Risk Modeling for Individualization of the FLAME Focal Boost Approach in External Beam Radiation Therapy for Patients With Localized Prostate Cancer

Purpose: The FLAME trial (NCT01168479) showed that isotoxic focal boosting to the intraprostatic lesion(s) in patients with intermediate-and high-risk prostate cancer improves 5-year disease-free survival (DFS). Although the near-minimum dose to the gross tumor volume (D98%) was associated with improved outcomes, a closer look suggested that this might not be the same for all patients. Therefore, we investigated whether risk factors that are associated with a benefit of focal boosting can be identified.Methods and Materials: We described the distribution of clinical characteristics and the number of high-risk factors with respect to the D98% in 526 FLAME trial patients. We used penalized Cox regression to develop a prediction model. To investigate a potential benefit in patient subgroups, we compared the model-based predictions of 5-year DFS assuming standard whole-gland radiation therapy of 77 Gy to the predictions assuming an additional focal boost with D98% of 95 Gy.Results: Patients with high-risk factors were well represented in the group of 120 patients that received D98% > 85 Gy and showed fewer recurrences compared with the group that received 77 Gy. Applying the model simulating a standard dose of 77 Gy, we predicted a high DFS for grade group (GG) 1 patients, whereas patients with high-risk characteristics appeared to show a low DFS. All risk groups showed a high level of DFS when simulating D98% of 95 Gy.Conclusions: Our results suggest that GG 1 patients already show a low level of failure at a standard dose of 77 Gy, limiting the additional benefit of focal boosting. In contrast, patients with high-risk characteristics, especially GG 4 or 5, show a low 5 year DFS, while focal boosting might improve this substantially. This suggests that reaching a high focal boost dose may be particularly beneficial for these patients. (c) 2023 Elsevier Inc. All rights reserved.