# What happened to DIPCAN's data after €7.7 million in public funding?

> What happened to DIPCAN's data? The project received €7.7 million in public funding and generated more than 60 TB of biomedical information
> By Observatorio Blockchain · Finance · 2026-08-26

DIPCAN (Digitalization and Management of Personalized Medicine in Cancer), an observational study carried out between 2021 and 2024 by a consortium of seven entities, received €7.7 million in public funding to integrate clinical, genomic, anatomopathological and radiomic data from patients with metastatic cancer and to develop artificial intelligence models.

The funding was granted under the 2021 R&D Missions in Artificial Intelligence program, through Spain's Ministry of Economic Affairs and Digital Transformation and the Secretary of State for Digitalization and Artificial Intelligence, within the Recovery, Transformation and Resilience Plan. The project also received funding from the European Union through NextGenerationEU, under the Recovery and Resilience Facility. The Government puts DIPCAN's budget at €7.7 million.

One of the project's main known results is a huge biomedical information infrastructure, which was also used to train artificial intelligence models. In September 2025, Genomcore, one of the participating companies, stated that DIPCAN had harmonized more than 60 terabytes of data from 130 Spanish hospitals, generated reports for more than 2,000 patients and trained artificial intelligence models to move toward personalized diagnosis.

However, as of today, the public information reviewed by Observatorio Blockchain does not make it equally clear who currently governs that dataset, which entities have access to it, under what conditions it may be reused or who may use it to develop and train new artificial intelligence models.

This does not mean the data are being misused. The question is what happens, once the initial funding has ended, to one of the main scientific assets created by a project financed with public money and built with information provided by cancer patients.

## A study with patient data

The Government defined [DIPCAN](https://digital.gob.es/content/dam/portal-mtdfp/ministerio/servicios-a-la-ciudadan%C3%ADa/procedimientos-y-servicios/digitalizacion-e-inteligencia-artificial/becas,-ayudas,-subvenciones-y-premios/2409016/20211220_Listado_PE_numero_de_expediente.pdf) as an observational study aimed at integrating information from 2,000 patients with non-hematological metastatic tumors. Its purpose was not to provide treatment within the study itself, but to carry out diagnostic tests, integrate the results and use them to generate algorithms. The MD Anderson Cancer Center Spain Foundation, the study's sponsor, explains that the recruitment phase ended on October 31, 2024. Since then, it says, the integration of the different types of data is being carried out with the aim of developing AI models for personalized medicine.

The European Commission states that around [2,000 patients from across Spain](https://spain.representation.ec.europa.eu/empresas-y-financiacion/nextgenerationeu/dipcan-un-estudio-pionero-de-lucha-contra-el-cancer-apoyado-por-la-union-europea_es) voluntarily participated in DIPCAN. Participants thus provided the project's fundamental raw material: information about their disease that could be linked to genomic, anatomopathological and imaging data.

### DIPCAN was developed through a consortium of seven entities with specialized roles.

Eurofins Megalab, a company dedicated to clinical analysis and diagnostics, coordinated the project and carried out the clinical analyses and the genomic and anatomopathological studies. The MD Anderson International Spain Foundation, dedicated to cancer research, was the study's sponsor and provided its scientific leadership. Pangaea Oncology, specialized in precision oncology and molecular diagnostics, worked on the analysis and interpretation of liquid biopsies and tumor genetic information.

Quibim, specialized in artificial intelligence applied to medical imaging, analyzed the radiomic information from MRI scans. Atrys Health, a medical services and precision diagnostics company, had among its duties the analytical and clinical validation of the AI platform's Big Data and quality control of the genomic data.

Artelnics, specialized in artificial intelligence and machine learning, was to generate the machine learning models from the collected data. And Genomcore, a Spanish technology company specialized in biomedical information and personalized medicine, was entrusted with biomedical data governance and bioinformatic analyses. It is precisely this last role that becomes especially important once the project's initial period has ended.

## More than 60 TB from more than 130 hospitals

Genomcore currently presents DIPCAN as one of the main use cases of its technology. According to the company, the project has harmonized more than 60 TB from more than 130 hospitals and integrated more than 29 different types of healthcare data sources. The technology infrastructure used is Genomcore BIMS (Biomedical Information Management System).

In June 2025, the company explained that with BIMS it had built for DIPCAN a multimodal repository that unifies clinical, molecular and imaging data to train artificial intelligence models. Later, in September, Genomcore received recognition from the European PRECISEU project in the Impact category precisely for its management of biomedical data in DIPCAN. In the information published on the occasion of the award, the company cited the more than 60 TB harmonized, the 130 hospitals, the reports for more than 2,000 patients and the training of AI models.

