Coronary Artery Calcium Scoring May Refine Heart Risk Estimate and Statin Decisions
Edited by Archita Rai
August 04, 2026J Am Coll Cardiol. 2026;doi:10.1016/j.jacc.2026.05.039
TOPLINE
Adding coronary artery calcium (CAC) scoring to risk estimates helped better identify adults at lower and higher risk for atherosclerotic cardiovascular disease (ASCVD) events. People with no detectable CAC had lower observed risk, while those with CAC had higher observed risk, especially when the decision about starting a statin was uncertain.
METHODOLOGY
Researchers analyzed data from a US community-based cohort (MESA) to evaluate how CAC relates to ASCVD event rates across risk categories and groups of statin eligibility.
The analysis included 5698 adults without clinical ASCVD at baseline (mean age, 61.5 years; 52.8% women).
All participants underwent cardiac CT CAC scoring at baseline, which was categorized as 0, 1-99, or ≥ 100. Participants were also grouped based on the presence or absence of CAC ( CAC = 0 vs CAC > 0).
Researchers calculated 10-year PREVENT-ASCVD risk, defined four risk tiers (< 3%, 3% to < 5%, 5% to < 10%, ≥ 10%), and categorized statin eligibility as not recommended, considered, or recommended per the 2026 dyslipidemia guideline of the American College of Cardiology/American Heart Association.
The outcome was incident ASCVD captured over 10 years.
TAKEAWAY
Participants not recommended for statins had 10-year ASCVD event rates of 1.2 per 1000 person-years with CAC = 0 vs 4.5 per 1000 person-years with CAC > 0.
Participants in the statin-considered group had 10-year ASCVD event rates of 2.7 per 1000 person-years with CAC = 0 and 5.2 per 1000 person-years with CAC > 0; those in the statin-recommended group had rates of 5.8 and 16.6 per 1000 person-years, respectively.
In participants with low-density lipoprotein-cholesterol levels of 70-189 mg/dL and no diabetes, CAC > 0 was linked to about twofold or higher event rates across PREVENT risk tiers.
Addition of CAC to PREVENT-based risk categories resulted in a categorical net reclassification improvement of 0.124, driven by appropriate reclassification of events and nonevents.
IN PRACTICE
“CAC may be most useful for contextualizing absolute risk and guiding the intensity of preventive efforts within contemporary prevention frameworks,” the researchers wrote.
SOURCE
This study was led by Rishi Rikhi, MD, MS, Wake Forest University School of Medicine, Winston-Salem, North Carolina. It was published online on July 15 in JACC.
LIMITATIONS
Researchers used baseline MESA data to calculate PREVENT risk, which may not have reflected current distributions of risk factors. CAC results were not blinded to clinicians or participants. MESA was a relatively healthy primary prevention cohort, which may have led to lower absolute rates of ASCVD events.
DISCLOSURES
This study received support from the National Heart, Lung, and Blood Institute and the National Center for Advancing Translational Science. Several authors reported receiving government research contracts and institutional or investigator grants from pharmaceutical, device, or biotechnology companies. Some authors reported receiving consulting fees, honoraria, participation on scientific or advisory boards, and industry-sponsored research support. One author reported owning equity in a health-related company.
Visceral Belly Fat Can Make You Age Faster, Experts Warn
Edited by Javed Choudhury
July 17, 2026Obesity (Silver Spring). 2026;doi:10.1002/oby.70239
TOPLINE
In middle-aged adults, greater visceral adipose tissue measured using DEXA was independently associated with higher phenotypic age acceleration in both men and women and with shorter leukocyte telomere length — another biomarker of biological aging — in women but not in men.
METHODOLOGY
Visceral adipose tissue is more metabolically active than subcutaneous fat, leading to greater risks for inflammation and cardiometabolic conditions. However, few studies have directly examined the association between DEXA-measured visceral adipose tissue and biological aging in middle-aged adults.
Researchers conducted a cross-sectional analysis using data from 4799 adults (mean age, 58 years; 2614 women) from a population-based cohort study conducted in Western Australia.
Whole-body DEXA was used to measure visceral adipose tissue mass, subcutaneous adipose tissue mass, and total body fat mass.
Phenotypic age acceleration was derived from chronological age and nine clinical biomarkers (lymphocyte percentage; mean cell volume; red blood cell distribution width; alkaline phosphatase, albumin, creatinine, glucose, and C-reactive protein levels; and white blood cell count).
Leukocyte telomere length was measured as the ratio of telomeric DNA to a single-copy control gene (T/S) in a subset of 1221 participants (51.8% women); shorter lengths indicated greater biological age.
TAKEAWAY
Each 100-g increase in visceral adipose tissue mass was associated with a 0.160-year (95% CI, 0.137-0.183) increase in phenotypic age acceleration in men and a 0.321-year (95% CI, 0.291-0.351) increase in women.
Each 1-SD increase in visceral adipose tissue was independently associated with a 1.397-year (95% CI, 1.197-1.596) increase in phenotypic age acceleration in men and a 1.919-year (95% CI, 1.738-2.099) increase in women.
In women, higher visceral adipose tissue mass was associated with shorter leukocyte telomere length, with each 100-g increase linked to a 0.007-unit decrease in the T/S ratio and each 1-SD increase linked to a 0.041-unit decrease in the ratio; no significant associations were found in men.
IN PRACTICE
“[The study] findings suggest that VAT [visceral adipose tissue] could be a specific marker of accelerated biological aging, independent of lifestyle factors and traditional anthropometric and other DXA-derived measures of obesity,” the authors of the study wrote.
SOURCE
The study was led by Riorden O’Shea, WA Country Health Service, Perth, Australia. It was published online in Obesity.
LIMITATIONS
The cross-sectional design prevented the establishment of causal relationships. The study population was highly homogeneous, which may have limited generalizability. Additionally, analysis of leukocyte telomere length was conducted only on a subset of participants.
DISCLOSURES
The cohort study received funding from the Federal Government of Australia, the Government of Western Australia, the National Health and Medical Research Council Equipment Grant scheme, and other sources. The authors declared no conflicts of interest.
Very strong evidence-based data have demonstrated the utility of lung cancer screening in individuals with well-defined risk due to past and current smoking behavior. The specific characteristics have been well laid out in the literature, and really advertised to the public.
Key Points
Lung cancer screening benefits strongest in well-defined smoking-risk populations.
US Preventive Services Task Force recommends screening eligible high-risk adults.
2024 screening uptake only ~20%-25%; some regions <10%.
Full uptake projected: 62,110 deaths prevented; 872,270 life-years gained/5 years.
Which interventions increase lung screening uptake?
How does screening adherence vary by US region?
What barriers limit primary care lung screening referrals?
The lung cancer screening strategy has been recommended in these well-defined populations, in fact, by the United States Preventive Service Task Force. Despite these facts, evidence very strongly suggests that, overall, only about 1 in 5, so 20% or maybe 25% of eligible individuals based upon risk factors were screened in 2024,and in certain regions of the country, that number was actually less than 10%.
The individuals in this report conducted a modeling study, addressing the question of what if we assumed that we could get the message to the public, also perhaps to primary care doctors and others involved in care? What would be the impact if this message got through and essentially 100% of those eligible in the United States for screening actually participated in these well-defined programs? What if we were able to encourage people to do this? What would be the impact?
Recognizing the limitations of any modeling exercise, this analysis revealed the following. If essentially 100% of eligible individuals were screened, the number of deaths prevented and life-years gained would increase threefold compared to the current situation. Overall, 62,110 lung cancer deaths would be prevented over 5 years, with a gain of 872,270 life-years total.
