In the field of hematopoietic stem cell transplantation (HSCT), early identification and precise intervention of graft-versus-host disease (GVHD) remain among the key challenges in clinical management. At the 2025 European Hematology Association (EHA) Annual Congress, a study jointly led by Professors Erlie Jiang, Junren Chen, and Yigeng Cao from the Institute of Hematology & Blood Diseases Hospital, Chinese Academy of Medical Sciences, was selected as an oral presentation (S261). The research focused on a novel autonomous artificial intelligence (AI) model—daGOAT, designed to precisely prevent acute GVHD (aGVHD). The model, in combination with low-dose prophylactic therapy, significantly reduced the incidence of severe aGVHD post-transplant, achieving an intelligent closed-loop system from risk prediction to clinical intervention. Oncology Frontier – Hematology Frontier conducted an exclusive interview with Professor Yigeng Cao to explore this new AI-powered advancement in hematopoietic transplantation.Oncology Frontier – Hematology Frontier: Congratulations on having your study selected for an oral presentation at the congress. The daGOAT model is a key highlight of your research. Could you introduce its design concept? How does it differ from traditional risk assessment tools?
Professor Yigeng Cao: Thank you. I’m delighted to share our research on the Oncology Frontier – Hematology Frontier platform, and we are honored to have this study selected for an oral presentation at EHA.
The daGOAT model has two major advantages over existing scoring systems and other AI models:
First, it enables real-time, dynamic data processing and analysis. You can think of it as an “intelligent scanner” that continuously collects and analyzes large volumes of clinical data following HSCT. The richness and timeliness of this information require AI to effectively integrate the data and identify patterns.
Second, it not only predicts risk but also stratifies it and actively triggers clinical decision-making. Between days 17 and 23 after transplantation, the model generates predictive results to assess whether the patient is likely to develop acute GVHD (aGVHD). Moreover, it provides graded risk stratification (e.g., moderate or severe). Based on this prediction, the model initiates direct action—if the risk is deemed high, the system automatically alerts clinicians to begin preventive interventions, such as prophylactic treatment. In this way, daGOAT functions as a truly autonomous AI system.
Oncology Frontier – Hematology Frontier: Under the guidance of the daGOAT model, the incidence of severe aGVHD has dropped to 5%, without an increase in drug-related adverse events. How do you interpret the clinical value of this “AI-based precision identification + low-dose intervention” approach? Do you foresee this model becoming a standard part of haploidentical (haplo) transplantation strategies in the future?
Professor Yigeng Cao: In this phase II clinical study, we enrolled a total of 110 patients undergoing hematopoietic stem cell transplantation. Risk prediction was performed using the daGOAT model, and combined with low-dose prophylactic treatment, the incidence of severe aGVHD in the study cohort was significantly reduced to just 5%, compared to 16% in the historical control group. This result suggests that the daGOAT model, when integrated with targeted intervention strategies, can effectively lower the risk of severe aGVHD, highlighting its clear clinical value.
More importantly, the model has already been successfully embedded into our clinical information system and is now being applied in real-world settings. Although the current study has concluded, we continue to use the daGOAT model to provide real-time risk assessments and decision-making support for every transplant patient. Next, we plan to initiate multicenter clinical studies to further validate the model’s applicability and generalizability, with the goal of expanding its use to more medical institutions and benefiting a broader patient population.
At this point, I’d also like to briefly respond to a common concern: “Will AI replace doctors?” As clinicians, of course we don’t wish to be replaced by AI. However, the reason we developed and implemented the daGOAT model is precisely because early identification and risk prediction of severe GVHD have long been major challenges. Although preventive strategies like PTCy and ATG are widely used, we still lack effective tools to determine which patients will develop severe GVHD. The emergence of the daGOAT model has significantly addressed this gap. It has gained strong recognition from clinicians and received positive feedback and support from patients.
Oncology Frontier – Hematology Frontier: This study marks a key step in applying autonomous AI to hematopoietic stem cell transplantation. Looking ahead, does your team plan to conduct multicenter validation studies or expand AI applications to predict and manage other transplant complications such as infections or immune dysregulation?
Professor Yigeng Cao: This phase II clinical study primarily aimed to evaluate the safety and efficacy of the daGOAT model. Based on the current data, the results are encouraging and demonstrate strong application potential. Our next step, following the formal publication of the study, is to promote the model’s broader application across institutions. We sincerely look forward to more experts and peers joining us to advance the real-world implementation and clinical translation of this autonomous AI tool.
We believe that the daGOAT model can significantly reduce the workload of clinicians in transplant management, while also improving patient outcomes and quality of life—delivering real value to clinical practice.
