Skip to main content

Nucleai Matches Cancer Patients to Immunotherapies With AI, Image Analysis

Premium

CHICAGO – Nucleai, a developer of a bioinformatics platform for precision oncology research and clinical decision support, may be a startup, but its technology has been proven over more than two decades.

CEO Avi Veidman and several colleagues started the company in 2017. He previously spent nearly 20 years in Israeli military intelligence, where he led a department that applied artificial intelligence and machine learning to the analysis of satellite and aerial images. Eventually, he decided that he wanted to establish a company at the nexus of big data and biology.

"When I retired from the service, I wanted to do something which was impactful, and healthcare seemed to be a very reasonable place to start with," Veidman said. "There are many problems that are unsolved, and I felt that AI and big data could really help in progressing those problems."

He also noted the prevalence of cancer, including in his own extended family. "At the end of the day, I thought that that was a good start for initiating a company that will help people," Veidman said.

Veidman assembled a multidisciplinary team including physicians, pharmaceutical industry veterans, technologists, and other product experts. They have created a platform to match patients to immunotherapy drugs by modeling the tumor microenvironment with a combination of imaging analysis and AI capabilities.

In putting the company together, Veidman mined his network of contacts among the extensive IT and biotech communities in Israel, and eventually found some supporters in the US. Nucleai's advisory board includes Alex Lazar from the University of Texas MD Anderson Cancer Center, Asaf Rotem of the Dana-Farber Cancer Institute, and Stanley Robboy of Duke University, who is a past president of the College of American Pathologists.

Nucleai, based in Tel Aviv, Israel, relies on AI to analyze large datasets of tissue images to model spatial characteristics of tumors and each patient's immune system to help predict patient response to various cancer therapies. The technology combines "computer vision" and machine learning, the company said, parsing hundreds of thousands of cells in a pathology slide image to determine how the tumor affects the immune system, then matches each patient to a specific therapy.

The company aims to support precision oncology research for pharmaceutical companies, provide clinical decision support for oncologists, and help payors make informed coverage decisions for immunotherapies. "We serve pharma companies' translational departments and biomarker development work," Veidman said.

In a poster presented at the 2020 American Society of Clinical Oncology (ASCO) virtual conference in late May, a team of researchers funded by Nucleai validated AI-based spatial analysis of tumor microenvironments for breast cancer. The poster described how the researchers were able to associate lymphocyte clusters and a high ratio between stromal lymphocyte density and tumor lymphocyte density with longer progression-free intervals in breast cancer patients.

The researchers, representing Chaim Sheba Medical Center and Kaplan Medical Center in Israel as well as MD Anderson in Houston, found that the Nucleai technology helped find treatments that significantly improved progression-free interval. The AI performed analysis of whole-slide images available from the Cancer Genome Atlas to look for tumor-infiltrating lymphocytes, which the scientists called a "potentially predictive immunotherapy biomarker."

Nucleai has a database including more than 20 million slides — compiled from its partners — as well as genomic test reports, radiology images, clinical records, and outcomes data. This allows the company to augment randomized clinical trials with real-world data. This represents one part of the firm's value proposition, according to Veidman, with the other being the maturity of its AI and image analysis technology that dates to the CEO's military days.

This month, the company announced that it had raised $6.5 million in the initial stage of a Series A funding round. The remaining part of the round should close by the end of the year, Veidman said.

Nucleai has brought in $11.5 million to date, including a $5 million seed round in 2018. 

Swiss biopharmaceutical company Debiopharm led the Series A initial stage, with participation from previous investors Vertex Ventures and Grove Ventures. As part of this new investment, Debiopharm will collaborate with Nucleai on unspecified efforts to accelerate the discovery of biomarkers for oncology drug response.

Nucleai now has a team of about 30, including employees and consultants. Veidman said that the new funding as well as the expected second phase of the Series A will allow the firm to expand both its R&D and business teams.

Nucleai currently only has a physical presence in Israel, but had been planning on opening a US office when the COVID-19 pandemic hit. Veidman said that the outbreak has, in some ways, increased the value of establishing an American base.

"[The pandemic] makes it even more urgent to open an office in the US, where you cannot fly to," he said. "So it makes sense that we'll have an office and a foot on the ground."

Veidman said that the company is eyeing the Boston area for its initial US presence, since it is a hub of pharma and biomedical research.

While only the relationship with Debiopharm has been announced, Veidman said that Nucleai has additional "commercial collaborations" with several large pharma companies that he declined to name. The firm also has an undisclosed health insurer in the US as a client.

"Payors have a lot of interest in treatment decisions regarding immunotherapy," Veidman noted, because expensive cancer immunotherapies often have a low efficacy rate.

"Payors are spending a lot of money on drugs that the patients will not respond to," Veidman said. "They would prefer that the right decision will be made and the right immunotherapy will be assigned to the patient."

The Nucleai CEO said that the company's pharma relationships could progress to development of companion diagnostics or biomarkers, perhaps in the next two years. There is no specific timetable for moving into other disease states.

Eventually, Veidman wants Nucleai to become a major player in treatment decisions based on the combination of diagnostic information and predictive modeling. "We should use all the information that's out there in order to get the best treatment decision that we can," he said.

The technology is adaptable to other areas of medicine, though oncology remains the primary thrust for the time being.

"The technology that we developed has platform capability. Biopsy analysis and histological analysis could be beneficial not only for oncology," Veidman said, specifically naming inflammatory bowel disease and nonalcoholic steatohepatitis.