Rwigema, Jean Claude

AbstractBackgroundDeep learning auto‐segmentation (DLAS) models have been adopted in the clinic; however, they suffer from performance deterioration owing to the clinical practice variability. Some commercial DLAS software provide an incremental retraining function that enables users to train a custom model using their institutional data to account for clinical practice variability.PurposeThis study was performed to evaluate and implement the commercial DLAS software with the incremental retraining function for definitive treatment of patients with prostate cancer in a multi‐user environment.MethodsCT‐based target organs and organs‐at‐risk (OAR) delineation of 215 prostate cancer patients were utilized. The performance of three commercial DLAS software built‐in models was validated with 20 patients. A retrained custom model was developed using 100 patients and evaluated on the remaining data (n = 115). Dice similarity coefficient (DSC), Hausdorff distance (HD), mean surface distance (MSD), and surface DSC (SDSC) were utilized for quantitative evaluation. A multi‐rater qualitative evaluation was blindly performed with a five‐level scale. Visual inspection was performed in consensus and non‐consensus unacceptable cases to identify the failure modes.ResultsThree commercial DLAS vendor built‐in models achieved sub‐optimal performance in 20 patients. The retrained custom model had a mean DSC of 0.82 for prostate, 0.48 for seminal vesicles (SV), and 0.92 for rectum, respectively. This represents a significant improvement over the built‐in model with DSC of 0.73, 0.37, and 0.81 for the corresponding structures. Compared to the acceptance rate of 96.5% and consensus unacceptable rate (i.e., both reviewers rated as unacceptable) of 3.5% achieved by manual contours, the custom model achieved a 91.3% acceptance rate and 8.7% consensus unacceptable rate. The failure modes of retrained custom model were attributed to the following: cystogram (n = 2), hip prosthesis (n = 2), low dose rate brachytherapy seeds (n = 2), air in endorectal balloon(n = 1), non‐iodinated spacer (n = 2), and giant bladder(n = 1).ConclusionThe commercial DLAS software with the incremental retraining function was validated and clinically adopted for prostate patients in a multi‐user environment. AI‐based auto‐delineation of the prostate and OARs is shown to achieve improved physician acceptance, overall clinical utility, and accuracy.

AbstractBackgroundThe objective of this study was to report acute changes in patient‐reported quality of life (PRQOL) using the 26‐item Expanded Prostate Index Composite (EPIC‐26) questionnaire in a prospective study using hypofractionated intensity‐modulated proton beam therapy (H‐IMPT) targeting the prostate and the pelvic lymph nodes for high‐risk or unfavorable intermediate‐risk prostate cancer.MethodsFifty‐five patients were enrolled. H‐IMPT consisted of 45 GyE to the pelvic lymph nodes and 67.5 GyE to the prostate and seminal vesicles in 25 fractions. PRQOL was assessed with the urinary incontinence (UI), urinary irritative/obstructive symptoms (UO), and bowel function (BF) domains of EPIC‐26 questionnaire. Mean changes in domain scores were analyzed from pretreatment to the end of treatment and 3 months posttreatment. A clinically meaningful change (or minimum important change) was defined as a score change > 50% of the baseline standard deviation.ResultsThe mean scores of UO, UI, and BF at baseline were 84.6, 91.1, and 95.3, respectively. At the end of treatment, there were statistically significant and clinically meaningful declines in UO and BF scores (−13.5 and −2.3, respectively), while the decline in UI score was statistically significant but not clinically meaningful (−13.7). A clinically meaningful decline in UO, UI, and BF scores occurred in 53.5%, 22.7%, and 73.2% of the patients, respectively. At 3 months posttreatment, all three mean scores showed an improvement, with fewer patients having a clinically meaningful decline in UO, UI, and BF scores (18.4%, 20.5%, and 45.0%, respectively). There was no significant reduction in the mean UO and UI scores compared to baseline, although the mean BF score remained lower than baseline and the difference was clinically meaningful.ConclusionsUO, UI, and BF scores of PRQOL declined at the end of H‐IMPT. UO and UI scores showed improvement at 3 months posttreatment and were similar to the baseline scores. However, BF score remained lower at 3 months posttreatment with a clinically meaningful decline.

248 Background: Hypofractionated prostate cancer radiation has showed similar results in several prior phase III studies (PCG GU 003, PACE-B, and Hypo PC RT). However, prospective phase II-III clinical trial data testing 5 tx after prostatectomy is scarce. Methods: Between 2018 and 2019, 41 patients were treated after postprostatectomy for high risk features. 5 patients were treated adjuvantly, 36 for salvage including 8 with oligometastatic disease. Indications for adjuvant RT included a PSA < 0.2 and +margins, SVI, or EPE. Salvage RT was offered for PSA ≥0.2. Oligometastatic RT for patients with ≤5 RT targets. Staging included C11 PET for all cases. Total dose to the prostate bed was 30-32 Gy in 5tx QOD with IMRT, conebeam IGRT, and MRI registration. All salvage patients received ADT for 6 months and oligometastatic patients for 18 months. Dose to the metastatic sites was 30 in 5tx QOD. Of the 41 patients 8 also received SBRT to the sites of oligometastatic disease. We looked at clinical outcomes defining biochemical failure as a PSA > 0.2 after treatment, baseline adjusted CTC AE V5.0, baseline adjusted patient reported toxicities (PRO CTC AE), QOL (EPIC, PROMIS), and AUA was used for all cases. Results: Median follow up was 23 months (range 10-37). Pre-RT T stage was T2-T3b, with 47% being T3a-b; Pre RT Median PSA of 0.4 (range < 0.1-1.9); Median GS 8 (6-9); and (+) margins in 48.8%. Sites of oligometastatic disease radiated included the LN and bone. Treatment related AE were grade 0-1 in all cases, except for one patient with G2 GU incontinence. Overall QOL remained high during follow including Promis 10 overall, mental, and physical scores; urinary bother, irritative/obstructive, and AUA scores; bowel overall, bowel bother, and bowel function scores; and overall sexual, sexual function, and sexual bother scores remained at baseline levels during follow up. Only hormonal overall, hormonal function, and hormonal bother had lower scores at 3 months that recovered by 12 months in patients treated with ADT for 6 months and by 24 months in patients treated with ADT for 18 months. A total of 3 clinical failure have been seen; 2 patients with regional failures alone; and one with axial skeleton bony failures for 93% clinical control at median follow up of 23 months. All 3 patients with clinical failure were salvaged successfully with SBRT and all patients remain disease free at last follow up. A total 5 patients with raising PSAs alone have been seen. All patients have been re-staged with C11 PET. No failures in the prostate bed or previously radiated sites have been seen. Conclusions: Toxicity for RT over 5tx is lower than expected with only one case of grade 2 urinary incontinence. QOL scores remained high during follow up, minor changes in hormonal scores were seen during ADT, but recovered after. 30-32 Gy over 5 tx provided 100% control in radiated targets and metastatic sites. Clinical trial information: NCT03570827.