Uruguay's Hospital de Clínicas will bring AI to mammography: the real challenge is digitizing before automating

By Radio Montecarlo · October 19, 2026.

The Hospital de Clínicas will set up an artificial intelligence node to help diagnose mammograms. Its director, Álvaro Villar, announced the plan on October 19, World Breast Cancer Awareness Day. Villar said the hospital will also replace its analog mammography machines with digital equipment.

The director added that the node will be able to review mammograms already reported as normal, to check for any early risk of the disease. Villar also said the number of surgeries has increased and the wait between a suspected case and the operation has shortened. The source gives no figures, launch timeline, system vendor or technical details.

The announcement coincides with a fundraising campaign. The Uruguayan Football Association will auction four pairs of pink socks worn by the national team against India. The money raised will go to the hospital's Breast Cancer program. The auction will take place on October 29, both in person and online.

Our take: the most important part of this story is not the AI but the switch from analog to digital mammography machines. An algorithm can only analyze images that exist as data. Digitization comes first. It is less eye-catching and more expensive, and it often decides whether a clinical AI project works or never gets past the headline.

The idea of reviewing mammograms already reported as normal deserves attention. It is a kind of automated “second reading.” This type of support generally aims to have the system flag doubtful cases for the radiologist to re-examine, without replacing them. It is a sensible and prudent use: it makes use of an archive the hospital already has and requires no new tests for patients. But it is worth waiting to see how it is validated: on what data, with what false-positive rate, and who is accountable for each decision.

There is also a deeper point. In healthcare, AI performs better when it draws on complete, well-maintained records. A public hospital that digitizes today is laying the foundation for a data asset that will stay valuable for years. If it is well governed, with clear privacy rules and standards, this is a realistic way to bring high-quality early diagnosis to more women, including outside the major medical centers.

In the short term, the risk is the usual one: announcing technology before securing funding, staff training and independent evaluation. In the long term, the direction is encouraging. Detecting disease earlier, more consistently and at lower cost is one of the concrete ways AI can move us toward a medicine that turns serious illnesses into problems treated in time. The thing to watch is whether Clínicas publishes results, timelines and metrics for this node.