Metastatic Chordoma Is Associated with Significantly Shorter Progression-Free Survival following Resection

Background: Metastasis from chordoma is relatively uncommon and understudied. When they occur, metastatic lesions are most commonly found in the lungs and bony structures around the skull base. Cases of metastasis to every organ system and those from iatrogenic seeding following endoscopic endonasal resection have been reported. Metastatic disease is associated with poor prognosis. Sample sizes of studies evaluating predictive factors for and outcomes of metastatic chordoma are small. The objectives of this analysis were to utilize a larger sample size with data from an experienced skull base center to evaluate proposed predictive factors and outcomes for metastatic chordoma.

Methods: Medical records of patients who underwent surgical resection for chordoma within the UPMC system from 2001 to 2020 were gathered through the electronic medical record software. The data were statistically analyzed using R Statistical Software. Patients who developed metastatic disease at any point in their treatment course were identified. Variables were entered into a Cox proportional hazards model for progression-free survival following surgery (PFSS). Kaplan–Meier survivorship curves were created to depict progression-free survival versus diagnosis of metastasis, and the log-rank test was used to calculate p values.

Results: This analysis included data from 194 patients who underwent a total of 270 operations. The mean age of diagnosis of our cohort was 43.6 years. Of these resections, 65% were for recurrent chordoma. 187 patients had chordoma of the skull and 20 patients had chordoma of the spine. Our cohort consists of 20 patients with metastatic chordoma (Table 1). Of these patients, 40% were males with a mean age of 40.9. 17 patients had a primary location of the skull and 8 had chordoma of the spine. The average time from diagnosis to metastasis was 70.3 months. The most common site of metastasis is the spine, followed by the lungs, then followed by intradural drop metastasis (Table 2). Ki67, 1p loss, and 10q loss showed statistically significant differences (p < 0.05). Univariate analysis showed that a diagnosis at any time point during treatment and a diagnosis of metastasis at the time of surgery showed increased risk (HR 3.8996 and 3.0391, respectively) for decreased PFSS (p < 0.05; Table 4). As evidenced by the Kaplan–Meier survivorship curve, patients diagnosed with metastasis at any time point during their treatment have decreased PFSS (HR: 3.9 [2.613–5.819], p < 0.05; [Fig. 1]). Patients who had a diagnosis of metastasis at the time of surgery ([Fig. 2]) also had a significantly decreased PFSS (HR: 3.039 [1.661–5.561], p < 0.05), but this trend disappeared when comparing within the group ([Fig. 3]) of patients who had ever been diagnosed with metastasis during their treatment course (HR: 0.9733 [0.4864–1.948], p = 0.93).

Conclusion: Metastatic chordomas are associated with significantly shorter progression-free survival and our data suggests that it is the phenotype of the potentially metastatic tumor that may contribute to poor outcomes and not the metastasis itself.

No conflict of interest has been declared by the author(s).

Publication History

Article published online:
12 February 2021