Construction and Validation of a Survival Prediction Model for Patients After Vertebroplasty

To analyze the risk factors that affect the survival of patients undergoing vertebroplasty and construct a predictive nomogram.

Retrospective analysis of the survival status for patients age ≥50 years who underwent vertebroplasty in our hospital from January 2013 to August 2022. Demographic information, inpatient data, laboratory examination results, medication records, and other information were extracted from the clinical scientific research database of our hospital. Through proportional hazards assumption, univariate and subsequent multivariate COX regression, the independent risk factors that affect the survival prognosis of patients after vertebroplasty were summarized. A survival prediction nomogram based on these independent risk factors were constructed and validated.

Three hundred fifty-nine patients were enrolled, 251 in the training set and 108 in the validation set. Multivariate COX regression showed that mean serum albumin (hazard ratio [HR] = 0.59565, 95% confidence interval [CI], 0.36160 to 0.9812), number of vertebroplasty (HR = 0.1978, 95% CI, 0.06529 to 0.2197), interval between the first two vertebroplasty procedures (HR = 0.05642, 95% CI, 0.02933 to 0.1085), and number of activating vitamin D prescriptions (HR = 0.34975, 95% CI, 0.19855 to 0.6161) were independent risk factors for the survival prognosis of patients after vertebroplasty. Based on these independent risk factors, a predictive nomogram was constructed. The area under the curve of the 5- and 8-year survival prediction models in the validation set was 0.889 and 0.760, respectively. The calibration curves of the nomogram in the training and validation sets were close to the ideal diagonal. The decision curve analysis showed that the predictive model exhibited good net benefit and predictive ability.

Mean serum albumin, number of vertebroplasty, interval between the first two vertebroplasty procedures, and number of activating vitamin D prescriptions were independent risk factors for the survival prognosis of patients after vertebroplasty. The predictive nomogram constructed based on these risk factors had a good predictive ability and certain potential for clinical decision making.