In addition, regarding the prediction and intervention of infections, chronic GVHD, and disease relapse—as you mentioned—these are indeed critical concerns. As is well known, GVHD, infection, and relapse are considered the “three major hurdles” after hematopoietic stem cell transplantation, and they remain key challenges we are actively working to overcome. That said, AI model development and validation require time and systematic support. We hope to carry out further research in these areas in the future and to continue sharing our progress and insights with our colleagues.
Expert Profile
Professor Erlie Jiang
- Director, Stem Cell Transplantation Center, Institute of Hematology & Blood Diseases Hospital, Chinese Academy of Medical Sciences
- MD, Chief Physician, Doctoral Supervisor and Postdoctoral Mentor
- Chair, Hematopoietic Stem Cell Transplantation Committee, China Primary Health Care Foundation
- Chair, Leukemia Committee, Chinese Society of Clinical Oncology (CSCO)
- Leader, Leukemia Autologous Transplantation Collaborative Group, China Hematology Specialty Alliance
- Deputy Leader, Hematopoietic Stem Cell Application Group, Hematology Branch, Chinese Medical Association
- Executive Member, Hematologic Oncology Committee, Chinese Anti-Cancer Association
- Deputy Leader, Hematopoietic Stem Cell Transplantation and Cellular Therapy Group
- Deputy Leader, Autologous Hematopoietic Stem Cell Transplantation Working Group, CSCO
- Executive Director, Tianjin Anti-Cancer Association
- Vice President, Tianjin Society of Hematology and Regenerative Medicine
- Editorial Board Member, Chinese Journal of Hematology, Leukemia & Lymphoma, and other journals
Institute of Hematology & Blood Diseases Hospital, Chinese Academy of Medical Sciences
- MD, PhD, Researcher, Doctoral Supervisor, and Distinguished Professor at Peking Union Medical College
- Chief Technology Officer, Information and Resource Center, Institute of Hematology & Blood Diseases Hospital
- Received his M.D. from National Taiwan University (1996–2002)
- Ph.D. in Molecular and Cellular Biology from Harvard University (2004–2008)
- MBA from Washington University in St. Louis (2011–2014)
- Worked in a Fortune 500 biotechnology company in the U.S. (2010.7–2016.6)
- Worked at China Seed Group and Sinochem Group (2016.10–2020.6)
- Joined the Institute of Hematology & Blood Diseases Hospital in July 2020
- Selected for Wuhan’s “Optics Valley 3551 Talent Plan” by the Wuhan Development and Reform Commission (2017)
- Selected for the Central Organization Department’s “National Long-Term Innovation Talent Program” (2018)
His current research focuses on applying data science to hematologic disease research by mining real-world data to address clinical questions. His work spans panoramic medical data research, covering diagnosis, treatment, and prognosis across various hematologic disease subtypes. Through data-driven approaches and interdisciplinary collaboration, he aims to solve shared scientific challenges, enabling breakthroughs in hematology and integrating data science with clinical medicine. In the past three years, he has published 13 papers as corresponding author in high-impact journals such as Nature Computational Science, American Journal of Hematology, Leukemia, and Blood.
Stem Cell Transplantation Center, Institute of Hematology & Blood Diseases Hospital, Chinese Academy of Medical Sciences
- Communist Party Member; MD; Visiting Scholar at Harvard University
- Secretary to the Center Director; Associate Chief Physician
- Master’s Supervisor, Peking Union Medical College
- Member of the 23rd “Doctoral Service Corps” to Qinghai, organized by the Organization Department of the CPC Central Committee and the Communist Youth League
- Performed the first allogeneic hematopoietic stem cell transplantation on the Qinghai-Tibet Plateau
- “Leading Talent” under Qinghai Province’s Kunlun Talents High-End Innovation Program
- Master’s Supervisor and Distinguished Expert at Qinghai University
Academic and Professional Roles
- Secretary, Hematopoietic Stem Cell Application Group, Chinese Society of Hematology
- Member, Transplantation and Immunotherapy Group, Chinese Society of Hematology Physiology
- Member, Integrative Rehabilitation Committee, Chinese Anti-Cancer Association
- Clinical Medicine Consultant, Highland Medicine Expert Advisory Committee
- Member, Tianjin Society of Hematology and Regenerative Medicine
- Member, Tianjin Health Education Association and its Hematology Branch
Research and Publications
- First author of over 10 SCI papers in journals including American Journal of Hematology, Nature Computational Science, and Biology of Blood and Marrow Transplantation
- Principal Investigator of 7 national and provincial-level projects, including the National Natural Science Foundation of China and Tianjin Health Science & Technology Projects
- Key participant in 8 additional national and institutional research projects
Reviewer For
- Chinese Journal of Hematology
- Hematological Oncology
- Frontiers in Immunology
