A new preoperative model with good predictive performance for overall survival among intrahepatic cholangiocarcinoma (iCC) patients was described in a recent study published in the Journal of Ultrasound in Medicine.
The newly developed method by Li et al consisted of a preoperative nomogram integrated with an ultrasound-based signature of both radiomic and radiographic features. This method had better performance in predicting overall survival in iCC patients than the Tumor Node Metastasis (TNM) staging system, the authors said.
“The preoperative nomogram with [carbohydrate antigen 19-9], sex, ascites, radiomics signature, and radiographic signature had [concordance indexes (C-indexes)] of 0.72 and 0.75 in the training and testing cohorts, respectively, and it had significantly higher predictive performance than the 8th TNM staging system in the testing cohort (C-index, 0.75 vs 0.67, P =.004) and a higher C-index than the radiomics nomograms (0.75 vs 0.68, P =.044),” the authors explained.
Read more about CC prognosis
The method also demonstrated better performance when compared with models based on radiographic or radiomic features alone.
To construct the nomogram, the authors selected predictors from the radiographic-radiomic signatures, as well as clinical and laboratory results from a cohort of 170 patients with iCC who underwent curative resection. The method was tested in a training cohort (n=127) and validated in an independent testing cohort (n=43) using the C-indexes.
To evaluate the performance of the new preoperative model, the authors compared the output with that of the TNM staging system and other previously developed radiographic and radiomic nomograms.
“The preoperative model based on imaging mining with [artificial intelligence] and human collaboration may maximize the benefits for treatment decision-making and can be easily and non-invasively applied before surgery,” according to Li et al.
Li M-D, Lu X-Z, Liu J-F, et al. Preoperative survival prediction in intrahepatic cholangiocarcinoma using a ultrasound-based radiographic-radiomics signature. J Ultrasound Med. Published online September 21, 2021. doi:10.1002/jum.15833