EIC Accelerator Report Card: Fimmic's AI Cancer Pathology

The company the European Commission listed as Fimmic Oy has traded as Aiforia Technologies since 2019 and has been listed on Nasdaq First North Growth Market Finland since 10 December 2021. In the EIC SME Instrument Phase 2 cut-off of 5 June 2019, it won a grant for AiforCancerDX, a project to bring its deep learning image analysis software for pathology from research use into clinical cancer diagnostics, according to the Commission's list of selected companies, which names no other beneficiary for the project. CORDIS records an EU contribution of €2,051,000 towards a total cost of €2,930,000, from 1 November 2019 to 30 April 2022.

The project promised two certified AI models, one for lung cancer and one for prostate cancer. Both exist: the lung cancer PD-L1 model received its CE-IVD mark in November 2021 and the prostate cancer Gleason model in May 2022, four weeks after the project's end date, according to Aiforia and Tissuepathology.com. The business built on them is still small: group revenue was €3.54 million in 2025 and the net loss €12.67 million, according to the company's 2025 financial statements bulletin.

One Company, Two Names

Fimmic was founded in 2013 by the researchers Johan and Mikael Lundin and the entrepreneur Kari Pitkänen as a spin-off of the Institute for Molecular Medicine Finland (FIMM), according to the University of Helsinki. The Finnish trade register shows the name Aiforia Technologies Oy registered on 15 January 2019, with Fimmic kept as an auxiliary name, and the conversion to a public limited company, Aiforia Technologies Oyj, on 19 July 2021. The Commission's June 2019 list still printed the old name and the old website, www.fimmic.com, which now redirects to another domain. Global Venturing also describes Aiforia as founded in 2013 as Fimmic. CORDIS lists the grant holder as Aiforia Technologies Oyj in Helsinki.

The Technology: Deep Learning on Scanned Tissue Slides

Pathologists diagnose cancer by looking at stained tissue under a microscope and scoring features such as how many tumour cells carry a biomarker, or how far a prostate tumour has lost its normal gland pattern. That scoring is manual and varies between observers. Aiforia's software runs in the cloud on digitised whole-slide images: a deep learning model finds the tumour areas and calculates the score, which the pathologist then reviews. The company also sells Aiforia Create, a tool with which research users train their own image analysis models, and that research business was all it had when the EIC project began. The project's final report says the company "did not have any software solutions or deep learning AI models validated and certified for diagnostic use" at the start.

One peer-reviewed validation came out of the project. A study in Histopathology (March 2022), by Liesbeth Hondelink, Sami Blom, Danielle Cohen and co-authors, developed and validated a supervised deep learning algorithm for the PD-L1 tumour proportion score in non-small cell lung cancer, and CORDIS lists it as the project's publication. The final report does not give accuracy figures for the certified products, so the claim of "unprecedented accuracy and consistency" in the project summary could not be checked against published head-to-head data.

What the EIC Project Promised, and What the Record Shows

The objective on CORDIS was to introduce the Aiforia Cloud platform to clinical use and "build the world's first ready-made deep learning algorithms for automated cancer diagnostics", for lung and prostate cancer. The final report covering November 2020 to April 2022 says the company set up an in-house validation and CE-IVD documentation process, certified the PD-L1 lung cancer and Gleason prostate cancer algorithms, built a clinical version of the platform and strengthened its sales force. Outside the project, it also launched three CE-IVD marked breast cancer models and a CE-IVD marked clinical viewer in the same period.

The external record matches the report. The PD-L1 lung model was CE-IVD marked in November 2021, according to Aiforia and Tissuepathology.com. The Gleason prostate model followed on 27 May 2022, bringing the portfolio to five CE-IVD marked clinical models, according to Aiforia and Tissuepathology.com. That was four weeks after the project's official end on 30 April 2022, a small delay against a 30-month plan.

1. Certified lung cancer AI modelDeliveredPD-L1 model CE-IVD marked, November 2021.
2. Certified prostate cancer AI modelDeliveredGleason model CE-IVD marked, May 2022, four weeks after the end date.
3. Clinical version of the platformDeliveredCE-IVD marked clinical viewer, per the final report.
4. Clinical uptakePartialHospital customers in Finland, France and the US; clinical revenue still small.
5. Become "the global leader" in tissue analysisUnverifiedNo independent market share data found.

What Came After the Grant

The money that followed the EIC grant was much larger than the grant itself. In June 2021 the company closed a first Series B tranche of €12.6 million led by Epredia, with Aktia Nordic Micro Cap, according to Global Venturing. Its initial public offering followed, with trading from 10 December 2021, gross proceeds of about €30 million if the over-allotment option was fully exercised and a market value of about €130 million at the offer price, according to Krogerus, the company's legal adviser, and Borenius, adviser to the joint global coordinators.

Since then the product range has grown. In February 2025 Aiforia obtained certification under the EU In Vitro Diagnostic Regulation (IVDR) for a set of breast and prostate cancer AI solutions, and it ended 2025 with eleven CE-IVD marked products, according to its 2025 bulletin and Tissuepathology.com. The bulletin names Fimlab Laboratories in Finland, the AP-HP hospital group in Paris, a Mayo Clinic collaboration in the US, a sales partnership with Siemens Healthineers and an extended preclinical contract with the drug maker Orion. The 2023 bulletin describes an exclusive licence from the Mayo Clinic for a colorectal cancer recurrence risk model.

Revenue has risen every year from 2022 to 2025, by 24% in 2025 alone, but the net loss has stayed at €11.9 million to €12.9 million a year, according to the bulletins for 2023, 2024 and 2025. Cash fell from €24.7 million at the end of 2022 to €9.5 million at the end of 2025. 2026 then started badly: on 21 July the company issued a profit warning, and its half-year report showed revenue down 46% to €753,000, an operating loss of €7.62 million and an order book of €3.46 million, which it blamed on delayed revenue recognition and customer agreements, according to Placera and Inderes.

Fresh money has come in each time. A directed share issue raised €4.2 million in December 2025 and another about €6.4 million in June 2026, at €1.772 per share, according to Inderes and Placera, against an IPO price of €5.01. On 31 July 2026 the company signed a venture debt facility of up to €20 million with the European Investment Bank, paid in three tranches of €5 million, €7 million and €8 million, each tied to revenue and other targets, according to Inderes and MFN. A search of all pages of the EIC Fund portfolio found neither Aiforia nor Fimmic; the 2019 award was grant only.

Other EU Projects

CORDIS lists Aiforia in two further EU research projects as a partner: DECIDER, on chemotherapy resistance in high-grade serous ovarian cancer (February 2021 to July 2026, €400,000 for Aiforia), and Thera4Care, a theranostics project running from October 2024 to September 2029 (€300,000 for Aiforia).

Why Europe Should Care

Earlier and more consistent cancer diagnosis is a core aim of Europe's Beating Cancer Plan, and pathology is a bottleneck where many laboratories lack specialists. Aiforia is a European supplier of clinical pathology AI with products certified under the In Vitro Diagnostic Regulation, and as a medical AI supplier it falls under the high-risk rules of the EU AI Act. The EU has now backed it twice, through the 2019 grant and the 2026 EIB facility. Its 2025 bulletin itself notes that in France "European AI players" are considered particularly relevant for health AI, a point about strategic autonomy in clinical software.

The Verdict

Aiforia and the EIC have not been asked for comment for this article. It is based entirely on public records and published statements, linked throughout.