Dosimetric predictors of acute and late gastrointestinal toxicities in stereotactic online adaptive magnetic resonance-guided radiotherapy (SMART) for locally advanced pancreatic adenocarcinoma.

BACKGROUND AND PURPOSE: A retrospective evaluation of dosimetric predictors and leveraged dose-volume data for gastrointestinal (GI) toxicities for locally-advanced pancreatic cancer (LAPC) treated with daily stereotactic MRI-guided online-adaptive radiotherapy (SMART). MATERIALS AND METHODS: 147 patients with LAPC were treated with SMART at our institution between 2018 and 2021. Patients were evaluated using CTCAE V5.0 for RT-related acute (≤3 months) and late (>3 months) toxicities. Each organ at risk (OAR) was matched to a ≥ grade 2 (Gr2+) toxicity endpoint composite group. A least absolute shrinkage selector operator regression model was constructed by dose-volumes per OAR to account for OAR multicollinearity. A receiver operator curve (ROC) analysis was performed for the combined averages of significant toxicity groups to identify critical volumes per dose levels. RESULTS: 18 of 147 patients experienced Gr2+ GI toxicity. 17 Gr2+ duodenal toxicities were seen; the most significant predictor was a V33Gy odds ratio (OR) of 1.69 per cc (95 % CI 1.14-2.88). 17 Gr2+ small bowel (SB) toxicities were seen; the most significant predictor was a V33Gy OR of 1.60 per cc (95 % CI 1.01-2.53). The AUC was 0.72 for duodenum and SB. The optimal duodenal cut-point was 1.00 cc (true positive (TP): 17.8 %; true negative (TN); 94.9 %). The SB cut-point was 1.75 cc (TP: 16.7 %; TN: 94.3 %). No stomach or large bowel dose toxicity predictors were identified. CONCLUSIONS: For LAPC treated with SMART, the dose-volume threshold of V33Gy for duodenum and SB was associated with Gr2+ toxicities. These metrics can be utilized to guide future dose-volume constraints for patients undergoing upper abdominal SBRT.

Inter-fractional portability of deep learning models for lung target tracking on cine imaging acquired in MRI-guided radiotherapy.

MRI-guided radiotherapy systems enable beam gating by tracking the target on planar, two-dimensional cine images acquired during treatment. This study aims to evaluate how deep-learning (DL) models for target tracking that are trained on data from one fraction can be translated to subsequent fractions. Cine images were acquired for six patients treated on an MRI-guided radiotherapy platform (MRIdian, Viewray Inc.) with an onboard 0.35 T MRI scanner. Three DL models (U-net, attention U-net and nested U-net) for target tracking were trained using two training strategies: (1) uniform training using data obtained only from the first fraction with testing performed on data from subsequent fractions and (2) adaptive training in which training was updated each fraction by adding 20 samples from the current fraction with testing performed on the remaining images from that fraction. Tracking performance was compared between algorithms, models and training strategies by evaluating the Dice similarity coefficient (DSC) and 95% Hausdorff Distance (HD95) between automatically generated and manually specified contours. The mean DSC for all six patients in comparing manual contours and contours generated by the onboard algorithm (OBT) were 0.68 ± 0.16. Compared to OBT, the DSC values improved 17.0 - 19.3% for the three DL models with uniform training, and 24.7 - 25.7% for the models based on adaptive training. The HD95 values improved 50.6 - 54.5% for the models based on adaptive training. DL-based techniques achieved better tracking performance than the onboard, registration-based tracking approach. DL-based tracking performance improved when implementing an adaptive strategy that augments training data fraction-by-fraction.