I. PROJECT SUMMARY
ID: PN-IV-P1-PCE-2023-0042
Title: Delta-radiomics for personalised treatment of head and neck cancer using artificial intelligence (DeltaRadHNC)
Duration: Jan 2025 – Dec 2027
The challenge: Head and neck cancer (HNC) is a complex disease with incidence on the rise worldwide requiring continuous research for treatment optimisation and personalisation. HNC presents an exclusive set of diagnostic and therapeutic challenges due to its complex radiobiological behaviour and tumour heterogeneity. To advance HNC management there is need for a more in-depth evaluation of tumour features responsible for local recurrence and/or second cancers, using radiomic and holomic approaches.
The goal: The employment of radiomic analysis and holomic data to predict treatment outcome in HNC patients, risk of recurrence/second cancer as well as timeline of recurrence. The radiomic features augmented with additional clinical information are used as input for deep learning networks trained and validated on retrospective data from a cohort of over 1,000 HNC patients. To achieve this goal, highly detailed and robust tumour imaging features (radiomic features) will be identified using deep learning, and further augmented with non-imaging data for a complex patient characterization (holomic features).
The outcome: The development of a software system to offer clinicians a risk evaluation tool for recurrence/second cancer with every follow-up image acquired (delta-radiomics), allowing for close monitoring of patients at risk. The research will also provide the clinician a rich set of data to enable treatment intensity adjustment for a more personalised and optimised therapy.
II. MAIN OBJECTIVE
The employment of radiomic analysis and clinical (holomic) data to predict treatment outcome and evaluate risk as well as timeline of possible recurrence in HNC patients. The radiomic features augmented with additional clinical information to be used as input for a deep learning network trained and validated on retrospective data from a cohort of over 1,000 HNC patients treated at the Amethyst Radiotherapy Centre, Cluj-Napoca.
The specific objectives of the project are the following:
- To connect a set of radiomic features to specific biological parameters (such as hypoxia, proliferative status, intrinsic radioresistance) in order to assist with treatment adaptation and optimisation of HNC patients.
- To develop a deep learning network to be trained and validated on radiomic features extracted from retrospective CT images of HNC patients. Enrich the radiomic information with patient-specific clinical data for a holomic
- To use the trained deep learning network to predict treatment outcome for individual patients.
- To develop a secondary deep learning network trained to detect patients with suboptimal post-treatment CT images that require more specific imaging (PET, MRI).
- To perform an early detection of patients with high risk of recurrence / secondary cancer using the variation in radiomic features between sequential CT images taken at different time points (delta-radiomics).
III. THE RESEARCH TEAM
PROJECT LEADER:
Loredana Gabriela MARCU (https://www.scopus.com/authid/detail.uri?authorId=7006835272)
MEMBERS:
Oreste STRACIUC (https://www.scopus.com/authid/detail.uri?authorId=35191261800)
David MARCU (https://www.scopus.com/authid/detail.uri?authorId=57188819985)
Renata ZAHU (https://www.scopus.com/authid/detail.uri?authorId=57880563900)
Ioana-Claudia COSTIN (https://www.scopus.com/authid/detail.uri?authorId=57860742300)
IV. PREVIOUS PUBLICATIONS RELATED TO THE PROJECT
- L. Marcu, D. Marcu
Pharmacogenomics and Big Data in medical oncology: developments and challenges
Therapeutic Advances Med Oncol 16:1-18 (2024)
- D. Marcu, C. Grava, L. Marcu
Current role of delta-radiomics in head and neck oncology
Int J Molecular Sciences 24(3):2214 (2023)
- L. Marcu, D. Marcu
Current omics trends in personalised head and neck cancer chemoradiotherapy
Journal of Personalized Medicine 11(11):1094 (2021)
- L. Marcu, D. Marcu
Points of view on Artificial Intelligence in radiology – one good, one bad and one fuzzy
Health and Technology 11:17-22 (2021)
- L. Marcu, C. Boyd, E. Bezak
Current issues regarding artificial intelligence in cancer and health care. Implications for medical physicists and biomedical engineers.
Health and Technology 9(4):375-381 (2019)
- L. Marcu, J. Forster, E. Bezak
The potential role of radiomics and radiogenomics in patient stratification by tumor hypoxia status
J American College Radiol 16(9PB):1330-1338 (2019)
- L. Marcu, C. Boyd, E. Bezak
Feeding the data monster: data science in head and neck cancer for personalized therapy
J American College Radiol 16(12):1695-1701 (2019)
- L. Marcu, P. Reid, E. Bezak
The promise of novel biomarkers for head and neck cancer from an imaging perspective
Int J Molec Sci 19(9):2511 (2018)
- L. Marcu, L. Moghaddasi, E. Bezak
Imaging of tumour characteristics and molecular pathways with PET: developments over the last decade towards personalised cancer therapy
Int J Radiat Oncol Biol Phys 102(4):1165-1182 (2018)
- L. Marcu, D. Marcu
In silico modelling of radiation effects towards personalised treatment in radiotherapy
AIP Conference Proceedings 1916(1): 040001 (2017)