303 - Impact of Castrate Sensitivity and Genetic Mutations on Local Control after Spine Stereotactic Body Radiation Therapy for Prostate Cancer Spinal Metastases
Presenter(s)
C. B. Jackson1, L. A. Boe2, J. Haseltine3, B. A. Mueller3, A. Schmitt4, M. Vaynrub1, W. C. Newman1, H. Nagar1, D. J. Gorovets3, J. Janopaul-Naylor1, S. M. McBride3, V. S. Brennan3, E. Lis1, O. Barzilai1, M. Bilsky1, Y. Yamada3, and D. S. Higginson1; 1Memorial Sloan Kettering Cancer Center, New York, NY, 2Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY, 3Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, 4Memorial Sloan Kettering, New York, NY
Purpose/Objective(s): Spinal metastases are common in patients with prostate cancer (PC). Relatively little data exist on the risk of local failure (LF) in spinal metastases from castrate resistant (CR) vs. castrate sensitive (CS) PC after spine stereotactic body radiation therapy (SBRT).
Materials/Methods: This is a retrospective cohort study of 229 PC patients treated to 300 spinal lesions. All patients received at least 2 months of follow-up MRI; patients with prior overlapping radiotherapy to the spine were excluded. The primary outcome was LF, defined radiographically. Secondary outcomes included vertebral compression fracture (VCF), defined as progressive or de novo fracture in the absence of LF, as well as overall survival (OS). Patients were classified as CR with rising prostate specific antigen or new/growing lesions with castrate levels of testosterone. The Benjamini-Hochberg false discovery rate correction was utilized to account for multiple hypothesis testing for genetic mutations associated with LF.
Results: The median OS was 46 months (95% confidence interval [CI] 32-60), and median follow-up after SBRT was 21 months (interquartile range [IQR] 11-39). One-hundred sixty-two out of 229 (71%) patients had somatic genetic sequencing. A total of 152 lesions received 27 Gy in 3 fractions (27/3; 51%), 97 lesions received 30 Gy in 3 fractions (30/3; 32%), and 51 lesions received 24 Gy in 1 fraction (24/1; 17%). There were 128 (43%) CS lesions, and the remainder were CR. The 2- and 5-year rates of LF, considering death as a competing risk, were 10 and 13%, respectively. There were 31 instances of LF (10.3%). Salvage treatment included SBRT in 10 instances (32%), surgery alone in 3 instances (9.6%), and surgery with SBRT in 1 instance (3.2%). For 27/3, 30/3, and 24/1, the 2-/5-year risks of LF were 16/20%, 3.7/3.7%, and 0/0%, respectively (p<0.001). For CR and CS lesions, the 2-year risk of LF after spine SBRT was 15 vs. 3.2%, respectively (p<0.001). On multivariable, competing risks regression adjusted for clustering at the patient level, the factors associated with LF included CR status (hazard ratio [HR] 4.41, 95% CI 1.39-14, p = 0.01), somatic PTEN mutant status (n=36; HR 3.82, 95% CI 1.75-8.35, p<0.001), and somatic KMT2C mutant status (n=17; HR 4.01, 95% CI 1.57-10.3, p=0.004). The 2-year risks of VCF and VCF requiring kyphoplasty or surgical fusion were 6.5 and 5.3%, respectively. Concurrent androgen deprivation therapy (ADT) was not associated with increased risk of VCF after spine SBRT (p=0.3).
Conclusion: Spine SBRT for PC is associated with excellent local control, 87% at 5 years. CR lesions are more likely to fail than CS lesions. For hypofractionated SBRT, 30 Gy in 3 fractions provides superior local control to 27 Gy in 3 fractions. Concurrent ADT does not increase the risk of VCF after spine SBRT. Dose escalation may be considered for CR, PTEN mutant, and KMT2C mutant lesions.