Recent Advances In Digital Pathology

Documenting the recent advances in digital pathology. This is part four of our report on the Digital Pathology & AI Congress: USA. For other parts, please see links toward the bottom of the article.

 

The conference showcased several recent breakthroughs and novel approaches in the field of digital pathology, highlighting the rapid pace of innovation and the potential for transformative impacts on patient care and research. The conference also featured industry perspectives and case studies, shedding light on emerging trends and real-world applications of digital pathology and AI in precision medicine and the pharmaceutical and biotechnology sectors.

Novel Methods for Slide-Free, Label-Free Pathology

Richard Levenson of the University of California, Davis, presented his work on novel methods for slide-free, label-free imaging techniques. He discussed the development of fluorescence imitating brightfield imaging (FIBI) and quantitative oblique back-illumination microscopy (qOBM), which allow for near-real-time histology results without tissue sectioning or staining.[1]

“Standard formalin fixation and paraffin embedding processes and microtomy, slide staining, coverslipping, labeling, and whole slide imaging take time and resources — at least hours if not days to weeks, depending on logistics. Being able to go directly from tissue to histology within minutes is a major advance,” noted Dr. Levenson. He added that slide-free technologies could help reduce the infrastructure needs of digital pathology.

FIBI employs a standard fluorescence microscope and rapid staining techniques to enhance contrast. The staining process is simple and can be applied on fresh or formalin-fixed tissues. The tissue needs to be 100 microns.

“Expose the surface to be imaged to eosin, and then hematoxylin, for 5 to 30 seconds each, followed by a brief wash, is all that is required,” he said. “Without the H&E stain, nothing is visible with FIBI.” He explained that the tissue is illuminated with blue light, and the combination of absorbance and fluorescence enables imaging of the tissue surface with high contrast and image quality similar to that obtained with traditional H&E staining of formalin-fixed paraffin-embedded tissues. However, irregular hand-cut surfaces result in out-of-focus FIBI images. As a solution, they use an extended depth of field to improve the image quality without deconvolution.

qOBM produces grayscale output that can be converted into virtual H&E images using AI-powered color conversion tools without the need for staining. Dr. Levenson explained that qOBM is intrinsically a depth-sectioning technology and that any given plane can be imaged within 100 µm of the surface of a specimen. The depth of the optical section is determined by adjusting the lens focal position.

These cost-effective and nondestructive methods hold promise for applications such as biopsy monitoring and intraoperative surgical guidance. Moreover, both FIBI and qOBM are designed to facilitate AI integration, enabling the development of intelligent systems for real-time analysis and decision support.

Dr. Levenson ended his talk by highlighting the versatility of these slide-free approaches and their potential use in diverse and resource-limited settings. FIBI is evolving for large-format, high-speed imaging (up to 10 × 10 cm2) for margin status determination in surgery and is being tested for use in Ghana for breast core-needle biopsy evaluation and diagnosis.

[1] Richard Levenson. Slide-free, label-free and pain-free (where pain = cost, complexity). Presented at the 10th Digital Pathology & AI Congress: USA; May 7-8, 2024; San Diego, CA.

Dr. Levenson, Professor and Vice Chair for Strategic Technology at the Department of Pathology and Laboratory Medicine, University of California, Davis.

From Pixels to Insights: Advancing Drug Development With Digital Pathology and AI

Albert Juan Ramón from Johnson & Johnson discussed how the company is leveraging digital pathology and AI to revolutionize drug development processes by obtaining actionable insights from tissues.[2] Through live demonstrations, Dr. Ramón presented AI models aiding in disease diagnosis, biomarker detection, patient stratification in clinical histopathology, and characterization of the tumor microenvironment for predicting patient responses to therapies.

Dr. Ramón showcased Microscopy Image Analysis (MIA) Pathology Insights, Johnson & Johnson’s Al-based pipeline for the characterization of histopathology images. The pipeline includes AI algorithms designed to characterize the tumor microenvironment (TME), assess disease severity, analyze biomarker expression, and predict patient responses to therapies. By examining the complex interplay between tumor cells and their surrounding environment, these algorithms can provide valuable insights into treatment efficacy, support the development of targeted therapies tailored to individual patients, inform patient screening for clinical trials, and guide patient selection for treatment. The MIA Biomarker has been used for patient prescreening in a recent multicenter clinical trial of an FGFR inhibitor for metastatic urothelial cancer. The investigators decided to cancel tissue testing based on the prescreening results of Al, leading to a nearly 30% reduction in the number of molecular tests conducted.

In the preclinical setting, Dr. Ramón demonstrated how AI facilitates image analysis and automated scoring for toxicity assessments, streamlining the drug development process. By automating toxicity scoring, MIA Preclinical can reduce the pathologist burden, enhance the efficiency of preclinical studies, and increase the objectivity of the process.

Dr. Ramón’s presentation underscored the pivotal role of digital pathology and AI in revolutionizing precision medicine and improving patient outcomes through enhanced diagnostic accuracy, identification of subgroups of patients who may benefit from specific treatments, and development of personalized therapies.

[2] Albert Juan Ramón. From pixels to insights: Advancing drug development with digital pathology and AI at Johnson & Johnson. Presented at the 10th Digital Pathology & AI Congress: USA; May 7-8, 2024; San Diego, CA.

Dr. Ramón, Principal Scientist at Johnson & Johnson.

Links To Other Parts Of The Series

Part 1: Highlights from the 10th Digital Pathology & AI Congress: USA

Part 2: Digital Pathology Implementation: Insights From Experts At DP&AI: USA

Part 3: Clinical Implementation Challenges And Potential Solutions

References

[1] Richard Levenson. Slide-free, label-free and pain-free (where pain = cost, complexity). Presented at the 10th Digital Pathology & AI Congress: USA; May 7-8, 2024; San Diego, CA.

[2] Albert Juan Ramón. From pixels to insights: Advancing drug development with digital pathology and AI at Johnson & Johnson. Presented at the 10th Digital Pathology & AI Congress: USA; May 7-8, 2024; San Diego, CA.

Share This Post

Leave a Reply