This is a remarkable achievement when you consider death, morbidity, mortality, pain, and suffering associated with these lung cancer events. All of this could occur if an established screening modality was more widely employed.
This is an important message, I think, as a community, physicians, public health officials, members of society in general, and family members with individuals who are potentially at risk. We need to figure out a better way of communicating the benefits of lung cancer screening.
Offering lung cancer screening to everyone with a 20-year smoking history could expand access to screening, identify more cancers, and reduce disparities, new research suggests.
In an analysis of nearly 1 million US veterans, researchers estimated that a simplified approach to lung cancer screening — based on smoking duration rather than pack-years — would expand screening eligibility by nearly 30% and reduce potentially missed lung cancers by over 70%.
Those shifts would be especially pronounced among women and Black individuals — two groups that are underserved by current screening criteria.
And they support smoking duration as a “simpler, more sensitive, and more equitable metric for screening eligibility,” researcher Brendan T. Heiden, MD, MPHS, Washington University School of Medicine in St. Louis, St. Louis, told meeting attendees.
Toward a Better Metric
Current guidelines from the US Preventive Services Task Force (USPSTF) recommend annual lung cancer screening with low-dose CT for adults aged 50-80 years who have at least a 20 pack-year smoking history and either currently smoke or quit within the past 15 years.
The 20 pack-year metric is equivalent to smoking a pack of cigarettes per day for 20 years. Because it requires patients to remember their smoking intensity over decades, it can be challenging to calculate and translate into care, Heiden said.
As it stands, few Americans who are eligible under current USPSTF guidelines actually undergo lung cancer screening, at about 15%-20%, Heiden noted. Meanwhile, mounting evidence suggests that many lung cancers occur in individuals who never meet those eligibility criteria.
Boosting screening uptake, Heiden said, is not enough: There’s a need to revisit eligibility itself to reach more high-risk individuals.
Some groups are already taking steps in that direction. Recently updated guidelines from the National Comprehensive Cancer Network (NCCN) added a category 2B recommendation supporting screening for individuals with at least a 20-year smoking history, regardless of pack-years. (The guidelines also say former smokers are eligible no matter how long ago they quit.)
For their study, Heiden’s team sought to estimate the performance of that smoking-duration metric against current USPSTF pack-year criteria. They used Veterans Health Administration data on over 980,000 veterans whose smoking histories were prospectively collected; lung cancer diagnoses were identified through the Veterans Affairs Central Cancer Registry.
Most of the included veterans (67%) had a smoking history; their mean age was 64 years, and 21% were Black.
Overall, the researchers found that basing eligibility on 20-year smoking duration would substantially expand access to screening: Among veterans with a smoking history, 68% qualified for screening under current USPSTF criteria compared with 87% using the smoking-duration approach.
The gains were especially pronounced among women and Black individuals (who, based on prior research, typically smoke less intensely than White males). Under USPSTF criteria, only about 55% of female and Black veterans qualified for screening compared with 83% for both groups under the smoking-duration criterion.
Importantly, Heiden said, people meeting the smoking-duration threshold remained at substantially elevated risk for lung cancer, suggesting the broader screening criteria were not merely capturing low-risk smokers.
The 5-year lung cancer incidence among veterans eligible under the smoking-duration approach was 1.59% — 11 times the rate of 0.14% among never smokers.
Perhaps most striking, Heiden said, the proportion of potentially missed cancers dropped from 13% with the pack-year metric to just 4% using the smoking-duration metric — a relative reduction of more than 70%.
Again, women and Black individuals would see the largest gains: Among Black veterans, potentially missed cancers fell from 25% to 6%, whereas among female veterans they declined from 22% to 7%.
Optimal Approach Still Unclear
The analysis had limitations, including a predominantly male veteran population whose smoking exposure was far greater than that of the general US population, indicating high inherent lung cancer risk.
But the results support what the NCCN has already done, according to Mary Reid, PhD, MSPH, BSN, a member of the group’s lung cancer screening guideline panel and chief of cancer screening, survivorship and mentorship at Roswell Park Comprehensive Cancer Center in Buffalo, New York.
“Doing the calculation for pack-years can be difficult,” Reid told Medscape Medical News. “Smoking duration is easier to calculate and really the way to go.”
The USPSTF does not comment on individual studies outside of its recommendation development process.
At the meeting, study discussant Katharine A. Rendle, PhD, called the work “impressive,” citing the size of the cohort and strength of the data.
It’s particularly noteworthy that the simpler screening criteria improved sensitivity for all veterans, while largely eliminating disparities, according to Rendle, of the Abramson Cancer Center at the University of Pennsylvania in Philadelphia.
Still, she said, further research could better define the optimal screening strategy.
“Smoking duration is a promising approach, but in my opinion, guidelines likely need to account for the underlying risk in the population,” Rendle said, noting that current smoking prevalence in the US population is about 10%.
She suggested future studies consider other smoking-duration thresholds, such as 30 or 40 years, and look at other outcomes, including life-years gained.
“It’s critical that we prioritize strategies that maximize potential benefit from screening — not just identify those at lung cancer risk — given downstream costs and burden on populations and healthcare systems,” Rendle said.
The study had no commercial funding. Heiden, Rendle, and Reid had no relevant disclosures.
Prostate cancer surgeons routinely use MRI to guide surgical planning, yet imaging findings are not used for risk stratification in standard prognostic models. A new meta-analysis suggests they should be.
The analysis, of 40 studies involving nearly 25,000 men undergoing radical prostatectomy, found that preoperative MRI findings offered independent prognostic value for biochemical recurrence, metastasis, and prostate cancer-specific death.
Specifically, extraprostatic extension, seminal vesicle invasion, and certain MRI quantitative metrics (such as high Prostate Imaging Reporting and Data System [PI-RADS] scores) were strongly associated with outcomes. That was after adjustment for established factors used in current prediction models from the National Comprehensive Cancer Network and other groups — including prostate-specific antigen level, digital rectal exam results, and biopsy Gleason grade.
“MRI appears to be a complementary prognostic layer,” researchers led by Georgios Agrotis, MD, PhD, of the Netherlands Cancer Institute, reported in JAMA Oncology.Article Key Points
The findings bolster the current use of MRI in surgical planning and support its formal incorporation into preoperative risk models to improve subsequent risk-based treatment decisions.
“This is the main message of our study,” Agrotis told Medscape Medical News. “The potential value is not only better prediction but also better personalization of care.”
Patients who appear similar based on traditional clinicopathological factors may in fact have different levels of risk that can be discerned by MRI, he said.
Adding MRI to widely used prognostic models could also standardize its use across centers, paving the way for multicenter studies to zero in on best practices, Agrotis noted.
Urologic oncologist Nicholas Pickersgill, MD, of the Memorial Sloan Kettering Cancer Center, New York City, also sees the potential.
In an accompanying editorial, he noted that while MRI is already used for nerve-sparing, lymph node dissection, radiation, and other decisions, “oncologists set aside the MRI results when predicting what the cancer will do.”
This meta-analysis, he added, “strengthens the case for reconsidering this omission,” though prospective studies are needed to confirm clinical benefit.
In the meantime, Pickersgill told Medscape Medical News that many centers, including his own, are developing in-house prognostic nomograms that include MRI findings to predict posttreatment recurrence and progression.
“The future of prostate cancer risk stratification almost certainly lies in models that integrate refined pathology, imaging, and molecular features alongside clinical variables to advance individualized treatment planning,” Pickersgill said.
