Sarasota Memorial Health Care System’s Brian D. Jellison Cancer Institute has delivered one of the strongest early-detection records for lung cancer anywhere in the country.
In 2025, the team diagnosed 53 lung cancers through its program and caught 75 per cent of them at Stage I or II — the most treatable stages. The hospital revealed this achievement on Apr. 14.
From 2022 to 2025 clinicians evaluated more than 9,000 patients and identified 144 lung cancers, with 67 per cent caught early, according to SMH.
These results stand out sharply against a tough backdrop. Throughout Florida, only 26 per cent of lung cancers are diagnosed at an early stage. This falls below the national 28.1 per cent mark and leaves 26.7 per cent of patients with no treatment at all.
Late-stage disease drops five-year survival to about 10 per cent while early detection lifts it to 65 per cent or higher. It can potentially hit nearly 90 per cent for Stage I.
By more than doubling the state average, Sarasota Memorial is proving that hospitals can save lives with the application of smart technology.
Read more: Breath Diagnostics advances pre-op pneumonia screening with FDA breakthrough designation
Hospital builds success with AI, a dedicated clinic, and national honours
Sarasota Memorial launched its low-dose CT screening program in 2016 and steadily grew participation through community education. Leaders soon realized screenings alone miss too many people, especially those outside strict eligibility rules or with no obvious risk factors.
They integrated Eon’s AI platform across all facilities to scan more than 430,000 radiology reports each year in real time. The system automatically flags incidental pulmonary nodules — suspicious spots spotted on scans ordered for unrelated issues like heart checks or accidents — then enriches the data with patient risk factors and alerts the lung team. Referrals jumped from one or two incidental cases per week to an average of 170.
Clinicians route every flagged patient into a dedicated lung-nodule clinic that follows evidence-based guidelines, holds weekly multidisciplinary reviews and keeps follow-up on track to avoid unnecessary tests.
Program coordinator Amie Miller, who has led the effort since 2015, says the AI finally let them reach people who “were missing the boat” and walking around unaware. Thoracic surgeon Dr. Blair Marshall noted that the multidisciplinary approach keeps care efficient and patient-focused.
This week’s announcement follows other significant accolades late last year. During Lung Cancer Awareness Month in November, the team earned two new national Center of Excellence designations from the non-profit GO2 for Lung Cancer. One was for overall cancer care and the other for incidental pulmonary nodule management.
Other promising screening innovations attract attention in the U.S.
Sarasota Memorial’s model joins a wave of creative approaches that aim to catch lung cancer sooner and reach more people.
Moffitt Cancer Center has deployed Florida’s first mobile low-dose CT unit, driving directly into underserved communities to boost screening rates where access and awareness lag. The unit screens up to 30 patients a day at events and workplaces, cutting travel barriers for high-risk residents.
Breath Diagnostics’ OneBreath platform asks patients for a single exhale, then uses patented microreactor technology to analyze volatile organic compounds in breath. Clinical trials show 94 per cent sensitivity for early-stage disease, delivering results in minutes without radiation or invasive procedures.
Additionally, the Median Technologies SA (EPA: ALMDT) (FRA: 4ZG) eyonis AI tool, recently FDA-approved, has been helping radiologists spot and characterize nodules on standard low-dose CT scans more accurately.
Advancements like these are expanding options beyond traditional screening and could give patients nationwide a better shot at early, lifesaving intervention.
Read more: Prestigious medtech intelligence firm recognizes Breath Diagnostics for innovation
Follow Rowan Dunne on LinkedIn
rowan@mugglehead.com
