Toxicity Analysis of a Modified Prostate Bed CTV Based on PSMA-PET Patterns of Failure.

PURPOSE/OBJECTIVE(S): After radical prostatectomy (RP), about 20-50% of prostate cancer patients experience biochemical failure, and many can be cured with salvage radiotherapy (sRT). Adding androgen deprivation and pelvic lymph node (PLN) RT to sRT decreases risk of disease progression at the cost of increased toxicity. PSMA PET patterns of local failure after RP show failures outside the RTOG prostate bed CTV (RTOG_CTV) occurring postero-laterally and inferiorly, while in much of the covered bladder local failures are not seen. A modified CTV (m_CTV) encompassing all sites of observed local failures and less of the bladder would potentially reduce dose to the bladder but increase dose to rectum. Herein, we compared sRT plans using RTOG_CTV vs. m_CTV and assessed predicted genitourinary (GU) and gastrointestinal adverse events (AEs).

MATERIALS/METHODS: Volumes for RTOG_CTV, m_CTV, PLN CTV, and OARs were retrospectively contoured on planning CT scans of 28 post-RP patients. Plans were generated for each RTOG_CTV and m_CTV. Doses to prostate bed PTV (CTV + 8mm expansion, 6mm posteriorly) and PLN CTV were 68 Gy and 52.7 Gy, respectively, all in in 34 fractions. Dose-volume histogram (DVH) data for bladder, rectum, and bowel of each plan were extracted. Normal Tissue Complication Probability (NTCP) curves were generated using Lyman-Kutcher-Burman model to predict risk of late G≥3 GU AEs, early GU urgency, leakage, frequency, nocturia, late GI bleed, and late G≥3 GI AEs. Data were compared using 2 sample t-test and Wilcox signed rank test. Correlation of bladder volume and differences in DVH and NTCP data between plans was assessed using Pearson coefficient.

RESULTS: The m_CTV and RTOG_CTV had a median volume of 70.56cc (IQR 59.72-82.17cc) and 132.14cc (IQR 109.33-155.84cc), respectively. Use of m_CTVs resulted in mean absolute decrease of 20.22% and 17.94% in bladder V65 Gy and V40 Gy, respectively (p<0.01 for both), without differences in rectum V65 Gy (p = 0.44), rectum V40 Gy (p = 0.11), or bowel V45 Gy (p = 0.29). Per NTCP curves, m_CTV plans had mean absolute decrease of 24.89% in predicted late G≥3 GU AEs (p<0.01). For predicted early GU AEs, there were mean absolute decreases ranging from 4.85 to 10.21% (p<0.01 for all). Predicted late GI bleed (p = 0.98) and late G≥3 GI AEs (p = 0.99) were not significantly different between plans. When assessed as a continuous variable, smaller bladder volumes had strong correlation with more benefit from m_CTV in bladder V65 Gy (r = -0.71, p<0.01) and moderate correlation with less early GU AEs (r -0.59 to -0.56, p<0.01 for all). There was no correlation with bladder V40 Gy (p = 0.27) or late G≥3 GU AEs (p = 0.14).

CONCLUSION: The m_CTV plans had significantly less dose to the bladder, no increased dose to the rectum and significantly lower predicted rates of early and late GU toxicity. Using a m_CTV encompassing observed local failures from PSMA PET could improve the therapeutic ratio for patients undergoing sRT. Further clinical validation is warranted.