A research team led by Prof. WANG Jiguang from the Division of Life Science and Department of Chemical and Biological Engineering at the Hong Kong University of Science and Technology (HKUST) has discovered a previously overlooked subtype of brain tumor termed IME IDH-mutant astrocytoma. Contrary to typical medical expectations, this subtype exhibits strong immune activity, classified as "immune-hot", yet tends to have poorer survival rates, and may respond differently to treatments compared to other types. In addition, the team has developed an AI framework to assist doctors in accurately identifying this subtype, enabling more personalized patient care.
Diffuse gliomas are the most common malignant brain tumors in adults, with IDH-mutant astrocytoma being a major subtype. However, these tumors can vary significantly between individuals. Even when receiving the same standard treatment, patients may respond differently, and many tumors recur quickly. A key reason for this is inter-tumor heterogeneity: each tumor has a unique composition of cancer cells and surrounding support cells (the tumor's microenvironment), which influences how each tumor behaves.
To improve patient outcomes, researchers need to understand the differences at a deep molecular level using large, multidimensional datasets. Prof. Wang's team achieved this by analyzing "multi-omics" data—spanning genes, messenger RNAs, proteins, and more—collected from different parts of the tumor and at different points in time. Their study, published online in Cancer Cell * in September, uncovered an "immune‑hot" subtype associated with poorer treatment outcomes, paving the way for more personalized care. By integrating multiple advanced techniques that measure protein abundance, mRNA expression and DNA methylation at both bulk and single-cell levels, the team identified four protein-based subtypes of IDH-mutant astrocytoma:
- AFM (Adipogenesis Fatty acid Metabolism): linked to fat acid and energy metabolism
- PPR (Proliferative Progenitor): characterized by rapid growth and proliferative features
- IME (Immune Mesenchymal-Enriched): immune-hot and mesenchymal cell signals
- NEU (Neuronal): features similar to neuronal cells
Approximately 13% of cases fell into the IME group. These tumors showed a distinctive cell appearance known as gemistocytic differentiation, along with heavy infiltration by immune cells, particularly by worn-out "exhausted" T cells and plasma cells. They also displayed strong interferon signaling, indicating an active immune response. Despite these immune characteristics, patients with IME tumors had poorer outcomes, with survival rates comparable to the fast-growing PPR tumors.
Laboratory studies revealed two factors contributing to this unfavorable outcome: an influx of plasma cells and activation of a gene called GBP1, which together appear to accelerate tumor growth and enable cancer cells to evade the immune system. Furthermore, through longitudinal patient observations, the team discovered a trend where many IDH-mutant astrocytomas tend to shift toward the IME or PPR subtypes when they come back, highlighting their aggressive nature.
"Our study sheds light on the hidden complexity of IDH-mutant astrocytoma,"said Prof. Wang."By combining AI with high-dimensional omics data, we gain a brand-new perspective for stratifying and managing patients. This not only uncovers new biological insights but also offers novel targets and strategies for future personalized treatments."To translate these discoveries into clinical tools, the team developed GUIDE (Generic Utility for IME Diagnostic Estimation), an AI platform designed to assist doctors in tumor classification. GUIDE combines microscope images of tumor tissue with genomic and proteomic data, providing rapid and accurate analysis even in cases with missing data. Validation studies on independent patient cohorts confirmed that GUIDE effectively identifies IME tumors, which are associated with poorer survival outcomes.
This research introduces a novel method for categorizing gliomas, challenges the conventional belief that"immune-hot"tumors indicate a positive prognosis, and provides a practical tool to enhance precision healthcare. Furthermore, it suggests innovative treatment approaches, such as targeted drugs and immunotherapy tailored to this aggressive tumor group.
*The study was led by Prof. WANG Jiguang and his HKUST PhD student TANG Jihong, with major contributions from Prof. JIANG Tao's team at Beijing Tiantan Hospital and Beijing Neurosurgical Institute. Collaborators included Prof. MAO Ying and Prof. HUA Wei (Shanghai Huashan Hospital), Prof. WANG Qianghu (Nanjing Medical University), and Prof. LIU Kai (HKUST). Co-first authors are HKUST PhD students TANG Jihong and RUAN Yuyan, and Dr. FAN Wenhua and Dr. LIU Xing from Beijing Tiantan Hospital. Prof. Wang and Prof. Jiang are the corresponding authors.
