Magnetic Resonance Histology for surgical cancer pathology: fast, non-destructive 3D imaging

Background: Pathology is a cornerstone of clinical oncology, and pathology reports exert substantial influence on clinical decision-making. Our growing understanding of tumor heterogeneity and microenvironments has highlighted the importance of how we sample tumors. While single-cell profiling is now possible, the logistics of the pathology lab have not changed. Staff must inspect, prepare, and sample a high volume of tissues daily. Unfortunately, comprehensive tumor profiling is persistently pitted against short turnaround times and workflow challenges. 3D imaging is proven to improve tissue analysis and interpretation, but many methods for 3D tissue microscopy are time consuming and require chemical alteration. To bridge the gap between traditional and comprehensive tissue sampling, we have developed high-throughput, non-destructive magnetic resonance histology (MRH) for clinical cancer pathology. METHODSRobust MRH in the clinic necessitates innovative engineering. We implemented a gradient system providing spatial encoding gradients up to 2,500 mT/m (25X stronger than clinical MRI) and built sensitive radiofrequency coils to capture the weak signal in tiny voxels. We wrote software for compressed sensing to accelerate acquisition, deceasing imaging time from days to hours. Finally, we built a high-performance computational environment to post-process and display massive image arrays remotely on the pathologist's computer. We routinely image clinical specimens of a variety of tumor types, and MRH datasets are registered to digitized pathology for direct comparison with pathologist markup and digital pathology analytics. RESULTSWe have successfully imaged fixed clinical specimens from surgical pathology at high resolutions. In small tissues, we can achieve a 3D tissue image at 40mm3 in ~1 hr. In large tissues like radical prostatectomy (up to 6cm3) we acquire whole-organ isotropic images at 100mm3 in ~1 hr 15 min. MRH demonstrates excellent tumor contrast in clinical specimens of prostate, head and neck, breast cancer, soft tissue, and multiple pediatric cancers. Imaging happens after standard fixation, so no obstruction to tissue fixation was seen. Subsequent workup showed no difference in tissue quality or standard stain intensity. Registration of MRH data to digitized pathology confirmed a variety of distinct anatomical and tumor features visible on MRH based on pathologist markups. CONCLUSIONS MRH is a fast, non-destructive method for 3D imaging of histological specimens that can be integrated into the clinical domain. Tumor specimen MRH does not interfere with standard protocols for surgical pathology workup, and can provide excellent tumor contrast over large volumes. While the full extent of novel information detectable in tumor MRH has yet to be uncovered, we have demonstrated the potential of MRH as a disruptive new tool in surgical cancer pathology.

Learning Objectives 

1. Describe how magnetic resonance histology uses MRI principles for pathology applications.

2. Identify a variety of applications for novel magnetic resonance histology imaging in clinical cancer pathology.

3. Understand how magnetic resonance histology fits into clinical pathology workflows without inhibiting standard care practices.