One of the challenges, he added, will be determining exactly how to integrate prognostic information from MRI.
“Guideline committees should now define the evidentiary threshold required,” Pickersgill said.
In the meta-analysis, MRI-detected extraprostatic extension was associated with increased risks of all three outcomes the researchers considered: a twofold higher risk for biochemical recurrence (pooled hazard ratio [HR], 2.16), a threefold higher risk for metastatic failure (HR, 3.18), and a nearly 11-fold increase in the risk for prostate cancer-specific mortality (HR, 10.93).
Seminal vesicle invasion, meanwhile, almost tripled the risk for biochemical recurrence (HR, 2.74) and increased the risk for metastatic failure nearly sixfold (HR, 5.58).
Several quantitative MRI parameters were also prognostic for biochemical recurrence. The risk was roughly doubled when men had PI-RADS scores of 4 or 5 (HR, 2.15), tumor diameter of 20 mm or more (HR, 2.35), or low apparent diffusion coefficient values (HR, 2.39).
Overall, the analysis found, MRI-based T staging was a substantially stronger predictor of biochemical recurrence than clinical T staging based on digital rectal exam.
Agrotis and his colleagues acknowledged some limitations of their analysis. For one, MRI use might have been influenced by institutional practices or clinical indication, creating selection bias toward higher-risk disease. In addition, most studies had a primary endpoint of biochemical recurrence, while relatively few assessed metastatic failure or mortality — which, the researchers wrote, “limits inference regarding long-term outcomes.”
There was no external funding for the work. Agrotis and Pickersgill had no disclosures.
M. Alexander Otto is a physician assistant with a master’s degree in medical science and a journalism degree from Newhouse. He is an award-winning medical journalist who worked for several major news outlets before joining Medscape. Alex is also an MIT Knight Science Journalism fellow. Email:aotto@medscape.net.
Coronary Artery Calcium Scoring Remains Predictive in Primary Prevention With High Lp(a)
Crystal Phend
March 23, 2026
Coronary artery calcium (CAC) screening holds up for cardiovascular riskprediction in a population with high lipoprotein(a) (Lp[a]), with the two serving as independent risk factors, an observational study showed.
“These findings strengthen the case for CAC as a pragmatic gatekeeper in deciding when to initiate or escalate preventive therapy in middle- to older-age individuals with elevated Lp(a), particularly in primary prevention settings where overtreatment is a concern,” Parveen K. Garg, MD, MPH, of the University of Southern California Keck School of Medicine in Los Angeles, and his colleagues wrote in an editorial accompanying the study.
The findings mesh with the updated lipid managementguideline released last week from the American College of Cardiology, American Heart Association, and other medical societies. The new guideline recommends universal Lp(a) testing at least once for adults, and CAC scoring in men aged at least 40 years and women aged at least 45 years as risk enhancing factors to guide treatment.
“We’re going to see a lot more people identified, especially people before having a first event,” said Harpreet S. Bhatia, MD, MAS, of the University of California San Diego. “This study helps us understand how a coronary calcium score, when appropriately used, can help…divide up that potentially 2 billion-person group of people with elevated Lp(a) into risk groups.”
While genetic, epidemiologic, and mechanistic data have firmly established Lp(a) as a causal risk factor for atherosclerotic cardiovascular disease (ASCVD), when to start and how intensively to target cardioprotective therapies in patients with high levels but without established ASCVD or traditional high-risk features has been unclear, as Garg and colleagues wrote in their editorial.
Whether CAC could meaningfully discriminate ASCVD risk in the setting of elevated Lp(a) was unclear, “particularly given Lp(a)’s known association with noncalcified and high-risk plaque phenotypes,” Garg’s group wrote.
Study Findings
Bhatia’s study pooled data from 11,319 largely middle-aged adults (mean age, 56 years; 54% women) in four US prospective cohorts: Multi-Ethnic Study of Atherosclerosis (MESA), Coronary Artery Risk Development in Young Adults Study, the Framingham Offspring Study, and Jackson Heart Study. These individuals initially free of baseline ASCVD were followed for a mean of 14.8 years, during which 1569 myocardial infarction, stroke, and coronary revascularization events occurred.
ASCVD risk correlated with both elevated Lp(a), defined as > 50 mg/dL (hazard ratio [HR], 1.24; 95% CI, 1.09-1.41) and a CAC score > 0 (HR, 2.44; 95% CI, 2.14-2.77), but the two were independent risk factors (P for interaction = .80).
“But the key is in the absolute event rates,” Bhatia told Medscape Medical News.
While ASCVD risk was higher with elevated Lp(a) and a zero CAC score than without elevated Lp(a), the absolute risk was low in both scenarios (4.9 vs 3.8 events per 1000 person-years; HR, 1.28; 95% CI, 1.01-1.60).
The risk rose with CAC to a peak 6.12-fold elevation in risk with a CAC score of at least 300 and Lp(a) over 50 mg/dL (HR, 6.12; 95% CI, 4.80-7.81).
Although results were consistent across age and sex subgroups, there was greater absolute risk past 50 years of age and in men. “These findings should not be interpreted as diminishing the relevance of elevated Lp(a), particularly in younger individuals and especially women,” Garg’s group cautioned. Absence of coronary calcification in younger adults may simply reflect insufficient time for calcific remodeling, they wrote.
Next Steps
While there are as yet no recommended risk prediction models that incorporate Lp(a) into a clinical algorithm, Bhatia said the study has shifted his practice to using CAC testing to move patients toward a more aggressive LDL cholesterolmanagement vs continued monitoring over time if they fit a certain profile: “someone 40 years of age or above, an appropriate candidate for calcium scoring, has a high Lp(a), maybe doesn’t have a lot of other cardiovascular risk factors, and we kind of want to understand more what their personal risk is.”
His group is now working on an observational study to identify the timeframe for repeat calcium scoring in this population, he said.
And the field may well move into using Lp(a) to guide other therapies as well, with antisense oligonucleotide and small interfering RNA therapies targeting Lp(a) advance through late-phase clinical trials.
“Trials evaluating cardiovascular risk reduction with these agents, on top of guideline-directed preventive therapy, may be especially informative when focused on individuals with both elevated Lp(a) and evidence of subclinical atherosclerosis, such as abnormal CAC scores,” Garg and colleagues noted.
This study was supported byUC San Diego BEACON. MESA was supported by the National Heart, Lung, and Blood Institute as well as by grants from the National Center for Advancing Translational Sciences. Bhatia reported being supported by a National Institutes of Health grant and serving as a consultant or advisor for Bayer, Abbott, Arrowhead, Kaneka, Novartis, and New Amsterdam. Garg’s group reported having no relevant conflicts of interest.
Crystal Phend is an award-winning medical journalist with decades of experience reporting on clinical research and healthcare developments across specialties. When not walking the halls at a medical conference, she can be found at a keyboard in upstate New York.
Lung cancer screening with low-dose computed tomography has been proven to reduce lung-cancer-specific and all-cause mortality. The UK launched the NHS England Targeted Lung Health Check Programme in 2019, which has now become the national Lung Cancer Screening Programme, with full coverage expected by 2030. Here we present the progress and outcomes of the program. People aged 55–74 were offered low-dose computed tomography of the thorax if they had ever smoked and if risk thresholds, as determined by multivariable models, were met. Delivery of the program is through regionally federated clinical infrastructure and leadership, with national strategic, clinical and economic frameworks. The program has invited over two million people, with 7,193 lung cancers diagnosed—63.1% at tumor, node, metastasis stage 1 and 12.6% stage 2—to March 2025. This has increased the early-stage proportion of lung cancer in England over 5 years, particularly in socioeconomically deprived regions. The NHS England Programme exemplifies how large-scale implementation can be achieved at speed through centralized protocols and effective project management. The program has demonstrated feasibility and scalability in reaching high-risk and underserved populations, but needs to further address inequalities in participation. These findings support adoption of lung cancer screening across the UK and globally, and offer practical tools for international adaptation.
