Converging on a major milestone in digital pathology 


Eye on Innovation

Eye on Innovation features exciting advances taking place at Mayo Clinic Laboratories. This monthly series shines a spotlight on recently developed tests and highlights how Mayo Clinic translates ideas and discoveries into testing resources that improve diagnosis and care for patients across the globe.

After a three-year collaborative effort, from 2019 to 2021, Mayo Clinic has converged several critical areas of pathology — Conventional Chromosomes, Molecular Cytology, and FISH — to MetaSystems, a cutting-edge digital slide scanning and analysis platform. The conversion, spanning five laboratories, represents a major milestone in Mayo’s digital pathology initiative, aimed at arming pathologists with more advanced, less labor-intensive tools that replace aging technology. 

“This conversion was one of the first projects launched after the digital pathology initiative was created in early 2019,” says Kurt Simon, senior project manager at Mayo Clinic’s Department of Laboratory Medicine and Pathology (DLMP). “MetaSystems is the flagship project for phase I of the initiative.”

Freedom to view and share

A manual capture station used to take pictures with the new system

The new system involves several innovative firsts, including the ability to export images in Digital Imaging and Communications in Medicine (DICOM) format. DICOM allows pathologists and laboratory technologists to view, store, retrieve, and share medical images more universally.

“The old system, as well as a lot of other scanners in the industry, was proprietary, meaning you’re stuck using their software to view images,” says Simon. “Now to accommodate consumers, the industry is trying to use more open-sourced formatting like TIF or JPEG. So once we’re done with a proprietor’s (scanner) software, we can then use any kind of software to view images along with their metadata (information about each image). MetaSystems has the capability for TIF and DICOM exporting, which can be used by other systems.”

Xinjie Xu, Ph.D., is a senior associate consultant and director of the Conventional Chromosome Laboratory in DLMP. “A proprietary format limits the potential for AI/machine learning outside of the software,” she says. “We worked with MetaSystems to enable the image export function to the DICOM format, which contains metadata and is a vendor-neutral image format. DICOM images can be archived in a vendor-neutral archiving solution and be viewed in a third-party image viewer.”

Simon adds, “What DICOM does is ‘standardize’ how metadata is presented. So it’s more efficient for data mining and for other applications. MetaSystems is our first system in digital pathology that has the capability to export in this normalized format.”   

Streamlined workflow

Another innovation highlight? With MetaSystems’ new Metafer scanners, laboratory technologists now have the capability to run dark field and bright field scanning using the same machine. This capability is a big advantage for chromosome and FISH analysis, two techniques of cytogenetic testing.

“Chromosome testing needs a bright field scanning, and FISH needs a dark field scanning. A significant proportion of patients that we test have both chromosome and FISH ordered and performed by our laboratories,” Dr. Xu says. “Results from both techniques often need to be correlated with each other. In daily practice, these two laboratories work very closely with each other.

Automated slide scanner with the capability to run both dark field and bright field scanning

Having a scanner platform that can run both dark field and bright field brings huge advantages to us. For example, during validation, the two teams shared scanners. Scanners in one lab can function as backup scanners for the other lab.”

The new scanner platform also provides a streamlined workflow for technologists and consultants, as they can use the same software for both techniques during analysis and interpretation.

Dr. Xu continues: “Implementing the scanner is our first step. We basically digitized our chromosome and FISH practice by using the scanner. After that, we will focus on how to develop accurate algorithms and deploy these algorithms to automate our workflow. Our goal is to replace a good proportion of our manual analysis by automation. The key is to have an accurate algorithm and a robust workflow to ensure both efficiency and quality.”

Reduced manual analysis

MetaSystems will also allow the labs to leverage deep neural network (DNN) technology for karyotyping, or chromosome testing. In other words, DNN has the potential to automate part of the labor-intensive work normally required of laboratory technologists — who must be rigorously trained and highly experienced to manually analyze the chromosome number and morphology.

A karyogram created with the application that uses deep neural network technology to separate and classify chromosomes

“We are starting to evaluate how to best utilize this algorithm and integrate it in our routine workflow,” Dr. Xu says. “I don’t think it will completely replace manual analysis. But leveraging the DNN technology has the potential to automatically analyze normal and straightforward chromosomes, and then our technologists can better focus their time on abnormal and challenging chromosomes. This has the potential to reduce manual labor significantly, which translates into higher efficiency. We will be able to accommodate higher volume without the need for more resources.”   

Using MetaSystems’ Metafer scanners for FISH probe testing almost eliminates the need for manual capture. For one FISH region with optimal quality, the scanner only needs several minutes to capture hundreds of nuclei images, and these images will be ready for downstream analysis right away.

“We can load lots of slides at once and walk away and let the scanner do all the image capturing,” says Dr. Xu. “This can even be left running overnight.” 

Success amid disruption

It is also worth noting that the MetaSystems conversion project was completed on schedule, despite being in process during the peak of the COVID-19 pandemic.

“DLMP was totally disrupted by having resources diverted to support COVID testing,” says Simon. “But the MetaSystems conversion was one of the priority projects we kept working on, and we managed to stay on schedule despite redeployment of lab resources during COVID.”

This was no small accomplishment considering that, early on, the project lost one of its key team members, a supervisor who was temporarily redeployed to manage COVID testing needs. But the Conventional Chromosome and FISH labs managed to recruit other staff and resources into the project.

“This project has been a great opportunity to grow the knowledge base and leadership skills of lab staff,” says Simon. “The labs have done a great job at recognizing and empowering employees to stand up and be a part of a project like this. And those employees have become great leaders themselves as a result.”    

Now that the pandemic is loosening its grip, DLMP can concentrate more resources and staff on the digital pathology initiative in the coming years. 

“We are in the process of transforming our practice,” Dr. Xu says. “We know AI/machine learning won’t completely replace manual analysis because of the complexity of medicine. But by utilizing the combination of these algorithms and human expertise, our goal is to provide more efficient and accurate test results for our patients.”

More from Eye on Innovation

 

Chris Bahnsen (@cjbahnsen)

Chris Bahnsen

Christoph Bahn covers emerging research and discovery for Mayo Clinic Laboratories. His writing has also appeared in The New York Times, Los Angeles Times, and Smithsonian Air & Space. He divides his time between Southern California and Northwest Ohio.