Evaluating quality of routine vs. tape‐transfer histology sections for 3D reconstruction of paraffin embedded brain tissue and correlation with 7T MRI

William Hai Dang Ho, Yixin Wang, Hossein Moein Taghavi, Phillip DiGiacomo, Jeffrey Nirschl, Pauline Chu, Marios Georgiadis, Michael ZeinehAlzheimer 2024 · DOI

TL;DR

This paper introduces a critical step in bridging the gap between high-resolution MRI and detailed histology for Alzheimer's research: optimizing tissue sectioning. They found that a "tape transfer" method significantly reduces tissue expansion and improves yield compared to traditional water bath techniques, making it more suitable for accurate 3D reconstruction and correlation with MRI data.

What Problem Does This Paper Address?

Accurately correlating microscopic histological features with macroscopic MRI scans is essential for developing and validating MRI biomarkers for neurodegenerative diseases like Alzheimer's. However, a major hurdle is the deformation that occurs during traditional histological sample preparation, particularly sectioning. These deformations distort the tissue, making it difficult to precisely align histological slices with their corresponding MRI data. This paper tackles the problem of optimizing the tissue sectioning and mounting process to minimize these deformations, thereby improving the accuracy of 3D reconstruction and subsequent correlation with high-resolution MRI.

Methods

The researchers compared two tissue mounting techniques for 7 µm thick human brain sections:

  • Traditional Water Bath: This standard method involves floating the tissue section on water before transferring it to a charged slide.
  • Tape Transfer: A newer technique where the section is captured directly onto an adhesive-coated slide using a tape-like material.

Both methods were applied to paraffin-embedded human brain tissue that had previously undergone 7T MRI scanning. Before sectioning, a "blockface image" of the tissue block was captured. After sectioning and staining (Hematoxylin and Eosin), the resulting tissue sections were imaged. The researchers then quantitatively analyzed:

  • Tissue Expansion: Measured by comparing the area of the stained tissue section to the original blockface image.
  • Internal Tissue Distortions: Calculated by comparing the tissue section mask to the original blockface mask.
  • Tissue Yield: Subjectively assessed as the percentage of tissue successfully transferred to the slide.

Datasets

The study used fixed human brain specimens obtained from the Stanford Alzheimer's Disease Research Center (ADRC). These specimens were first scanned with a 7T MRI, then processed for histology. No new datasets were introduced; the study focused on evaluating a processing technique using existing biological samples.

Key Insights and Ideas

  • Deformation is a Major Bottleneck: The study highlights that tissue deformation during sectioning is a significant challenge for accurate 3D reconstruction and MRI-histology correlation.
  • Tape Transfer Offers Superior Precision: The tape transfer method dramatically reduces tissue expansion, which is crucial for maintaining the spatial integrity of the tissue.
  • Yield Matters: The tape transfer method offers a near-complete tissue yield, meaning more of the valuable brain tissue is preserved on the slide compared to the traditional method.
  • Trade-offs Exist: While tape transfer excels in reducing expansion and improving yield, it can introduce more internal microfractures, which might pose challenges for image registration.

Key Results

  • Tissue Expansion: Traditional water bath mounting resulted in an average tissue expansion of 16.7% relative to the blockface image, while tape transfer showed a significantly lower expansion of 0.9%.
  • Tissue Damage: Both methods exhibited similar average total tissue damage (approximately 3%), referring to areas where tissue was present in the blockface but missing on the slide.
  • Tissue Yield: Tape transfer was subjectively estimated to have a near 100% tissue yield, compared to approximately 60-70% for the traditional method.
  • Distortion Patterns: Traditional capture showed distortion concentrated at the periphery, while tape transfer exhibited linear striations within internal structures.

Limitations and Future Work

The authors acknowledge that the tape transfer technique, while reducing overall expansion, can introduce more internal microfractures. These microfractures could potentially affect the accuracy of image registration with MRI. Future work could focus on optimizing the tape transfer process to minimize these internal defects while retaining its benefits of reduced expansion and improved yield. Further investigation into the impact of these microfractures on quantitative analysis and registration accuracy would also be valuable.

Summary updated

Abstract

Abstract Background Specimen analysis is crucial for identifying imaging and neuropathological signatures. Histology is the gold‐standard, but sample preparation and sectioning induce tissue deformations which hinder quantitative analysis or registration of histology to 3D MRI providing a challenge to the development of MRI biomarkers. Overall, we aim to develop a workflow to correlate histology with high‐resolution MRI at a microscopic level (Figure 1), Here, we evaluate a critical step in this process ‐ the section quality from tissue mounting techniques, comparing: A) traditional water bath (Figure 1F), and B) tape transfer (Figure 1G), for the purpose of image segmentation and correlation with high‐resolution MRI. Methods Fixed human brain specimens were MRI scanned, dehydrated with graded ethanol, xylene‐cleared, and paraffin‐embedded. Before sectioning, the specimen surface was captured by a brightfield microscope (Figure 2, top left, “Blockface image”). Embedded tissue was sectioned at 7 μm using a Leica 100 microtome and mounted on slides using either A) traditional water bath capture onto standard charged slides (Figure 1F), or B) tape‐transfer capture onto adhesive‐coated slides (Alphametrix) (Figure 1G). Slides were hematoxylin and eosin stained followed by imaging using the same microscope settings as the blockface image, and total stained tissue area was calculated. Area of expansion was determined by change from the blockface image (Figure 2, 3rd column). Internal tissue distortions were calculated by subtracting the tissue section mask from the original blockface mask (Figure 2, 2nd column). Results The tissue area measurements show that traditional capture resulted in average expansion of 16.7% relative to the blockface image, whereas tape transfer had 0.9% expansion (Figure 3). Both methods had similar amounts of average total tissue damage (approximately 3%), representing regions with tissue in the blockface image where no tissue was present on the slide. Subjectively, tissue distortion was concentrated at the periphery on traditional capture, whereas there were linear striations within internal structures on tape transfer (red arrows). Tissue yield for tape transfer was subjectively near 100% compared to traditional at approximately 60‐70%. Conclusion Tape transfer had a more consistent yield and better edge preservation with blockface image but had more internal microfractures
BibTeX
@inproceedings{ho2024evaluating,
  author = {William Hai Dang Ho and Yixin Wang and Hossein Moein Taghavi and Phillip DiGiacomo and Jeffrey Nirschl and Pauline Chu and Marios Georgiadis and Michael Zeineh},
  title = {{Evaluating quality of routine vs. tape‐transfer histology sections for 3D reconstruction of paraffin embedded brain tissue and correlation with 7T MRI}},
  booktitle = {Alzheimer},
  year = {2024},
  doi = {10.1002/alz.093158},
}
Figure