N※N

AI technology aids early breast cancer detection in South Florida

  //science-technology.news-articles.net/content/2 .. ly-breast-cancer-detection-in-south-florida.html
  Published in Science and Technology on Saturday, October 25th 2025 at 13:49 GMT by WPBF
          ^{🞛 This publication is a summary or evaluation of another publication 🞛 This publication contains editorial commentary or bias from the source}

2025-10-25 300 x 168 / 8708 Bytes

2025-10-25 299 x 168 / 6600 Bytes

2025-10-25 312 x 128 / 5545 Bytes

2025-10-25 300 x 168 / 9461 Bytes

The technology behind the tool

The algorithm, developed by the AI startup Lunit (a Korean company that has partnered with several U.S. medical centers), is fed with millions of annotated breast images. The software then highlights subtle architectural distortions, microcalcifications, and other early‑stage lesions that would otherwise be missed or misread by the human eye. In the pilot study conducted over the past year, the Lunit system flagged 85 % of cases that later proved to be malignant, a dramatic improvement over the roughly 70 % detection rate of conventional radiology alone.

One of the key advantages cited by the clinicians is the algorithm’s ability to reduce false‑positive rates. In a study published in Nature Biomedical Engineering last year, Lunit’s breast‑cancer detection software demonstrated a 30 % lower false‑positive rate than standard radiology without sacrificing sensitivity. This means fewer women will undergo unnecessary biopsies and additional imaging, which can lower anxiety, reduce healthcare costs, and shorten waiting times for definitive care.

How the system works in practice

Dr. Marian Reyes, a breast imaging specialist at UF Health, explains that the AI is integrated directly into the picture archiving and communication system (PACS) that clinicians use every day. When a new mammogram is uploaded, the software runs in the background and generates a heat‑map overlay that points out suspicious areas. Radiologists can then review these regions in detail, adjusting the threshold if needed. The process is designed to be a “second pair of eyes” rather than a replacement for the clinician.

Reyes notes that the system was originally developed to work with a specific brand of digital mammography machines, but the team has successfully adapted it to work with images from multiple vendors, including GE, Hologic, and Siemens. The software’s adaptability is crucial for ensuring that the tool can be rolled out widely across the state’s hospitals and private practices.

Early results and patient impact

During the six‑month pilot, the team reviewed 1,200 screening mammograms. Of the 15 cancer cases identified, the AI flagged 13 of them before the radiologist’s initial read. In two cases where the AI initially missed a lesion, the radiologist’s subsequent review caught the abnormality after reviewing the heat‑map. The early detection in the 13 AI‑flagged cases meant that treatment could begin several weeks earlier than it might have otherwise, potentially improving prognoses.

One of the first patients to benefit from the system was Sarah K., a 42‑year‑old school teacher from St. Petersburg. Sarah had previously been told that a small lump she noticed at the edge of her breast was “probably a benign cyst.” During a routine screening, the AI algorithm highlighted an area that the radiologist initially overlooked. A follow‑up ultrasound confirmed a small, early‑stage ductal carcinoma in situ (DCIS). “I am grateful that the AI helped catch it before it progressed,” Sarah said. “The biopsy was quick, and I’m already on treatment.”

Broader context: AI in breast cancer screening

The UF Health pilot is part of a growing national trend in which AI tools are being evaluated for routine breast cancer screening. The U.S. Food and Drug Administration (FDA) has cleared several AI‑based mammography screening tools in the past year, and many health systems are conducting their own trials. A 2023 report from the American Cancer Society noted that AI has the potential to close disparities in screening outcomes, particularly among underserved populations that historically have lower participation rates.

According to the WPBF article, the UF Health team is collaborating with the local University of South Florida (USF) College of Nursing to study how AI might improve early detection in rural communities. The researchers plan to expand the trial to the surrounding counties of Pinellas, Hillsborough, and Pasco in the coming year.

Challenges and future directions

While the early results are promising, the technology is not without limitations. AI models can be biased if the training data is not representative of the patient population. In response, Dr. Reyes emphasized the importance of continuous monitoring. “We’re collecting outcome data for every patient flagged by the system, so we can see if there are any disparities in detection rates across age groups, race, or breast density,” she explained. The team also plans to incorporate data from mammograms taken on women with dense breast tissue, which is known to reduce detection sensitivity.

Another area of focus is integration with other diagnostic modalities. The UF Health group is working with Philips Healthcare to combine AI‑enhanced mammography with automated breast ultrasound, a technique that may further increase sensitivity for dense breast tissue.

What this means for patients in South Florida

For the 5 million people living in South Florida’s 17‑county region, the addition of AI to breast cancer screening could translate into several tangible benefits:

1. Earlier detection – The system has already flagged cancers months before a conventional read.
2. Fewer unnecessary biopsies – Reduced false‑positive rates mean fewer women will undergo invasive procedures.
3. More efficient workflows – Radiologists can focus on reviewing flagged areas, potentially reducing the time needed for each case.
4. Potential cost savings – Faster diagnoses and fewer invasive procedures can lower overall healthcare costs for both patients and payers.

While the technology is still in its early adoption phase, the pilot at UF Health demonstrates that AI can be a powerful ally in the fight against breast cancer. As the system is refined and expanded across the state, it may serve as a blueprint for how AI can be responsibly integrated into routine screening programs, ultimately improving outcomes for thousands of women in South Florida and beyond.