DONGHENG MA

Dr. DONGHENG MA

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 July 2025 in Cancers
Verified
Journal Article
Open access
PDF file
HNHinano Nishikubo
Hinano Nishikubo
KKKyoka Kawabata
Kyoka Kawabata
SKSaki Kanei
Saki Kanei
RARika Aoyama
Rika Aoyama
DMDONGHENG MA
DONGHENG MA
MYMasakazu Yashiro
Masakazu Yashiro
Background/Objectives: The neurotrophic tropomyosin receptor kinase (NTRK) genes NTRK1, NTRK2, and NTRK3 encode tyrosine kinase receptors, and their fusion genes are known as the oncogenic driver genes for cancer. This study aimed to compare the diagnostic ability of NTRK fusion among five types of multi-cancer genome profiling tests (multi-CGP tests) and determine a useful multi-CGP test for NTRK fusion, recorded in the Center for Cancer Genomics and Advanced Therapeutics (C-CAT) database in Japan. This study aimed to compare the diagnostic results for NTRK fusions among the five different CGP tests. Methods: A total of 88,688 tumor cases were enrolled in the C-CAT profiling database from 2019 to 2024. The detection frequency of NTRK fusion genes was compared to the results for five multi-CGP tests: NCC Oncopanel, FoundationOne CDx (F1), FoundationOne Liquid (F1L), GenMineTOP (GMT), and Guardant360. Results: NTRK fusion genes were detected in 175 (0.20%) of the 88,688 total cases. GMT, which is equipped with RNA sequencing function, frequently detected NTRK fusion genes (20 of 2926 cases; 0.68%) in comparison with the other four multi-CGP tests that do not have RNA sequencing analysis. GMT showed significantly (p < 0.05) higher diagnostic ability for NTRK fusions compared with the other four multi-CGP tests. Especially, NTRK2 fusion was significantly (p < 0.001) more highly determined by GMT than it was by the other four multi-CGP tests. The detection rates for FGFR1 and FGFR3 were significantly higher in GMT than in other multi-CGP tests. In contrast, the detection rates of the ALK and RET fusion genes were significantly higher in F1L. Conclusions: GMT, which is equipped with RNA sequencing analysis, might show a useful diagnostic ability for NTRK fusions, especially for NTRK2 fusion genes.

Verified
Journal Article
Open access
PDF file
HGHongdong Gao
Hongdong Gao
HNHinano Nishikubo
Hinano Nishikubo
DMDONGHENG MA
DONGHENG MA
JPJuncheng Pan
Juncheng Pan
TSTomoya Sano
Tomoya Sano
MYMasakazu Yashiro
Masakazu Yashiro
Cancer-associated fibroblasts (CAFs) are a key constituent of the tumor microenvironment. CAFs may affect the development of tumor cells. The critical role of CAFs in the tumor microenvironment is linked to their epigenetic modifications, as a stable yet reversible regulation of cellular phenotypes. Current evidence indicates that their formation and function are closely linked to epigenetic mechanisms. Existing research indicates that the epigenetic alteration abnormalities are triggered by metabolic cues and stabilize the acquired phenotype of CAFs. This process is associated with transcriptional changes and patient outcomes in various tumors, providing a biological rationale and translational potential for reprogramming CAFs. Understanding of epigenetic modifications in CAFs remain insufficient, while DNA methylation in CAFs can alter CAF states through multiple pathways and thereby influence tumor progression. It is necessary to investigate the unique, identifiable epigenetic signatures of CAF. As an epigenetic reader couple histone acetylation to high-output oncogenic transcription; meanwhile, noncoding RNAs modulate CAF formation and therapeutic responses via bidirectional crosstalk between tumor cells and stroma. The interactions between different epigenetic modifications and their underlying regulatory logic may play a crucial role in developing new therapeutic strategies. This review focuses on the roles of DNA methylation, histone acetylation, and enhancer reprogramming in CAFs.

Verified
Journal Article
Open access
PDF file
HNHinano Nishikubo
Hinano Nishikubo
DMDONGHENG MA
DONGHENG MA
TSTomoya Sano
Tomoya Sano
DIDaiki Imanishi
Daiki Imanishi
TSTakashi Sakuma
Takashi Sakuma
MYMasakazu Yashiro
Masakazu Yashiro
Genetic abnormalities of the fibroblast growth factor receptor 2 (FGFR2) gene, including amplification, fusions, and mutations, have been reported in various solid tumors. While molecular targeted therapies against FGFR2 fusion have been proved to be useful in cholangiocarcinoma, the therapeutic significance of FGFR2 inhibitors remains unclear in other various solid cancers. Genomic and clinical information from solid tumor cancer gene panel testing cases is consolidated in the Center for Cancer Genomics and Advanced Therapeutics (C-CAT) database in Japan. This study aimed to utilize the C-CAT database to clarify the clinical–pathological significance of FGFR2 abnormalities. A total of 101,231 patients with solid cancer have been registered in the C-CAT database between June 2019 and June 2025. Of the 101,231 cases, 1312 cases with FGFR2 gene abnormalities were analyzed. FGFR2 alterations included amplification in 515 cases, fusion in 280 cases, and mutations in 568 cases. They were detected most frequently in the biliary tract (271 cases), esophagus/stomach (231 cases), and breast (211 cases). Amplification was frequent in the esophagus/stomach (205 cases) and breast (105 cases). Mutations were frequent in the uterus (111 cases), breast (89 cases), and biliary tract (86 cases). Among 515 FGFR2 alteration cases, FGFR2 inhibitors were administered in 85 cases. Of the 85 cases, disease control was achieved in 49 cases, 44 cases of which were biliary tract cancer. FGFR2 might be a promising therapeutic target not only for cholangiocarcinoma with fusion but also for esophagus/stomach cancer and breast cancer with FGFR2 alterations.

