Development and Validation of Nomograms to Provide Individualized Predictions of Survival Benefits from Surgery in Patients with Intermediate/advanced Hepatocellular Carcinoma
Xin-Fei Xu,Ju-Dong Li,Jiong-Jie Yu,Hao Xing,Zheng Wang,Jun Han,Ming-Da Wang,Chao Li,Han Zhang,Han Wu,Wen-Tao Yan,Jia-He Wang,Bing Quan,Ya-Hao Zhou,Wei-Min Gu,Hong Wang,Ting-Hao Chen,Feng Shen,Meng-Chao Wu,Tian Yang
DOI: https://doi.org/10.1200/jco.2018.36.15_suppl.e16112
IF: 45.3
2018-01-01
Journal of Clinical Oncology
Abstract:e16112 Background: According to the BCLC treatment guidelines, surgery is not recommended for intermediate/advanced hepatocellular carcinoma (HCC). In real world, however, liver resections are often performed in patients with intermediate/advanced but resectable HCC, especially in the East. Methods: We retrospectively evaluated multicentric data of 1,325 patients newly diagnosed with intermediate/advanced HCC who underwent curative resection. We randomly divided the subjects into development (n = 875) and validation (n = 450) samples. Multivariate Cox proportional hazards models were developed and separately validated on the basis of patients’ clinicopathological variables assessed for associations with 1-year recurrence and 3-year mortality. The discriminatory accuracy of these models was compared with conventional tools by analyzing receiver operating characteristic (ROC) curves. Results: The statistical nomograms built based on performance status, Child-Pugh grade, portal hypertension, preoperative alpha-fetoprotein level, tumor rupture, largest tumor diameter, tumor number, macrovascular and microvascular invasion, and satellites had good calibration and discriminatory abilities, with c-indices of 0.70 (1-year recurrence) and 0.68 (3-year survival), respectively. These models showed satisfactory goodness-of-fit and discrimination abilities in the validation cohort (c-index, 0.68 for 1-year recurrence and 0.69 for 3-year survival). The areas under the ROC curve using these nomograms exceeded those of traditional staging systems, indicating superior discriminatory capability (c-indices, 0.60-0.63 and 0.56-0.62, respectively). Conclusions: Our proposed online nomograms, which present graphically postoperative prognostic models for recurrence and survival in patients with intermediate/advanced but resectable HCC, offer valuable guidance to surgeons and hepatologists for individually predicting survival benefits from surgery and planning recurrence surveillance and adjuvant therapy.