A Deep Learning Approach Validates Genetic Risk Factors for Late Toxicity After Prostate Cancer Radiotherapy in a REQUITE Multi-National Cohort

Published on Oct 15, 2020in Frontiers in Oncology4.848
· DOI :10.3389/FONC.2020.541281
Michela Carlotta Massi2
Estimated H-index: 2
(Polytechnic University of Milan),
Francesca Gasperoni (Medical Research Council)+ 43 AuthorsTiziana Rancati27
Estimated H-index: 27
Sources
Abstract
Background: REQUITE (validating pREdictive models and biomarkers of radiotherapy toxicity to reduce side effects and improve QUalITy of lifE in cancer survivors) is an international prospective cohort study. The purpose of this project was to analyse a cohort of patients recruited into REQUITE using a deep learning algorithm to identify patient-specific features associated with the development of toxicity, and test the approach by attempting to validate previously published genetic risk factors. Methods: The study involved REQUITE prostate cancer patients treated with external beam radiotherapy who had complete 2-year follow-up. We used five separate late toxicity endpoints: ≥grade 1 late rectal bleeding, ≥grade 2 urinary frequency, ≥grade 1 haematuria, ≥ grade 2 nocturia, ≥ grade 1 decreased urinary stream. Forty-three single nucleotide polymorphisms (SNPs) already reported in the literature to be associated with the toxicity endpoints were included in the analysis. No SNP had been studied before in the REQUITE cohort. Deep Sparse AutoEncoders (DSAE) were trained to recognize features (SNPs) identifying patients with no toxicity and tested on a different independent mixed population including patients without and with toxicity. Results: One thousand, four hundred and one patients were included, and toxicity rates were: rectal bleeding 11.7%, urinary frequency 4%, haematuria 5.5%, nocturia 7.8%, decreased urinary stream 17.1%. Twenty-four of the 43 SNPs that were associated with the toxicity endpoints were validated as identifying patients with toxicity. Twenty of the 24 SNPs were associated with the same toxicity endpoint as reported in the literature: 9 SNPs for urinary symptoms and 11 SNPs for overall toxicity. The other 4 SNPs were associated with a different endpoint. Conclusion: Deep learning algorithms can validate SNPs associated with toxicity after radiotherapy for prostate cancer. The method should be studied further to identify polygenic SNP risk signatures for radiotherapy toxicity. The signatures could then be included in integrated normal tissue complication probability models and tested for their ability to personalize radiotherapy treatment planning.
References33
Newest
#1Sarah L. Kerns (URMC: University of Rochester Medical Center)H-Index: 27
#2Laura FachalH-Index: 24
Last. Catharine M L West (MAHSC: Manchester Academic Health Science Centre)H-Index: 77
view all 40 authors...
BACKGROUND: 10-20% of patients develop long-term toxicity following radiotherapy for prostate cancer. Identification of common genetic variants associated with susceptibility to radiotoxicity might improve risk prediction and inform functional mechanistic studies. METHODS: We conducted an individual patient data meta-analysis of six genome-wide association studies (n = 3,871) in men with European ancestry who underwent radiotherapy for prostate cancer. Radiotoxicities (increased urinary frequenc...
23 CitationsSource
#1Edmond Pui Hang Choi (HKU: University of Hong Kong)H-Index: 13
#2Eric Yuk Fai Wan (HKU: University of Hong Kong)H-Index: 15
Last. Cindy L. K. Lam (HKU: University of Hong Kong)H-Index: 30
view all 5 authors...
Background Even though the negative impacts of nocturia on sleep quality and health-related quality of life (HRQOL) have been documented in previous research, their interrelationship has been poorly studied. This study aimed to explore whether nocturia would affect sleep quality, which in turn affects HRQOL.
9 CitationsSource
#1Petra Seibold (DKFZ: German Cancer Research Center)H-Index: 27
#2A Webb (University of Leicester)H-Index: 9
Last. William Li (The Queen's Medical Center)H-Index: 1
view all 128 authors...
Abstract Purpose REQUITE aimed to establish a resource for multi-national validation of models and biomarkers that predict risk of late toxicity following radiotherapy. The purpose of this article is to provide summary descriptive data. Methods An international, prospective cohort study recruited cancer patients in 26 hospitals in eight countries between April 2014 and March 2017. Target recruitment was 5300 patients. Eligible patients had breast, prostate or lung cancer and planned potentially ...
15 CitationsSource
#1Sangkyu Lee (MSK: Memorial Sloan Kettering Cancer Center)H-Index: 6
#2Sarah L. Kerns (URMC: University of Rochester Medical Center)H-Index: 27
Last. Jung Hun Oh (MSK: Memorial Sloan Kettering Cancer Center)H-Index: 26
view all 6 authors...
Purpose Late genitourinary (GU) toxicity after radiation therapy limits the quality of life of prostate cancer survivors; however, efforts to explain GU toxicity using patient and dose information have remained unsuccessful. We identified patients with a greater congenital GU toxicity risk by identifying and integrating patterns in genome-wide single nucleotide polymorphisms (SNPs). Methods and Materials We applied a preconditioned random forest regression method for predicting risk from the gen...
