Development and validation of the NOVA score for surgical triage in metastatic spinal cord compression.

STUDY

DESIGN

Retrospective diagnostic model development and validation study.

OBJECTIVES

To develop and validate the NOVA (Nottingham Oncologic Vertebral Algorithm) Score, a pragmatic, surgeon-oriented tool for predicting 12-month survival in patients with metastatic spinal cord compression (MSCC), and to compare its performance to established prognostic scores and oncologist estimates.

METHODS

Two independent cohorts of patients with MSCC referred for surgical consideration were included from a tertiary spine center: a derivation cohort (n = 184) and a validation cohort (n = 100). Feature selection was performed using random forest analysis, multivariable logistic regression, and permutation-based variable importance. The final NOVA Score incorporated three clinical parameters: Karnofsky Performance Status (KPS), primary tumor category, and presence of extraspinal metastases. The score ranges from 0 to 10, with a threshold of ≥ 7 used to predict > 12-month survival. Predictive performance was assessed using accuracy, F1-score, area under the receiver operating characteristic curve (AUC), Cohen's kappa, and calibration plots.

RESULTS

In the validation cohort, the NOVA Score achieved an accuracy of 71.0%, AUC of 0.693, F1-score of 47.3% for > 12-month survival, and Cohen's kappa of 0.28. Calibration plots showed good agreement in the mid-to-high probability range. Compared to the Revised Tokuhashi Score, OSRI, Modified Bauer Score, and oncologist estimates, NOVA demonstrated superior sensitivity for identifying long-term survivors using only three readily available variables.

CONCLUSIONS

The NOVA Score is a simple, reproducible tool for early prognostication in MSCC. Its minimal input requirements and balanced performance support its utility for surgical triage in multidisciplinary settings. Further external validation is warranted.