Occult metastases and survival of lung cancer by clinical diagnosis and CT screening: A simulation study.

Objectives: It is significant to know how much early detection and screening could reduce the proportion of occult metastases and benefit NSCLC patients.

Methods: We used previously designed and validated mathematical models to obtain the characteristics of LC in the population including undetectable metastases at the time of diagnosis. The survival was simulated using the survival functions from Surveillance, Epidemiology and End Results (SEER) data stratified by stage.

Results: Based on the simulations, 35.3% of patients diagnosed with stage N0M0 and 56.9% of those diagnosed with stage N1M0 had nodal or distant metastases that were not discovered at the time of diagnosis. Among clinically detected Stage I lung cancers with tumor diameter 1-2 cm, 78% were true stage N0M0 (no occult metastases) while it was only 37% for patients with tumor diameters of 2-3 cm. This size threshold can be translated into a 0.75-year the "window of opportunity" for the curable disease. In a comparative analysis of two simulated groups of individuals: (1) clinically diagnosed (2) diagnosed by screening with a varying screening frequency (quarterly, biannual, annual and biennial), it was estimated that, once the screening intervals become shorter, substantially more cancers are found, but at an expense of a higher radiation exposure. The simulation projected that the mortality reduction in screened patients depending on the frequency, ranged from 15.04% to 18.82%.

Conclusions: The probability of occult metastases significantly increases when the primary tumor exceeds 2 cm in diameter. Effective screening measures that detect smaller tumors will considerably benefit asymptomatic LC patients.