Impact of brain segmentation methods on regional metabolism quantification in 18F-FDG PET/MR analysis

Abstract

Background: Accurate analysis of quantitative PET data plays a crucial role in studying small, specific brain structures. The integration of PET and MRI through an integrated PET/MR system presents an opportunity to leverage the benefits of precisely aligned structural MRI and molecular PET images in both spatial and temporal dimensions. However, in many clinical workflows, PET studies are often performed without the aid of individually matched structural MRI scans, primarily for the sake of convenience in the data collection and brain segmentation possesses. Currently, two commonly employed segmentation strategies for brain PET analysis are distinguished: methods with or without MRI registration and methods employing either atlas-based or individual-based algorithms. Moreover, the development of artificial intelligence (AI)-assisted methods for predicting brain segmentation holds promise but requires further validation of their efficiency and accuracy for clinical applications. This study aims to compare and evaluate the correlations, consistencies, and differences among the above-mentioned brain segmentation strategies in quantification of brain metabolism in ¹⁸F-FDG PET/MR analysis.

Results: Strong correlations were observed among all methods (r = 0.932 to 0.999, P < 0.001). The variances attributable to subject and brain region were higher than those caused by segmentation methods (P < 0.001). However, intraclass correlation coefficient (ICC)s between methods with or without MRI registration ranged from 0.924 to 0.975, while ICCs between methods with atlas- or individual-based algorithms ranged from 0.741 to 0.879. Brain regions exhibiting significant standardized uptake values (SUV) differences due to segmentation methods were the basal ganglia nuclei (maximum to 11.50 ± 4.67%), and various cerebral cortexes in temporal and occipital regions (maximum to 18.03 ± 5.52%). The AI-based method demonstrated high correlation (r = 0.998 and 0.999, P < 0.001) and ICC (0.998 and 0.997) with FreeSurfer, substantially reducing the time from 8.13 h to 57 s on per subject.

Conclusions: Different segmentation methods may have impact on the calculation of brain metabolism in basal ganglia nuclei and specific cerebral cortexes. The AI-based approach offers improved efficiency and is recommended for its enhanced performance.

Keywords: Artificial intelligence; Magnetic resonance imaging; Metabolism; Positron emission tomography.

Fig. 1

Block diagram of atlas-based (A) and individual-based (B) segmentation

Fig. 2

The architecture of the VB-Net method for brain ROI segmentation. VB-Net consists of two input channels, which are the original MRI images, and the segmentation maps of gray matter, white matter, and cerebrospinal fluid obtained from the VBM analysis of SPM. VB-Net consists of four levels with an encoding path followed by four levels of the corresponding decoding path. On the left side of the network, the encoding path reduces the size of the input by downsampling. The network was divided into four blocks, which comprise several convolutional blocks and residual blocks. A skip connection was introduced to improve the segmentations, and bottleneck layers were introduced to decrease the memory consumption. Similarly, on the right side of the network, the decoding path recovered the semantic segmentation image by deconvolution

Fig. 3

The segmentation results of different methods overlaid on the PET images of one subject. A The results at the whole-brain level, and B depicts the results at the cerebral cortex level. Different methods produced brain segmentations with similar appearances. The methods based on SPM and FSL divided the cortical brain regions into block-shaped areas that contain cerebrospinal fluid; in contrast, the method based on FreeSurfer and Neural Network divided the cortical brain regions into surface-shaped areas that do not contain cerebrospinal fluid

Fig. 4

The Pearson correlation coefficients of SUVs calculated by eight different segmentation methods in the 21 brain regions at the whole-brain level (A) and 48 cortical regions at the cerebral cortex level (B)