It is a centuries-old image of good science: the lab-coated expert, hunched over a microscope analyzing a glass slide. Proscia wants to resign it to history – at least, when it comes to battling cancer. The Baltimore-based startup, founded out of Johns Hopkins University in 2014, wants to bring artificial intelligence to dermatopathology, the study of skin-born diseases. It has developed an AI module that can greatly enhance the accuracy and scope of tissue analysis. That means better results, and lower costs.
David West, Proscia’s CEO, has been with the project from the beginning. West, a computational biology graduate, had been researching how machine learning could predict prostate cancer outcomes using images of biopsies. Computing power meant “magnitudal leaps and bounds,” he tells Red Herring. “It was night and day.”
From there, Proscia was born. Today it has a dedicated staff of experts, a glittering advisory board and is backed by almost $2 million of venture capital. That is little surprise given the huge problem it is addressing.
Over half a million people die from cancer in the United States each year. It is the second most common cause of death in the county, behind heart disease. The means to fight cancer improve each year. But diagnoses are still made with the human eye. And that is hugely problematic.
“Somewhat like art, people see things differently,” says West. “As a result, subjectivity leads to error and leads to patient over- or under-diagnosis.” The cost of those errors are huge: $750bn in the US alone. Proscia’s early tests have brought back an accuracy of 98.7%. “Computers do not have bad days,” West adds.
Proscia’s proprietary algorithm can learn more from slides in a week than a human can view in a lifetime. But, as with all new technology, there has been blowback from those who’d rather keep the status quo. Some pathologists have viewed the platform as a threat to their job, rather than a digital assistant.
“We had to adjust our messaging and value proposition to make sure pathologists view Proscia an as asset, providing more robust data to provide better results,” says West.
Ultimately, though, Proscia is about more than placating scientists about robots taking their salaries. It’s about saving lives. Pathologists are among the most critical decision makers in the cancer process, from diagnosis to treatment.
Taking away subjectivity removes the reliance on diagnostic concurrence. It is also quicker, meaning laboratories get a double-win: beat cancer; save cash. “For the laboratory, it is massive gains in efficiencies,” says West. “The pathologist will have additional information to help them make a better diagnosis to the patient. The advantages for the patient will be better, more consistent outcomes.”