Abstract:Histopathology currently acts as a gold standard for human prostate cancer (PCa) diagnosis. However, the subjective nature of histopathology leads to inevitable discordance among pathologists. Specifically, the inter-observer discordance could be up to 40% for the differentiation between Gleason score 6 (low-grade) and 7 (high-grade) of PCa. According to clinical guidelines, patients with high-grade PCa need to be actively treated, while patients with low-grade PCa could undergo active surveillance due to its slowly growing feature. Therefore, differentiating high-grade and low-grade PCa is an urgent clinical need. By integrating stimulated Raman scattering microscopy and confocal Raman spectroscopy, our previous study found the aberrant cholesteryl ester (CE) accumulation in human PCa tissues. However, no significant difference in CE accumulation between the low-grade and high-grade PCa was found, primarily because the previous study only analyzed the composition of manually selected lipid droplets (LDs) without quantitative analysis of the whole field of view. Here, we employed hyperspectral stimulated Raman scattering (HSRS) microscopy to test the hypothesis of CE as a marker for differentiation of low-grade and high-grade human PCa. First, lipid, lipofuscin, and protein were quantitatively mapped in human prostate tissues based on HSRS imaging of C–H vibrational region and multivariate curve resolution analysis. Then, within the lipid channel, CE percentage and unsaturation level of LDs were quantitatively mapped according to the height ratio between Raman bands at 2870 and 2850 cm−1, and between Raman bands at 3006 and 2850 cm−1, respectively. In total of 6 normal prostate tissues, 9 low-grade and 9 high-grade PCa tissues from patients, we found lipofuscin accumulation in all the normal prostate but none in PCa. While all the high-grade PCa tissues had LD accumulation, only 3 low-grade PCa tissues had LD accumulation. Notably, among all the PCa tissues with LD accumulation, no significant difference in LD amount was found between low-grade and high-grade PCa. Fortunately, both CE percentage and unsaturation level of the LDs were significantly different between low-grade and high-grade PCa. Furthermore, it was shown that CE percentage could differentiate low-grade and high-grade PCa with high sensitivity and specificity. Taken together, our study may provide a new opportunity towards more accurate PCa diagnosis.

Prostate Cancer Histopathology Using Label-Free Multispectral Deep-Uv Microscopy Quantifies Phenotypes of Tumor Aggressiveness and Enables Multiple Diagnostic Virtual Stains

Prostate cancer histopathology with label-free multispectral deep UV microscopy quantifies phenotypes of tumor grade and aggressiveness

Clinical-Grade Validation of an Autofluorescence Virtual Staining System with Human Experts and a Deep Learning System for Prostate Cancer

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

Automated Whole Slide Imaging for Label-Free Histology using Photon Absorption Remote Sensing Microscopy

Autofluorescence Virtual Staining System for H&E Histology and Multiplex Immunofluorescence Applied to Immuno-Oncology Biomarkers in Lung Cancer

Feature Fusion of Raman Chemical Imaging and Digital Histopathology using Machine Learning for Prostate Cancer Detection

Prostate Cancer Identification Via Photoacoustic Spectroscopy and Machine Learning.

Label-free and non-destructive histology of unprocessed biological tissues with ultraviolet single-plane illumination microscopy

Molecular mapping of tumor heterogeneity on clinical tissue specimens with multiplexed quantum dots.

Label-free Monitoring of Colonic Cancer Progression Using Multiphoton Microscopy

Virtual Histology with Photon Absorption Remote Sensing using a Cycle-Consistent Generative Adversarial Network with Weakly Registered Pairs

Hyperspectral Stimulated Raman Scattering Microscopy Facilitates Differentiation of Low-Grade and High-Grade Human Prostate Cancer

Defining Prostate Cancer with High Gleason Score by Prostate Fluid Metabolic Fingerprint Based Multi-Modal Recognition

Label-Free Identification of Early Stages of Breast Ductal Carcinoma via Multiphoton Microscopy

A new optical density granulometry-based descriptor for the classification of prostate histological images using shallow and deep Gaussian processes

Three-dimensional Histo-Morphometric Features from Light Sheet Microscopy Images Result in Improved Discrimination of Benign from Malignant Glands in Prostate Cancer.

Quantitative Microscopy in Medicine

Deep neural network analysis employing diffusion basis spectrum imaging metrics as classifiers improves prostate cancer detection and grading

Image Registration of In Vivo Micro-Ultrasound and Ex Vivo Pseudo-Whole Mount Histopathology Images of the Prostate: A Proof-of-Concept Study

Advancing automated digital pathology by rapid spectral imaging and AI for nuclear segmentation