Skull Base Chordoma and Chondrosarcoma: Changes in National Radiotherapy Patterns and Survival Outcomes

Henry S. Park, MD, MPH; Kenneth B. Roberts, MD; James B. Yu, MDDepartment of Therapeutic Radiology, Yale University School of Medicine

Table 1: Demographic and clinical information of patients receiving RT vs. those not receiving RT.

Characteristic RT(n, %)

(n=170)

No RT (n=%) (n=196)

P-value

Age (median ± SD, years) 49 ± 17 43 ± 16 0.011 Sex 0.946 Male 87 (46) 101 (54) Female 83 (47) 95 (53) Race 0.434 White 144 (47) 160 (53) Non-White 26 (42) 36 (58) Hispanic Origin 0.548 Hispanic 28 (43) 37 (57) Non-Hispanic 142 (47) 159 (53) Year 0.699 1988-2000* 60 (45) 73 (55) 2001-2004 52 (46) 60 (54) 2005-2008 58 (48) 63 (52) Histology 0.195 Chordoma NOS* 87 (50) 87 (50) Chondroid / Dedifferentiated

10 (50) 10 (50)

Chondrosarcoma 73 (42) 99 (58) Surgery 0.672 Gross total resection* 39 (45) 47 (55) Subtotal resection 53 (49) 55 (51) Surgery NOS 8 (47) 9 (53) None / Biopsy only 70 (45) 85 (55) * Reference category RT = radiation therapy; NOS = not otherwise specified Note: RT data missing for 10 patients

BACKGROUND

METHODS

RESULTS

• Study design: Retrospective cohort analysis using prospectively collected data from the National Cancer Institute’s Surveillance Epidemiology, and End Results (SEER) database, which captures ~28% of all U.S. cancer diagnoses.

• Subjects: 376 adult patients with histologically-diagnosed chordoma not otherwise specified (NOS) [ICD-O-3 code 9370], chondroid or dedifferentiated chordoma [9371, 9372] or chondrosarcoma [9220, 9221, 9230, 9240, 9242, 9243] located in the bones of the skull or face (C410) diagnosed from 1988-2008.

• Statistical Analysis: Chi-square tests, Wilcoxon rank sum test, logistic regression analysis, Kaplan-Meier analysis, and Cox proportional hazards model analysis.

• Skull base chordoma and chondrosarcoma are rare tumors that arise from embryonic remnants along the neuroaxis.

• Indolent course typical, but local recurrences common despite maximal surgical resection and high-dose radiotherapy (RT), especially in skull base.

• Single-institution retrospective evidence indicates that proton beam radiotherapy (PBRT) may improve long-term outcomes, but meta-analyses do not conclusively demonstrate progression-free survival or overall survival benefit for PBRT.

• We sought to assess the evolving role of RT and its effect on survival of skull base chordoma and chondrosarcoma.

• Among 57% of patients who underwent surgical resection, 42% received gross total resection (GTR).

• RT was utilized in 46% of patients overall and in 45% of patients following GTR.

• Utilization of surgical resection or RT was not significantly different by sex, race, diagnosis year, or histology on univariate or multivariate analysis.

• After adjustment, patients ≥ 65 years were less likely to receive GTR (OR 0.39, p=0.018) and more likely to receive RT (OR 1.75, p=0.048).

• After adjustment, diagnosis after 2001 (HR 0.42, p=0.002) and chondrosarcoma histology (HR 0.47, p=0.003), but not surgical extent or RT utilization, were associated with improved survival.

Table 2: Multivariate logistic regression analysis identifying predictors of gross total resection and radiation therapy utilization.

Gross Total Resection  Odds Ratio

95% CI Lower

95% CI Upper P

Age ≥ 65 0.39 0.18 0.85 0.018Sex Female 1.26 0.77 2.05 0.353Race Non-White 0.92 0.47 1.80 0.804Hispanic Origin Non-Hispanic 0.81 0.43 1.51 0.504Diagnosis Year ≥ 2001 0.65 0.39 1.07 0.093Histology Not Chordoma NOS 1.33 0.81 2.18 0.260

Radiation Therapy  Odds Ratio

95% CI Lower

95% CI Upper

P

Age ≥ 65 1.75 1.01 3.05 0.048Sex Female 1.04 0.69 1.59 0.843Race Non-White 0.79 0.45 1.39 0.410Hispanic Origin Non-Hispanic 1.24 0.71 2.16 0.457Diagnosis Year ≥ 2001 1.19 0.77 1.85 0.431Histology Not Chordoma NOS 0.75 0.49 1.14 0.181Surgery No GTR 0.97 0.59 1.59 0.90095% CI = 95% confidence interval; NOS = not otherwise specified. 

Overall Survival  Hazards

Ratio95% CI Lower

95% CI Upper P

Age 1.04 1.03 1.06 <0.001Sex Female 0.86 0.53 1.38 0.523Race Non-White 1.37 0.71 2.65 0.353Hispanic Origin Non-Hispanic 0.85 0.45 1.61 0.615Diagnosis Year ≥ 2001 0.42 0.24 0.74 0.002Histology Not Chordoma NOS 0.47 0.28 0.78 0.003Surgery No GTR 1.09 0.63 1.91 0.752RT No RT 0.90 0.56 1.54 0.66495% CI = 95% confidence interval; NOS = not otherwise specified.  

Table 3: Cox proportional hazards model identifying predictors of overall survival.

Figure 1: Kaplan-Meier survival curves by (A) year of diagnosis and (B) histology.

A

Log-rank p < 0.001

B

Log-rank p < 0.001

CONCLUSIONS

• On a national scale, overall survival from skull base chordoma and chondrosarcoma has improved significantly over time, though rates of surgical resection and RT have not increased over time.

• Older patients less likely to receive GTR and more likely to receive RT.

• Lack of association between GTR/RT utilization and improved survival may be confounded by unmeasured variables.

• Future research is needed to evaluate potential determinants of this improved survival, including PBRT utilization, imaging technology, surgical technique, and supportive care.

For reprints, please contact: Henry Park, henryspark@gmail.com