Many companies developing artificial intelligence software for healthcare are trying to improve doctors’ ability to diagnose disease. Fewer emerging healthcare AI businesses are focused on what happens after a diagnosis, aiming to figure out how treatments are working.
One of them, AIQ Solutions, has developed software that analyzes medical imaging data to identify all of a cancer patient’s lesions—areas of abnormal tissue—and quantify how they have changed after treatment. CEO Eric Horler claims his company’s technology, focused initially on metastatic bone disease caused by prostate or breast cancer, can provide such insights faster and at a more granular level than competing tools.
“One of the things that’s very difficult for a physician to understand is whether that treatment is really working, and the degree to which it’s working,” Horler says.
Madison, WI-based AIQ received FDA clearance last year to begin marketing its first product, dubbed Quantitative Total Bone Imaging. Now, thanks to an injection of venture capital, the company plans to boost its commercialization efforts and continue developing additional products for various diseases, including bone marrow disorders.
AIQ has raised a $3.2 million Series A funding round, Horler tells Xconomy. The equity investment comes from a group of Wisconsin-based investors: Capital Midwest Fund led the round, along with Rock River Capital Partners and 30Ventures; Wisconsin Investment Partners also put in money, AIQ says.
The fresh capital will help AIQ try to establish a foothold in the market as it competes against other cancer imaging software companies, including Cleveland-based MIM Software, Sweden-based Exini Diagnostics, and UK-based Mirada Medical. Other companies, including UK-based Cambridge Cancer Genomics, are calculating tumor response by analyzing tissue samples from liquid biopsies.
Horler says AIQ’s technology aims to be more comprehensive and precise than other available treatment assessment products. Say a patient’s cancer has spread from his prostate to his bones. The patient may have hundreds of lesions across his skeleton, and the response of each individual lesion to drug treatment will likely vary, Horler says. AIQ has found that a tiny percentage of malignant bone lesions—perhaps two or three out of hundreds—might be resistant to the initial therapy. But doctors don’t usually know which ones are the problem, so they put the patient on a more aggressive treatment plan, often involving drugs with a higher toxicity and higher cost.
“It’s definitely worse for the patient,” Horler says.
AIQ wants to offer an alternative. If its software can identify the culprit lesions, doctors might be able to implement a more targeted and less costly treatment, such as a “localized ablation technique” to zap specific cancerous tissue, Horler says.
AIQ’s software scrutinizes medical scans of a patient’s lesions at different points in time, and—using machine learning and other advanced data analytics techniques—identifies all the cancerous lesions. It detects the ones that are not responding to treatment, quantifies the degree of treatment resistance, and maps the locations of the problem lesions, Horler says.
The underlying technology was developed at the University of Wisconsin Carbone Cancer Center by professors Robert Jeraj and Glenn Liu. The company was incorporated in 2015. Two years later, it struck a deal with the Wisconsin Alumni Research Foundation for an exclusive license to the technology’s two patents and the software code for the bone imaging product, according to AIQ’s website. Horler points to studies published in 2016 and 2017 that provided clinical evidence that supported the FDA clearance in 2018.
In addition to Jeraj and Liu, the company’s other founders include Dona Alberti, a former UW Carbone Cancer Center assistant director of translational research, and Guy Starbuck, previously a software architect and development manager at Yahara Software and Stericycle. Horler joined as CEO last September. He previously led Madison-based medical device firm Swallow Solutions, and was a manager at GE Healthcare before that. (The leadership team is pictured above, left to right: chairman Rick Lesniewski; Liu, the chief medical officer; Jeraj, the chief scientific officer; Alberti, the chief operating officer; Starbuck, the chief technical officer; and Horler.)
Although it’s still early for AIQ, Capital Midwest general partner Eli Blee-Goldman says he was impressed with the scientific background of the startup’s founding team, the initial impact the product has had on patients, and the early commercialization strategy. That helped AIQ stand out from the crowd of AI-related healthcare startups that Blee-Goldman encounters at large industry conferences, where many of the companies look the same, he says.
AIQ is selling its product initially to biotech startups and large pharmaceutical companies that use the software primarily to assess the impact of their drugs, Horler says. He declined to name the customers. The plan is to eventually sell the software to hospitals for routine clinical care, he says.
Blee-Goldman says that’s a smart approach. Many healthcare AI startups have tried to sell directly to hospitals from the beginning, and a lot of them have “flamed and died,” he says. “It’s tough,” he adds.
It can be difficult to demonstrate the value of AI-based programs. Horler says some companies developing AI applications for healthcare have opted to “just throw tons and tons and tons of data” at machine learning and deep learning systems, and “hope that something useful spits out the back end.”
“That’s where the results don’t seem to be living up to the promises,” Horler says. “I believe the value in AI is going to be in the really targeted applications. I think you’re starting to see a shift toward that.”