Japanese researchers have developed an AI-powered prognostic model that significantly improves treatment planning for patients with spinal metastasis by accurately predicting one-year survival using modern clinical data and machine learning.
Spinal metastasis, a common complication in advanced cancer, often leads to severe pain, neurological deficits, and reduced quality of life. Treatment decisions—ranging from aggressive surgical intervention to palliative care—depend heavily on expected survival. However, widely used prognostic scoring systems are based on outdated data and fail to reflect recent advances in oncology, including targeted therapies and immunotherapy that have extended patient survival.
In a study published in Spine, researchers from Nagoya University Graduate School of Medicine reported the development of a simplified yet highly accurate survival prediction model built on large-scale, prospective clinical data from patients receiving contemporary cancer treatments.
The model was derived from a multicenter prospective study involving 401 patients who underwent surgery for spinal metastasis across 35 medical institutions in Japan between 2018 and 2021. Using a machine learning technique known as Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression, the researchers identified variables most strongly associated with one-year survival. Model performance was assessed using standard statistical measures, including the area under the receiver operating characteristic curve and calibration analysis.
Five preoperative factors emerged as key predictors of survival. These included the vitality index reflecting patient motivation and psychological condition, age 75 years or older, Eastern Cooperative Oncology Group (ECOG) performance status indicating functional ability, the presence of bone metastases outside the spine, and preoperative opioid use, which may be associated with immunosuppression and accelerated tumor progression. All five factors can be assessed clinically without the need for specialized equipment.
The model demonstrated strong predictive performance, with an AUROC of 0.762, and enabled patients to be stratified into three distinct risk categories. Patients in the low-risk group showed an 82.2 percent one-year survival rate, those in the intermediate-risk group had a 67.2 percent survival rate, and those classified as high-risk had a 34.2 percent one-year survival rate.
By offering a real-time, evidence-based survival estimate grounded in modern treatment outcomes, the model supports more informed surgical decision-making and individualized postoperative care planning.
While the system was developed using Japanese clinical data, the research team plans to validate and adapt the model using international datasets, with the goal of improving spinal metastasis care globally.
Also Read