(S042) Factors Influencing Brain Recurrence After PCI Among Patients With Limited Small-Cell Lung Cancer

Emma B. Holliday, Pamela K. Allen, Brett Carter, Xiong Wei, Stephen D. Bilton, Ritsuko Komaki; UT MD Anderson Cancer Center

PURPOSE: The brain is the most common site of distant failure for patients with small-cell lung cancer (SCLC) after complete response (CR) to treatment. Prophylactic cranial irradiation (PCI) was established as the standard of care for limited-stage disease after definitive chemoradiation (CRT). PCI confers an overall survival (OS) advantage but is not without toxicity. Therefore, we sought to identify a low-risk population for whom PCI may not be necessary. We hypothesized that patients with small tumor size may be at lower risk for brain failure, even in the absence of PCI.

METHODS: We identified 577 patients with limited-stage SCLC who were treated with definitive CRT at a single institution from 1986–2009 to a dose ≥ 45 Gy. The chi-square test was used for between-group comparisons for categorical variables, and the median test was used for between-group comparisons for continuous variables. Kaplan-Meier estimates were constructed for brain metastasis–free survival (BMFS) and OS. Analysis was performed using competing risk regression with BMFS as the primary endpoint. Factors identified as significantly associated with improved BMFS were entered into the multivariate model, and log-rank test was performed. Tumor size was recorded as the largest dimension of the primary parenchymal tumor or the largest regional node, if the primary tumor could not be visualized.

RESULTS: Of the 577 patients identified, 307 (53.2%) received PCI. Patients who did not receive PCI were older (median age 64 years [range: 27–95 yr] vs 61 years [range: 31–79 yr]; P = .021) and had a higher rate of those with Karnofsky performance status (KPS) score < 80 (n = 49 [18.1%] vs n = 24 [7.8%]; P < .001). Fifty-four (17.6%) patients who received PCI and 71 (26.3%) patients who did not receive PCI ultimately developed brain metastases. Tumor size was available for 520 patients. Among patients who received PCI, those with tumor size ≥ 5 cm trended towards increased brain failure (n = 21 [13.5%] vs n = 25 [22.3%]; P = .058). On multivariate competing risk analysis with other distant metastases DM as the competing risk, those with a tumor ≥ 5 cm were 50% more likely to develop brain failure (P = .035). For patients who received PCI, tumor size was still significantly associated with increased brain failure (subdistribution hazard ratio [SHR] = 1.79; 95% confidence interval [CI], 1.01–3.18; P = .048). For patients who did not receive PCI, tumor size was not associated with increased brain failure. There was no difference in OS based on tumor size.

CONCLUSIONS: Tumor size appears to be a significant prognostic factor for brain recurrence in patients with limited-stage SCLC after PCI. Further investigation is warranted to in order to best individualize treatment for limited-stage SCLC patients with large tumors.

Proceedings of the 97th Annual Meeting of the American Radium Society - americanradiumsociety.org