Lung cancer screening with low-dose computed tomography (LDCT) of the thorax has been shown by several randomized controlled trials to reduce lung-cancer-specific and all-cause mortality1,2. Globally, national-level progress in lung cancer screening varies from continued research, through evaluation of the clinical and cost-effectiveness of small pilot projects to full-scale implementation of national screening programs. The National Health Service (NHS) England Lung Cancer Screening Programme, known in its earlier, large-scale pilot phase as the Targeted Lung Health Check (TLHC) Programme, was launched in 2019, with the aim to improve the proportion of cancer detected at an early stage in England. NHS England had recently set a target of 75% of people with any cancer being diagnosed at stage 1 and 2 by 2028. The pilot was also expected to provide real-world evidence ahead of a revised evaluation by the UK National Screening Committee. The goal was to create a national strategic, clinical and economic framework, within which to achieve full-scale deployment. The pilot constituted an important contribution to the national early cancer diagnosis agenda, while identifying challenges and solutions to clinical referral pathways for lung cancer and other clinically relevant findings. This would also inform commissioning budgets for a full rollout. The program adheres to a detailed national protocol3 and quality assurance standard4, through regionally federated clinical infrastructure and leadership.
The progress toward implementation began with the UK Lung Screen pilot randomized trial5, which showed that lung cancer screening was feasible and able to achieve both a higher lung cancer detection rate and higher early-stage lung cancer proportion than the US National Lung Screening Trial, the first trial to confirm a mortality reduction6. Small nonrandomized pilot programs demonstrated similar results and, notably, very low levels of harm7,8,9,10,11. A large-scale pilot, the TLHC was approved by NHS England in 2019. In September 2022, lung cancer screening for high-risk individuals aged 55–74 years was formally recommended by the UK National Screening Committee after a detailed health economics evaluation12. The English government announced this in June 2023, with an acknowledgment that the annual cost would be £270 M per year at full rollout. On 1 February 2025, the TLHC was formally renamed as the NHS England Lung Cancer Screening Programme.
This Article reports on the progress and outcomes of the screening program to date, including the process of phased rollout, participant uptake, lung cancer detection and downstream management.
Results
Table 1 summarizes the total number of invitations, lung health checks, baseline LDCTs and lung cancers diagnosed for both the whole program and initial phase. Whole-program data represent total activity in the program until March 2025 with over 4.5 times the number of LDCT and lung cancer diagnoses than the initial phase data.
Table 1 Comparison of initial-phase and whole-program samples
In the whole-program data from April 2019 to March 2025, 2,510,092 participants had been invited for a baseline Lung Health Check (LHC) across all Cancer Alliances (Fig. 1, consort diagram). This amounts to an invitation (coverage) of 32.4% of the total estimated, potentially eligible population of people who have ever smoked, aged 55–74 (7,743,437). Of those invited, 49.0% (1,229,714) have undergone an LHC (uptake of offer), and of these 16.8% (206,516 of 1,229,714) were performed face to face, and 83.2% (1,023,198) were performed by telephone, followed by face-to-face confirmation for those participants who qualified for LDCT screening; 47.5% (584,095) of participants met the multivariable model risk threshold (Liverpool Lung Projectv2 (LLPv2) or Prostate Lung Colorectal Ovarym2012 (PLCOm2012) models) and 43.0% (528,686 of 1,229,714 LHCs) underwent a baseline LDCT, while 4.7% (27,236 of 584,095 who met the LHC risk threshold and qualified for LDCT) did not attend or canceled their LDCT and 2.3% (13,231 of 584,095) met the risk threshold but were ineligible for LDCT on the basis of the exclusion criteria. A total of 2.6% (14,942 of 584,095) had no record of having had a scan, with reason unknown. A total of 79,338 three-month surveillance scans were performed, 34,797 twelve-month scans and there were 136,194 scans undertaken at 24, 48 and 72 months. The latter includes participants in whom incident round scans would not have been performed by the time of data extraction, that is a participant may have had a baseline scan but not yet have reached the point of having a 24-month scan. Therefore, the number of surveillance and incident round scans may be underrepresented. Extended Data Figs. 1 and 2summarize the program geography and proportional national rollout as a marker of coverage, according to Cancer Alliance. Figure 2a shows monthly (noncumulative) numbers of invites, LHCs completed and LDCT performed. Early efforts to initiate the program that were hindered by COVID-19 restrictions are notable until Spring 202113.
a, Lung cancer screening delivery trajectories. The number of first invites (people invited), LHCs and scans (across every scanning round) per month for the whole program, from May 2019 to March 2025, are shown. b, National lung cancer early diagnosis rates. Early diagnosis rates (% diagnosed at tumour node metastasis (TMN) stages 1 and 2), as a proportion of all lung cancer diagnoses, have increased for all deprivation quintiles after the pandemic, with the biggest change in stage (low to high rate) seen among those living in the most deprived areas. Source: NHS England Analysis of Rapid Registration Data.
By March 2025, 7,193 lung cancers had been diagnosed (1.4% of baseline LDCT participants), 2,228 in the last year. The stage distribution was 63.1% stage 1, 12.6%, 12.6% and 8.8 % stage 2, 3 and 4, respectively, with 2.8% stage not specified (Fig. 1). National Cancer Registration Data (NCRD) show that lung cancer early stage detection rates across the UK have increased steadily since the pandemic and are now well above pre-pandemic levels. Furthermore, since the lung screening program started, the proportion of stages 1 and 2 lung cancers in the most deprived socioeconomic quintile has increased from the lowest to the highest quintile (Fig. 2b). This has not been reported in other cancers.
Of 582,700 people in the initial phase who were eligible for an LHC, 216,985 (37.2%) attended. The proportion of different groups of the eligible population, participation rates, LDCT attendance and cancer diagnoses are presented in Extended Data Table 1 , while the odds ratios (ORs) are presented in Fig. 3; 303,825 of 582,700 (52%) of the eligible population were male, and 275,335 of 582,700 (48%) were female. LHC uptake in men and women as a proportion of the eligible male and female populations, respectively, were equivalent (113,670 of 303,825 (37.4%) versus 103,310 of 275,335 (37.5%)), but fewer women underwent an LDCT scan as a proportion of those attending an LHC (48.4% females versus 56.7% males). Specifically, of those assessed as high risk, women were also less likely to attend an LDCT than men (OR = 0.87, P < 0.001).
Fig. 3: Patient demographic determinants of lung screening engagement and outcomes.
OR plots of initial-phase data assessing the associations between demographic characteristics. a, LHC attendance, of those in the eligible population. b, Participants assessed as high risk. c, LDCT attendance of those attending an LHC and high risk. d, Lung cancers detected of those who attended LDCT with ≥185-day follow-up. ORs represent point estimates and the accompanying error bars show the 95% confidence interval (CI). The size of each square is proportional to N, meaning that larger squares correspond to a greater number of individuals included at the start of the regression. Estimates were obtained using multivariable logistic regression. P values were calculated using two-sided Wald tests based on the standard normal distribution, with no adjustment for multiple comparisons. N refers to the total number of individuals included at the start of the regression (for example, people in the eligible population) and n refers to the number who experienced the outcome of interest (for example, those who attended an LHC). Counts were rounded to the nearest multiple of five and values below ten were suppressed to protect confidentiality.
