A nomogram to predict the progression-free survival of clival chordoma

Yixuan Zhai,Jiwei Bai,Mingxuan Li,Shuai Wang,Chuzhong Li,Xinting Wei,Yazhuo Zhang
DOI: https://doi.org/10.3171/2019.10.jns192414
IF: 5.408
2021-01-01
Journal of Neurosurgery
Abstract:OBJECTIVE Chordoma shows poor patient prognosis because of its high recurrence rate. Even though many clinical factors and biomarkers are reported to be associated with prognosis, no prediction model has been applied clinically. Thus, the authors aim to derive and validate a prognostic nomogram to predict progression-free survival (PFS) of chordoma. METHODS A total of 201 patients were randomly divided into a derivation group (151 cases) and a validation group (50 cases). The expression levels of biomarkers were quantified using tissue microarray analysis. A nomogram was established via univariate and multivariate Cox regression analysis in the derivation group. The predictive performance of the nomogram was then tested in the validation group. RESULTS The mean follow-up interval was 57 months (range 26–107 months). One clinical factor and 3 biomarkers were confirmed to be associated with PFS, including degree of resection, E-cadherin, Ki-67, and VEGFA. The nomogram with these prognostic factors had areas under the receiver operating characteristic curve of 0.87 and 0.95 in the derivation group at 3 years and 5 years, respectively, compared with 0.87 and 0.84 in the validation group. Calibration and score-stratified survival curve were good in the derivation group and validation group, respectively. CONCLUSIONS The established nomogram performs well for predicting the PFS of chordoma and for risk stratification, which could facilitate prognostic evaluation and follow-up.
surgery,clinical neurology
What problem does this paper attempt to address?