2705 - Clinical Predictors of 1-Year Mortality After Stereotactic Body Radiotherapy (SBRT) for Metastatic Cancer
Presenter(s)
C. D. Brown1, T. Joshi2, R. Darawsheh3, C. Shen4, and E. M. Steele4; 1University of North Carolina Hospitals, Chapel Hill, NC, 2University of North Carolina, Chapel Hill, NC, 3University of California, San Francisco, School of Medicine, San Francisco, CA, 4Department of Radiation Oncology, University of North Carolina, Chapel Hill, NC
Purpose/Objective(s): Stereotactic body radiotherapy (SBRT) can prolong disease control and preserve quality of life in patients with metastatic cancer. However, some patients obtain limited benefit due to early mortality. Identifying predictors of death within 1 year of SBRT can help guide patient selection and optimize treatment timing toward the delivery of high-value care.
Materials/Methods: Patients receiving SBRT to extracranial sites of metastatic cancer between 2012-2022 at an academic center were identified and these medical records were reviewed to build a database of demographic, disease, and treatment details. The cohort was divided into two groups: those who did versus did not die within 1 year of SBRT. Descriptive statistics were performed then chi-square, Fisher’s exact test, and t-tests were used to compare the two groups. Significant variables (p<0.2) on univariable analysis were included in a multivariable logistic regression model predicting mortality within 1 year of SBRT. Model fit was assessed with Nagelkerke pseudo-R2 and performance of the model was assessed using area under the ROC curve (AUC). Kaplan-Meier survival analysis was used to report the median overall survival (OS).
Results: In total, 546 patients were identified and 182 (33.3%) died within 1 year. Multiple factors were significantly associated with mortality (see Table 1). On multivariable analysis, renal cell carcinoma and prostate cancer had lower odds of 1-year mortality relative to NSCLC (p=0.019 and <0.001 respectively) and worse performance status was associated with increased odds of 1-year mortality (see Table 1). An AUC of 0.737 was achieved, indicating high predictive discrimination of the model. Median OS for this cohort was 24.2 months.
Conclusion: Clinical factors such as disease site, performance status, and number of progressing lesions are meaningfully associated with 1-year mortality following SBRT. With increasing utilization of SBRT for metastatic disease, optimal patient/treatment selection is essential to improve both care delivery and resource utilization.
Table 1: Univariable analysis (UVA) and multivariable regression model (MRM) predicting 1-year mortality after SBRT to metastases
|
| UVA | MRM | ||||
| Variable | Group | Death < 1y n=182 (%) | Death > 1y n=364 (%) | (p) | Odds Ratio | (p) |
| Age |
| 59.7 | 58.2 | 0.717 |
|
|
| Months from metastasis diagnosis to SBRT (median) |
| 8.1 | 14.8 | <0.001 | 0.985 | 0.014 |
| Disease site | NSCLC | 53 (29) | 59 (16) | <0.001 | -- |
|
| Breast | 36 (20) | 61 (17) | 0.668 | 0.186 | ||
| Prostate | 4 (2) | 46 (13) | 0.119 | <0.001 | ||
| Colorectal | 13 (7) | 46 (13) | 0.554 | 0.137 | ||
| Melanoma | 21 (12) | 29 (8) | 1.003 | 0.995 | ||
| RCC | 16 (9) | 51 (14) | 0.423 | 0.019 | ||
| Other | 39 (21) | 72 (20) | 0.758 | 0.344 | ||
| # of sites treated | 1 | 146 (80) | 322 (89) | 0.009 | -- |
|
| 2-3 | 36 (20) | 42 (12) | 1.351 | 0.315 | ||
| # of progressing lesions | 1 | 73 (40) | 207 (57) | <0.001 | -- |
|
| 2-5 | 86 (47) | 145 (40) | 0.871 | 0.731 | ||
| >5 | 23 (12) | 12 (3) | 2.134 | 0.180 | ||
| Symptomatic | No | 134 (74) | 298 (82) | 0.033 | -- |
|
| Yes | 48 (26) | 66 (18) | 1.312 | 0.254 | ||
| Performance Status (KPS) | High (80-100) | 110 (60) | 297 (82) | <0.001 | -- |
|
| Moderate (60-70) | 57 (31) | 59 (16) | 2.438 | <0.001 | ||
| Poor (=50) | 15 (8) | 8 (2) | 4.761 | 0.002 | ||