Among those assessed as high risk, older individuals were more likely to undergo LDCT compared to the reference group aged 55–64 years (65–74 years, OR = 1.09, P < 0.001; 75+ years, OR = 1.39, P < 0.001).
Ethnicity data were not known for 32.6% of the 582,700 individuals eligible for an LHC because of incomplete primary care records. The further a participant progressed through the pathway, the more likely ethnicity status would be recorded. A smaller proportion of participants invited for an LHC in the ‘other’ ethnic group attended than those of the ‘white’ ethnic group (18,295 of 97,265, 18.8% versus 184,765 of 295,410, 62.5%; OR = 0.15, 95% CI = 0.14-0.15, P < 0.001). In the subset of people assessed as high risk at LHC, those from ethnic groups other than white were significantly less likely to attend LDCT compared to those from the white group (OR = 0.79, P < 0.001).
People in areas of least deprivation (quintile 5) had a higher LHC uptake than the most deprived (quintile 1) (15,050 of 36,195, 41.6% versus 85,285 of 256,965, 33.2%, OR = 1.29, 95% CI = 1.25–1.32, P < 0.001). However, 37.6% (5,665 of 15,050) of LHC participants from the least deprived areas underwent LDCT scanning versus 57.2% (48,760 of 85,285) people from quintile 1. After a high-risk assessment, those living in the most deprived quintile (quintile 1) were less likely to attend LDCT than those in quintiles 2 (OR = 1.05, P < 0.05), 3 (OR = 1.19, P < 0.001) and 4 (OR = 1.29, P < 0.001), but not significantly different to quintile 5 (OR = 1.02, P = 0.6).
Lung cancer detection in the initial phase
By January 2023, 74,202 participants had undertaken a baseline LDCT (Table 2). Three-month and 12-month nodule surveillance LDCTs were completed in 9,995 (13.5%) and 6,689 (9.0%), respectively. Three-month and 12-month scan data are not mutually exclusive (nodule assessment often requires both time points) and scans were censored at March 2024. Cancers (censored at August 2023) diagnosed from the baseline, 3-month and 12-month time point scans were 890, 135 and 70 (74.4%, 11.3% 5.9%) of 1,196 cancers diagnosed in the initial-phase data at all time points. This equates to a cumulative cancer conversion rate of 1.2%, 1.4%, and 1.5%, respectively of 74,202 participants. Twenty-four-month scans (most would represent incident round scans, and a minority of nodule surveillance scans) were documented in 24,933 participants of whom 36 were diagnosed with cancer (3.0% of 1,196 screen-detected cancers across both rounds); 2,393 scans at ‘other’ (undefined) time points detected 65 cancers. The total cancer detection proportion was 1.6% across both rounds.
Table 2 Initial-phase LDCT scan timepoints and lung cancer detection (baseline scans before January 2023)
Initial-phase data restricted to a follow-up period of at least 185 days from LDCT were available for 53,430 people; 2.9% (1,565 of 53,430) of participants had a lung cancer diagnosis in the Cancer Outcomes and Services Data (COSD)-linked data within 185 days of LDCT. Cancer outcomes were censored in August 2023, representing 185 days after LDCT (including those for nodule surveillance). The objective of this was to ensure that all cancer diagnoses resulted from the LDCT. This time was chosen after an analysis confirming that this did not include cancers diagnosed that were not related to the program. This approach was only for the initial-phase data. In this subset, cancers diagnosed were significantly higher in women than men (835 of 23,415, 3.6% in women, 730 of 29,995, 2.4% in men; OR = 1.48, 95% CI = 1.34–1.64, P < 0.001). Older (for example, 65–74 years) participants were also more likely to have a cancer diagnosis than younger (55–64 years) participants (OR = 1.88 CI = 1.66–2.14, P < 0.001) (Fig. 3 and Extended Data Table 1).
Incidental findings
Assessment of non-lung-cancer LDCT scan findings was made in 114,430 participants selected from the initial phase who underwent baseline LDCT scanning (Table 3); 54,695 (47.8%) had documented coronary artery calcification (severity grading was not collected), 36,745 (32.1%) had aortic valve calcification and 13,830 (12.1%) had emphysema (radiologically moderate or severe). Of note, 525 (0.46%) other (non-lung) cancers were diagnosed.
Table 3 Number of people with incidental findings reported (n = 77,185) in the initial-phase projects (April 2019–March 2024), as a proportion of total who had an LDCT (N = 114,430)
The NHS England Lung Cancer Screening Programme (formerly TLHC) is a large, publicly funded national lung screening program that has scaled up at a fast pace by adopting a federated delivery model using a single mandated protocol and quality assurance standard. Both the delivery model and the program data provide a rich resource of real-world evidence for similar large-scale national programs that are at the stage of implementation. The scale of the program, with almost one-third of England’s total estimated eligible population having received invitations, demonstrates that implementation at scale is feasible. The whole-program results imply that the protocol is working, with 1.4% of participants diagnosed with lung cancer, of which 76% were at stages 1 and 2. Although the proportion of participants with lung cancer is lower than that observed in the British pilots, which was an average of 2% at the baseline round, it is in keeping with, or above, the proportion seen in the large randomized trials1,5,8,9,10,11,14.
The program has already shown an impact on the early-stage proportion in national lung cancer registry data, which has risen as screen-detected lung cancer has risen. No similar trend has been seen in other cancers in England. This demonstrates the impact that can be expected and was highlighted in an independent review of the NHS in England in 202415. The latest UK National Lung Cancer Audit reports increases in lung cancer incidence, stage 1 rates and surgical resection rates16. While data currently only show an increase in stages 1 and 2, the similar stage distribution to randomized trials showing a reduction in mortality suggests that a mortality reduction can be expected.
Most (83.2%) LHCs were delivered by telephone, an innovation adapted into the Standard Protocol in response to the COVID-19 pandemic, when it was found to be feasible and efficient. Although following the Standard Protocol and structural quality assurance standard is mandated, innovations can be proposed for review by the national operations team and clinical advisory committee. This flexibility is important to facilitate iterative improvements identified by the separate sites, while maintaining uniformity, which is felt to be a strength of the program.
The uptake of the offer for an LHC, 49.0% overall, is lower than in other screening programs but screening was initially targeted in the areas of highest socioeconomic deprivation, where uptake is lower in all screening programs. Our analysis of the initial-phase data confirmed that participants in the most deprived socioeconomic group and in non-white communities were less likely to respond to the invite. However, uptake is improving as the program progresses, with the latest data indicating over 60% (Fig. 2), comparing very favorably to other international experience in lung cancer screening. This may reflect the strong public engagement and careful steps taken to ensure that effective invitation methods were followed based on evidence from lung cancer and other screening programs17. Despite these encouraging findings, there should be concern about the people who choose not to respond. Even if it is assumed that the same proportion would be eligible for screening, this means that half of the population do not currently have the chance to benefit. As socioeconomic disadvantage is associated both with lower participation and lung cancer, it could be that an even greater proportion of those who benefit are being missed. In-service evaluation of changes to the program are important. An example is the recent change to the name of the program from ‘targeted lung health check’ to ‘lung cancer screening’. Before this, NHS England undertook a behavioral science-led assessment of a number of new names for both the program and the assessment. Surprisingly, participants preferred lung cancer screening for the program name because it was less ambiguous than ‘targeted lung health check’ but preferred ‘lung health check’ for the risk assessment by telephone or face to face. Further patient experience data are available from NHS England18.
