Ataraxis AI Secures $4 Million Seed Funding to Launch Advanced Breast Cancer Diagnostic Test

Ataraxis AI (NY,US), a New York City-based AI-driven precision medicine startup, has emerged from stealth with $4 million in seed funding co-led by Giant Ventures (London, UK) and Obvious Ventures (CA, US).

Development of Ataraxis Breast and Promising Clinical Results

The funding, raised in 2023 but recently announced, has powered the development of Ataraxis Breast, the world’s first AI-native test for breast cancer, designed to improve patient outcome predictions and deliver more personalized treatments.

According to the company, this diagnostic tool is currently the most advanced clinically validated option available for breast cancer prognosis.

So far, has shown promising results in a multi-site study involving over 7,500 patients across 15 institutions including Gundersen Health System (WI, US); Karmanos Cancer Institute (MI, US); and NYU Langone Health (NY, US).

Published in a preprint by Cornell University (NY, US), the study shows that Ataraxis Breast not only achieves 30% higher accuracy in predicting breast cancer recurrence compared to standard diagnostics but also helps reclassify patients previously categorized as “intermediate risk.”

By providing more precise risk assessments, the test can identify patients who might safely avoid chemotherapy, ensuring they receive less aggressive treatments while still addressing their cancer effectively.

Kestrel and Advancements in Diagnostic Accuracy

To develop Ataraxis Breast, the startup used its foundational AI model for digital pathology – called Kestrel. The company claims that the Kestrel model uncovers complex, novel features beyond human interpretation, including by physicians, and applies across all disease types.

Ataraxis Breast uses AI to analyze digital pathology images, predicting cancer recurrence 30% more accurately than traditional tissue-based tests. Its Kestrel model identifies complex biomarkers, such as subtle cellular patterns and molecular signals in digital pathology images, that are strongly correlated with cancer recurrence and treatment outcomes. Ataraxis says these insights, uncovered using deep learning, set it apart from standard diagnostics.

The AI-driven diagnostic test also works across all breast cancer subtypes and could benefit 100,000 additional patients yearly in the U.S., according to the company. Further testing has been planned to validate these claims.

Traditional molecular diagnostics, such as polymerase chain reaction and sanger sequencing, though standard in personalized oncology, can be limited due to their lower accuracy, lengthy development and high costs. These limitations stem from their reliance on predefined genetic targets, which reduces their ability to detect unexpected or novel mutations.

To address these limitations, Jan Witowski and Krzysztof Geras founded Ataraxis AI to create AI-driven diagnostic tests that predict cancer risks and outcomes.

The company has ambitious plans to develop diagnostic tests for at least 50% of the 26 million new cancer patients expected to be diagnosed globally by 2030.

“As a medical doctor, I was initially excited about the potential of precision medicine, having seen firsthand the urgent need for solutions to help hundreds of thousands of patients. Unfortunately, over the past decade, that promise has largely gone unfulfilled,” said Witowski.

“Building on our advisor Yann LeCun’s pioneering work in modern AI, we are establishing a completely new category of tests that will phase out molecular diagnostics, transitioning the industry to AI precision medicine—an approach that is more accurate, efficient, and cost-effective, with the potential to save countless more lives,” he continued.

“Ataraxis has the potential to revolutionize cancer treatment, pioneering in a new era of AI-driven precision medicine that will elevate patient care and outcomes,” added Madelene Larsson, principal at Giant Ventures.

Rohan Ganesh, partner at Obvious Ventures, also emphasized that tissue biopsies are a valuable source of medical data, noting that Jan Witowski and his team have created a groundbreaking multi-modal AI approach that enables doctors to identify treatment options beyond chemotherapy.