European Radiology； How can imaging be used to predict the ATRX mutation status of low-grade gliomas?

Glioma is the most common type of primary brain tumor and originates from neoplastic glial cells or glial cells
.
According to the World Health Organization's (WHO) classification of central nervous system (CNS) tumors, gliomas are histopathologically classified into grades
1, 2, 3, and 4.
Patients with low-grade gliomas (LGG, CNS grades 1 and 2) have lower mortality and better prognosis
compared with patients with high-grade gliomas (HGGs, CNS WHO grades 3 and 4).
However, low-grade gliomas sometimes undergo malignant transformation, which greatly shortens the patient's survival time
.
Therefore, early diagnosis and timely intervention are very critical and important
.

Building on the 2016 WHO Central Nervous System Classification, the new 2021 WHO Central Nervous System Classification introduces significant changes that advance the role of
molecular diagnostics in the classification of central nervous system tumors.
The mutation in ATRX was first identified
in patients with α thalassemia X-linked intellectual disability syndrome.
ATRX mutations may lead to telomerase inactivation of glioma cells, thereby affecting the biological behavior of astrocytes tumor cells and inducing the malignant progression
of LGG.
Unfortunately, the status of ATRX mutations, which is primarily detected by surgery or biopsy, is invasive
.
Therefore, non-invasive prediction of ATRX mutation status is important for providing diagnostic and prognostic information before surgery
.

Recently, a study published in the journal European Radiology established a comprehensive clinical radiomics model based on [18F]FDG PET and multimodal MRI to achieve noninvasive preoperative assessment
of ATRX mutation status of IDH-mutant LGG.

This study included 122 patients (47 ATRX mutants, 55 ATRX wild) diagnosed with IDH mutant LGG (CNS WHO Grade 1 and 2).

From structural MR (sMR) images (enhanced T1-weighted imaging, CE-T1WI; T2-weighted imaging, T2WI), functional MR (fMR) images (apparent diffusion coefficient, ADC; A total of 5540 radiomics features
were extracted from cerebral blood volume, CBV) and metabolic PET images ([18F]FDG PET).
A comprehensive clinical radiomics model was established using the random forest algorithm, combining the best multimodal radiomics model with three clinical parameters
.
The predictive effect
of the model was evaluated by receiver operation characteristics (ROC) and decision curve analysis (DCA).

The best multimodal model combines sMR (CE-T1WI), fMR (ADC), and metabolic ([18F]FDG) images ([18F]FDG PET+ADC+CE-T1WI), with areas under the curve (AUCs) of 0.
971 and 0.
962
in the training and test groups, respectively 。 The comprehensive clinical radiology model integrating [18F]FDG PET+ADC+CE-T1WI and three clinical parameters (KPS, SFSD and ATGR) showed the best predictive effect in the training and test groups (0.
987 and 0.
975, respectively).

Figure a, c LASSO.
b [18F]Characteristic coefficients in the FDG PET+CE-T1WI+ADC radiology model.
d, e model [18F]FDG PET+CE-T1WI+ADC ROC curves in the training and test groups

In this study, a multicenter clinical radiology comprehensive model based on [18F]FDG PET and multimodal MRI was established, which can simultaneously provide structural, functional, and metabolic information and can be applied to predict the ATRX mutation status
of IDH-mutated LGG.

Original source:

Liqiang Zhang,Hongyu Pan,Zhi Liu,et al.
Multicenter clinical radiomics-integrated model based on [ 18 F]FDG PET and multi-modal MRI predict ATRX mutation status in IDH-mutant lower-grade gliomas.
DOI:10.
1007/s00330-022-09043-4