Dan Wu's group published in Neuro-Oncology

        On January 7, 2023, Professor Dan Wu's team from the College of Biomedical Engineering and Instrument Science at Zhejiang University, in collaboration with Dr. Hongxi Zhang's team from the Department of Radiology at Zhejiang University Children's Hospital, published a paper titled Histological and Molecular Classifications of Pediatric Glioma with Time-dependent Diffusion MRI based Microstructural Mapping in the journal Neuro-Oncology. The team introduced a non-invasive microstructural mapping technique based on the theory of time-dependent diffusion MRI (-dMRI) to quantify tumor cell characteristics and demonstrated its initial efficacy in histological grading and H3K27 molecular subtyping of pediatric gliomas.

Central nervous system (CNS) tumors are the most common solid tumors in children and a leading cause of mortality. Among CNS tumors, gliomas are the most prevalent type, histologically classified into low-grade (LGG) and high-grade gliomas (HGG), with the latter having a long-term survival rate of less than 10%. Diffuse midline glioma (DMG) is a particularly devastating category within pediatric gliomas, typically located in critical areas such as the thalamus, brainstem, or spinal cord, making surgical intervention impractical. These DMGs have an extremely low survival period of only 9-11 months. Regardless of the histological characteristics, patients with DMG carrying H3K27 alterations exhibit significantly lower overall survival. Thus, H3K27 alteration serves as a crucial molecular biomarker for DMG in the 2021 World Health Organization (WHO) classification of CNS tumors. Accurate histological and molecular grading is essential for prognosis determination and treatment strategies in children with DMG.

-dMRI, based on diffusion time, exhibits unique advantages in mapping cellular microstructure. By utilizing specialized diffusion encoding schemes, it captures restricted diffusion across multiple diffusion times. The time-dependent diffusion can be used to construct biophysical models for quantifying microstructural features. -dMRI -based microstructural imaging is considered a step towards achieving "virtual pathology."

The team recruited 75 pediatric patients and acquired -dMRI images of gliomas using the house-made oscillating gradient (OGSE) and pulsed gradient spin echo (PGSE) sequences. Subsequently, the data was fitted to the IMPULSED model and correlated with pathological slices.

The team discovered that, for histological grading of HGG, cell density and intracellular fraction were higher than those of LGG, while cell diameter, diffusion coefficients, T1, and T2 values were lower, exhibiting significant differences andcell density showed the best classification performance. Regarding molecular subtyping, H3K27-altered tumors showed significantly lower T1, T2, apparent diffusion coefficient (ADC), extracellular diffusion coefficient, and cell diameter, as well as higher intracellular fraction and cell density compared to wild-type DMG. Among these features, cell diameter showed the best classification performance. Combining cell diameter and extracellular diffusion coefficient further improved the classification performance. Finally, by correlating model parameters with pathological images obtained from H&E-stained sections, the team demonstrated good consistency.

As emphasized by the WHO 2021 guidelines, combining histological and molecular information is crucial for accurate prognosis and optimal treatment selection in CNS tumors. The team designed a prospective study to explore the clinical application of newly proposed cellular microstructural markers based on -dMRI theory in histological and molecular identification of pediatric gliomas. Preliminary results indicated that the cell volume index exhibited favorable performance in histological grading, while the cell diameter index demonstrated high discriminatory power in distinguishing H3K27-altered and wild-type DMGs, underscoring the value of these non-invasive microstructural features in pediatric gliomas. Particularly, distinct microstructural characteristics were observed among different histological and molecular subtypes, highlighting the importance of utilizing pathology indices that specifically reflect cellular microstructure rather than relying solely on simple ADC measurements, which can be accurately provided by -dMRI technology.

Reference：

Hongxi Zhang, Kuiyuan Liu, Ruicheng Ba, Zelin Zhang, Yi Zhang, Ye Chen, Weizhong Gu, Zhipeng Shen, Qiang Shu, Junfen Fu, Dan Wu, Histological and molecular classifications of pediatric glioma with time-dependent diffusion MRI-based microstructural mapping, Neuro-Oncology, 2023;