Prospective external validation of a three-predictor frailty model for 90-day survival and complications following spinal metastasis surgery

Abstract

Background context

Surgical decision-making in patients with spinal metastases remains complex due to the need to balance potential surgical benefits with limited survival and common frailty. Predictive models can assist in this process, but their clinical utility is often limited by complexity and lack of validation.

Purpose

To externally validate a simple three-predictor frailty model for 90-day survival and complications, and to compare its performance with other commonly used tools.

Study design/setting

Prospective external validation study conducted at a single tertiary cancer center.

Patient sample

A consecutive cohort of 126 patients who underwent open posterior surgery with instrumentation for spinal metastases from solid tumors between 2018 and 2024.

Outcome measures

Primary outcomes were 90-day survival and the occurrence of postoperative complications. Secondary outcomes included 30-day, 180-day and overall survival. Model performance was evaluated through discrimination (AUC), risk stratification, accuracy for surgical indication and calibration.

Methods

The Anzuategui model (three predictors: tumor growth rate, comorbidities, and lymphocyte count) was applied preoperatively, along with four other three-predictor models (Tomita, Modified Bauer, Van der Linden, and Sioutos). Discrimination was assessed using ROC curves. Risk stratification was evaluated using predefined low-, moderate-, and high-risk categories, analyzed through Kaplan–Meier curves and complication rates. Model accuracy for surgical indication was calculated using a 90-day survival threshold as the reference. Calibration for both 90-day survival and postoperative complications was performed by comparing category-specific predicted probabilities derived from the development cohort with observed event rates in the validation cohort.

Results

The Anzuategui model demonstrated predictive performance for the primary outcomes comparable to the other models under evaluation. It achieved an AUC of 0.78 (95% CI: 0.70–0.85) for 90-day survival and 0.68 (95% CI: 0.59–0.76) for postoperative complications. Risk stratification showed clear separation between survival curves across the three predefined categories. Accuracy for predicting appropriate surgical indication was 70% (95% CI: 61–78), with a sensitivity of 64% and specificity of 85%. Tomita and Modified Bauer models showed comparable accuracy (75% and 74%, respectively) but lower specificity. Calibration indicated overestimation of 90-day mortality (intercept –1.75; slope 2.05) and modest miscalibration for postoperative complications (intercept –0.40; slope 0.67).

Conclusions

The Anzuategui model demonstrated acceptable external performance, with greater validity for predicting 90-day survival than for postoperative complications. Its simplicity and frailty-centered structure make it a practical bedside tool, particularly in urgent or resource-limited settings. Integrating this approach with established prognostic models may support more balanced decision-making across diverse clinical scenarios.