External validity of the Tokuhashi score in patients with vertebral metastasis

To calculate the accuracy of the Tokuhashi score (TS) in recent patients with vertebral metastasis (VM), candidates or not to surgical treatment, and thus to assess the external validity of TS.

Retrospective analysis of prospectively collected data from 90 patients (55 men, 35 women) with VM between 2004 and 2006. For each patient, data on the primary tumor (PT), date of diagnosis, TS at the time of VM diagnosis and date of death were retrieved from the electronic medical records and civil registry. True survival time and TS survival time were estimated to calculate the accuracy rate of the TS. A Kaplan–Meier analysis was used to study the survival function by prognostic groups. A correlation study between survival time and other variables was performed.

PT distribution was as follows: breast (22.2 %), lung (20 %), prostate (17.8 %), rectum (10 %), unknown (11 %), and others (18 %). Average overall survival after the VM diagnosis was 11.8 months (SD, 11 m): breast, 20 months (SD, 20 m); lung, 5.8 months (SD, 5.9 m); prostate, 14.5 months (SD, 13.4 m); rectum, 9.4 months (SD, 9.3); and unknown tumors, 2.7 months (SD, 5 m). Survival time was accurately predicted with the TS in 63 % of patients with a short life expectancy (survival, <6 months; TS, 0–8), 16 % of patients in the intermediate group (survival, 6–12 months; TS, 9–11), and 77 % of patients with a good prognosis (survival >12 months; TS, 12–15). By specific PT, the accuracy rate of the TS was low for breast cancer metastasis (35 %). The Kaplan–Meier curves show a significant separation among the prognostic groups (p < 0.05), but the log-rank test showed a statistically significant difference in survival only between short expectancy group and good prognostic group. Age at PT diagnosis and at VM diagnosis negatively correlated with survival (r = 0.22; p = 0.032 and r = 0.3, p = 0.04).

The TS was not highly accurate for predicting survival in patients with VM, treated or not surgically, and it was particularly imprecise in patients with an intermediate score (9–11 points) and those with breast cancer, so it is possible that the TS currently has a poor external validity.