Genomcore continues to present DIPCAN in 2026 as one of its reference projects in biomedical data management. All of this supports the conclusion that Genomcore has played a central technological role in integrating and managing the information.

## Who then governs the data?

However, it is important to stress that managing data technologically does not necessarily mean owning it, being its legal controller or having the capacity to decide unilaterally on its subsequent uses. The public information reviewed by [Observatorio Blockchai](https://observatorioblockchain.com/ia/por-que-ias-terminaran-financiando-periodismo/)n does not allow those powers to be attributed to it.

That is the main question DIPCAN leaves open. The project's public pages show who took part, what role each partner had, what technology was used and what volume of information came to be processed, but it is considerably harder to determine the current governance rules for the dataset.

For example, who currently decides who can access DIPCAN's information? The MD Anderson Foundation as the study's sponsor, the consortium, or is there a specific committee responsible for authorizing access? It is also worth asking what role the hospitals where the data originated play and what capacity Genomcore has as the provider of the technology infrastructure used to manage them.

## Can the data be used to train new AIs?

Another question is whether the other companies participating in DIPCAN can continue to access that information and under what conditions, or whether researchers, universities, hospitals and scientific centers that were never part of the consortium can request access to develop new research. These questions are different from knowing where the files are technically hosted. The important question is who currently has the authority to permit their use, for what purpose and under what conditions.

The issue becomes even more important because DIPCAN was conceived precisely to develop artificial intelligence. Genomcore states that its multimodal repository is used to train AI models, so we are not talking only about preserving information from a completed research project. The data may continue to have scientific and technological value as raw material for developing new algorithms.

## Several open questions

This raises several questions. Can Artelnics, Quibim, Genomcore, Atrys or the other consortium entities continue to use the data to develop new artificial intelligence models? Is their use limited to DIPCAN's original objectives, or can they also be used in subsequent research? It is also relevant to know whether external entities, such as a pharmaceutical company, a public university, a hospital or a research center, can request access and what procedure they would have to follow to obtain it.

If the data can be used to develop technologies that later reach the market, the question takes on an additional dimension. Can they be used to train models intended for commercial products and, if so, who can do so and under what conditions? The public documentation located does not allow these questions to be fully answered.

The fact that DIPCAN was financed with public funds does not mean patients' medical data should be public. On the contrary, as healthcare, genetic and cancer-related information, its processing and use are subject to special privacy and protection safeguards.

## Knowing the conditions of patients' consent

Scrutinizing the project therefore does not mean calling for the publication of its more than 60 TB of information, but rather knowing how that data is governed: who holds it in custody, who authorizes access, which researchers or entities may use it, for what purposes it may be reused and what limits exist on doing so.

A fundamental part of those answers should be found in the participants' informed consent. The approximately 2,000 patients voluntarily participated in the study, and the European Commission itself highlighted their altruistic contribution when presenting DIPCAN as an example of projects funded through NextGenerationEU.

Knowing the conditions of that consent would make it possible to determine what subsequent uses the patients authorized: future research, development of new algorithms, participation in other scientific projects or access by third parties and, in each case, under what conditions. It would also make it possible to know how long the responsible entities may retain the data and what limits participants set for possible commercial development.

## From data to products

In the public documentation reviewed, we did not locate DIPCAN's complete informed consent form. Therefore, we cannot know precisely what subsequent uses the participants authorized or how far that consent allows their data to be reused once the study's initial phase has ended. DIPCAN brought together companies specialized in genomics, artificial intelligence, molecular diagnostics, medical imaging and biomedical data management. That private participation is consistent with one of the goals of public R&D: turning research into useful technology and products.

Precisely for that reason, to assess the return on the €7.7 million in public funding it is relevant to know what happened to the project's results. If the data have served or can serve to develop algorithms, products, patents or licenses, it matters to know who owns those developments, under what conditions they are exploited and whether there is any return for the scientific or healthcare system that helped generate them.

## Assessing the legacy of the €7.7 million invested in DIPCAN

DIPCAN has generated infrastructure and artificial intelligence models and, according to Genomcore, has harmonized more than 60 TB of information from more than 130 hospitals. That dataset can continue generating scientific and technological value long after the initial funding has ended. That is why the central questions have to do with who governs that data today, who can access it, what it can be reused for and under what rules.

The public information makes it possible to reconstruct how they were generated. What it does not make equally clear is what happens to them after the grant. That is one of the key aspects for assessing the legacy of the €7.7 million invested in DIPCAN.
