Harnessing non-destructive 3D pathology

Liu, Jonathan; Glaser, Adam K.; Bera, Kaustav; True, Lawrence D.; Reder, Nicholas P.; Eliceiri, Kevin W.; Madabhushi, Anant

doi:10.1038/s41551-020-00681-x

Cited by 103 publications

(74 citation statements)

References 228 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…As high-resolution biomedical imaging technologies continue to evolve and generate increasingly larger datasets, computational techniques are needed to derive clinically actionable information, ideally through explainable approaches that generate new insights and hypotheses. Interpretable feature-based analysis strategies in digital pathology generally hinge upon obtaining high-quality segmentations of key structural primitives 10,29 (e.g., nuclei, glands, cells, collagen). However, a common bottleneck to achieving accurate segmentations is the need for large amounts of manually annotated datasets 40 .…”

Section: Discussionmentioning

confidence: 99%

“…in conjunction with high-throughput 3D lightsheet microscopy, a number of groups have been exploring the value of non-destructive 3D pathology of clinical specimens for diagnostic pathology 13,[16][17][18][19][20] . Compared to conventional slidebased histology, non-destructive 3D pathology can achieve vastly greater sampling of large specimens along with volumetric visualization and quantification of diagnostically significant microstructures, all while maintaining intact specimens for downstream molecular assays 10 .…”

Section: Introductionmentioning

confidence: 99%

“…A multi-stage computational pipeline was desired for classifying patient outcomes based on interpretable "hand-crafted" features [21][22][23] (i.e., glandular features) rather than an end-to-end deep-learning (DL) strategy for risk classification based directly on the imaging data [24][25][26] . This was motivated by: (1) the attractiveness of an intuitive feature-based approach as an initial strategy to facilitate hypothesis-testing and clinical adoption of an emerging modality in which datasets are currently limited and poorly understood 10,27 , and (2) the observation that when case numbers are limited, a hand-crafted feature-based approach can be more reliable than an end-to-end DL classifier 28,29 .…”

Section: Introductionmentioning

confidence: 99%

“…With conventional histopathology, only a small fraction of a clinical specimen (e.g., a biopsy) is typically sampled, and key prognostic microstructures are viewed as 2D cross-sections rather than in their native 3D configurations 10 . Here, the term “sampled” refers to the volume of an excised specimen that is converted into slide-mounted histology sections and viewed by a pathologist.…”

Section: Introductionmentioning

confidence: 99%

“…in conjunction with high-throughput 3D light-sheet microscopy, a number of groups have been exploring the value of non-destructive 3D pathology of clinical specimens for diagnostic pathology 13,16-20 . Compared to conventional slide-based histology, non-destructive 3D pathology can achieve vastly greater sampling of large specimens along with volumetric visualization and quantification of diagnostically significant microstructures, all while maintaining intact specimens for downstream molecular assays 10 . However, since the associated information content of a 3D pathology dataset of a biopsy is >100× larger than a 2D whole-slide image representation (in terms of total number of pixels), computational tools are necessary to analyze these large datasets efficiently and reproducibly for diagnostic and prognostic determinations.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Prostate cancer risk stratification via non-destructive 3D pathology with annotation-free gland segmentation and analysis

Xie

Reder

Koyuncu

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Prostate cancer treatment planning is largely dependent upon examination of core-needle biopsies. In current clinical practice, the microscopic architecture of the prostate glands is what forms the basis for prognostic grading by pathologists. Interpretation of these convoluted 3D glandular structures via visual inspection of a limited number of 2D histology sections is often unreliable, which contributes to the under- and over-treatment of patients. To improve risk assessment and treatment decisions, we have developed a workflow for non-destructive 3D pathology and computational analysis of whole prostate biopsies labeled with a rapid and inexpensive fluorescent analog of standard H&E staining. Our analysis is based on interpretable glandular features, and is facilitated by the development of image-translation-assisted segmentation in 3D (ITAS3D). ITAS3D is a generalizable deep-learning-based strategy that enables tissue microstructures to be volumetrically segmented in an annotation-free and objective (biomarker-based) manner without requiring real immunolabeling. To provide evidence of the translational value of a computational 3D pathology approach, we analyzed ex vivo biopsies (n = 300) extracted from archived radical-prostatectomy specimens (N = 50), and found that 3D glandular features are superior to corresponding 2D features for risk stratification of low- to intermediate-risk PCa patients based on their clinical biochemical recurrence (BCR) outcomes.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Prostate cancer risk stratification via non-destructive 3D pathology with annotation-free gland segmentation and analysis

Xie

Reder

Koyuncu

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Far‐Red Aggregation‐Induced Emission Hydrogel‐Reinforced Tissue Clearing for 3D Vasculature Imaging of Whole Lung and Whole Tumor

Gong,

Zhuang,

Chong

et al. 2024

Advanced Materials

View full text Add to dashboard Cite

Understanding the vascular formation and distribution in metastatic lung tumors is a significant challenge due to autofluorescence, antibody/dye diffusion in dense tumor, and fluorophore stability when exposed to solvent‐based clearing agents. Here, an approach is presented that redefines 3D vasculature imaging within metastatic tumor, peritumoral lung tissue, and normal lung. Specifically, a far‐red aggregation‐induced emission nanoparticle with surface amino groups (termed as TSCN nanoparticle, TSCNNP) is designed for in situ formation of hydrogel (TSCNNP@Gel) inside vasculatures to provide structural support and enhance the fluorescence in solvent‐based tissue clearing method. Using this TSCNNP@Gel‐reinforced tissue clearing imaging approach, the critical challenges are successfully overcome and comprehensive visualization of the whole pulmonary vasculature up to 2 µm resolution is enabled, including its detailed examination in metastatic tumors. Importantly, features of tumor‐associated vasculature in 3D panoramic views are unveiled, providing the potential to determine tumor stages, predict tumor progression, and facilitate the histopathological diagnosis of various tumor types.

show abstract

Engineering the future of 3D pathology

Liu,

Chow,

Colling

et al. 2023

The Journal of Pathology CR

Self Cite

View full text Add to dashboard Cite

In recent years, technological advances in tissue preparation, high‐throughput volumetric microscopy, and computational infrastructure have enabled rapid developments in nondestructive 3D pathology, in which high‐resolution histologic datasets are obtained from thick tissue specimens, such as whole biopsies, without the need for physical sectioning onto glass slides. While 3D pathology generates massive datasets that are attractive for automated computational analysis, there is also a desire to use 3D pathology to improve the visual assessment of tissue histology. In this perspective, we discuss and provide examples of potential advantages of 3D pathology for the visual assessment of clinical specimens and the challenges of dealing with large 3D datasets (of individual or multiple specimens) that pathologists have not been trained to interpret. We discuss the need for artificial intelligence triaging algorithms and explainable analysis methods to assist pathologists or other domain experts in the interpretation of these novel, often complex, large datasets.

show abstract

Harnessing non-destructive 3D pathology

Cited by 103 publications

References 228 publications

Prostate cancer risk stratification via non-destructive 3D pathology with annotation-free gland segmentation and analysis

Prostate cancer risk stratification via non-destructive 3D pathology with annotation-free gland segmentation and analysis

Far‐Red Aggregation‐Induced Emission Hydrogel‐Reinforced Tissue Clearing for 3D Vasculature Imaging of Whole Lung and Whole Tumor

Engineering the future of 3D pathology

Contact Info

Product

Resources

About