Dialogue is a proven diagnostic tool. For conditions such as Parkinson’s disease and amyotrophic lateral sclerosis, patient assessment includes interviews during which cognitive abilities, language use, and muscle movement are observed. Such assessments have been part of neurological exams for years, but David Suendermann-Oeft, CEO of medical technology startup Modality.ai, believes they are overdue.
Modality has developed software that can replace human interviews with virtual ones. Patients speak and answer questions as usual, but they converse with Tina, a conversational AI system. In addition to assessing the patient’s speech, the technology uses computer vision to monitor facial movements. Tina objectively collects measurements that can be used to diagnose neurological disorders, assess disease progression, and measure drug efficacy. The FDA recently classified Modality technology as a Class II medical device. Software as a medical device category.
A clinician can only interview one patient at a time. The process is not scalable, says Suendermann-Oeft. Tina can handle thousands of assessments. But for patients, the experience is designed to be a personal one. Modality designed Tina to be self-explanatory, so no special instructions are needed. She guides patients through the entire assessment, which users can complete at their leisure on any consumer device with a camera and microphone.
“People love interacting with Tina because she’s always there for you and she’s not judgmental,” says Suendermann-Oeft. “She doesn’t care if your pronunciation is bad.”
Founded by experts in dialogue systems, San Francisco-based Modality is now more commonly known as a virtual agent or virtual assistant. Suendermann-Oeft said the research involved building industrial-scale systems for tech companies’ phone interactions — work that was done before Amazon’s Alexa and other consumer-facing assistants. Modality stems from efforts to bring the capabilities of dialogue systems to the medical field. Modality was one of the companies recently featured in the Pitch Perfect competition MedCity INVEST Conference in Chicago. The startup was named a co-winner in the pharma tech category.
“As we start to see more neurointerventions coming to market, the solutions provided by Modality.ai become increasingly important,” said Jessica Federer, a partner at Millennia Partners in Boston and one of the judges in the category. “Reliable assessments as needed to support diagnosis and monitoring drive the field forward.”
Suendermann-Oeft stressed that Modality is not trying to replace clinicians. Diagnosis and treatment decisions still rest with the doctor. But Tina helped inform those decisions. Patient responses were recorded and categorized. This technique analyzes the speed and loudness of the response. Facial movements during these responses were also analyzed. Making these assessments virtual can free up clinicians for other tasks.
“Time is definitely an issue for clinicians, as is the lack of objective measures in many cases,” Suendermann-Oeft said. “Right now, ALS patients are tested for language only once. There is no ability to do more, not even to observe. Or measuring any therapy, any positive or negative impact.”
Tina can also be used to evaluate approved and experimental drugs. A key pain point for pharmaceutical companies is the clinical testing of experimental CNS drugs, which have been plagued by research failures. Modality is still building awareness in the pharmaceutical world, but Suendermann-Oeft says Tina offers the prospect of more effective objective assessments. These features could help speed up clinical trials and potentially reduce failure rates, he said.
Modality has raised $2.6 million since its founding in 2018, most recently in a seed round last year. The company is currently in the middle of a funding round with a target of $5 million. Suendermann-Oeft said his company will use the funding to expand commercialization of the technology. Modality does not provide its software directly to clinicians and researchers. The startup works with channel partners who already have relationships with clinical sites or pharmaceutical companies; Tina is integrated into a larger software product. For example, if a company was offering technology for decentralizing clinical trials, Tina would be part of that product.
Modality serves Tina under a software-as-a-service business model, charging a monthly fee per patient. Tina’s recent FDA classification of a medical device is key because it enables suppliers to claim reimbursement for neurological assessments using the technology. Modality.ai has a company in Germany that is now also expanding into Finland. The company will seek a CE mark, laying the groundwork for marketing the software to European healthcare websites.
Since Modality launched, Suendermann-Oeft said tens of thousands of people representing different age groups have spoken to Tina as part of some sort of assessment. Feedback from these users has shaped and refined the software, which is now in version 30. Many of the proposed changes help make the technology easier to use, Suendermann-Oeft said. The FDA’s approval of Tina was not disease-related. The technology has been selected to monitor patients in clinical trials for ALS, Parkinson’s, autism, depression and schizophrenia. Suendermann-Oeft said the technology is particularly popular with anxiety sufferers.
“These people usually don’t speak to anyone, let alone a psychiatrist,” he said. “But they did speak to Tina, which was an amazing thing.”
Screenshot of Tina by Modality.ai
