Lumbosacral chordoma. Prognostic factors and treatment.

STUDY DESIGN: Retrospective analysis. OBJECTIVES: To analyze the prognostic factors in patients with chordomas, the success of various treatments, the diagnostic value of open versus needle biopsy, the neurologic impairment after sacral nerve resection, and the clinical presentation and site of origin. SUMMARY OF BACKGROUND DATA: Staging of chordomas has not been of much value, compared with other bone tumors, because for chordomas, grade is similar, metastasis is infrequent at presentation, and the prognostic significance of size is uncertain. METHODS: A review of patients with chordoma from 1965 through 1996 found 23 cases (mean age of patients, 55 years). The mean follow-up was 84 months. Mean tumor size was 81 mm (range, 35-135 mm), location was lumbar (n = 6), S1 (n = 4), S2 (n = 3), S3 (n = 7), S4 (n = 2), and S5 (n = 1). RESULTS: No tumors were found in the higher sacrum (S1-S2) alone, without involvement of the lower sacrum. Survival analysis at 5 years showed overall survival (OS) 86%, continuous disease-free survival (CDFS) 58%, and local recurrence-free survival (LRFS) 60%. The location of tumor, defined by highest level of involvement (lumbar vs. sacrum) was of prognostic significance for OS (P = 0.01; log-rank test), CDFS (P = 0.036), but not for LRFS (P = 0.189). Results of multivariate regression showed that location was significant for OS (P = 0.007), CDFS (P = 0.008), and LRFS (P = 0.001). For patients with positive margins (n = 16), initial radiation correlated with longer CDFS (P = 0.002; Mantel-Cox) and LRFS (P = 0.005, Mantel-Cox), but was not significant for OS (P = 0.41). For patients who received no radiation, a positive margin correlated with a shorter CDFS (P = 0.04), a trend to shorter LRFS (P = 0.08), but no difference in OS. Therefore, both a tumor-free margin and initial radiation correlated with a longer survival. No patients had urinary or bowel dysfunction when both S3 nerves were preserved. If one S3 nerve was preserved, 1 of 3 patients had partial urinary incontinence and 2 of 3 patients required bowel medications. If both S3 nerves were resected, all patients required intermittent urinary catheterization and bowel medications. If both S2 nerves were resected, there was complete urinary and bowel incontinence. CONCLUSIONS: The highest level of tumor involvement was prognostically significant for OS, CDFS, and LRFS. Radiation was of value when complete excision was not achieved. Bilateral S3 nerve preservation is necessary to ensure retention of normal urinary and bowel function.