Salvage radiotherapy is an established treatment option for patients who develop biochemical or locoregional recurrence of prostate cancer after radical prostatectomy. In the postoperative setting, however, daily changes in bladder and rectal filling can alter the position of the prostate bed and nearby organs at risk, potentially affecting both target coverage and normal-tissue exposure.
Daily adaptive radiotherapy offers a way to account for these anatomical changes by recalculating and reoptimizing the treatment plan before each fraction. The prospective DART-PHASER study evaluated this approach in patients receiving moderately hypofractionated salvage radiotherapy after prostatectomy.
The study, titled “Daily adaptive radiotherapy in postoperative hypofractionated salvage radiothERapy for prostate cancer patients (DART-PHASER): Early clinical and dosimetric results,” was published in Clinical and Translational Radiation Oncology.
Authors: Luca Nicosia, Riccardo Filippo Borgese, Nicola Bianchi, Carolina Orsatti, Andrea Gaetano Allegra, Jacopo Balduzzi, Margherita Corsi, Chiara De-Colle, Antonio De Simone, Matilde Fiorini, Niccolò Giaj-Levra, Davide Gurrera, Cristina Mazzi, Stefania Naccarato, Edoardo Pastorello, Francesco Ricchetti, Michele Rigo, Andrea Romei, Gianluisa Sicignano, Gloria Guardia, Patricia Salgado Gonzàles, Ruggero Ruggieri, and Filippo Alongi.
Daily Adaptation During Salvage Radiotherapy
DART-PHASER is a prospective observational study (4181CESC; NCT05884632) evaluating CT-guided daily adaptive salvage radiotherapy in patients with prostate cancer following radical prostatectomy.
The current interim analysis included 42 patients and 840 treatment fractions. Median age was 64 years, and median follow-up was 14 months. Sixteen patients received 55 Gy to the prostate bed in 20 fractions, while 26 received 59 Gy to the prostate bed in 20 fractions. According to the treatment protocol, 55 Gy was prescribed for biochemical-only relapse and 59 Gy for macroscopic relapse. Pelvic irradiation was delivered to 17 patients, and 18 received concomitant androgen deprivation therapy.
All patients had no evidence of distant metastases on PSMA PET/CT. Treatment was delivered using the Ethos platform. Before each fraction, cone-beam CT was obtained and two plans were generated. The scheduled plan consisted of dose recalculation on the daily synthetic CT generated from CBCT, while the adapted plan was reoptimized according to the daily anatomy. The treating physician then selected the plan considered preferable based on dose distribution and dose-volume histogram parameters. The adapted plan was selected for treatment in 837 of 840 fractions.
Improved Target Coverage With Adaptation
Daily adaptation produced a clear improvement in target coverage. Median prostate bed planning target volume (PTV) V95% increased from 92.1% with the scheduled plan to 98.5% with the adapted plan (p<0.0001).
Among patients receiving pelvic treatment, median pelvic nodal PTV V95% increased from 96.55% to 98.8% (p<0.0001). Pelvic nodal PTV V107%, reflecting higher-dose regions within the target, was also reduced with adaptation.
Additional analyses accounting for the repeated treatment fractions within individual patients continued to show significant improvements in prostate bed and pelvic target coverage, supporting the consistency of the main findings.
Lower Rectal Exposure
One of the clearest organ-at-risk benefits was observed for the rectum. Median rectum V52.8 Gy decreased from 16.9% with the scheduled plan to 12.8% with the adapted plan (p<0.0001), corresponding to a 4.1-percentage-point reduction in high-dose rectal exposure.
Median rectal mean dose per fraction also decreased from 1.54 Gy to 1.42 Gy (p<0.0001). Importantly, the median value for the scheduled plan exceeded the prespecified 1.5 Gy-per-fraction threshold, whereas the median adapted-plan value remained below it.
Bland-Altman analyses further showed substantially closer agreement between the adapted plan and the approved offline reference plan. For prostate bed PTV V95%, the scheduled plan showed a bias of −7.50% relative to the reference plan, compared with only −0.06% for the adapted plan. For rectum V52.8 Gy, the corresponding biases were +5.43% and −0.39%.
These findings suggest that online reoptimization was particularly useful for maintaining intended prostate bed coverage and controlling rectal exposure as anatomy changed between treatment fractions.
Bladder and Bowel Findings Were More Modest
The effect of adaptation was less pronounced for some other organs at risk. Median bladder mean dose was 1.33 Gy per fraction with the scheduled plan and 1.23 Gy with the adapted plan, with no statistically significant difference (p=0.353).
Bladder V40.8 Gy was slightly higher with the adapted plan, at 31.8% versus 31.4% with the scheduled plan. Although the fraction-level difference was statistically significant, the absolute difference was small, both values remained well below the 50% constraint, and significance was lost when the analysis was repeated at the patient level.
For bowel dose, the median maximum dose was 2.35 Gy with the scheduled plan and 2.34 Gy with adaptation. The difference reached statistical significance but was considered clinically negligible. The investigators noted that daily variation in bladder filling and bowel position remained an important contributor to dosimetric variability even with adaptive planning.
Early Toxicity
Acute toxicity remained limited during the preliminary analysis. At the end of radiotherapy, overall grade 2 gastrointestinal toxicity occurred in 4.8% of patients. Grade 2 proctitis and grade 2 diarrhea were each reported in 2.4%. For genitourinary toxicity, grade 2 cystitis was reported in 2.4% at the end of treatment and in 9.5% at the 3-month assessment. No grade 3 or higher gastrointestinal or genitourinary toxicity was reported.
Pelvic irradiation was the only factor significantly associated with diarrhea. Grade 1 diarrhea occurred in 41.2% of patients receiving pelvic treatment compared with 4% of those without pelvic irradiation, while grade 2 diarrhea occurred in 5.9% versus 0%, respectively (p=0.002).
No significant association was observed between the 55-Gy versus 59-Gy treatment dose and toxicity at 2 weeks, the end of treatment, or 3 months.
What Do the Findings Mean?
The early DART-PHASER results show that AI-assisted daily adaptive planning improved target coverage during moderately hypofractionated postoperative salvage radiotherapy, particularly for the prostate bed and pelvic target volumes.
A clinically meaningful dosimetric improvement was observed for the rectum, with reductions in both high-dose rectal exposure and mean rectal dose. Changes in bladder and bowel dosimetry were more modest; the fraction-level difference in bladder V40.8 Gy did not remain significant in the patient-level analysis, while the bowel Dmax difference remained statistically significant but clinically negligible.
These findings should be viewed as preliminary. The analysis included 42 patients, and the current follow-up is insufficient to draw conclusions regarding late gastrointestinal or genitourinary toxicity or biochemical relapse-free survival. The study is continuing recruitment.
Overall, DART-PHASER provides early prospective evidence that daily online adaptation can improve target coverage and reduce rectal exposure during postoperative hypofractionated salvage radiotherapy. Longer follow-up is needed to determine whether these dosimetric improvements translate into reductions in late toxicity or other clinical benefits.
The full article is available in Clinical and Translational Radiation Oncology.