The Lung Cancer Screening Programme is one of few to use multivariable risk models to define eligibility. This approach was chosen based on evidence showing that risk models are more efficient at identifying participants with lung cancer. More recent research confirms this finding19,20,21. The lung cancer detection proportion of 1.4% seen in this program is lower than that observed in the UK Lung Screening trial (~2%) and UK pilots, and may reflect the wider coverage of the population. Furthermore, the proportion was calculated using the prevalent round CT number as the denominator, which may overestimate compared with the pilots. However, at this point, most lung cancers were detected at baseline or during surveillance CTs. Nevertheless, the cancer proportion exceeds that reported from the USA, even before eligibility criteria were expanded further in the USA22. The decision to use two rather than a single risk model was made in view of the lack of prospective head-to-head comparison between these two models in terms of cancer yield and cost-effectiveness in the UK population at the time the program commenced. Whole-program data showed that of those attending an LHC 47% were assessed as high risk and 90.5% of those underwent LDCT.
The participant-level analysis in the ‘initial-phase’ data (Extended Data Table 1 and Fig. 3) showed that slightly fewer females attended an LHC overall, and fewer were assessed as high risk. This merits further analysis to establish whether this is simply explained by fewer high-risk women responding to an LHC, as seen in some trials23, or failure of risk prediction models to correctly predict risk. In addition, there was a slightly lower chance of undergoing LDCT in females who were assessed as high risk. LHC uptake was significantly lower in the more deprived, although these participants were more likely to be both at high risk and to have screen-detected cancer. The latter is expected and probably explained by the higher rates of smoking that would increase the risk estimates from both multivariable models. People from white ethnic backgrounds were markedly more likely to attend their LHC than people with other ethnicities (62.5% versus 18.8%). These data should be interpreted with caution because they represent participants in the earlier wave of invitations and consequently may differ from whole-program aggregated data where overall participation is higher. However, screening uptake is recognized to be affected by ethnicity in other screening programs24, and the marked difference for ethnicity demands close monitoring and should inform elements of future LHC design addressing inequities25,26. Additional gains in uptake and cancer detection may then be possible by delivering engagement strategies targeted at these underserved groups. Projects have trialed a wide range of approaches, with mixed results, but there has been insufficient assessment of efficacy, making this a topic for future research.
The increase in early-stage proportion shown by national lung cancer registration data in line with screening activity and the finding that this was most marked in the most deprived socioeconomic quintile of the population shows what countries can expect from lung cancer screening. The change in socioeconomic distribution of early-stage disease may reduce as the program is rolled out to areas with lower incidence of lung cancer, but the targeted nature of the program and the use of multivariable models should mitigate this by identifying those at higher risk who will inevitably be more likely to come from the more deprived sectors of society.
Early stage detection is subject to overdiagnosis; what is needed to confirm efficacy, is a reduction in late-stage rate and mortality. Both of these take longer to become apparent. However, a report from Manchester, one of the earliest sites, confirmed a 25% difference in late stage between geographical areas with and without screening27, which is correlated with mortality reduction28.
An important challenge has been the management of non-lung-cancer diagnoses and other incidental findings. From the outset, this was supported by an incidental findings protocol as part of the quality assurance standard4. The principle underlying the protocol is to act only on those findings where there is likely to be benefit. For example, the most common finding, coronary artery calcification, is managed assuming that most people eligible for screening will also be eligible for primary prevention. Many patients have not had this; therefore, the findings of moderate and severe calcification prompt a reminder to the participant and primary care team. Currently no additional action is recommended for mild calcification because the evidence for benefit in the context of screening is not confirmed29.
The program has quantified how common incidental findings are; while this provides opportunities to improve outcomes, management must also address overdiagnosis, physical and psychological harm, risk to program delivery and impact on workload in primary and secondary care. A study showed that despite a protocol being in place, many participants did not receive the recommended interventions30. In response, the latest version of the protocol includes nationally standardized pathways for the management of incidental findings. The incidental findings protocol is tailored to the UK healthcare system and, after legal opinion, addresses the medico-legal implications of the management of incidental findings. This was an important step in reassuring radiologists about the safety of following the protocol. It is an important country-specific aspect of management of incidental findings.
The screening program has put existing services under some pressure and required an increase in surgical and oncology capacity to manage the additional curative-intent treatment. As the program expands further, new sites are able to learn from earlier adopters about the requirement for expansion of services and workforce. The program has used locally outsourced capacity for LHC call provision and LDCT scanning, including mobile unit provision and radiology reporting. Computer-aided detection is already mandated for lung nodules and often includes automated volumetry. A future challenge is to ensure that lung screening meets the standards set for screening in the NHS, in particular robust quality assurance of performance. This in turn requires improved data systems bespoke to lung screening.
The main limitation of the analyses presented is that data collection is ‘real world’ and may not be as rigorously verified as in a research trial. However, the very large dataset presented mitigates this as there are unlikely to be systematic errors; any individual inaccuracies are unlikely to affect the overall analysis. Detailed participant record-level data with cancer diagnoses linked to national registration data were only available for the initial phases of the program, so this means that findings on demographic impact on screening participation are less reliable. For participant-level data, we censored dates for LDCT follow-up at 185 days for diagnoses related to screening, recorded through linked national registry data accepting that this may result in some underrepresentation of scan totals and attributable cancer diagnoses.
Missing data on ethnicity and problems with completeness of smoking records in primary care are an important challenge for successful implementation of lung cancer screening; this is often a greater challenge in other countries. Although NHS primary care records have relatively high levels of smoking record accuracy, for some individuals direct invitation to establish smoking status is required31.
Limited data were available for some important measures of performance. These include data on recall rate, incidental findings referral rate, details of nodule findings required to measure false positives, interval cancers and data on outcomes. This detail will be part of future publications, but some radiology reporting consortia collect these data and use this to feedback on individual radiology performance. NHS England has developed 14 effectiveness standards, currently in consultation, which are designed to provide a basis for performance management and quality assurance of outcomes.
The program does not currently have a bespoke, end-to-end participant-level information technology system, although this is recommended in the Standard Protocol3. This is expected in mid-2026.
The program presents an important opportunity for research and several studies have developed alongside the implementation, using datasets considerably larger than available from trials25,32.
In conclusion, the NHS England national Lung Cancer Screening Programme has shown how large-scale implementation can be achieved at speed in high-risk groups through the application of a single protocol and effective project management, informed by previous research and smaller-scale pilots. Almost a third of the eligible population has been invited, with evidence of downstream impacts on the stage of diagnosis in participants and national lung cancer data. In doing so it demonstrates what can be expected and provides practical tools adaptable for use in other countries. The program has achieved these results despite other significant challenges in the NHS, including workforce and economic strain, industrial action and a respiratory pandemic. These data, and the material available freely online, and as supplementary material, provide evidence of feasibility, of the impact of a larger-scale program and a blueprint to assist others implementing this essential element of lung cancer care.
Non-Risk-Based Lung Cancer Screening Tied to Better Survival
Liam Davenport
March 31, 2026
COPENHAGEN, Denmark — Screening individuals for lung cancer with low-dose CT without preselection based on their risk profile is associated with a substantial reduction in lung cancer-specific mortality vs no screening, suggested a Chinese analysis.
The cancer detection rate was similar to that in risk-based screening trials, and the study results underscore that early-stage lung cancer detection via screening is associated with markedly improved survival, said Caichen Li, MD, while presenting the research at European Lung Cancer Congress (ELCC) 2026on March 27.
