ESTRO 2024 - Abstract Book

S1214

Clinical - Head & neck

ESTRO 2024

Conclusion:

This single-centre study analyses the patterns of locoregional failure in a UK clinical setting. The majority of HNC LRRs in this dataset originate within the high-dose volumes (Type A and Type B), receiving a maximal treatment dose of 65 Gy in 30 fractions with bilateral neck irradiation. Type A, high-dose, classifications are suggestive of radiobiological resistance to treatment dose. The majority of LRR residing in high-dose volumes is consistent with findings from the original classification methodology [2]. There were no significant associations between clinical covariates and classification type. Further investigation into radiomic features of LRR patients and classification types in the CUH dataset is ongoing.

Keywords: registration, spatial mapping, recurrence

References:

[1] Rwigema JC, Heron DE, Ferris RL, Andrade RS, Gibson MK, Yang Y, et al. The impact of tumor volume and radiotherapy dose on outcome in previously irradiated recurrent squamous cell carcinoma of the head and neck treated with stereotactic body radiation therapy. Am J Clin Oncol. 2011;34(4):3729. https://doi.org/10.1097/COC.0b013e3181e84dc0. [2] Mohamed AS, Rosenthal DI, Awan MJ, Garden AS, Kocak-Uzel E, Belal AM, El-Gowily AG, Phan J, Beadle BM, Gunn GB, Fuller CD. Methodology for analysis and reporting patterns of failure in the Era of IMRT: head and neck cancer applications. Radiat Oncol. 2016 Jul 26;11(1):95. doi: 10.1186/s13014-016-0678-7. PMID: 27460585; PMCID: PMC4962405.

641

Poster Discussion

Explainable AI to identify cut-off of muscle loss associated with survival in oral cavity cancer

Jie Lee 1 , Jhen-Bin Lin 2 , Wan-Chun Lin 3 , Kun-Pin Wu 3 , Yu-Jen Chen 1 , Shih-Hua Liu 1

1 MacKay Memorial Hospital, Radiation Oncology, Taipei, Taiwan. 2 Changhua Christian Hospital, Radiation Oncology, Changhua, Taiwan. 3 National Yang Ming Chiao Tung University, Institute of Biomedical Informatics, Taipei, Taiwan

Purpose/Objective:

Muscle loss during radiotherapy is associated with poorer survival outcomes in patients with oral cavity cancer. However, the cut-off of muscle loss associated with survival outcomes remains unclear. Machine learning (ML) algorithms, which can handle complex interactions and non-linear relationships, may demonstrate favorable performance in predicting survival. Explainable artificial intelligence (XAI) can provide visualizations of the inner workings of ML models and potentially identify the cut-off of muscle loss. This study aimed to identify the cut-off of muscle loss associated with survival in oral cavity cancer by using XAI.