiomics signature based on 16 selected radiomics features showed good performance in metastasis assessment in both primary [area under the curve (AUC) = 0.945, 95% confidence interval (CI) 0.892-0.998] and validation cohorts (AUC = 0.754, 95% CI 0.570-0.938). The multiscale nomogram model contained radiomics features signatures, four-gene expression related to cell cycle pathway, and CA 19-9 level. The multiscale model showed good discrimination performance in the primary cohort (AUC = 0.981, 95% CI 0.953-1.000), the validation cohort (AUC = 0.822, 95% CI 0.635-1.000), and the independent-test cohort (AUC = 0.752, 95% CI 0.608-0.896) and good calibration. Decision curve analysis confirmed the clinical application value of the multiscale model. Conclusion: This study presented a multiscale model that incorporated the radiological eigenvalues, genomics features, and CA 19-9, which could be conveniently utilized to facilitate the individualized preoperatively assessing metastasis in CRC patients.

Keywords: carbohydrate antigen 19-9; colorectal cancer; genomics; metastasis; radiomics.

Figure 1

Flowchart of study. (A) The data collection pipeline in this study. (B) Flowchart of patients' inclusion.

Figure 2

Radiomics signatures selection and model performance. (A) Tuning parameter (λ) was selected in the LASSO model by 10-fold cross-validation via minimum criteria. The area under the receiver operating characteristic curve was plotted versus log(λ). A log (λ) = −2.91 was chosen. (B) LASSO coefficient profiles of the 854 texture features. A coefficient profile plot was produced against the log (λ) sequence. Vertical line was plotted at the optimal λ value, which resulted in 16 non-zero coefficients. (C) The radiomics score for each patient in the primary (AUC) cohort (n = 62) and validation cohort (n = 28). (D) ROC curve of radiomics model in the primary cohort and validation cohort. LASSO, least absolute shrinkage and selection operator.

Figure 3

Genomics features selection. (A) Protein–protein interaction (PPI) network and (B) Gene Ontology (GO) enrichment analysis of differentially expressed genes between metastasis group and non-metastasis group. Red node represents proteins enriched in cell cycle pathway. (C) Statistical analysis of CDKN2A mRNA expression level in metastasis group and non-metastasis group. n = 90. ^**P < 0.01.

Figure 4

ROC curve of multiscale radiogenomics model. (A) ROC curve of multiscale radiogenomics model in the primary cohort and validation cohorts. (B) ROC curve of multiscale radiogenomics model in the independent-test cohort.

Figure 5

Nomogram, calibration curve, and decision curve derived from the multiscale radiogenomics model. (A) Nomogram developed with the radiomics score, genomics, and CA-19-9. (B) Calibration curve analysis for the multiscale radiogenomics model. The y axis represents the actual probability of metastasis. The x axis represents the predicted probability of metastasis. (C) Decision curve analysis for the multiscale radiogenomics model.