Research shows a 12% drop in interval cancers with the use of artificial intelligence in screening
The implementation of advanced algorithms in public health protocols has demonstrated promising results in the early detection of breast pathologies. A large randomized clinical trial carried out at Suécia, called the MASAI trial, revealed that technological support in imaging exams can significantly reduce the incidence of diagnoses made between regular screening cycles.
The study, coordinated by Lund University, involved a significant cohort of more than 100 thousand women. The data indicates that the technology not only helps identify tumors at early stages, but also demonstrates high sensitivity for more aggressive biological types. The integration of these tools aims to optimize the accuracy of medical reports without compromising patient safety.
Experts observed that the application of artificial intelligence in screening programs maintained the specificity of exams at high levels. Direct comparison with the traditional double reading method by radiologists showed consistent benefits, covering different age groups and breast density profiles, which suggests a universal applicability of the methodology.
So-called interval cancers, which appear in the period between one mammogram and another, tend to present more unfavorable clinical characteristics and complex prognoses. The reduction in these cases, as highlighted by the research, represents a crucial advance for oncology and public health strategies, enhancing the chances of less invasive treatments and greater survival.
Detailed statistics from the MASAI study
Analysis of the data collected during the clinical trial brought to light shocking numbers about the effectiveness of assisted screening. Houve an overall reduction of 12% in cases of interval cancer in the group undergoing technological intervention. Além In addition, there was a 16% drop in the late detection of invasive tumors, a determining factor for therapeutic success.
Another highlight was the 21% decrease in diagnoses of large tumors, classified as T2 or higher. The technology also proved to be efficient in identifying specific molecular subtypes, recording 27% fewer cases of non-luminal A tumors, which are biologically more aggressive and fast-growing.
In the direct comparison of occurrences, the intervention group recorded 82 interval cancers, while the control group, submitted only to the conventional method, recorded 93 cases. Essa statistical difference reinforces the hypothesis that the hybrid approach, combining human expertise and computational capacity, offers an extra layer of protection to patients.
Classification methodology and workflow
The operation of the technology is based on the automated analysis of mammographic images, where the system assigns a risk score on a scale ranging from 1 to 10. Essa stratification allows the efforts of the medical team to be directed more strategically and efficiently within diagnostic centers.
Exams classified as low risk are sent for a single reading performed by a radiologist. On the other hand, images that receive a high score, indicating a greater likelihood of anomalies, must undergo a double assessment, ensuring rigorous scrutiny where it is most needed.
The system also works by highlighting suspicious regions in the images, serving as an immediate second opinion to support clinical decisions. During the study, diagnostic sensitivity in the AI-assisted group reached 80.5%, surpassing the 73.8% observed in the control group, while specificity remained stable at 98.5% for both.
Impact on radiologists’ routine
In addition to the direct clinical benefits for patients, the introduction of artificial intelligence has generated positive impacts on the operational management of radiology departments. The study documented a 44% reduction in the burden of reading exams for professionals, freeing up valuable time in the medical schedule.
Leia Também
This workflow optimization is vital in a scenario where the volume of exams in population programs is growing. With the automated triage of lower-risk cases, specialists can dedicate more attention and time to analyzing complex cases and multidisciplinary discussions on therapeutic approaches.
The technology has demonstrated the ability to alleviate the demand on healthcare systems without increasing false positive rates. The balance achieved avoids the generation of unnecessary anxiety in patients and reduces the need for invasive procedures, such as biopsies, in situations where there is no confirmed malignancy.
Clinical results and patient profile
The effectiveness of the model was tested on a group with a median age of approximately 54 years, reflecting the target audience of the tracking campaigns. The majority of tumors identified early with the aid of AI were small in size and lacked lymph node involvement, essential conditions for a favorable prognosis.
The detection rate in the screening stage was 81% in the group with technological support, compared to 74% in the conventional method. The tool’s superior sensitivity remained consistent regardless of breast tissue density, even benefiting women with dense breasts, which have historically represented a challenge for traditional mammography.
Professionals involved in the study reported greater confidence in the analyses, as the tool easily integrates with existing hospital IT systems. Diversas regions of Suécia have already started to adopt this model in their daily routine, requiring only software updates compatible with the digital mammography equipment in use.
Perspectives for global health
The success of the MASAI trial positions artificial intelligence as an indispensable ally in the modernization of healthcare systems. Países Europeans are already evaluating the implementation of similar protocols, aiming to mitigate the shortage of specialists and expand the coverage of breast cancer screening programs.
Considering that the disease records millions of new cases annually around the world, the ability to increase early detection rates is essential. Quando identified in the early stages, cure rates can exceed 90%, transforming the prognosis of the disease and reducing long-term mortality.
Despite the optimistic results, the researchers emphasize the need for constant human supervision. Combining algorithmic sensitivity with clinical judgment represents the most promising model for the future of radiology, ensuring responsible and ethical adoption of technological innovation.
