Using AI to predict radiotherapy toxicity risk based on patient germline genotyping Book Section
| Authors: | Oh, J. H.; Lee, S.; Thor, M.; Deasy, J. O. |
| Editors: | Mun, S. K.; Dieterich, S. |
| Article/Chapter Title: | Using AI to predict radiotherapy toxicity risk based on patient germline genotyping |
| Abstract: | The collateral irradiation of normal tissues can result in damage that reduces the quality of life for cancer survivors. The variability of toxicity risk has been increasingly recognized as multifactorial, involving patient-specific genetics, dose-volume levels, and other risk factors. The association between genetics and radiotherapy (RT)-induced toxicity, referred to as radiogenomics, has received increasing attention. Traditional statistical analyses have mainly focused on testing the effect of individual genetic variants without considering non-linear interactions of variants. We have shown that artificial intelligence (AI) methods, including machine learning approaches, can efficiently leverage large-scale genetic variants (e.g., single nucleotide polymorphisms [SNPs]), taking into account the complex interactions among genetic markers. In addition, novel post-modeling analyses, employing bioinformatics network techniques, can identify key genes associated with tissue-specific toxicity. The next challenge of genetic prediction models will be to integrate genetic and RT dose-volume factors. Such models have the potential to identify patients at high risk for the development of toxicity and thus offer individualized risk-specific treatment planning. In this chapter, we review results for multiple endpoints yielding usable stratifications of odds ratios for a significant fraction of patients treated with RT. Progress in radiogenomics has been slow primarily due to a lack of data sets and other analytical obstacles. We discuss these issues that need to be addressed when handling genome-wide variants. We conclude by looking to a future when germline genomics is combined with RT dose-volume factors to personalize RT-induced toxicity risk. © 2023 by World Scientific Publishing Co. Pte. Ltd. |
| Book Title: | Artificial Intelligence in Radiation Oncology |
| ISBN: | 978-981-126-353-8 |
| Publisher: | World Scientific Publishing Co. |
| Publication Place: | Hackensack, NJ |
| Date Published: | 2022-01-01 |
| Start Page: | 249 |
| End Page: | 272 |
| Language: | English |
| DOI: | 10.1142/9789811263545_0012 |
| PROVIDER: | scopus |
| DOI/URL: | |
| Notes: | Book Chapter: 12 -- Export Date: 1 May 2023 -- Source: Scopus |
Altmetric
Citation Impact
BMJ Impact Analytics
Related MSK Work