Li, of the First Affiliated Hospital of Guangzhou Medical University, National Center for Respiratory Medicine, Guangzhou, China, and colleagues launched the LUNG-CARE Project, a prospective interventional cohort study involving individuals aged 40-74 years from the same geographic region without a prior lung cancer diagnosis or treatment within 5 years, a chest CT within the past year, or significant cancer-related symptoms.
The participants were enrolled between December 1, 2015, and July 31, 2021, and underwent a single low-dose CT scan and were followed up for incident lung cancer, as well as for survival endpoints.
They were then compared with a “natural control” cohort of individuals undergoing routine medical care and who had not received systematic screening, obtained from the Guangzhou CDC Registry.
Presenting updated and expanded results from those published last year, Li reported that 11,708 individuals were included in the screening group and 114,392 in the control group. The screening group was “slightly older,” he said, with 48.2% vs 46.5% aged 60 years or older, and were more likely to be current smokers, at 17.7% vs 10.9%, and heavy smokers (≥ 20 pack-years), at 12.4% vs 7.6%.
Lung cancer was detected in 227 (1.9%) individuals from the screening group, whereas a lung cancer was diagnosed in 1105 (1.0%) individuals from the unscreened control group.
Li said they saw a “clear stage shift between the two groups,” with around 81.5% of lung cancers in the screening group diagnosed as stage I vs 25.1% in the nonscreening group. Advanced stage disease (III and IV) accounted for 69.7% of cases among control individuals vs 13.7% with screening.
Almost all (93.4%) of the cases in the screening group were adenocarcinoma. Histology heterogeneity was greater in the nonscreening group, with 11.7% of cases squamous cell carcinoma and 5.9% small-cell carcinoma, Li said.
After a median follow-up of 7.0 years, low-dose CT was associated with a significant reduction in lung cancer-specific mortality vs no screening, at a hazard ratio [HR] of 0.45 (P < .001). The effect was more pronounced in women at an HR of 0.28 (P < .001) vs 0.55 (P = .004) in men.
Overall survival at 5 years was also greater with screening, at 87% vs 39% without, or an HR of 0.13 (P < .001).
Li said high-risk patients had significantly poorer survival, regardless of whether using the US National Comprehensive Cancer Network (age ≥ 50 years and ≥ 20 pack-year smoking history) or the more expansive Chinese risk definitions (age > 40 years and one or more risk factors from: 20 pack-year smoking history; environmental/occupational exposure; history of chronic obstructive pulmonary disease/pulmonary fibrosis/tuberculosis; or family history of lung cancer), or when restricting the analysis to patients with respiratory comorbidities or to heavy smokers.
Comparing their results with those from screening trials such as National Lung Screening Trial (NLST), TALENT in Taiwan, and the Female Asian Nonsmoker Screening Study (presented at WCLC 2025), Li said that, despite the broader eligibility of their study, the lung cancer detection rate was comparable, suggesting that these selection criteria for traditional high-risk based screening may be unnecessary.
The results call for the “evaluation of screening strategies beyond conventional risk-based criteria,” and “highlight the need for randomized trials in a nonhigh risk population before broader implementation is considered,” Li specified.
Study Limitations
Li acknowledged that the study was limited by being a nonrandomized comparison and having an imbalance in baseline and disease characteristics between the screened and unscreened cohorts.
Randomized control studies such as the NLST and NELSON have “demonstrated that low-dose CT screening can reduce lung cancer mortality among individuals with a heavy smoking history,” Li noted.
Li also pointed out that current lung screening trials have employed selective strategies targeting high-risk individuals, “rather than evaluating screening in a non-risk-based population.”
Joachim GJV Aerts, MD, PhD, Department of Pulmonary Diseases, Erasmus MC, Rotterdam, Netherlands, who was not involved in the new research, highlighted that the new study started out as a single-arm prospective cohort study with participants recruited, among others means, via social media, and that that control cohort was added later.
Discussing the results on behalf of the European Society for Medical Oncology, he continued that, as the screening population was slightly older and more likely to be smokers, they would likely have been included in a regular risk-based screening study.
Moreover, without follow-up screening, the impact of screening is diluted after about 3-5 years by the normal occurrence of incident lung cancer, as shown in the NELSON trial, in which the lung cancer incidence detection rates were similar between the screening and control groups after approximately 7 years.
Overall, Aerts said that the study “shows, again, that CT screening detects lung cancer in an early stage,” that it increases lung cancer survival, and that the impact is greater in a more healthy population, with the effect lasting “for over 4 years,” although longer follow-up data are required.
Results May Be a ‘Game-Changer’ for Asian Populations
“Current screening guidelines were built around smoking history and in doing so, they leave behind a large and growing group of people who develop lung cancer despite never having smoked,” said Marina Garassino, MD, professor of medicine, hematology and oncology, The University of Chicago Medicine, Chicago, who was not involved in the study, in a press release.
“In Asia, this is not a marginal issue: never-smoking women represent a substantial share of all lung cancer cases, driven by factors like air pollution and genetic risk rather than tobacco.”
Garassino underlined that the current results may be a “game-changer for Asian populations, but we should resist the temptation to over-generalize.”
“Lung cancer in Asia follows a different epidemiological playbook…guidelines built on Western smoking-based data simply do not serve these populations,” she explained.
“What this study does demand, urgently, is updated criteria that recognize Asian ancestry as an independent risk factor for screening eligibility,” she said.
Li declared having no relevant financial relationships. Aerts declared relationships with Eli-Lilly, MSD, AstraZeneca, Accord, PharmaMar, CureVac, Amphera, Nutricia, Genmab, Vivace, Summit, and Amphera. Garassino declared having relationships with AstraZeneca/MedImmune, Bristol‐Myers Squibb, GlaxoSmithKline, MSD, Roche, Takeda, Bayer, Blueprint Medicines, Celgene, Daichii Sankyo, Incyte, Inivata, Janssen, Lilly, Novartis, Pfizer, Regeneron, Sanofi, Seattle Genetics, and Tiziana Life Sciences.
MRI-Based ENZIAN Shows Promise in Diagnosing Deep Endometriosis
Edited by Mamta Pawara
December 12, 2025
TOPLINE:
MRI using the ENZIAN and #ENZIAN scoring systems showed strong reliability in evaluating endometriosis in the ovarian (O) compartment and deep endometriosis in the vaginal/rectovaginal space (A), uterosacral ligaments (B), and rectosigmoid colon (C), with the best diagnostic accuracy in compartment O.
METHODOLOGY:
Researchers conducted a meta-analysis, analysing 12 studies from January 2005 to January 2025 to compare MRI-based ENZIAN (five studies) and #ENZIAN (seven studies) classification scoring systems with surgery or laparoscopy (nine studies) for endometriosis.
They assessed interobserver and intraobserver agreement per the ENZIAN and #ENZIAN systems across 11 compartments, including the peritoneum (P); tubo-ovarian compartment (T); compartments A, B, C, and O; and F locations — uterus (adenomyosis; FA), bladder (FB), intestine other than the rectum (FI), ureter (FU), and all other locations (FO).
Diagnostic performance was evaluated by calculating the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) using surgery or laparoscopy as reference standards.
The analysis included 2285 women aged 32-36 years with deep endometriosis, with chronic pelvic pain being the predominant symptom (n = 340).
TAKEAWAY:
MRI achieved exceptional diagnostic accuracy for compartments O (sensitivity, 97.5%; specificity, 96.4%; PPV, 97.3%; NPV, 98.3%) and T (sensitivity, 94.1%; specificity, 89.1%; PPV, 76.3%; NPV, 96.8%).
