Validation of clinical acceptability of deep-learning-based automated segmentation of organs-at-risk for head-and-neck radiotherapy treatment planning

Introduction

Organ-at-risk segmentation for head and neck cancer radiation therapy is a complex and time-consuming process (requiring up to 42 individual structure, and may delay start of treatment or even limit access to function-preserving care. Feasibility of using a deep learning (DL) based autosegmentation model to reduce contouring time without compromising contour accuracy is assessed through a blinded randomized trial of radiation oncologists (ROs) using retrospective, de-identified patient data.

Methods

Two head and neck expert ROs used dedicated time to create gold standard (GS) contours on computed tomography (CT) images. 445 CTs were used to train a custom 3D U-Net DL model covering 42 organs-at-risk, with an additional 20 CTs were held out for the randomized trial. For each held-out patient dataset, one of the eight participant ROs was randomly allocated to review and revise the contours produced by the DL model, while another reviewed contours produced by a medical dosimetry assistant (MDA), both blinded to their origin. Time required for MDAs and ROs to contour was recorded, and the unrevised DL contours, as well as the RO-revised contours by the MDAs and DL model were compared to the GS for that patient.

Results

Mean time for initial MDA contouring was 2.3 hours (range 1.6-3.8 hours) and RO-revision took 1.1 hours (range, 0.4-4.4 hours), compared to 0.7 hours (range 0.1-2.0 hours) for the RO-revisions to DL contours. Total time reduced by 76% (95%-Confidence Interval: 65%-88%) and RO-revision time reduced by 35% (95%-CI,-39%-91%). All geometric and dosimetric metrics computed, agreement with GS was equivalent or significantly greater (p<0.05) for RO-revised DL contours compared to the RO-revised MDA contours, including volumetric Dice similarity coefficient (VDSC), surface DSC, added path length, and the 95%-Hausdorff distance. 32 OARs (76%) had mean VDSC greater than 0.8 for the RO-revised DL contours, compared to 20 (48%) for RO-revised MDA contours, and 34 (81%) for the unrevised DL OARs.

Conclusion

DL autosegmentation demonstrated significant time-savings for organ-at-risk contouring while improving agreement with the institutional GS, indicating comparable accuracy of DL model. Integration into the clinical practice with a prospective evaluation is currently underway.

Intensity Modulated Proton Therapy for Bilateral Breast or Chest Wall and Comprehensive Nodal Irradiation for Synchronous Bilateral Breast Cancer: Initial Clinical Experience and Dosimetric Comparison

Purpose

Synchronous bilateral breast cancer (SBBC) poses distinct challenges for radiation therapy planning. We report our proton therapy experience in treating patients with SBBC. We also provide a dosimetric comparison of intensity modulated proton therapy (IMPT) versus photon therapy.

Methods and Materials

Patients with SBBC who received IMPT at our institution were retrospectively analyzed. The clinical target volume (CTV) included the breast or chest wall and comprehensive regional lymph nodes, including axilla, supraclavicular fossa, and the internal mammary chain. Intensity modulated proton therapy and volumetric modulated arc therapy (VMAT) plans were generated with the goal that 90% of the CTV would recieve at least 90% of the prescription dose (D90>=90%). Comparisons between modalities were made using the Wilcoxon signed rank test. Physician-reported acute toxic effects and photography were collected at baseline, end of treatment, and each follow-up visit.

Results

Between 2015 and 2018, 11 patients with SBBC were treated with IMPT. The prescription was 50 Gy in 25 fractions. The median CTV D90 was 99.9% for IMPT and 97.6% for VMAT (P = .001). The mean heart dose was 0.7 Gy versus 7.2 Gy (P = .001), the total lung mean dose was 7.8 Gy versus 17.3 Gy (P = .001), and the total lung volume recieving 20 Gy was 13.0% versus 27.4% (P = .001). The most common acute toxic effects were dermatitis (mostly grade 1-2 with 1 case of grade 3) and grade 1 to 2 fatigue. The most common toxic effects at the last-follow up (median, 32 months) were grade 1 skin hyperpigmentation, superficial fibrosis, and extremity lymphedema. No nondermatologic or nonfatigue adverse events of grade >1 were recorded.

Conclusions

Bilateral breast and/or chest wall and comprehensive nodal IMPT is technically feasible and associated with low rates of severe acute toxic effects. Treatment with IMPT offered improved target coverage and normal-tissue sparing compared with photon therapy. Long-term follow-up is ongoing to assess efficacy and toxic effects.

Comparative analysis of acute toxicities and patient reported outcomes between intensity-modulated proton therapy (IMPT) and volumetric modulated arc therapy (VMAT) for the treatment of oropharyngeal cancer

BACKGROUND AND PURPOSE

IMPT improves normal tissue sparing compared to VMAT in treating oropharyngeal cancer (OPC). Our aim was to assess if this translates into clinical benefits.

MATERIALS AND METHODS

OPC patients treated with definitive or adjuvant IMPT or VMAT from 2013 to 2018 were included. All underwent prospective assessment using patient-reported-outcomes (PROs) (EORTC-QLQ-H&N35) and provider-assessed toxicities (CTCAEv4.03). End-of-treatment and pretreatment scores were compared. PEG-tube use, hospitalization, and narcotic use were retrospectively collected. Statistical analysis used the Wilcoxon Rank-Sum Test with propensity matching for PROs/provider-assessed toxicities, and t-tests for other clinical outcomes.

RESULTS

46 IMPT and 259 VMAT patients were included; median follow-up was 12 months (IMPT) and 30 months (VMAT). Baseline characteristics were balanced except for age (p = 0.04, IMPT were older) and smoking (p < 0.01, 10.9% IMPT >20PYs, 29.3% VMAT). IMPT was associated with lower PEG placement (OR = 0.27; 95% CI: 0.12-0.59; p = 0.001) and less hospitalization ≤60 days post-RT (OR = 0.21; 95% CI:0.07-0.6, p < 0.001), with subgroup analysis revealing strongest benefits in patients treated definitively or with concomitant chemoradiotherapy (CRT). IMPT was associated with a relative risk reduction of 22.3% for end-of-treatment narcotic use. Patients reported reduced cough and dysgeusia with IMPT (p < 0.05); patients treated definitively or with CRT also reported feeling less ill, reduced feeding tube use, and better swallow. Provider-assessed toxicities demonstrated less pain and mucositis with IMPT, but more mucosal infection.

CONCLUSION

IMPT is associated with improved PROs, reduced PEG-tube placement, hospitalization, and narcotic requirements. Mucositis, dysphagia, and pain were decreased with IMPT. Benefits were predominantly seen in patients treated definitively or with CRT.