1237 - Somatic Genetic Variants as Predictors for Radiation-Induced Short and Long-Term Toxicity in Head and Neck Cancers
Presenter(s)
A. D. Thompson1, M. H. Brown2, Y. Song3, D. Rajabi Abhari3, S. S. Valluri4, A. Shetty5, W. Mendes6, R. Mehra7, K. Hatten8, R. Taylor8, K. F. Moyer8, J. S. Wolf8, T. N. Tyer1, K. Steacy1, C. Eggleston2, C. Datnow-Martinez2, D. Gaykalova9, P. T. Tran10, M. J. Ferris2, and J. K. Molitoris2; 1Department of Radiation Oncology, University of Maryland Medical Center, Baltimore, MD, 2Department of Radiation Oncology, University of Maryland School of Medicine, Baltimore, MD, 3University of Maryland School of Medicine, Baltimore, MD, 4Univ of Maryland School of Medicine, Radiation Oncology, Baltimore, MD, 5Department of Radiation Oncology, Division of Translational Radiation Sciences, University of Maryland Baltimore, School of Medicine,, Baltimore, MD, 6University of Maryland School of Medicine, Baltimore, MD, USA, Baltimore, MD, 7Division of Hematology/Oncology, University of Maryland School of Medicine, Baltimore, MD, 8Department of Otorhinolaryngology - Head & Neck Surgery, University of Maryland School of Medicine, Baltimore, MD, 9University of Maryland School of Medicine, BALTIMORE, MD, 10Division of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX
Purpose/Objective(s): Radiation (RT)-induced toxicity impacts quality of life in head and neck cancer (HNC), yet genomic determinants of RT morbidity remain poorly characterized. While tissue radiosensitivity is largely host mediated, tumor alterations in DNA repair and chromatin remodeling may reflect shared mechanisms, germline predisposition, or tumor-host interactions influencing toxicity. This study evaluates somatic tumor mutation associations with RT toxicities in HNC.
Materials/Methods: Next-generation sequencing (648 gene panel) was performed on 143 HNC tumor samples. Toxicities, defined as grade 1+ per CTCAE v5, were retrospectively collected from medical records across 16 acute and 19 late endpoints. Univariable Fisher’s exact test assessed mutation-toxicity associations. Somatic mutations were annotated via COSMIC as oncogenes or tumor suppressor genes.
Results: The most common subsites were oropharynx 46.9%, oral cavity 18.9%, and larynx 13.3%; 46.2% were p16 positive. Acute toxicity occurred in 97.9%, including dermatitis (88.1%), pain (67.8%), and dysphagia (53.1%). Late toxicity occurred in 77.2%, including xerostomia (51.5%), dysphagia (44.6%), dysgeusia (25.7%), and trismus (16.8%). Overall, 184 unadjusted associations (p<0.05) were identified involving 72 genes across 37 endpoints. Mutation associations clustered into distinct toxicity groups. ATM was associated with acute anorexia (OR 30.0, p=0.004) and dehydration (p=0.010). ASXL1 and WNK2 were each associated with anorexia, dehydration, and feeding tube placement. KMT2C was associated with 7 late endpoints including trismus (OR 19.7, p=0.002), dysphagia (OR 10.5, p=0.029), and xerostomia (OR 7.3, p=0.035). CBLB had 20 late associations including dysgeusia (OR 47.8, p=0.002), xerostomia (p=0.010), fibrosis (p=0.029), and edema (p=0.023). CDH1 was protective for dermatitis (OR 0.1, p=0.037) and pain (OR 0.1, p=0.040).
Conclusion: To our knowledge, this is the largest study exploring associations between somatic tumor mutations and RT-induced toxicities in HNC. ATM, a DNA damage response mediator, clustered with acute nutritional decline. KMT2C, involved in chromatin remodeling, clustered with late structural toxicities including trismus, dysphagia, and xerostomia. CBLB, a negative immune regulator, had the broadest late toxicity profile consistent with immune mediated tissue damage. CDH1 was protective for dermatitis, consistent with its role in skin regeneration. These patterns suggest tumor genomic alterations may correlate with toxicity, potentially reflecting shared DNA repair biology, germline overlap, or tumor-host interactions. As associations were univariable without adjustment for confounders or multiple comparisons, validation with multivariable adjustment and paired germline analysis is needed before clinical application can be considered.