AI improves vision of metastatic breast cancer

A new study promoted by the Daiichi Sankyo Alliance with AstraZeneca, and developed by IQVIA, has managed to offer a “more complete” understanding of metastatic breast cancer care in Spain, based on natural language processing (NLP), a branch of Artificial Intelligence (AI).

This tool has allowed extract pathological anatomy information and analyze data such as prevalence, tumor subtypes, comorbidities and progression times, including a complete detail of the care process, taking into account activities, treatments and times.

He CamON studio has shown that, of the 11,060 patients evaluated, a 10.5% (1,166) had metastatic disease, being the 99.3% of the cases women with an average age of 61.6 years to the diagnosis.

Of the total, almost half (47.6%) had de novo metastasis, while the 52.4% ended up developing metastases after an initial diagnosis of breast cancer, with a mean time to progression of 2.7 years.

After analyzing the tumor subtypes, the researchers have revealed that the 29.6% were HER2 negative/Hormone Receptor positive, the 27.5% HER2-low, the 15.6% HER2+ and the 4.3% triple negative. He 64.6% has been ER positive, and the 56.2% positive PR.

Likewise, it has been observed that andl 22.9% of patients did not have information about their tumor subtype, a “key piece of information” for clinical decision-making, although the researchers have recognized that this figure could be due to a limitation of the natural language processing methodology used.

“This study highlights the importance of correct identification of tumor subtypes as a basis for implementing personalized therapeutic strategies that optimize the management of metastatic breast cancer,” said the lead author of the study, Dr. Luis Manso.

The medical oncologist at the 12 de Octubre University Hospital in Madrid has considered that these data offer a “real and updated vision” of breast cancer in Spain, and that they provide “valuable information to guide future therapeutic strategies and lines of research that drive improvements” in the multidisciplinary approach of the disease.

Likewise, it has been pointed out that the underrepresentation of the HER2-low subtype reflects both methodological limitations and its progressive incorporation into clinical practice during the study period.

Most frequent locations of metastasis

Among the most frequent locations of metastases are the bone, with the 51.2% of cases, followed by the liver (35.2%) and the respiratory system (33.9%).

Meanwhile, the most common comorbidities have been hypertension (40%), dyslipidemia (32.3%) and diabetes (17.4%), showing the need for a multidisciplinary approach in these patients.

The data have also revealed that the average time from diagnosis to the start of treatment has been 35.4 dayswhile the time until surgery has been 43.8 days without neoadjuvant. Most surgeries have been conservative (54.7%), and the 27.1% They have been radical.

All these findings offer “key information” for optimize diagnoses, personalize treatments, identify patients eligible for clinical trials and plan therapeutic strategies based on comorbidities and location of metastases.

The research, presented during the National Congress of the Spanish Society of Medical Oncology (SEOM2025), has also generated evidence with a “direct impact” on participating hospitals, and has made it possible to consolidate a tool that offers dynamic information on each hospital, avoiding manual data entry and allowing clinicians to make the most of the recorded information.

This is a retrospective observational study of the electronic medical records of thousands of people with breast cancer treated in eight Spanish hospitals between 2016 and 2023, including 12 de Octubre, the Jiménez Díaz Foundation and the Torrejón Hospital, in Madrid; the La Fe Hospital, in Valencia; the Duran i Reynals Hospital and the Bellvitge University Hospital, in Barcelona; the Son Espases Hospital, in Palma de Mallorca; the Vinalopó Hospital, in Alicante; and Ribera Povisa, in Pontevedra.

“CamON reflects our commitment to innovation based on real data and collaboration between industry, hospitals and experts in advanced analytics such as IQVIA, we are transforming clinical information into useful knowledge to improve healthcare practice and move towards more personalized care in metastatic breast cancer,” stated the VP Head of Oncology at Daiichi Sankyo Spain, Ana Zubeldia.

For his part, the Healthcare director at IQVIA Spain, Carles Illaexplained that the use of methodologies based on real-world data, such as those applied in the CamON project, allows us to generate a “much more representative view” of what happens in normal clinical practice.

“Thanks to the application of natural language processing, it is possible to analyze a very high volume of electronic medical records. “This approach offers a more complete understanding of metastatic breast cancer care and reinforces the value of artificial intelligence as a tool to support clinical and health management decision-making,” he concluded.