Digital Pathology in Drug Development: Opportunities and Challenges
Posted 20th April 2020 by Liv Sewell
We recently caught up with Dr Olulanu Aina, Scientific Director and Veterinary and Toxicological Pathologist at Janssen, about digital pathology applications in drug discovery...
How did you become an early adopter of digital pathology in drug development?
I was introduced to digital slide scanning by Dr Robert Cardiff, first adopter of the Aperio Imaging system, during my PhD. I then completed two years of veterinary pathology residency at UC Davis and a third year of internship at Genentech. While at Genentech, I learned to perform image analysis on digitized slides using Definiens software.
I returned to UC Davis as a Fellow in the Cardiff Lab and subsequently Associate Director of the Center for Genomic Pathology. The benefits of applying digital pathology techniques to gather, organize and share data became really apparent, especially in the analysis of biomarker expression in breast cancer tissue slides with Aperio Genie software. I spent the last five years integrating digital pathology and image analysis into daily workflow at Regeneron as a Senior Staff Pathologist.
How are you using digital pathology to enhance drug development?
As a toxicologic pathologist, part of my role is evaluating the safety of drug candidates and comparing potency, efficacy, and toxicity of newly generated compounds, sometimes to the current clinical standard-of-care. Highly accurate quantitative measurements are useful in this analysis and help in better decision making in the development of new therapies. Quantitative image analysis and digital pathology methods can aid tremendously in this process.
An example of this is the work I did while at Regeneron on a bispecific molecule developed for the treatment of B cell lymphomas. I applied quantitative image analysis to the monkey toxicology study slides to measure the efficacy of the molecule. From the analysis, we could compare the efficacy of the new drug to the current clinical standard-of-care treatment with precision. We went on to quantitate the efficacy of the drug and identify biomarkers of clinical response in the samples from patients undergoing clinical trials with this molecule using both fluorescent and chromogenic IHC staining techniques and Indica HALO image analysis platform.
What is the potential impact of digital pathology in drug discovery?
Digital pathology is making a very big impact in drug discovery, and it will continue to do so. Digital pathology and artificial intelligence are going to be of tremendous value, not just for pathologists in industry, but also for pathologists in academia as well as contract research organizations.
It makes us more efficient in the use of our time and energy. In toxicologic pathology, we generate a lot of slides, and even if we can just use digital imaging and AI to eliminate looking at slides without disease, that cuts our time by a factor of maybe seven or eight – enabling us to focus on slides that do have lesions. These are the slides we need to spend time on to really understand the mechanism of toxicity. So digital pathology frees us up to spend our time where it is most needed. It also gives us a lot of flexibility in the work that we do – the ability to gather, organize and share data in ways we never could with a microscope.
I strongly believe that if AI becomes something that is well integrated into our daily work it is going to make a tremendous difference in the amount of quality work that we can do under the time pressures we face.
Pathologists are trained to give semi-quantitative analysis of tissues. With quantitative image analysis, we can now begin to give real statistical data associated with biology. The ability to be able to use AI and Image Analysis to produce statistical data in toxicological pathology is something I’m really excited about.
As an early adopter of digital pathology, what are the challenges you’ve faced as you have integrated digital technology?
One of the biggest challenges in toxicologic pathology is the volume of slides that need to be digitized and organized with the appropriate metadata. Another challenge is in building confidence and encouraging pathologists who have been trained on, and have used, a microscope their entire career to try digital pathology. There can be a reluctance to get on board with new technology when it is novel and there are uncertainties. A third challenge is the IT infrastructure. It is one thing to digitalize a slide and it is another to be able to view and analyze at a rapid pace (from different physical locations). A lot goes into digitalizing, viewing and analyzing slides and good IT support is essential and critical.
How have you overcome these challenges?
First, I try to engage IT colleagues very early on, when we evaluate new equipment, processes or systems. Second, I’ve sought buy-in from our leadership – because these things cost a substantial amount of money and require adequate resources. When colleagues see that the technology works well, and enables greater flexibility and efficiency, they typically get on board.
Olulanu Aina is Scientific Director at Janssen, a Johnson & Johnson company.
The 6th Digital Pathology and AI Congress: USA will feature a track dedicated to digital pathology applications in pharma and industry. Find out more here.
Leave a Reply