Moderate-to-high sensitivity (76.1%-83.4%) and strong specificity (87.8%-94.8%) were noted in compartments A, B, and C, with reasonably high PPV and NPV.
Sensitivity was notably lower for three F locations: 64.6% for FA, 56.9% for FB, and 66.8% for FI.
Inter and/or intraobserver agreement was excellent or substantial for compartments O, A, B, and C in most studies but was inconclusive for other compartments.
IN PRACTICE:
For endometriosis, “MRI showed high diagnostic values for compartments O (ovary), T (tuba-ovarian condition), A (vagina/rectovaginal space), B (utero-sacral ligaments), and C (rectum/sigmoid colon). Substantial to high agreement was observed in compartments O, A, B, and C. The diagnostic values of the P and F-compartments were either low, or could not be pooled, due to limited data, and therefore no conclusions could be drawn. Clear definitions and training might improve the role of MRI in the #ENZIAN score,” the authors wrote.
SOURCE:
The study was led by Yusra Mohamed, University of Amsterdam, Amsterdam, Netherlands. It was published online on November 29, 2025, in European Radiology.
LIMITATIONS:
The study was limited by potential bias from incomplete patient selection; reduced comparability as not all patients underwent a reference standard or did so within a 6-month interval; reduced confidence in diagnostic interpretations because radiologist experience was unspecified in approximately half of the studies; and the inability to extract definitions of F locations, hampering assessment of these compartments. The inclusion of studies solely from Europe limited generalisability.
DISCLOSURES:
The authors reported that this study did not receive any funding and declared having no conflicts of interest.
AI tool predicts Alzheimer’s disease with nearly 93% accuracy using brain scans
Could an AI tool analyze MRI scans and identify patterns associated with Alzheimer’s disease? Image credit: Bisual Studio/Stocksy
Researchers have developed a machine-learning model that analyzes MRI brain scans to detect Alzheimer’s disease, achieving 92.87% accuracy in distinguishing mild cognitive impairment or Alzheimer’s.
The model identified structural patterns associated with cognitive decline, with volume loss in specific brain regions emerging as a possible early biomarker of the disease.
Researchers also found sex-related differences in brain changes, suggesting that biological factors, such as hormonal changes, may influence how Alzheimer’s develops.
Alzheimer’s disease is a progressive condition that can cause memory loss and cognitive decline. Detecting Alzheimer’s typically requires a comprehensive medical evaluation, which a person may only receive after presenting symptoms that could suggest a decline in brain functioning.
Alzheimer’s disease slowly worsens over time, and an early, accurate diagnosis can be beneficial for treating the progression of the disease. However, early diagnosis is often challenging because initial symptoms may resemble typical age-related changes in memory or thinking
As such, diagnostic methods to diagnose Alzheimer’s disease early, or even predict the onset of symptoms, could be critical for maximizing the effectiveness of emerging, disease-modifying treatments and manage the condition.
A new study, published in Neuroscience, highlights an artificial intelligence (AI) tool that analyzes MRI scans and identifies patterns of brain volume loss associated with Alzheimer’s disease.
Findings indicate that the model could accurately predict the condition, suggesting that machine-learning techniques may help detect the disease earlier than traditional diagnostic approaches.
Training AI to recognize Alzheimer’s
To develop the predictive model, researchers from Worcester Polytechnic Institute analyzed 815 MRI scans from participants aged 69 to 84.
The scans came from the Alzheimer’s Disease Neuroimaging Initiative, a large research project that collects imaging data from people with normal cognition, mild cognitive impairment, and Alzheimer’s disease.
As Alzheimer’s disease injures neurons and leads to a loss of brain tissue, these scans could contain subtle changes that may indicate early disease development.
The team used a machine learning model to measure brain volume across 95 different regions. An algorithm then compared these measurements to identify patterns distinguishing healthy brains from those affected by cognitive impairment or Alzheimer’s disease.
The results showed that the model could reliably classify the scans and predict disease presence with nearly 93% accuracy.
Key brain regions linked to Alzheimer’s
From analyzing the large dataset of brain scans, the researchers also identified several brain regions where structural changes were strongly associated with the disease.
Notably, volume loss in the hippocampus, amygdala, and entorhinal cortex were among the strongest indicators of Alzheimer’s disease across age and sex groups.
The hippocampus plays a key role in memory and learning, the amygdala regulates emotions, and the entorhinal cortex is involved in memory, navigation, and perception, and among the first parts of the brain to be affected by Alzheimer’s disease.
Interestingly, researchers also found that individuals aged 69 to 76, the youngest group studied, commonly showed volume loss in the right hippocampus, suggesting this region may serve as an early biomarker for the disease.
Medical News Today spoke with Dung Trinh, MD, internist for the MemorialCare Medical Group and chief medical officer of the Healthy Brain Clinic in Irvine, CA, about the possible role of the right hippocampus.
“The paper points to the hippocampus as one of the earliest and most consistently structures in Alzheimer’s affecting memory, with rapid tissue loss occurring early in the disease process,” Trinh told us.
“In this dataset, the 69 to 76 age group showed substantial right hippocampal volume decreases, which likely means that this region was sensitive to subtle early-stage neuro degeneration before more widespread cortical changes became dominant,” he detailed.
“I would frame it as a promising signal rather than a definitive standalone biomarker because the study is still based on one cohort and internal validation only,” noted Trinh.
Differences between male and female brains
Additionally, the study also uncovered some differences in how Alzheimer’s may affect male and female brains.
In female brain scans, volume loss was more prominent in the left middle temporal cortex. This is a region associated with language and visual processing.
However, in male brain scans, changed were more pronounced in the right entorhinal cortex.
The researchers propose that these differences may relate to hormonal changes linked with aging, such as declines in estrogen and testosterone, which have been previously associated with Alzheimer’s risk in females and males.
Trinh noted that he found the sex-specific asymmetries interesting and agrees with the authors assumption as a possible reason:
“The authors discuss a biologically credible framework involving hormonal change, especially reduced estradiol after menopause, genetic risk such as [the] APOE-e4[genetic variant], and neuroinflammatory processes interacting with amyloid and tau pathology. Those factors though were not directly measured in this study, so they should be viewed as possible explanations rather than proven causes.”
Next steps for the research
The research team plans to continue refining their predictive models using more advanced deep-learning approaches.
Trinh cautioned that while the study shows promise, further validation is still necessary:
“AI-based imaging can detect multiregional structural patterns that may be hard to appreciate by eye, and this study suggests those patterns may emerge across the transition from cognitively normal to mild cognitive impairment to Alzheimer’s. If future validation occurs, it could help clinicians identify higher-risk patients earlier, monitor progression more closely, and eventually tailor treatment plans around an individual’s neuroanatomical profile.”
“In practice, that could mean earlier intervention, better patient selection for disease-modifying therapies, and closer monitoring of those most likely to decline. But I would stress that this paper shows promise, not clinical readiness,” he noted.
“It would help to combine MRI with other biomarkers — for example amyloid, tau, blood-based biomarkers, genetics, and longitudinal follow-up — to show whether the model predicts real-world progression, not just classification within one dataset,” added Trinh.
The research team also aim to investigate other factors that may influence Alzheimer’s development, including conditions such as diabetes.
If validated in larger populations, the research team suggests that AI-based tools could eventually help clinicians identify individuals at risk for Alzheimer’s disease earlier, improving both diagnosis and the ability to test new therapies.
Jump to Key Points
