Room: Room 202
Purpose: To develop and validate a radiomics nomogram based on radiomics features and clinical factors for preoperatively prediction early recurrence (ER) of intrahepatic cholangiocarcinoma (ICC) after partial hepatectomy.
Methods: In this study, a prediction model was developed by the training cohort involving 139 ICC patients recorded from January 2010 to June 2014. Radiomics features were extracted from contrast-enhanced magnetic resonance imaging (MRI) arterial-phase images. Feature selection was preformed in two steps of Spearmanâ€™s rank correlation test and Least absolute shrinkage and selection operator (LASSO) regression. The radiomics signature was constructed from the selected features weighted by LASSO coefficients. Combining with clinical characteristics, a radiomics nomogram was developed by the multivariable logistic regression model. An independent validation cohort involved 70 patients, recorded from July 2014 to March 2016. The nomogram performance was evaluated in terms of discrimination and calibration using area under the curve (AUC) and Hosmerâ€“Lemeshow test.
Results: The radiomics signature, consisting of 9 features, showed significant differences between ER group and non-ER group (P < 0.01, both in training and validation cohorts). The AUC of the radiomics signature for the training and validation cohort was 0.82 (95% CI, 0.74 to 0.88), 0.77 (95% CI, 0.65 to 0.86), respectively. The AUC of the radiomics nomogram combining the radiomics signature and clinical stage was 0.90 (95% CI, 0.83 to 0.94) in the training cohort, 0.86 (95% CI, 0.76 to 0.93) in the validation cohort. The calibration curves of the nomogram showed good agreement of ER probability between prediction and observation in both training and validation cohorts.
Conclusion: The radiomics nomogram, a non-invasive preoperative prediction model, is capable in predicting early recurrence of ICC after partial hepatectomy.