ESTRO 2021 Abstract Book

S983

ESTRO 2021

Conclusion We found a significant benefit in RR and PR by adding RT to Ch as NT, without significantly increasing toxicity. Longer follow-up is necessary to assess the impact on clinical outcomes.

PO-1185 Development of a nomogram for predicting brain metastasis of small cell lung cancer Q. hou 1,2 , B. Sun 1 , N. Yao 3 , L. Wei 1 , S. Xu 4 , J. Cao 5,2 1 Shanxi Medical University, Department of Medical Imaging, Taiyuan, China; 2 Shanxi Provincial Cancer Hospital, Department of Radiobiology, Taiyuan, China; 3 Shanxi Provincial Cancer Hospital, Department of Raiobiology, Taiyuan, China; 4 Shanxi Province Children's Hospital, Department of Radiation, Taiyuan, China; 5 Shanxi Provincial Cancer Hospital, Department of Radiotherapy, Taiyuan, China Purpose or Objective Brain metastasis (BM) is usually a fatal event in patients with small-cell lung cancer (SCLC). However, there is no effective tool to predict BM in SCLC patients. In this study, we aimed to develop and validate a widely accepted nomogram based on clinical and laboratory data to predict brain BM in patients with SCLC. Materials and Methods From January 2012 to December 2019, 332 (221/111 in training/validation cohort) consecutive SCLC patients without BM at initial diagnosis in our hospital were analyzed retrospectively. Univariate Cox regression analyses were conducted to screen for factors significantly related to BM. A least absolute shrinkage and selection operator (LASSO) regularized Cox regression model was used to select the most relevant factors. Then a nomogram for predicting BM was established on the final multivariate Cox proportional hazards model. To contrast the utility of the nomogram against the widely used TNM, we used the concordance index (C- index) and a calibration curve to determine its predictive and discriminatory capacity, and used decision curve analysis (DCA) and integrated Brier score (IBS) to determine its clinical utility. Results Prophylactic cranial irradiation (PCI; HR: 0.22, 95% CI 0.11-0.46), chemotherapy cycles[1] (HR: 0.65, 95% CI: 0.45-0.96), M[2] stage (HR: 1.40, 95% CI: 1.20-1.64), lactate dehydrogenase (LDH; HR: 1.52, 95% CI: 1.02- 2.27) and protein progastrin releasing peptide (ProGRP; HR: 1.51, 95% CI: 1.04-2.17) were the strong independent factors of BM. Based on the above factors, a nomogram model predicting BM was established. The nomogram provided good discrimination, with C-index for prediction of BM in SCLC patients was 0.825 and 0.806 in the training and validation cohorts, respectively. The calibration curve illustrated the good agreements between nomogram prediction and actual observations. DCA and IBS demonstrated that the nomogram was clinically useful. Conclusion We have developed a robust tool that can predict brain metastasis in SCLC patients. Selection of an enriched patient population at high risk for brain metastasis will facilitate the design of trials aiming at its prevention.