Effect of lumbar bone marrow composition on bone marrow diffusion signal intensity and apparent diffusion coefficient in healthy adult women / 重庆医学

Objective To analyze the effect of lumbar bone marrow composition on bone marrow diffu-sion-weighted imaging(DWI)in healthy adult women.Methods Retrospective analysis was performed on up-per abdominal MRI of 103 adult women.Bone marrow fat fraction of lumbar vertebra was measured according to two-point water-lipid separation technique,and apparent diffusion coefficient(ADC)value of lumbar verte-bra was measured according to DWI image(b=800 s/mm2).The subjects were divided into the high-signal group and the equal-low-signal group according to the signal intensity of lumbar vertebra and adjacent erector spine muscles.The effects of age,lumbar bone marrow fat fraction and menstrual status on the signal intensity and ADC value of lumbar bone marrow diffusion were analyzed.Finally,the correlation between lumbar bone marrow fat fraction and ADC value was analyzed.Results Univariate analysis showed that the lumbar bone marrow diffusion signal intensity and ADC value were affected by age,lumbar bone marrow fat fraction and menstrual status(P＜0.001).Multivariate analysis showed that age(P=0.046)and lumbar bone marrow fat fraction(P=0.005)were the influencing factors of lumbar bone marrow diffusion signal intensity,but men-strual status(P=0.242)was not the influencing factor.In addition,lumbar bone marrow fat fraction(P＜0.001)was the factor influencing the ADC value of lumbar bone marrow,and the two were negatively correla-ted(r=-0.607,P＜0.001),but age(P=0.497)and menstrual status(P=0.082)were not the influencing factors.Conclusion The bone marrow composition of lumbar vertebrae in healthy adult women has significant effects on the signal intensity and ADC value of bone marrow diffusion.

Differences in specific abdominal fat depots between metabolically healthy and unhealthy children with overweight/obesity: The role of cardiorespiratory fitness.

OBJECTIVES: Fat depots localization has a critical role in the metabolic health status of adults. Nevertheless, whether that is also the case in children remains under-studied. Therefore, the aims of this study were: (i) to examine the differences between metabolically healthy (MHO) and unhealthy (MUO) overweight/obesity phenotypes on specific abdominal fat depots, and (ii) to further explore whether cardiorespiratory fitness plays a major role in the differences between metabolic phenotypes among children with overweight/obesity. METHODS: A total of 114 children with overweight/obesity (10.6 ± 1.1 years, 62 girls) were included. Children were classified as MHO (n = 68) or MUO. visceral (VAT), abdominal subcutaneous (ASAT), intermuscular abdominal (IMAAT), psoas, hepatic, pancreatic, and lumbar bone marrow adipose tissues were measured by magnetic resonance imaging. Cardiorespiratory fitness was assessed using the 20 m shuttle run test. RESULTS: MHO children had lower VAT and ASAT contents and psoas fat fraction compared to MUO children (difference = 12.4%-25.8%, all p < 0.035). MUO-unfit had more VAT and ASAT content than those MUO-fit and MHO-fit (difference = 34.8%-45.3%, all p < 0.044). MUO-unfit shows also greater IMAAT fat fraction than those MUO-fit and MHO-fit peers (difference = 16.4%-13.9% respectively, all p ≤ 0.001). In addition, MHO-unfit presented higher IMAAT fat fraction than MHO-fit (difference = 13.4%, p < 0.001). MUO-unfit presented higher psoas fat fraction than MHO-fit (difference = 29.1%, p = 0.008). CONCLUSIONS: VAT together with ASAT and psoas fat fraction, were lower in MHO than in MUO children. Further, we also observed that being fit, regardless of metabolic phenotype, has a protective role over the specific abdominal fat depots among children with overweight/obesity.

Fully automated segmentation of lumbar bone marrow in sagittal, high-resolution T1-weighted magnetic resonance images using 2D U-NET.

BACKGROUND: We investigated a 2-dimensional (2D) U-Net model to delineate lumbar bone marrow (BM) using a high resolution T1-weighted magnetic resonance imaging. METHOD: Healthy controls (n = 44, 836 images) and patients with hematologic diseases (n = 56, 1064 images) received MRI of the lumbar spines. Lumbar BM on each image was manually delineated by an experienced radiologist as a ground-truth. The 2D U-Net models were trained using a healthy lumbar BM only, diseased BM only, and using healthy and diseased BM combined, respectively. The models were validated using healthy and diseased subjects, separately. A repeated-measures analysis of variance was performed to compare segmentation accuracies with 2 validation cohorts among U-Net trained with healthy subjects (UNET_HC), U-Net trained with diseased subjects (UNET_HD), U-Net trained with all subjects including both healthy and diseased subjects (UNET_HCHD), and 3-dimensional Grow-Cut algorithm (3DGC). RESULTS: When validated with the healthy subjects, UNET_HC, UNET_HD, UNET_HCHD and 3DGC achieved the mean and standard deviation of the Dice Similarity Coefficient (DSC) of 0.9415 ± 0.07056, 0.9583 ± 0.05146, 0.9602 ± 0.0486 and 0.9139 ± 0.2039, respectively. When validated with the diseased subjects, DSCs of UNET_HC, UNET_HD, UNET_HCHD and 3DGC were 0.8303 ± 0.1073, 0.9502 ± 0.0217, 0.9502 ± 0.0217 and 0.8886 ± 0.2179, respectively. The U-Net models segmented BM better than the semi-automatic 3DGC (P < 0.0001), and UNET_HD produced better results than UNET_HC (P < 0.0001). CONCLUSIONS: We successfully constructed a fully automatic lumbar BM segmentation model for a high-resolution T1-weighted MRI using U-Net, which outperformed most of the previously reported approaches and the existing semi-automatic algorithm.