Read time: 5 minutes.
Lung cancer screening has come a long way since low-dose CT (LDCT) scans first became available in 2015 in the United States. However, LDCTs can still miss some lung cancers while causing unnecessary anxiety with false positives. New research from the World Congress on Lung Cancer 2025 reveals how cutting-edge technologies and expanded screening approaches are transforming the early detection of lung cancer.
Making Low-Dose CT More Precise
Current lung cancer screening guidelines for a LDCT scan:
- Aged 50-80 years old, and
- Smoking history of 20 pack-years or more, and
- Currently smoke, or quit within the last 15 years
Unfortunately, these screening criteria only identify 25% of individuals with tobacco exposure as eligible for LDCT screening. In addition, racial and ethnic disparities persist due to differences in access, adherence, and eligibility criteria that are derived from clinical trials with largely non-Hispanic White participants.
New approaches to screening that involve biomarkers and artificial intelligence (AI) are poised to dramatically change the way we think about and implement lung cancer screening.
Protein-based risk models: The INTEGRAL-Risk model combines 13 specific blood proteins with age and smoking history to identify high-risk individuals. Unlike the current screening guidelines, the INTEGRAL-Risk model identified 85% of lung cancers as eligible for screening —more than triple the current rate—while screening the same number of people. The model works equally well across different racial groups and can therefore address longstanding screening disparities.
Enhanced nodule assessment: When LDCT detects a lung nodule, determining if it's cancerous remains challenging. The INTEGRAL-PEN model uses circulating protein biomarkers along with imaging features to predict malignancy risk. This approach achieved 87% accuracy in distinguishing cancerous from benign (non-cancerous) nodules, significantly outperforming current methods like LungRADS (65% accuracy). It is particularly effective for small nodules under 10 mm, where current tools are limited in distinguishing cancer from non-cancer.
AI-powered prediction tools: Artificial intelligence models like Sybil can analyze existing CT scans to predict future lung cancer risk. In a diverse patient population, this multimodal AI process achieved 82% accuracy with 96% specificity and therefore has the potential to identify high-risk individuals from routine imaging studies.
Incidental Pulmonary Nodule Programs Can Complement LDCT Efforts
Hospitals perform thousands of CT scans monthly for various reasons, such as chest pain, injury evaluation in an emergency room, and heart disease screening. Many of these scans accidentally capture incidental lung nodules that could be early lung cancers.
The two studies described below showcase how incidental nodule programs can increase our ability to detect early lung cancers.
AI-based detection systems: Memorial Healthcare System in Florida used AI to analyze over 200,000 radiology reports, identifying 6,685 people with lung nodules that required follow-up work. This approach detected 127 lung cancers compared to just 10 found through traditional LDCT screening in the same period—more than 12 times as many cancers. Most of these cancers were caught at early, treatable stages.
Systematic management programs: The Mayo Clinic implemented automated systems to ensure proper follow-up of incidentally found nodules. By systematically managing 25,944 patients with lung nodules, they identified eight additional cancers that might otherwise have been missed due to inadequate follow-up.
Other Risk Factors for Lung Cancer
Experts now acknowledge that tobacco exposure is one of several risk factors for lung cancer. Screening programs are therefore expanding beyond traditional tobacco-based criteria to capture the growing population of lung cancers in people without any tobacco exposure (never-smokers) and those with different risk profiles.
Age considerations: Current guidelines often exclude older people from screening programs. Screening data from the United Kingdom shows that though people aged 75-80 have higher overall mortality from screen-detected lung cancer, those who undergo surgery have survival rates identical to younger people. This suggests screening decisions should focus on surgical fitness rather than biological age alone.
Family history and genetic predisposition: Taiwan's national program screens both individuals with heavy tobacco exposure (current guidelines) and those without any tobacco exposure but with family history of lung cancer. Those with family history actually had higher detection rates (1.7%) than heavy tobacco users (0.8%), with a projected 67% reduction in lung cancer deaths.
These data from Taiwan challenge the traditional tobacco-only screening criteria. A study from India shows that among people with lung cancer and a strong family history of cancer, 15% carried inherited genetic variants in DNA repair genes like BRCA1/2 and ATM. These findings support genetic testing for lung cancer patients and those with significant family history to help identify relatives who may benefit from lung cancer screening. It is important to note that both these studies were conducted in Asia, and the data may reflect the unique biology of lung cancer in these regions.
Air pollution: Air pollution, particularly PM2.5 particles, increases lung cancer risk in people without any tobacco exposure through tumor-promoting mechanisms. Studies show exposure levels between 20-25 μg/m³ create peak cancer risk, with laboratory evidence that PM2.5 can directly transform normal lung cells into cancerous ones.
The WCLC 2025 discussions on early detection represent a fundamental shift from one-size-fits-all screening to personalized risk assessment. The future of early detection will probably be a combination approach of biomarkers, AI analysis, genetic testing, and environmental factors.
Make sure to read all of our 2025 World Conference on Lung Cancer coverage: