Background

Radiotherapy(RT) is a key curative treatment modality of Head and Neck Cancers(HNC) that, despite technological advances and due to unavoidable irradiation of nearby normal tissues, leads to significant side effects, including permanent debilitating long-term toxicities such as difficulty swallowing, dry mouth or necrosis of the mandible(1). These toxicities are inter-related, with patients experiencing more than one type(2). As such, multi-toxicity prediction models using multimodal data can reflect more accurately the reality and interplay of side effects potentially leading to more effective damage-limiting approaches that are essential in the personalisation of HNC RT. 

Novelty & Importance

The novelty of our work resides on developing multi-toxicity prediction models using multi-modal data that also includes spatial information from 3D dose maps. This work will prove a key step in the subsequent development of AI-based adaptive RT protocols to reduce toxicity and improve outcomes. 

Aims & Objectives

To develop DL-based multi-toxicity prediction models using clinical, demographic and image-based data(dose maps, annotated CT images). 

Planned Research Method

We will evaluate different multimodality fusion strategies of non-imaging (patient, disease, toxicity, DVH) and imaging (dose maps, CT images) data as inputs, assess biases and implement DL intepretability methods to contribute to explainability.

6.Quality Control

We will adhere to TRIPOD guidance(3) and leverage existing national and international collaborations such as the PREDMORN study(4), that our group is leading, to facilitate external evaluation, including making all of our code publicly available to facilitate wide implementation.

References

1.        H.P. van der Laan, et al. Impact of radiation-induced toxicities on quality of life of patients treated for head and neck cancer. Radiotherapy and Oncology, 160:47–53, 7 2021

2.        A.T.T. Wong et al. Symptom burden and dysphagia associated with osteoradionecrosis in long-term oropharynx cancer survivors: A cohort analysis. Oral Oncology, 66:75–80, 3 2017.

3.        Moons KG, et al. Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med. 2015 Jan 6;162(1):W1-73. doi: 10.7326/M14-0698. PMID: 25560730.

4.        L. Humbert-Vidan, et al. Protocol letter: A multi-institutional retrospective case-control cohort investigating PREDiction models for Mandibular OsteoRadioNecrosis in head and neck cancer (PREDMORN). In Radiotherapy and Oncology, 176:99-100, 11, November 2022