28 CitationsSource
#1Issam El Naqa (UM: University of Michigan)H-Index: 58
#2Sarah L. Kerns (UR: University of Rochester)H-Index: 27
Last. Randall K. Ten Haken (UM: University of Michigan)H-Index: 79
view all 8 authors...
: Advances in patient-specific information and biotechnology have contributed to a new era of computational medicine. Radiogenomics has emerged as a new field that investigates the role of genetics in treatment response to radiation therapy. Radiation oncology is currently attempting to embrace these recent advances and add to its rich history by maintaining its prominent role as a quantitative leader in oncologic response modeling. Here, we provide an overview of radiogenomics starting with gen...
27 CitationsSource
#1Tiziana RancatiH-Index: 27
#2Federica PaloriniH-Index: 9
Last. Riccardo Valdagni (University of Milan)H-Index: 50
view all 5 authors...
One of the most relevant achievements of Professor Gianni Bonadonna was the implementation of the methodology of controlled clinical trials in medical oncology. It is valid for all cancer types, on...
8 CitationsSource
#1David Azria (French Institute of Health and Medical Research)H-Index: 56
#2Ariane Lapierre (French Institute of Health and Medical Research)H-Index: 1
Last. Barry S. Rosenstein (ISMMS: Icahn School of Medicine at Mount Sinai)H-Index: 35
view all 19 authors...
The ability to stratify patients using a set of biomarkers that predicts toxicity risk would allow for radiotherapy modulation and serve as a valuable tool for precision medicine and personalized radiotherapy (RT). For patients presenting with tumors with a low risk of recurrence, modifying RT schedules to avoid toxicity would be clinically advantageous. Indeed, for the patient at low risk of developing radiation associated toxicity, use of a hypofractionated protocol could be proposed leading t...
19 CitationsSource
#1Jung Hun Oh (MSK: Memorial Sloan Kettering Cancer Center)H-Index: 26
#2Sarah L. Kerns (URMC: University of Rochester Medical Center)H-Index: 27
Last. Joseph O. Deasy (MSK: Memorial Sloan Kettering Cancer Center)H-Index: 64
view all 6 authors...
Computational methods using genome-wide association studies to predict radiotherapy complications and to identify correlative molecular processes
23 CitationsSource
#1Christopher I. Amos (Dartmouth College)H-Index: 132
#2Joe Dennis (University of Cambridge)H-Index: 55
Last. Douglas F. Easton (University of Cambridge)H-Index: 171
view all 93 authors...
Background:Common cancers develop through a multistep process often including inherited susceptibility. Collaboration among multiple institutions, and funding from multiple sources, has allowed the development of an inexpensive genotyping microarray, the OncoArray. The array includes a genome-wide backbone, comprising 230,000 SNPs tagging most common genetic variants, together with dense mapping of known susceptibility regions, rare variants from sequencing experiments, pharmacogenetic markers a...
171 CitationsSource
#1Dirk De Ruysscher (UM: Maastricht University)H-Index: 84
#2Gilles Defraene (Katholieke Universiteit Leuven)H-Index: 12
Last. Catharine M L West (University of Manchester)H-Index: 77
view all 13 authors...
Abstract The optimal design and patient selection for interventional trials in radiogenomics seem trivial at first sight. However, radiogenomics do not give binary information like in e.g. targetable mutation biomarkers. Here, the risk to develop severe side effects is continuous, with increasing incidences of side effects with higher doses and/or volumes. In addition, a multi-SNP assay will produce a predicted probability of developing side effects and will require one or more cut-off threshold...
9 CitationsSource
Cited By2
Newest
#1Nicola Rares Franco (Ghent University Hospital)
#1Nicola Rares Franco (Ghent University Hospital)H-Index: 1
Last. Tiziana RancatiH-Index: 27
view all 43 authors...
AIM To identify the effect of single nucleotide polymorphism (SNP) interactions on the risk of toxicity following radiotherapy (RT) for prostate cancer (PCa) and propose a new method for polygenic risk score incorporating SNP-SNP interactions (PRSi). MATERIALS AND METHODS Analysis included the REQUITE PCa cohort that received external beam RT and was followed for 2 years. Late toxicity endpoints were: rectal bleeding, urinary frequency, haematuria, nocturia, decreased urinary stream. Among 43 li...
Source
#1Michela Carlotta Massi (Polytechnic University of Milan)H-Index: 2
#2Francesca IevaH-Index: 15
Last. Anna Maria PaganoniH-Index: 18
view all 4 authors...
Class imbalance is a common issue in many domain applications of learning algorithms. Oftentimes, in the same domains it is much more relevant to correctly classify and profile minority class observations. This need can be addressed by Feature Selection (FS), that offers several further advantages, s.a. decreasing computational costs, aiding inference and interpretability. However, traditional FS techniques may become sub-optimal in the presence of strongly imbalanced data. To achieve FS advanta...