Personalised screening intensity based on existing lung cancer risk and spirometry

Area of Science:

• Pulmonology
• Oncology
• Radiology
• Biostatistics

Background:

• Current lung cancer screening programs typically use fixed annual or biennial intervals.
• Risk models may inform personalized screening intervals to enhance cost-effectiveness and decrease interval cancers.
• Spirometric airflow obstruction is a potential biomarker for lung cancer risk that warrants investigation for predictive value.

Purpose of the Study:

• To evaluate the performance of existing risk models in predicting next-round risk for lung cancer screening.
• To assess whether incorporating spirometric data (FEV1 % pred) improves risk prediction and screening efficiency.
• To determine if personalized screening intervals based on risk models can optimize lung cancer screening programs.

Main Methods:

• Risk models (PLCOm2012, LCRAT+CTneg) were tested for predicting next-round risk in Manchester and National Lung Screening Trial (NLST) cohorts.
• Models were adapted in an NLST sub-study (ECOG-ACRIN) to include FEV1 % pred.
• Model performance was evaluated using discrimination metrics like Area Under the Curve (AUC).

Main Results:

• Higher risk scores correlated with increased lung cancer detection rates at the next screening interval.
• The PLCOm2012 model showed an AUC of 0.72, while LCRAT+CTneg achieved an AUC of 0.75 on external validation.
• Adding FEV1 % pred provided only a modest improvement in discrimination (ΔAUC 0.01–0.04).

Conclusions:

• Baseline risk assessment can identify lower-risk individuals for extended screening intervals, reducing costs and harms.
• Focusing resources on high-risk individuals can prevent delayed detection of lung cancer.
• FEV1 % pred is unlikely to significantly enhance next-round risk prediction in lung cancer screening.