Oncological diagnostics is gradually entering a new industrial phase, where value no longer relies solely on genetic sequencing equipment or specialized laboratories, but increasingly on artificial intelligence models capable of extracting biological information directly from medical images. It is on this positioning that Moonlight AI AG, a young Swiss company specializing in image analysis applied to clinical diagnosis, is developing, which announces a Seed fundraising of 2.8 million euros.
Founded around expertise combining hematology, pathology and artificial intelligence, the startup develops software capable of transforming images of blood smears or cytological samples into usable diagnostic data. The objective is to allow laboratories and doctors to more quickly identify certain genomic biomarkers or pathological signatures associated with cancers, without systematically resorting to next generation genetic sequencing (NGS).
This approach responds to a structural limit of precision medicine. If targeted therapies rely increasingly on the genetic analysis of tumors, sequencing remains expensive, complex to deploy and often slow in hospital environments. Moonlight AI seeks to circumvent this barrier by using computer vision to detect molecular information directly from images already produced in daily clinical workflows.
“Our technology allows laboratories to generate actionable and immediate results from slides they already use in their basic workflows,” explains Christian Ruiz, CEO and co-founder of Moonlight AI. “By removing the need for expensive equipment or manual processes, we are empowering laboratories to increase their diagnostic capacity and deliver results to patients faster. »
Technically, the company is part of the rise in power of AI models applied to digital pathology. In recent years, several academic studies have shown that it is becoming possible to infer certain genetic mutations or tumor characteristics from histopathological images, opening the way to a new generation of AI-assisted diagnoses. Moonlight AI is now trying to translate these scientific advances into clinical infrastructure that can be used on a large scale.
The startup is currently focusing its developments on several pathologies with high diagnostic complexity, including myelodysplastic syndromes (MDS), non-small cell lung cancer and chronic lymphocytic leukemia. Part of the financing should also make it possible to accelerate regulatory work and the marketing of the solutions developed.
One of the company’s key strategic assets is its proprietary database. Moonlight AI is currently building a library combining cytopathological imaging and genomic data, with the ambition of constituting one of the first datasets of its type internationally. In the field of medical AI, the quality and diversity of data are becoming a determining factor, both for training models and for their future clinical validation.
“By collaborating with an international consortium of clinical partners, we are building a dataset designed to ensure models are robust in real-world laboratory environments and across diverse patient populations,” said Nicole H. Romano, CTO and co-founder of Moonlight AI.
The startup is also looking to expand this consortium to increase the clinical and geographic diversity of the data used to train its models. “The success of AI-based diagnostics fundamentally depends on the quality and diversity of clinical data,” emphasizes Dr. Stefan Habringer, Chief Medical Officer and co-founder of Moonlight AI. “We are therefore opening our consortium to other hospitals and laboratories wishing to help shape the next generation of diagnostics. »
The funding round was co-led by Lotus One Investment, VP Venture Partners and MEDIN Fund, with the participation of N&V Capital as well as historic investor QAI Ventures.
