OpenKBP-Opt: an international and reproducible evaluation of 76 knowledge-based planning pipelines.

Objective.To establish an open framework for developing plan optimization models for knowledge-based planning (KBP).Approach.Our framework includes radiotherapy treatment data (i.e. reference plans) for 100 patients with head-and-neck cancer who were treated with intensity-modulated radiotherapy. That data also includes high-quality dose predictions from 19 KBP models that were developed by different research groups using out-of-sample data during the OpenKBP Grand Challenge. The dose predictions were input to four fluence-based dose mimicking models to form 76 unique KBP pipelines that generated 7600 plans (76 pipelines × 100 patients). The predictions and KBP-generated plans were compared to the reference plans via: the dose score, which is the average mean absolute voxel-by-voxel difference in dose; the deviation in dose-volume histogram (DVH) points; and the frequency of clinical planning criteria satisfaction. We also performed a theoretical investigation to justify our dose mimicking models.Main results.The range in rank order correlation of the dose score between predictions and their KBP pipelines was 0.50-0.62, which indicates that the quality of the predictions was generally positively correlated with the quality of the plans. Additionally, compared to the input predictions, the KBP-generated plans performed significantly better (P< 0.05; one-sided Wilcoxon test) on 18 of 23 DVH points. Similarly, each optimization model generated plans that satisfied a higher percentage of criteria than the reference plans, which satisfied 3.5% more criteria than the set of all dose predictions. Lastly, our theoretical investigation demonstrated that the dose mimicking models generated plans that are also optimal for an inverse planning model.Significance.This was the largest international effort to date for evaluating the combination of KBP prediction and optimization models. We found that the best performing models significantly outperformed the reference dose and dose predictions. In the interest of reproducibility, our data and code is freely available.

Combined brachytherapy and ultra-hypofractionated radiotherapy for intermediate-risk prostate cancer: Comparison of toxicity outcomes using a high-dose-rate (HDR) versus low-dose-rate (LDR) brachytherapy boost.

PURPOSE/OBJECTIVE: To compare toxicity profiles of low-dose rate (LDR) and high-dose rate (HDR) brachytherapy boost combined with ultra-hypofractionated external beam radiation therapy (UH-EBRT). MATERIALS/METHODS: 99 patients with intermediate-risk prostate cancer underwent an HDR (n = 59) or LDR (n = 40) boost combined with UH-EBRT (5 Gy x 5) . HDR (Ir-192) was delivered a single dose (15 Gy) and LDR (Pd-103) prescription dose was 100 Gy. Median baseline IPSS was 5 for both cohorts. Median follow-up was 29.3mos. Cumulative incidences were calculated for toxicity. Fisher exact tests were used to evaluate associations. RESULTS: Overall incidence of grade 2 genitourinary toxicity for the entire cohort at 12 and 24 months was 21% and 29%, respectively. The incidence of grade 2 genitourinary toxicity at 12 and 24 months was higher for LDR cohort compared with HDR cohort (45% vs 5.1% and 55% vs 11%; p<0.001). On MVA, only treatment regimen (LDR versus HDR) was associated with grade 2+ genitourinary toxicity (p<0.001). Two patients experienced grade 2 rectal toxicity in each cohort. No grade > 3 toxicities were observed. CONCLUSIONS: Both LDR and HDR brachytherapy combined with UH-EBRT had favorable toxicity profiles, but significantly less grade 2+ genitourinary toxicity was observed in patients receiving HDR.

Quantifying clinical severity of physics errors in high-dose rate prostate brachytherapy using simulations.

PURPOSE: To quantitatively evaluate through automated simulations the clinical significance of potential high-dose rate (HDR) prostate brachytherapy (HDRPB) physics errors selected from our internal failure-modes and effect analysis (FMEA). METHODS AND MATERIALS: A list of failure modes was compiled and scored independently by 8 brachytherapy physicists on a one-to-ten scale for severity (S), occurrence (O), and detectability (D), with risk priority number (RPN) = SxOxD. Variability of RPNs across observers (standard deviation/average) was calculated. Six idealized HDRPB plans were generated, and error simulations were performed: single (N = 1722) and systematic (N = 126) catheter shifts (craniocaudal; -1cm:1 cm); single catheter digitization errors (tip and connector needle-tips displaced independently in random directions; 0.1 cm:0.5 cm; N = 44,318); and swaps (two catheters swapped during digitization or connection; N = 528). The deviations due to each error in prostate D90%, urethra D20%, and rectum D1cm3 were analyzed using two thresholds: 5-20% (possible clinical impact) and >20% (potentially reportable events). RESULTS: Twenty-nine relevant failure modes were described. Overall, RPNs ranged from 6 to 108 (average ± 1 standard deviation, 46 ± 23), with responder variability ranging from 19% to 184% (average 75% ± 30%). Potentially reportable events were observed in the simulations for systematic shifts >0.4 cm for prostate and digitization errors >0.3 cm for the urethra and >0.4 cm for rectum. Possible clinical impact was observed for catheter swaps (all organs), systematic shifts >0.2 cm for prostate and >0.4 cm for rectum, and digitization errors >0.2 cm for prostate and >0.1 cm for urethra and rectum. CONCLUSIONS: A high variability in RPN scores was observed. Systematic simulations can provide insight in the severity scoring of multiple failure modes, supplementing typical FMEA approaches.