334 - Validation of a Poissonian LQ-Based Population TCP Model for Breast Cancer Radiotherapy

Purpose/Objective(s): Validate a Poissonian linear-quadratic (LQ) population tumor control probability (TCP) model that incorporates tumor volume heterogeneity, inter-patient cell radiosensitivity variability, and time-to-failure, using published data on invasive breast carcinomas from major radiotherapy (RT) trials.

Materials/Methods: Local control (LC) data were obtained from the START A/B, Fast Forward, and Early Breast Cancer Trialists Collaborative Group studies. We integrated the additive impact of RT after surgery with RT-alone data to encompass varied clinical scenarios. Model parameters (a, ß, a/ß, sa (inter-patient heterogeneity), tumor volume scaling, and exponential time-to-failure), were fitted to observed LC rates. An overall treatment time effect for loco–regional relapse and the average residual number of clonogens after surgery were also estimated. In-breast relapse rates for neoadjuvant tumor-directed RT dose-fractionation schemes were predicted for various tumor sizes.

Results: The model reproduced 5-year LC outcomes in both adjuvant and RT-alone settings (Table 1). The average residual clonogen count after surgery was 300. The estimated overall treatment time effect for local–regional relapse was 0.58 Gy/day, which is in close agreement with the 0.60 Gy/day published from the START trials. Predicted 5-year in-breast relapse rates following tumor-directed neoajuvant RT plus surgery were 0.1%, 0.6%, 1.9%, and 10.0% for 23Gy/1 fx, and 1.0%, 6.6%, 13.3%, and 15.8% for 21Gy/1 fx, for primary tumors diameters of 5 mm (T1a), 10 mm (T1b), 15 mm (T1c), and 30 mm (T2) tumors, respectively.

Conclusion: By incorporating tumor size heterogeneity, radiosensitivity variability, and a treatment time effect, this Poissonian LQ-based population TCP model provides a robust framework for predicting local control in breast cancer radiotherapy. Following further validation with expanded data sets, it may serve as a valuable tool to tailor dose regimens to individual patient and tumor characteristics, potentially improving local control rates and optimizing treatment strategies.