Multi-Modality Artificial Intelligence for Involved-Site Radiation Therapy: Clinical Target Volume Delineation in High-Risk Pediatric Hodgkin Lymphoma.
Authors
Affiliations (12)
Affiliations (12)
- Department of Radiology, School of Medicine and Public Health, University of Wisconsin, Madison, WI, USA. Electronic address: [email protected].
- Department of Radiation Oncology, University of Colorado School of Medicine, Aurora, CO, USA.
- Department of Radiation Oncology, BC Cancer, Vancouver Center, Vancouver, Canada.
- Department of Radiation Oncology, Centre Hospitalier de l'Universite de Montreal, Centre Hospitalier Universitaire Sainte-Justine, Montreal, Quebec, Canada.
- Department of Radiation Oncology, University of Pennsylvania, Philadelphia, PA, USA.
- Department of Radiology, School of Medicine and Public Health, University of Wisconsin, Madison, WI, USA.
- Department of Radiology, School of Medicine and Public Health, University of Wisconsin, Madison, WI, USA; University of Wisconsin Carbone Comprehensive Cancer Center, Madison, WI, USA.
- Department of Pediatric Oncology, Roswell Park Comprehensive Cancer Center, Buffalo, NY, USA; Department of Pediatrics, University at Buffalo Jacobs School of Medicine and Biomedical Sciences, Buffalo, NY, USA.
- Radiation Medicine Program, Princess Margaret Cancer Centre, Toronto, Ontario, Canada.
- Department of Pediatrics, Emory University School of Medicine, Atlanta, GA, USA; Aflac Cancer and Blood Disorders Center, Children's Healthcare of Atlanta, Atlanta, GA, USA.
- Department of Radiation Oncology, Mayo Clinic, Jacksonville, FL, USA.
- Department of Radiology, School of Medicine and Public Health, University of Wisconsin, Madison, WI, USA. Electronic address: [email protected].
Abstract
Clinical target volume (CTV) delineation for involved-site radiation therapy (ISRT) in Hodgkin lymphoma (HL) is time-consuming due to the need to analyze multi-time-point PET/CT scans co-registered to the planning CT. Our goal was to develop automated CTV segmentation algorithms that integrated multi-modality imaging to facilitate ISRT planning. This study included planning CT, baseline PET/CT (PET1), and interim PET/CT (PET2) scans from 288 pediatric patients with high-risk HL enrolled in the [redacted] trial. Data from 58 patients across 24 institutions were held out for external testing, while the remaining 230 cases from 95 institutions were used for model development. We investigated three deep learning (DL) architectures (SegResNet, ResUNet, and SwinUNETR) and evaluated the impact of incorporating PET1 and PET2 images along with the planning CT. Performance was assessed using the Dice similarity coefficient (DSC) and 95% Hausdorff distance (HD95). Inter-observer variability (IOV) was estimated by comparing original institutional CTVs with those newly delineated by four radiation oncologists on 10 cases. The quality of CTVs generated by the top-performing model was independently assessed by radiation oncologists on 40 other cases using a 5-point Likert scale and compared against the original institutional CTVs. On the external cohort, a SwinUNETR model incorporating planning CT, PET1, and PET2 images achieved the highest performance, with a DSC of 0.72 and HD95 of 34.43 mm. All models incorporating PET/CT images were significantly better (P<0.01) than planning CT-only models. IOV analysis yielded a DSC of 0.70 and HD95 of 30.14 mm. In clinical evaluation, DL-generated CTVs received a mean quality score of 3.38 out of 5, comparable to original physician-delineated CTVs (3.13; P = 0.13) CONCLUSION: The DL model was able to generate clinically useful CTVs with quality comparable to manually delineated CTVs, suggesting its potential to improve physician efficiency in ISRT planning.