Predictive factors of late genitourinary toxicity after CyberKnife re-irradiation for locally recurrent prostate cancer.

Aims: Re-irradiation with stereotactic body RT (re-SBRT) is a valid treatment option for local relapse in prostate cancer after postoperative or definitive radiotherapy (RT). However, we need predictive factors to prevent the onset of late adverse events. Here we present a retrospective analysis conducted on a cohort of patients treated with re-SBRT through Cyberknife ® robotic system. The correlation between dosimetric data and incidence of late genitourinary (GU) toxicity was evaluated to establish a model to predict late GU Grade >2 adverse events in this population. Methods: We collected data of 50 consecutively patients treated from June 2012 to February 2016. All patients were affected by biochemical relapse defined by European Urology Association Criteria after definitive or postoperative radiotherapy, and macroscopic evidence of intra-prostatic or prostate bed recurrence was detected by 18F-choline PET/CT and MRI. Patients with metastatic or regional nodal disease were excluded. All patients underwent reSBRT using the CyberKnife ®robotic system, for a total dose of 30 Gy in 5 fractions every other day. Toxicity was assessed by the Common Terminology Criteria for Adverse Events (CTCAE) toxicity scale v.5. Relationship between late GU G>2 and Gross Target Volume (ccGTV), Dose to 50% of urinary bladder volume (DB50), Maximum dose within Planning Target Volume (Dmax), urethra Dmax (UDmax) and Total Equivalent Dose (tEQD2) administered to prostate or prostate bed was explored with logistic regression. A receiver operating characteristic (ROC) curve was used to find the optimal cut-off point for continuous variables significantly predictive for late GU adverse events. Results: After a median follow up of 48.2 months (6.4-86.3), late G> 2 GU toxicity occurred in 13 (26%) patients. At Univariate Analysis, no significant impact of ccGTV (p=0.38), DB50 (p=0.25), Dmax (p=0.88), and tEQD2 total was detected (p=0.76 ). Only UDmax showed significant association with Late G> 2 GU toxicity (p=0.02). ROC analysis showed that UDmax >34.12 Gy best predicted GU toxicity, with a positive and negative likelihood ratio of 3.56 (95% CI 1.1-11.3) and 0.69 (95% CI 0.4-1.1), respectively (AUC 0.66, p=0.06). Conclusion: These data suggest that higher urethral dose may be associated to higher risk of late G> 2 GU toxicity. Thus, due to the low number of events, larger series with long term follow up would be needed to better identify a model to predict late GU adverse events after re-SBRT. [ABSTRACT FROM AUTHOR]