Statistical analysis of prognostic factors for survival in patients with spinal metastasis

• Published in: Acta medica Okayama Vol. 66; no. 3; pp. 213 - 219
• Main Authors: Kataoka, Masaki, Kunisada, Toshiyuki, Tanaka, Masato, Takeda, Ken, Itani, Satoru, Sugimoto, Yoshihisa, Misawa, Haruo, Senda, Masuo, Nakahara, Shinnosuke, Ozaki, Toshifumi
• Format: Journal Article
• Language: English
• Published: Japan 01-06-2012
• Subjects: Adolescent; Adult; Age; Aged; Aged, 80 and over; Bone; Chemotherapy; Child; Child, Preschool; Female; Humans; Male; Metastases; Middle Aged; Neurological diseases; Pain; Prognosis; Proportional Hazards Models; Retrospective Studies; Spinal Neoplasms - mortality; Spinal Neoplasms - secondary; Statistical analysis; Survival; Survival Rate; Tumors
• ISSN: 0386-300X

There are a variety of treatment options for patients with spinal metastasis, and predicting prognosis is essential for selecting the proper treatment. The purpose of the present study was to identify the significant prognostic factors for the survival of patients with spinal metastasis. We retrospectively reviewed 143 patients with spinal metastasis. The median age was 61 years. Eleven factors reported previously were analyzed using the Cox proportional hazards model:gender, age, performance status, neurological deficits, pain, type of primary tumor, metastasis to major organs, previous chemotherapy, disease-free interval before spinal metastasis, multiple spinal metastases, and extra-spinal bone metastasis. The average survival of study patients after the first visit to our clinic was 22 months. Multivariate survival analysis demonstrated that type of primary tumor (hazard ratio [HR] = 6.80, p < 0.001), metastasis to major organs (HR = 2.01, p = 0.005), disease-free interval before spinal metastasis (HR = 1.77, p = 0.028), and extra-spinal bone metastasis (HR = 1.75, p = 0.017) were significant prognostic factors. Type of primary tumor was the most powerful prognostic factor. Other prognostic factors may differ among the types of primary tumor and may also be closely associated with primary disease activity. Further analysis of factors predicting prognosis should be conducted with respect to each type of primary tumor to help accurately predict prognosis.