
==== Front
medRxiv
MEDRXIV
medRxiv
Cold Spring Harbor Laboratory

10.1101/2024.09.11.24313485
preprint
1
Article
Computed tomography radiomics-based cross-sectional detection of mandibular osteoradionecrosis in head and neck cancer survivors
MD Anderson Head and Neck Cancer Symptom Working Group
Kamel Serageldin http://orcid.org/0000-0002-0046-4337

Humbert-Vidan Laia http://orcid.org/0000-0002-8005-6770

Kaffey Zaphanlene http://orcid.org/0009-0008-7999-5245

Abusaif Abdulrahman
Fuentes David T. A. http://orcid.org/0000-0002-2572-6962

Wahid Kareem http://orcid.org/0000-0002-0503-0175

Dede Cem http://orcid.org/0000-0002-0543-9325

Naser Mohamed A. http://orcid.org/0000-0003-1020-4966

He Renjie http://orcid.org/0000-0001-9166-6286

Moawad Ahmed W.
Elsayes Khaled M.
Chen Melissa M. http://orcid.org/0000-0002-3274-2653

Otun Adegbenga O. http://orcid.org/0000-0001-7599-1628

Rigert Jillian http://orcid.org/0000-0001-5419-980X

Chambers Mark http://orcid.org/0000-0001-8901-6583

Hope Andrew http://orcid.org/0000-0001-6793-2863

Watson Erin http://orcid.org/0000-0002-2096-7732

Brock Kristy K. http://orcid.org/0000-0001-9364-5040

Hutcheson Katherine http://orcid.org/0000-0003-3710-5706

van Dijk Lisanne http://orcid.org/0000-0002-9515-5616

Moreno Amy C. http://orcid.org/0000-0001-6762-6807

Lai Stephen Y. http://orcid.org/0000-0001-8301-7286

Fuller Clifton D. http://orcid.org/0000-0002-5264-3994

Mohamed Abdallah S. R. http://orcid.org/0000-0003-2064-7613

12 9 2024
2024.09.11.24313485https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use.
http://medrxiv.org/lookup/doi/10.1101/2024.09.11.24313485
nihpp-2024.09.11.24313485.pdf
Abstract

Purpose

This study aims to identify radiomic features extracted from contrast-enhanced CT scans that differentiate osteoradionecrosis (ORN) from normal mandibular bone in patients with head and neck cancer (HNC) treated with radiotherapy (RT).

Materials and Methods

Contrast-enhanced CT (CECT) images were collected for 150 patients (80% train, 20% test) with confirmed ORN diagnosis at The University of Texas MD Anderson Cancer Center between 2008 and 2018. Using PyRadiomics, radiomic features were extracted from manually segmented ORN regions and the corresponding automated control regions, the later defined as the contralateral healthy mandible region. A subset of pre-selected features was obtained based on correlation analysis (r > 0.95) and used to train a Random Forest (RF) classifier with Recursive Feature Elimination. Model explainability SHapley Additive exPlanations (SHAP) analysis was performed on the 20 most important features identified by the trained RF classifier.

Results

From a total of 1316 radiomic features extracted, 810 features were excluded due to high collinearity. From a set of 506 pre-selected radiomic features, the optimal subset resulting on the best discriminative accuracy of the RF classifier consisted of 67 features. The RF classifier was well calibrated (Log Loss 0.296, ECE 0.125) and achieved an accuracy of 88% and a ROC AUC of 0.96. The SHAP analysis revealed that higher values of Wavelet-LLH First-order Mean and Median were associated with ORN of the jaw (ORNJ). Conversely, higher Exponential GLDM Dependence Entropy and lower Square First-order Kurtosis were more characteristic of normal mandibular tissue.

Conclusion

This study successfully developed a CECT-based radiomics model for differentiating ORNJ from healthy mandibular tissue in HNC patients after RT. Future work will focus on the detection of subclinical ORNJ regions to guide earlier interventions.
==== Body
pmc
