Advancing Normal Tissue Complication Probability Modeling with Supervised Contrastive Learning for Predicting Osteoradionecrosis

Eric Ababio Anyimadu; Xinhua Zhang; Clifton David Fuller; G. Elisabeta Marai; Guadalupe Canahuate

doi:10.1145/3748522.3779983

Back

Advancing Normal Tissue Complication Probability Modeling with Supervised Contrastive Learning for Predicting Osteoradionecrosis

Conference proceeding

Open access

Peer reviewed

Advancing Normal Tissue Complication Probability Modeling with Supervised Contrastive Learning for Predicting Osteoradionecrosis

Eric Ababio Anyimadu, Xinhua Zhang, Clifton David Fuller, G. Elisabeta Marai and Guadalupe Canahuate

Proceedings of the 41st ACM/SIGAPP Symposium on Applied Computing, Vol.2026, pp.140-147

ACM Conferences

SAC '26: 41st ACM/SIGAPP Symposium on Applied Computing

06/09/2026

DOI: 10.1145/3748522.3779983

PMCID: PMC13249462

PMID: 42272749

Appears in UI Libraries Support Open Access

Files and links (1)

url

https://doi.org/10.1145/3748522.3779983View

Published (Version of record) Open Access

Abstract

Normal tissue complication probability (NTCP) modeling using dose-volume histograms (DVHs) is fundamentally challenged by high dimensionality, severe multicollinearity, and substantial overlap between DVH profiles of patients with differing toxicity outcomes, limiting the effectiveness of classical classification approaches. We introduce SC-NTCP, a supervised contrastive learning framework that transforms DVH data into a compact, separable latent representation optimized for predicting osteoradionecrosis (ORN) in head and neck cancer patients. Rather than relying on raw, high-dimensional DVH features, SC-NTCP explicitly maximizes intra-class similarity and inter-class separability within the embedding space, enabling more accurate downstream classification. Using a cohort of head and neck cancer patients, we benchmarked SC-NTCP against logistic regression, support vector machines, multilayer perceptrons, and convolutional neural networks. SC-NTCP demonstrated superior discrimination (AUC = 0.77), improved calibration, and enhanced interpretability via gradient-based feature attribution, while the integration of clinical covariates further augmented predictive performance. By addressing the inherent limitations of DVH data, SC-NTCP offers a principled and interpretable approach for robust radiation toxicity prediction, with the potential to inform personalized treatment planning and improve clinical outcomes.

contrastive learning

normal tissue complication

osteoradionecrosis

dose-volume histogram

UIOWA OA Agreement

Details

Title: Subtitle: Advancing Normal Tissue Complication Probability Modeling with Supervised Contrastive Learning for Predicting Osteoradionecrosis
Creators: Eric Ababio Anyimadu - University of Iowa
Xinhua Zhang - University of Illinois Chicago
Clifton David Fuller - The University of Texas MD Anderson Cancer Center
G. Elisabeta Marai - University of Illinois Chicago
Guadalupe Canahuate - University of Iowa
Resource Type: Conference proceeding
Publication Details: Proceedings of the 41st ACM/SIGAPP Symposium on Applied Computing, Vol.2026, pp.140-147
Conference: SAC '26: 41st ACM/SIGAPP Symposium on Applied Computing
Series: ACM Conferences
DOI: 10.1145/3748522.3779983
PMID: 42272749
PMCID: PMC13249462
Publisher: Association for Computing Machinery (ACM)
Number of pages: 8
Language: English
Date published: 06/09/2026
Academic Unit: Electrical and Computer Engineering
Record Identifier: 9985174809402771

Metrics

1 Record Views