9 February 2026 in Metabolites
Verified
Journal Article
Open access
PDF file
DIDaiki Imanishi
Daiki Imanishi
HNHinano Nishikubo
Hinano Nishikubo
DMDONGHENG MA
DONGHENG MA
HGHongdong Gao
Hongdong Gao
TSTomoya Sano
Tomoya Sano
MYMasakazu Yashiro
Masakazu Yashiro
The tumor heterogeneity that is frequently observed in cancer tissues comprises not only cancer cells but also stromal cells in the tumor microenvironment. One of the major components of tumor stroma, i.e., cancer-associated fibroblasts (CAFs), play crucial roles in tumor progression and the tumor response to chemotherapy. The known subtypes of CAFs are antigen-presenting CAFs (apCAFs), myofibroblastic CAFs (myCAFs), and inflammatory CAFs (iCAFs). It has been speculated that (i) the heterogeneity of CAF subtypes might contribute to tumor progression; (ii) cell-to-cell interactions among CAF subtypes in tumors might be associated with the development of various types of carcinomas, and (iii) juxtracrine and/or paracrine signaling from CAFs may play important roles in this development. A clarification of the mechanisms that underlie the tumoral heterogeneity of CAFs could contribute to cancer treatment as precision medicine. This review explains the significance of CAF heterogeneity in tumor microenvironments, especially concerning the CAF subtypes.

28 January 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
DMDONGHENG MA
DONGHENG MA
HNHinano Nishikubo
Hinano Nishikubo
TSTomoya Sano
Tomoya Sano
MYMasakazu Yashiro
Masakazu Yashiro
Tumor mutational burden (TMB) is a key pan-cancer biomarker for immunotherapy selection, but its routine assessment by whole-exome sequencing (WES) or large next-generation sequencing (NGS) panels is costly, time-consuming, and constrained by tissue and DNA quality. In parallel, advances in computational pathology have enabled deep learning models to infer molecular biomarkers directly from hematoxylin and eosin (H&E) whole-slide images (WSIs), raising the prospect of a purely digital assay for TMB. In this comprehensive review, we surveyed PubMed and Scopus (2015–2025) to identify original studies that applied deep learning directly to H&E WSIs of human solid tumors for TMB estimation. Across the 17 eligible studies, deep learning models have been applied to predict TMB from H&E WSIs in a variety of tumors, achieving moderate to good discrimination for TMB-high versus TMB-low status. Multimodal architectures tended to outperform conventional CNN-based pipelines. However, heterogeneity in TMB cut-offs, small and imbalanced cohorts, limited external validation, and the black-box nature of these models limit clinical translation.

25 March 2026 in Genes
Verified
Journal Article
Open access
PDF file
DMDONGHENG MA
DONGHENG MA
HNHinano Nishikubo
Hinano Nishikubo
TSTomoya Sano
Tomoya Sano
MYMasakazu Yashiro
Masakazu Yashiro
Background: Predicting genomic alterations from routine hematoxylin and eosin (H&E) whole-slide images (WSIs) may help triage molecular testing. Methods: We retrospectively enrolled 437 patients at Osaka Metropolitan University Hospital across 26 cancers, matched with clinical gene-panel data. We curated 1023 binary endpoints across SNV, CNV, and SV categories. We extracted slide embeddings from five pathology foundation models (Prism, GigaPath, Feather, Chief, and Titan) using a unified feature extraction pipeline and benchmarked them using a lightweight downstream Multi-Layer Perceptron (MLP) classifier. Using the best-performing patch feature system, we trained a multi-instance learning model to assess incremental benefit. Results: Titan achieved the highest and most stable transfer performance, with a median endpoint-wise Area Under the Receiver Operating Characteristic curve (AUROC) of 0.77 in the slide benchmarking; at the patch-level, prediction of APC_SNV reached an AUROC of 0.916, and prediction of KRAS_SNV reached an AUROC of 0.811 on the held-out test set. Conclusions: In a heterogeneous clinical gene-panel setting, pathology foundation models can provide strong baseline genomic-prediction signals without additional fine-tuning. We propose a practical, deployment-oriented two-stage workflow: rapid slide-embedding screening to prioritize robust representations and candidate endpoints, followed by patch-level training for high-value tasks where additional performance gains and interpretable regions are clinically worthwhile.