By Benjamin Ross
March 8, 2017 | The founders of Proscia, a Baltimore, Maryland-based digital pathology provider, know full well the pain and turmoil that comes with having a close family member struggle with cancer. David West, Jr., Proscia’s CEO and co-founder, grew up watching his mother fight breast cancer; Hunter Jackson, the company’s VP of research, experienced a similar trial when his sister developed a tumor at a young age. These experiences continue to invigorate the team at Proscia, and has inspired them to take a focus on making a difference in the cancer realm.
“Our angle on pathology has always been cancer as a whole,” David West, Jr. told Diagnostics World. “Obviously, pathology is important for other diseases besides cancer, but we feel like that’s where we can make the biggest impact and that’s what inspires us as a team.”
Proscia’s platform is currently used by over 600 researchers and pathologists for data management, telepathology, and image analysis to augment their current work. Most recently, Proscia has entered a partnership with Samsung Medical Center, Seoul, South Korea, hoping to develop diagnostic tests that are computer- and image-based. The goal of the collaboration is to leverage mathematical oncology and novel deep-learning approaches to provide predictive insight into the likelihood of lung cancer lymph node metastasis, which would improve a physician’s ability to prescribe targeted cancer therapies.
“We’re moving very, very fast to getting this to the commercial stage. . . from our end it’s similar technology across, so it’s very easy for us to scale this to different cancers,” said West, Jr. “Our goal is take our technology and build a very strong portfolio. . . As a software company, we’re in a really good position to develop and make really strong computer-based diagnostics, and we intend to commercialize them using our platform the same way we’ve commercialized our other image analysis capabilities.”
West, Jr. was studying biomedical engineering at Johns Hopkins when he became fascinated by the digital pathology space. “I had been connected to researchers at Johns Hopkins who were looking at images of prostate cancer biopsies and measuring with computers features of the nuclei and other histologies, and predicting biochemical recurrence for prostate cancer,” he recalled. “This was when I really caught the digital pathology bug.” In 2014 West, Jr., along with 3 other technologists from Johns Hopkins, the Moffitt Cancer Center, and the University of Pittsburgh, founded Proscia, excited by the possibilities of the digital pathology space and the algorithms that researchers use to measure features in biopsies of cancer cells.
According to West, Jr., the digital pathology community has an issue of minimizing subjectivity, with pathologists only agreeing with each other on the results of a study around 55% of the time. “Even if you have the same pathologist look at the same slide on multiple dates, they disagree with themselves.”
Proscia’s work has been in the realm of machine learning-based diagnostics where they can integrate an image, such as a scanned biopsy slide, and measure patterns and use those measurements to predict metastasis or other critical cancer-related outcomes. They do none of the scanning, instead focusing entirely on software that fits on top of their platform modules: data management, which is used to help pathologists and researchers store their slides and manage their data easily using cloud-based storage; telepathology modules that make it easy for pathologists to share cases and facilitate remote workflows; and image analysis modules, which can be broken down into a couple of stages that focus on specific stains.
Samsung reached out to Proscia three months ago after they had seen the digital pathology company’s work with breast cancer. The South Korean clinical and biomedical research institute wanted Proscia to use their platform to validate lung cancer biopsies. “There’s no solution [Samsung] had for accurately predicting lymph node metastasis based on the primary tumor,” said West, Jr. “It’s a problem that they had and they knew that our technology had a good shot of being able to do this.”
West, Jr. explained that Samsung’s problem required the same technology that Proscia used with their breast cancer test, but with a slight twist; the most important thing is that each unique form of cancer requires a different kind of dataset. Lung cancer in a computer is going to look different than breast cancer, and from a clinical standpoint the outcomes that the oncologists and pathologists need to develop a treatment plan are going to be different.
Samsung Medical Center will be uploading and annotating relevant regions of interest on tumor tissue samples. This data is then stored and shared with Proscia on their platform to leverage their image analysis and deep learning assays.
West, Jr. would be the first to say that these innovations don’t happen overnight, and in fact he stressed how many months and years of validation are imperative for platforms in the digital pathology space. Pending the outcome of the study, West, Jr. is confident that this partnership will be beneficial for both companies. “We intend to scale up these studies, specifically these lung cancer studies, if these results look promising,” he said. “Our hope is that we can expand that study with Samsung, and perhaps recruit other partners as well.”