Image-Based Mandibular and Maxillary Parcellation and Annotation Using Computed Tomography (IMPACT): A Deep Learning-Based Clinical Tool for Orodental Dose Estimation and Osteoradionecrosis Assessment

Authors

Laia Humbert-Vidan
Austin H Castelo
Renjie He
Lisanne V van Dijk
Dong Joo Rhee
Congjun Wang
He C Wang
Kareem A Wahid
Sonali Joshi
Parshan Gerafian
Natalie West
Zaphanlene Kaffey
Sarah Mirbahaeddin
Jaqueline Curiel
Samrina Acharya
Amal Shekha
Praise Oderinde
Alaa M S Ali
Andrew Hope
Erin Watson
Ruth Wesson-Aponte
Steven J Frank
Carly E A Barbon
Kristy K Brock
Mark S Chambers
Muhammad Walji
Katherine A Hutcheson
Stephen Y Lai
Clifton D Fuller
Mohamed A Naser
Amy C Moreno
OPC-SURVIVOR Program
MD Anderson Head and Neck Cancer Symptom Working Group

Publication Date

7-1-2025

Journal

Physics and Imaging in Radiation Oncology

DOI

10.1016/j.phro.2025.100817

PMID

40894269

PMCID

40894269

PubMedCentral® Posted Date

7-25-2025

PubMedCentral® Full Text Version

Author MSS

Abstract

Background and purpose: Accurate delineation of orodental structures on radiotherapy computed tomography (CT) images is essential for dosimetric assessment and dental decisions. We propose a deep-learning (DL) auto-segmentation framework for individual teeth and mandible/maxilla sub-volumes aligned with the ClinRad osteoradionecrosis staging system.

Materials and methods: Mandible and maxilla sub-volumes were manually defined on simulation CT images from 60 clinical cases, differentiating alveolar from basal regions; teeth were labelled individually. For each task, a DL segmentation model was independently trained. A Swin UNETR-based model was used for mandible sub-volumes. For smaller structures (e.g., teeth and maxilla sub-volumes) a two-stage model first used the ResUNet to segment the entire teeth and maxilla regions as a single ROI used to crop the image input for Swin UNETR. In addition to segmentation accuracy and geometric precision, a dose-volume comparison was made between manual and model-predicted segmentations.

Results: Segmentation performance varied across sub-volumes - mean Dice values of 0.85 (mandible basal), 0.82 (mandible alveolar), 0.78 (maxilla alveolar), 0.80 (upper central teeth), 0.69 (upper premolars), 0.76 (upper molars), 0.76 (lower central teeth), 0.70 (lower premolars), 0.71 (lower molars) - with limited applicability in segmenting sub-volumes absent in the data. The maxilla alveolar central sub-volume showed a statistically significant dose-volume difference in both Dmean and D2%.

Conclusions: We present a novel DL-based auto-segmentation framework of orodental structures, enabling spatial localization of dose-related differences. This tool may enhance image-based bone injury detection and improve clinical decision-making in radiation oncology and dental care for head and neck cancer patients.

Keywords

Head and neck cancer, Radiotherapy, Deep-learning auto-segmentation models, Orodental structures, Osteoradionecrosis

Published Open-Access

yes

Recommended Citation

Humbert-Vidan, Laia; Castelo, Austin H; He, Renjie; et al., "Image-Based Mandibular and Maxillary Parcellation and Annotation Using Computed Tomography (IMPACT): A Deep Learning-Based Clinical Tool for Orodental Dose Estimation and Osteoradionecrosis Assessment" (2025). Faculty, Staff and Student Publications. 4845.
https://digitalcommons.library.tmc.edu/uthgsbs_docs/4845