2.4 million new diagnoses. Nearly 700,000 deaths. That is what a single year of breast cancer looks like worldwide. Every 13 seconds, another woman learns she has it. How her story ends often depends on time, and on a detail in an image just a few millimetres in size. An early-stage tumour can hide: in dense breast tissue it blends into its surroundings, and the only sign may be a subtle change that is easy to miss. So what happens when artificial intelligence looks at the image too? A real case from Germany offers an answer.

Breast cancer is the most common cancer in women, both worldwide and in Europe. Here, doctors diagnose it in roughly 540,000 women each year, and about 140,000 die of the disease. In the Czech Republic, there are more than 7,000 new cases every year, and approximately 1,700 women die of breast cancer. And yet it would take so little: every woman aged 45 and over is entitled to free mammography screening once every two years. However, only about 60% of women aged 45 to 69 attend regularly.

Petra Steyerová, MD, PhD, a radiologist specialising in breast imaging at Mammacentrum Zahradní Město, explains why catching the disease as early as possible matters so much:

“Breast cancer is most treatable and curable when it is detected at an early stage, ideally before it can be felt. Mammography screening makes it possible to detect breast cancer early, before the patient can feel the tumour herself or it shows itself in any other way.”

A subtle finding that made the difference

Breast cancer tends to be hardest to detect when it is hidden in dense breast tissue. On a mammogram, dense tissue can look similar to suspicious lesions, which makes any changes within it much harder to distinguish. “The biggest challenge usually isn’t finding the tumour, but telling overlapping breast tissue apart from a genuine pathological lesion,” says Dr. Steyerová.

A case from Germany, where the tumour was hidden in exactly this kind of dense tissue, shows this clearly. Carebot AI MMG marked a suspicious area on the image and alerted the radiologist to a region that deserved a closer look. Further examination then confirmed breast cancer. This is where AI can serve as an important second pair of eyes. It does not replace the radiologist’s expertise, but it helps them focus on small findings that might otherwise be difficult to recognise. Because behind every mammogram is a woman who deserves the most accurate result possible.

So that no lesion goes unnoticed: how Carebot AI MMG helps radiologists

Can artificial intelligence hold its own against experienced doctors when reading mammograms? That is exactly what clinical studies evaluating the diagnostic performance of Carebot AI MMG set out to examine. One of them included 222 mammography examinations, 48 of which had confirmed breast cancer. The AI’s results were compared with the assessments of six radiologists and, in the study dataset, showed comparable diagnostic performance.

According to Dr. Steyerová, AI already plays a role today, and will play an even greater one in the future, mainly by making radiologists’ work easier and faster. It helps identify images where the likelihood of a tumour is low, so the radiologist can spend less time on them. Conversely, it flags those that require more detailed analysis or further examination. At the same time, she adds that because of the complexity of mammographic findings, AI will not replace assessment by the human eye any time soon. The final decision always remains with the doctor.

Every image matters

Cases like this are more common than you might think. Many women have dense breast tissue, and a tumour can hide very well within it. That makes regular screening all the more important. And it helps all the more when the radiologist is not alone in reading the images. Artificial intelligence will not replace them, but it can help make sure nothing important is missed. Pink October is a good opportunity to take the first step and book a mammogram.