Preoperative prediction of postsurgical outcomes in mass-forming intrahepatic cholangiocarcinoma based on clinical, radiologic, and radiomics features
- Authors
- Park, Hyo Jung; Park, Bumwoo; Park, Seo Young; Choi, Sang Hyun; Rhee, Hyungjin; Park, Ji Hoon; Cho, Eun-Suk; Yeom, Suk-Keu; Park, Sumi; Park, Mi-Suk; Lee, Seung Soo
- Issue Date
- Nov-2021
- Publisher
- SPRINGER
- Keywords
- Cholangiocarcinoma; Image processing; Multidetector computed tomography; Precision medicine; Prognosis; computer-assisted
- Citation
- EUROPEAN RADIOLOGY, v.31, no.11, pp.8638 - 8648
- Indexed
- SCIE
SCOPUS
- Journal Title
- EUROPEAN RADIOLOGY
- Volume
- 31
- Number
- 11
- Start Page
- 8638
- End Page
- 8648
- URI
- https://scholar.korea.ac.kr/handle/2021.sw.korea/135833
- DOI
- 10.1007/s00330-021-07926-6
- ISSN
- 0938-7994
- Abstract
- Objectives Current prognostic systems for intrahepatic cholangiocarcinoma (IHCC) rely on surgical pathology data and are not applicable to a preoperative setting. We aimed to develop and validate preoperative models to predict postsurgical outcomes in mass-forming IHCC patients based on clinical, radiologic, and radiomics features. Methods This multicenter retrospective cohort study included patients who underwent curative-intent resection for mass-forming IHCC. In the development cohort (single institution data), three preoperative multivariable Cox models for predicting recurrence-free survival (RFS) were constructed, including the clinical-radiologic, radiomics, and clinical-radiologic-radiomics (CRR) models based on clinical and CT findings, CT-radiomics features, and a combination of both, respectively. Model performance was evaluated in the test cohort (data from five institutions) using Harrell's C-index and compared with postoperative prognostic systems. Results A total of 345 patients (233, development cohort; 112, test cohort) were evaluated. The clinical-radiologic model included five independent CT predictors (infiltrative contour, multiplicity, periductal infiltration, extrahepatic organ invasion, and suspicious metastatic lymph node) and showed similar performance in predicting RFS to the radiomics model (C-index, 0.65 vs. 0.68; p = 0.43 in the test cohort). The CRR model showed significantly improved performance (C-index, 0.71; p = 0.01) than the clinical-radiologic model and demonstrated similar performance to the postoperative prognostic systems in predicting RFS (C-index, 0.71-0.73 vs. 0.70-0.73; p >= 0.40) and overall survival (C-index, 0.68-0.71 vs. 0.64-0.74; p >= 0.27) in the test cohort. Conclusions A model integrating clinical, CT, and radiomics information may be useful for the preoperative assessment of postsurgical outcomes in patients with mass-forming IHCC.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - College of Medicine > Department of Medical Science > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholar.korea.ac.kr/handle/2021.sw.korea/135833)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.