Efficient and accurate non-invasive tumor monitoring and diagnosis by interpretable deep learning

Summary: Detecting tumor-specific DNA methylation in circulating tumor DNA (ctDNA) offers a non-invasive method for tumor detection. The primary challenge lies in identifying the extremely low abundance of ctDNA in cell-free blood plasma (cfDNA). In this study, we present Oncoder, an interpretable d...

Full description

Saved in:
Bibliographic Details
Main Authors: Youpeng Yang, Jiaying Liu, Yutong He, Yingjie Yang, Tao Jiang, Jia Tang, Xin Li
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:iScience
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2589004225014191
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849708297448325120
author Youpeng Yang
Jiaying Liu
Yutong He
Yingjie Yang
Tao Jiang
Jia Tang
Xin Li
author_facet Youpeng Yang
Jiaying Liu
Yutong He
Yingjie Yang
Tao Jiang
Jia Tang
Xin Li
author_sort Youpeng Yang
collection DOAJ
description Summary: Detecting tumor-specific DNA methylation in circulating tumor DNA (ctDNA) offers a non-invasive method for tumor detection. The primary challenge lies in identifying the extremely low abundance of ctDNA in cell-free blood plasma (cfDNA). In this study, we present Oncoder, an interpretable deep learning-based tool for economical and accurate non-invasive tumor monitoring and diagnosis. Unlike other methods, Oncoder learns scientifically sound reference methylation atlases from patient blood to provide additional diagnostic insights, fostering trust among clinicians and patients, and continuously improves its accuracy through iterative learning. In simulations, Oncoder reduced prediction errors of tumor signals in blood by at least 30% compared to existing methods and showed the highest prediction correlation, indicating more accurate tumor progression monitoring. We also evaluated Oncoder’s performance in various real-world applications. Oncoder sensitively detected changes in ctDNA levels during tumor development and treatment and exhibited superior diagnostic potential even in the earliest stages of cancer.
format Article
id doaj-art-51bf5865e39d4448b73e686b7687bc84
institution DOAJ
issn 2589-0042
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series iScience
spelling doaj-art-51bf5865e39d4448b73e686b7687bc842025-08-20T03:15:42ZengElsevieriScience2589-00422025-08-0128811315810.1016/j.isci.2025.113158Efficient and accurate non-invasive tumor monitoring and diagnosis by interpretable deep learningYoupeng Yang0Jiaying Liu1Yutong He2Yingjie Yang3Tao Jiang4Jia Tang5Xin Li6School of Medicine, Shenzhen Campus of Sun Yat-sen University, Shenzhen, ChinaSchool of Medicine, Shenzhen Campus of Sun Yat-sen University, Shenzhen, ChinaSchool of Medicine, Shenzhen Campus of Sun Yat-sen University, Shenzhen, ChinaSchool of Medicine, Shenzhen Campus of Sun Yat-sen University, Shenzhen, ChinaSchool of Medicine, Shenzhen Campus of Sun Yat-sen University, Shenzhen, ChinaNHC Key Laboratory of Male Reproduction and Genetics, Guangdong Provincial Reproductive Science Institute (Guangdong Provincial Fertility Hospital), Guangzhou, China; Department of Public Health and Preventive Medicine, School of Medicine, Jinan University, Guangzhou, China; Corresponding authorSchool of Medicine, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China; Guangdong Provincial Key Laboratory of Digestive Cancer Research, The Seventh Affiliated Hospital of Sun Yat-sen University, Shenzhen, China; Corresponding authorSummary: Detecting tumor-specific DNA methylation in circulating tumor DNA (ctDNA) offers a non-invasive method for tumor detection. The primary challenge lies in identifying the extremely low abundance of ctDNA in cell-free blood plasma (cfDNA). In this study, we present Oncoder, an interpretable deep learning-based tool for economical and accurate non-invasive tumor monitoring and diagnosis. Unlike other methods, Oncoder learns scientifically sound reference methylation atlases from patient blood to provide additional diagnostic insights, fostering trust among clinicians and patients, and continuously improves its accuracy through iterative learning. In simulations, Oncoder reduced prediction errors of tumor signals in blood by at least 30% compared to existing methods and showed the highest prediction correlation, indicating more accurate tumor progression monitoring. We also evaluated Oncoder’s performance in various real-world applications. Oncoder sensitively detected changes in ctDNA levels during tumor development and treatment and exhibited superior diagnostic potential even in the earliest stages of cancer.http://www.sciencedirect.com/science/article/pii/S2589004225014191CancerArtificial intelligence
spellingShingle Youpeng Yang
Jiaying Liu
Yutong He
Yingjie Yang
Tao Jiang
Jia Tang
Xin Li
Efficient and accurate non-invasive tumor monitoring and diagnosis by interpretable deep learning
iScience
Cancer
Artificial intelligence
title Efficient and accurate non-invasive tumor monitoring and diagnosis by interpretable deep learning
title_full Efficient and accurate non-invasive tumor monitoring and diagnosis by interpretable deep learning
title_fullStr Efficient and accurate non-invasive tumor monitoring and diagnosis by interpretable deep learning
title_full_unstemmed Efficient and accurate non-invasive tumor monitoring and diagnosis by interpretable deep learning
title_short Efficient and accurate non-invasive tumor monitoring and diagnosis by interpretable deep learning
title_sort efficient and accurate non invasive tumor monitoring and diagnosis by interpretable deep learning
topic Cancer
Artificial intelligence
url http://www.sciencedirect.com/science/article/pii/S2589004225014191
work_keys_str_mv AT youpengyang efficientandaccuratenoninvasivetumormonitoringanddiagnosisbyinterpretabledeeplearning
AT jiayingliu efficientandaccuratenoninvasivetumormonitoringanddiagnosisbyinterpretabledeeplearning
AT yutonghe efficientandaccuratenoninvasivetumormonitoringanddiagnosisbyinterpretabledeeplearning
AT yingjieyang efficientandaccuratenoninvasivetumormonitoringanddiagnosisbyinterpretabledeeplearning
AT taojiang efficientandaccuratenoninvasivetumormonitoringanddiagnosisbyinterpretabledeeplearning
AT jiatang efficientandaccuratenoninvasivetumormonitoringanddiagnosisbyinterpretabledeeplearning
AT xinli efficientandaccuratenoninvasivetumormonitoringanddiagnosisbyinterpretabledeeplearning