Exploring clinical and demographic determinants of time to treatment in spinal metastases: A Large-Scale cohort study of more than 1000 patients

Authors:
Publisher website
Google Scholar
Abstract
Objective To identify predictors for time to first treatment modality (TTT) in patients diagnosed with spinal metastatic disease. Methods Our radiology database was retrospectively queried to identify patients with spinal metastases (01/2017-12/2021). Demographics, setting of initial imaging (inpatient/outpatient/ED), histology, treatment modality, and TTT were collected. Several different early TTT thresholds were explored (≤ 30, ≤ 21, ≤14, and ≤ 7 days). After controlling for age, gender, diagnostic setting, pathology, type of treatment multivariable linear and logistic regressions were performed for the primary and secondary thresholds at different time-points. Results A total of 1009 patients with spinal metastases were included, with a median age of 65 years (IQR 15) and 529 (52.4%) males. Diagnoses occurred in the inpatient (37%), outpatient (50%), and emergency department (13%) settings. The most common primary tumors were breast (22.4%), prostate (19.5%), and non-small cell lung cancer (12.8%). Treatment modalities included surgery (14.6%), chemotherapy (54.5%), and radiation therapy (43.5%). The median time to treatment (TTT) was 13 days (SD 27, range 0–676), with no significant difference by setting (p = 0.2). In multivariable linear regression, surgery was associated with shorter TTT (p = 0.037). Logistic regression using different early-TTT thresholds identified age, surgical intervention, lymphoma, and outpatient setting as significant predictors across models. Sensitivity analyses excluding TTT outliers showed that surgery was no longer a significant predictor, while other covariates remained stable. Conclusions Surgical intervention and lymphoma independently predicted earlier TTT of any modality in spinal metastases, whereas older age and outpatient diagnosis were associated with delays.

This publication has 43 references indexed in Scilit: