December 7, 2022
Keio University Faculty of Pharmacy
Tokyo Medical University
National Cancer Center
National Center for Geriatrics and Gerontology
A joint research group—including the National Cancer Center, the National Center for Geriatrics and Gerontology, Toray Industries, Inc., and Preferred Networks, Inc.—led by Associate Professor Juntaro Matsuzaki of the Keio University Faculty of Pharmacy and Professor Takahiro Ochiya of the Tokyo Medical University Institute of Medical Science, was the first in the world to demonstrate that different types of cancer can be distinguished with high accuracy using data from a comprehensive analysis of microRNAs in the serum of cancer patients.
This research was conducted at the National Cancer Center as the "Project for the Development of Foundational Technology for Measuring microRNAs in Body Fluids" (Project Leader: Professor Takahiro Ochiya), with support from the Japan Agency for Medical Research and Development (AMED) under its Project for the Development of Core Technologies for Drug Discovery to Realize Next-Generation Treatment and Diagnosis. Utilizing the National Cancer Center Biobank and the National Center for Geriatrics and Gerontology Biobank, the team simultaneously analyzed the serum microRNA profiles of 9,921 cases of 13 types of solid cancer, including pancreatic and ovarian cancer, along with 5,643 non-cancer controls and 626 cases of various benign diseases.
A machine learning model was trained on microRNA data from four-fifths of the total samples. When the model was used to predict cancer types on the remaining one-fifth of the data, it achieved a high diagnostic prediction accuracy of 0.88 (95% confidence interval: 0.87–0.90) across all stages. The performance was even higher when limited to stages 0–II, where early diagnosis is particularly significant, with an accuracy of 0.90 (95% confidence interval: 0.88–0.91). This performance varies significantly depending on the machine learning algorithm, and the results also demonstrated the superiority of the algorithm developed in this study.
In recent years, expectations have been growing for Multi-cancer Early Detection (MCED) tests as a new cancer screening strategy, and the results of this study indicate that blood microRNA testing is a promising approach. To coincide with this, all the microRNA data obtained in this study and the machine learning code used for the analysis have been made public. It is hoped that these resources will be utilized to further invigorate this research area.
The results of this research were published in the online edition of "JNCI Cancer Spectrum," an academic journal of the US National Cancer Institute (NCI), on November 25, 2022 (US Eastern Time).
Please see below for the full press release.