Quantitative and Functional Assessment of Red and Yellow Bone Marrow Using PET-MR Imaging
Chuan Huang1,2, Anuradha Janardhanan1,3, and Mark Schweitzer1

1Radiology, Stony Brook Medicine, Stony Brook, NY, United States, 2Psychiatry, Stony Brook Medicine, Stony Brook, NY, United States, 3Diagnostic Imaging, Kuala Lumpur Hospital, Kuala Lumpur, Malaysia

Synopsis

Understanding the distribution of red marrow is important for various hematopoetic diseases and especially osseous metastases as areas of red marrow are the primary sites for hematogenous seeding of tumor cells, accounting for approximately 90% of skeletal metastases. Using a simultaneous PET-MR we sought to evaluate voxel of red marrow in the femora and pelvis using fat/water sequences correlated with FDG PET uptake. This quantitative assessment of red and yellow marrow was done in specific anatomic subregions. The bone marrow composition and metabolism were found to be symmetric in each individual. Good correlation between SUV and %red were found for each ROI among all subjects. The metabolism (FDG uptake) was found to be different for the ROIs with the same amount of red marrow. Further research will study whether this leads to higher chance of tumor seeding.

Purpose

Understanding the distribution of red marrow is important for various hematopoetic diseases and especially osseous metastases as areas of red marrow are the primary sites for hematogenous seeding of tumour cells, accounting for approximately 90% of skeletal metastases [1]. Using a simultaneous PET-MR we sought to evaluate voxel of red marrow in the femora and pelvis using fat/water sequences correlated with FDG PET uptake. This quantitative assessment of red and yellow marrow was done in specific anatomic subregions. We looked for symmetry of both MR and PET findings. We postulate that, in the future, that baseline red marrow distribution will help to identify areas that might be predisposed to metastases, especially if they have higher metabolic activity.

Methods

In this study, 9 volunteers with no known cancer or marrow disorder, getting FDG PET/MR scans for other research studies, were enrolled after obtaining written consent as approved by IRB. Ten-minute simultaneous PET-MR acquisitions of the pelvis were acquired approximately 1 hour after the FDG injection (~5mCi dose). For accurate fat quantification, a 6-echo 3D GRE DIXON sequence was used along with a graph-cut based fat quantification algorithm [2]. Fourteen 1cm2 representative region-of-interest (ROIs) were drawn manually on the corresponding PET and MR images using Osirix software in (bilateral, i.e. left and right sides) sacral ala, iliac bone, acetabulum, femoral head, femoral neck, lesser trochanter, and proximal femoral shaft by a radiologist.

Results and Discussion

We found the percentages of red marrow (%red) and FDG standardized uptake value (SUV) are fairly symmetric for each ROI pairs (left vs. right) within individual subjects (correlation coefficient R=0.96 for %red and R=0.95 for FDG SUV). There is also a clear linear correlation between the percentage of red marrow and its FDG uptake (R=0.71). Of all the areas that were examined, we found the highest ratio between FDG uptake to %red to be in the sacrum, iliac bones in the pelvic ring and lumbar spine. The neck and lesser trochanter of the femur were found to have the least amount of FDG uptake to %red. This agrees with existing SPECT literature [3]. With this, we can postulate that these high ratio areas would naturally indicate areas of potential tumor seeding.

Conclusion

The bone marrow composition and metabolism were found to be symmetric in each individual. Good correlation between SUV and %red were found for each ROI among all subjects. The metabolism (FDG uptake) was found to be different for the ROIs with the same amount of red marrow. Further research will study whether this leads to higher chance of tumor seeding.

Acknowledgements

No acknowledgement found.

References

[1] Basu S, Houseni M, Bural G, Chamroonat W, Udupa J, Mishra S, Alavi A: Mol Imaging Biol (2007) 9:361-365

[2] Hernando, D., Kellman, P., Haldar, J. P., & Liang, Z. P. (2010). Robust water/fat separation in the presence of large field inhomogeneities using a graph cut algorithm. Magnetic Resonance in Medicine, 63(1), 79-90.

[3] Sixpatients, H. (1991). In Vivo SPECT Quantitation of Bone Metabolism in Hyperparathyroidismand. JOURNAL OF NUCLEAR MEDICINE 1991;32:1157-1116

Figures

Figure 1. ROIs used in this study. CT Volume rendered in 3D Slicer.

Figure 2. The symmetry between the left and right ROIs. The correlation coefficients are R=0.96 (left) and R=0.95 (right).

Figure 3. The linear correlation between FDG PET SUV and %red. The correlation coefficient is R=0.71.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
0681