In New Orleans, shoppers often request “lagniappe,” pronounced lan-yap, to express their desire for something extra, the baker’s dozen, a 13th donut. Added benefit was a major topic of interest for the 500 participants in the combined International Early Lung Cancer Action Program (IELCAP), International Early Lung Cancer and Respiratory Tract (IELCART), and AGILE 3 conferences held March 19 to 21, 2026, in Gdansk, Poland. Although this topic ranged freely across benefits derived from multiple incidental findings on low-dose, non-contrast computed tomography (LDCT) scans, screening tests for lung cancer (CTLCS), there was a strong focus on “opportunistic screening” for atherosclerotic coronary artery disease (CAD).

Although CAD is the foremost contributor to mortality in the U.S. (one death every 34 seconds), no population screening exists. We do assess risk factors for CAD, including cigarette smoking, obesity, hypertension, diabetes, and hyperlipidemia, but offer no screening test for CAD, per se, other than a medical history of symptoms suggesting myocardial ischemia. Current Medicare guidelines specify that, only when the patient is symptomatic, is an electrocardiogram (ECG), an echocardiogram, a stress test, or a coronary angiogram indicated. A core weakness is that the first symptom of CAD is often unstable angina, myocardial infarction, or sudden death.

The prognostic value of coronary artery calcium

In the 1990s, coronary artery calcium (CAC) measurement, derived from gated CT scans, was established as substantially better at quantifying risk of future CAD than risk scores based upon clinical parameters, e.g., from the Framingham study. Current guidelines state that “Coronary artery calcium is the single most reliable predictor for long-term, cause-specific mortality.”

Over time, multiple studies demonstrated that CAC scores, from non-gated LDCT, provide results comparable to formal Agatston scores (AS), but neither method is approved for population screening or insurance coverage.

In Gdansk, speakers provided data from multiple large studies showing that CAC scores from LDCT accurately predict future risk of CAD mortality. There was general agreement that presence of any CAC identifies increased risk and that risk escalates with increasing CAC scores. Consideration of prescription of statins and anti-hypertensive therapy, where appropriate, was advised.

Shared benefits in lung cancer screening

Multiple investigators demonstrated that a substantial majority of CTLCS-eligible adult smokers have CAC scores above zero and many have scores more than 1,000. Fewer than half are currently treated with statins or antihypertensives. This helps to explain why almost as many individuals in CTLCS programs die from CAD as from lung cancer (LC) itself and why overall and CAD mortality is also improved.

The identification in an individual eligible for CTLCS thus offers a serendipitous opportunity to detect the presence, location, and extent of atherosclerotic plaques in the coronary arteries of asymptomatic persons, who are not otherwise eligible for CAC studies.

In Gdansk, multiple centers in the U.S., Europe, and Asia reported experience with CAC detection and reporting in CTLCS. There was general agreement that as many as 70 percent of those CTLCS eligible are also high risk for CAD. Lecturers concurred that LDCT scans allow provision of an AS or ordinal CAC score comparable to gated CT scans, providing detailed information on the location and extent of calcified plaque. Programs provided long-term follow-up of patients with CAC scores obtained from CTLCS. Not only did CAC scores accurately predict increasing mortality, CAD mortality decreased when scores were reported and preventive care provided.

These findings suggest an intriguing opportunity for shared benefit for primary care physicians (PCPs) and cardiologists, since CTLCS is already a covered benefit for up to 14 million Americans and there is no added procedural cost or radiation exposure.

Artificial intelligence in opportunistic screening

Reporting CAC routinely on all LDCT, however, exacerbates radiologist workload. At present, CAC is “systematically underreported” on non-cardiac CTs, despite recommendations from the American College of Cardiology and the American Heart Association support opportunistic screening for CAC reporting on chest CTs.

Does artificial intelligence (AI) offer a palliative? Multiple research studies reported on the feasibility and accuracy of AI, not just aiding in detection of lung nodules, calculating tumor growth, volume doubling times, and predicting risk of LC in detected nodules but also generating accurate CAC scores.

English investigators reported that “Implementing AI as first-reader to rule out negative CT scans, shows considerable potential to reduce CT-reading workload and does not lead to missed LC.” A report from the iDNA group asserted that “tumor volume doubling assessment performance in follow-up screening outperforms human readers in the early identification of rapid growth in histologically-confirmed cancers.”

Investigators from the European Four in the Lung Study reported that “analysis of one-run, non-ECG triggered dual-source computed tomography (DSCT) data with fully automated AI analysis provides highly accurate Agatston score measurements with a minimal clinical relevant downgrade classification rate.”

Moving toward widespread integration

Identification of CAC is actionable. The Notify 1-Project showed that notification of caregiver and patient when CAC is seen on an unrelated CT chest, has a marked effect on statin prescription. “At six months, the statin prescription rate was 51.2 percent (44 of 86) in the notification arm versus 6.9 percent (6 of 87) with usual care (p-value < 0.001).”

Most important, Italian investigators reported that “ITALUNG showed a significant drop in cardiac death at 11 years of follow-up (15.6 versus 34 per 10,000)” and that “There was a trend toward lower mortality when reporting of CAC occurred and with use of cardiovascular (CV) drugs.”

Current AI modules are capable of providing detailed reports, competitive with human readers, in detection and growth of lung nodules, and reducing radiologist workload. Dr. Claudia Henschke announced that IELCAP scans are now read first by an AI module and a detailed report generated. After radiologists read the scan, they can modify the reports, but the time-consuming production of complex reports has already been completed by machine. In addition, AI generates reports of CAC that generally agree with those produced by human readers. Since 2023, automated AI CAC scoring was live on all University of California San Francisco non-contrast CTs.

iDNA investigators estimate radiologist workload reduction with AI at an estimated 67 to 79 percent. This new information, combined with increasing interest from cardiologists internationally, perhaps also offers a way forward, to palliate the very low uptake of population CTLCS in the U.S. The ability to generate more accurate information on future risk of CAD offers an incentive for cardiologists to participate actively in advising patients eligible for CTLCS to be screened. PCPs, who have, to date, shown little enthusiasm for referral of their patients at risk to CTLCS programs, might also be convinced to reconsider. This would represent a win-win scenario for all parties. CT screening has now shown the ability to not only greatly reduce mortality and morbidity from LC, it now can help reduce deaths from CAD. Two of the largest contributors to premature mortality in the U.S. might accordingly be improved with a simple low-risk study, at no cost, covered by federal and private medical insurance plans.

There were multiple further discussions in Gdansk of combined screening for “the big three.” The role of CTLCS in the detection of early-stage obstructive lung disease and emphysema, which cause 5 percent of deaths in the U.S., is a more complex topic to be addressed in a future post.

Frederic W. Grannis, Jr. is a thoracic surgeon.