A Predictive Nomogram for Postoperative Hydrocephalus After Intra- and Paraventricular Tumor Resection: A Retrospective Study of 196 Patients.

BACKGROUND:Hydrocephalus is a common complication of intra- and paraventricular tumors both before and after tumor resection. We investigated the risk factors for postoperative hydrocephalus and established a predictive nomogram to improve surgical planning and adjust the frequency of postoperative clinical and radiographic monitoring. METHODS:A retrospective study of 196 intra- and paraventricular tumor patients with follow-up data was conducted to investigate the risk factors for postoperative hydrocephalus via univariate and multivariate Cox regression analyses. We developed a nomogram incorporating these factors based on multivariate analysis and according to scores calculated by the predictive model; a hazard function curve was used to compared risk of hydrocephalus. RESULTS:Among the 196 patients, 33 had postoperative hydrocephalus. Intraventricular tumor (P = 0.005), glioblastoma (P = 0.010), preoperative hydrocephalus (P = 0.007), and radiotherapy (P = 0.033) were independent risk factors for postoperative hydrocephalus. The nomogram including these independent risk factors had moderate predictive accuracy, with a concordance index of 0.716 (95% confidence interval [CI]: 0.605-0.828), while the area under the curve values at 6, 12, and 24 months were 0.708 (95% CI: 0.563-0.853), 0.763 (95% CI: 0.656-0.870) and 0.861 (95% CI: 0.779-0.943), respectively. The hazard function showed differences between the lower and higher nomogram score groups. The higher the nomogram score, the higher the risk of postoperative hydrocephalus (P < 0.001). CONCLUSIONS:The established nomogram performs well for predicting postoperative hydrocephalus. Clinicians can use this nomogram to review their practice regarding hydrocephalus associated with intra- and paraventricular tumors, plan surgical treatment and adjust the frequency of postoperative clinical and radiographic monitoring.