Predicting Skeletal-related Events Using SINS

Kazuo Nakanishi; Yasukazu Hijikata; Kazuya Uchino; Yoshihisa Sugimoto; Hideaki Iba; Seiya Watanabe; Shigeru Mitani

doi:10.1097/BRS.0000000000004983

Predicting Skeletal-related Events Using SINS

Spine (Phila Pa 1976). 2024 Mar 13. doi: 10.1097/BRS.0000000000004983. Online ahead of print.

Authors

Kazuo Nakanishi¹, Yasukazu Hijikata², Kazuya Uchino¹, Yoshihisa Sugimoto¹, Hideaki Iba¹, Seiya Watanabe³, Shigeru Mitani³

Affiliations

¹ Kawasaki Medical School Department of Orthopedics, Traumatology and Spine Surgery.
² Doctor of Public Health Kyoto University Section of Clinical Epidemiology, Department of Community medicine.
³ Kawasaki Medical School Department of Bone and Joint Surgery.

PMID: 38475677
DOI: 10.1097/BRS.0000000000004983

Abstract

Study design: Predictive study utilized retrospectively collected data.

Objective: The primary objective was to evaluate the predictive association between the Spine Instability Neoplastic Score (SINS) and Skeletal-related events (SREs). Secondary objectives included examining characteristics of cases with SINS < 6 among those who developed SRE, and evaluating the impact of additional predictors on prediction accuracy.

Summary of background data: Advances in cancer treatment have prolonged the lives of cancer patients, emphasizing the importance of maintaining quality of life. Skeletal-related events from metastatic spinal tumors significantly impact quality of life. However, currently, there is no scientifically established method to predict the occurrence of SRE. SINS, developed by the Spine Oncology Study Group, assesses spinal instability using six categories. Therefore, the predictive performance of SINS for SRE occurrence is of considerable interest to clinicians.

Methods: This predictive study utilized retrospectively collected data from a single-center registry comprising over 1,000 patients with metastatic spinal tumors. SINS and clinical data were collected. Logistic regression was used to create a prediction equation for SRE using SINS. Additional analyses explored factors associated with SRE in patients with SINS < 6.

Results: The study included 1,041 patients with metastatic spinal tumors. SRE occurred in 121 cases (12%). The prediction model for SRE using SINS demonstrated an area under the curve (AUC) of 0.832. Characteristics associated with SRE included lower female prevalence, surgeries to primary sites, bone metastases to non-spinal sites, and metastases to other organs. A post hoc analysis incorporating additional predictors improved the AUC to 0.865.

Conclusion: The SINS demonstrated reasonable predictive performance for SRE within one month of the initial visit. Incorporating additional factors improved prediction accuracy. The study emphasizes the need for a comprehensive clinical prediction model for SRE in metastatic spinal tumors.