AI arrives in mammography in Santiago del Estero: the value lies in a second pair of eyes, not in replacing the doctor

By 385 Noticias · October 18, 2026.

The Diagnóstico Médico por Imágenes (DMI) center in Santiago del Estero, Argentina, has adopted Transpara, an artificial intelligence software for mammography. According to the local outlet, the system analyzes digital images and produces reports that flag suspicious findings such as nodules and microcalcifications, which are small calcium deposits that sometimes accompany early lesions.

The tool was presented during Pink October, the breast cancer awareness month. Dr. Samanta Boyanovsky, a specialist in diagnostic and breast imaging, was clear about its role: “The interpretation is up to me; the final diagnostic decision always rests with the specialist physician.”

According to the source, DMI already had Hologic direct digital mammography, high-resolution medical monitors, breast ultrasound, magnetic resonance imaging, multislice CT and PET-CT, as well as a system in which two doctors read each report. The specialist restated the usual guideline: an annual mammogram from age 40, or earlier screening by ultrasound if a doctor recommends it.

The article gives no performance data for Transpara, such as sensitivity, false positives or workload reduction, and cites no local validation studies. So we cannot judge how much it really improves care. We only know that it has been deployed and how the center describes it.

Our take: the design is more interesting than the algorithm. The AI is added to an existing double reading by doctors, so it serves as an extra safety check rather than a replacement. Breast screening is one of the areas where medical imaging AI has the most potential, because the task is repetitive, the volume is high and a human error caused by fatigue can have serious consequences. It is a case where expert judgment and the machine complement each other.

The location also matters. When tools like this reach a province in northern Argentina, not just the large hospitals in major cities, AI begins to reduce a real inequality: unequal access to diagnosis with good equipment. In the short term, fair questions remain: validation in the local population, cost to patients, procedures for when the algorithm and the radiologist disagree, and how to keep the machine's alerts from leading to overdiagnosis.

In the long term, this trend supports our core view: detection that keeps getting earlier, cheaper and more widely available can turn serious diseases into problems that are treated in time. Until then, the practical message is still the human one: go get screened.