Abstract:BACKGROUND CONTEXT: Accurately predicting the survival of patients with spinal metastases is important for guiding surgical intervention. The SORG machine-learning (ML) algorithm for the 90-day and one-year mortality of patients with metastatic cancer to the spine has been multiply validated, with a high degree of accuracy in both internal and external validation studies. However, prior external validations were conducted using patient groups located on the east coast of the United States, representing a generally homogeneous population. The aim of this study was to externally validate the SORG algorithms with a Taiwanese population. STUDY DESIGN/SETTING: Retrospective study at a single tertiary care center in Taiwan PATIENT SAMPLE: Four hundred and twenty-seven patients who underwent surgery for metastatic spine disease from November 1, 2010 to December 31, 2018 OUTCOME MEASURES: 90-day and one-year mortality METHODS: The baseline characteristics of our validation cohort were compared with those of the previously published developmental and external validation cohorts. Discrimination (c-statistic and receiver operating curve), calibration (calibration plot, intercept, and slope), overall performance (Brier score), and decision curve analysis were used to assess the performance of the SORG ML algorithms in this cohort. RESULTS: Ninety-day and one-year mortality rates were 110 of 427 (26%) and 256 of 427 (60%), respectively. The external validation cohort and the developmental cohort differed in body mass index (BMI), preoperative performance status, American Spinal Injury Association impairment scale, primary tumor histology and in several laboratory measurements. The SORG ML algorithm for 90-day and 1-year mortality demonstrated a high level of discriminative ability (c-statistics of 0.73 [95% confidence interval [CI], 0.67-0.78] and 0.74 [95% CI, 0.69-0.79]), overall performance, and had a positive net benefit throughout the range of threshold probabilities in decision curve analysis. The algorithm for 1-year mortality had a calibration intercept of 0.08, representing a good calibration. However, the 90-day mortality algorithm underestimated mortality for the lowest predicted probabilities, with an overall intercept of 0.81. CONCLUSIONS: The SORG algorithms for predicting 90-day and 1-year mortality in patients with spinal metastatic disease generally performed well on international external validation in a predominately Taiwanese population. However, 90-day mortality was underestimated in this group. Whether this inconsistency was due to different primary tumor characteristics, body mass index, selection bias or other factors remains unclear, and may be better understood with further validative works that utilize international and/or diverse populations. (C) 2021 Elsevier Inc. All rights reserved.

Evaluation of Different Scoring Systems for Spinal Metastases Based on a Chinese Cohort.

Advancement in Prognostic Scoring Systems for Patients with Spine Metastasis

Scoring Algorithms for Predicting Survival Prognosis in Patients With Metastatic Spinal Disease <i>The Current Status and Future Directions</i>

Risk factors and prognosis of surgery for spinal metastasis

The emergence of new prognostic scores in lung cancer patients with spinal metastasis: A 12-year single-center retrospective study

Which Prognostic Model Best Predicts Poor Prognosis in Patients with Spinal Metastases? A Comparative Analysis of 8 Scoring Systems

The Predictive Value of the Tokuhashi Revised Scoring System for the Survival Time of Patients with Spinal Metastases

Comparison of three predictive scoring systems for morbidity in oncological spine surgery

[Predictive Value of Three Surgical Scoring Systems for Estimation of Life Expectancy in Patients with Extradural Spinal Metastasis].

Can Survival Scoring Systems for Spinal Metastases be Used to Predict Postoperative Neurologic Recovery? A Retrospective Study on 204 Patients With Thoracolumbar Metastases Treated at a Tertiary Center

Development and Validation of a Novel Survival Prediction Model in Patients with Spinal Metastasis from Non-small Cell Lung Cancer.

Comparison of eight modern preoperative scoring systems for survival prediction in patients with extremity metastasis

Prediction of Survival Prognosis for Spinal Metastasis from Cancer of Unknown Primary: Derivation and Validation of a Nomogram Model.

Modified Score Based on Revised Tokuhashi Score is Needed for the Determination of Surgical Intervention in Patients with Lung Cancer Metastases to the Spine

The significance of Tomita scoring system in spinal metastases

Comparison of Classically and Machine Learning Generated Survival Prediction Models for Patients with Spinal Metastasis - A Meta-Analysis of Two Recently Developed Algorithms.

Validation Of A Scoring System Predicting Survival And Function Outcome In Patients With Metastatic Epidural Spinal Cord Compression (Mescc): A Prospective And Multicenter Study

International External Validation of the SORG Machine Learning Algorithms for Predicting 90-Day and One-Year Survival of Patients with Spine Metastases Using a Taiwanese Cohort

[Estimation of Life Expectancy for Selecting Surgical Procedure and Predicting Prognosis of Extradural Spinal Metastases].

Establishment and validation of a novel survival prediction scoring algorithm for patients with non-small-cell lung cancer spinal metastasis

A Machine Learning Algorithm for Predicting 6-Week Survival in Spinal Metastasis: an External Validation Study Using 2,768 Taiwanese Patients.