Related Experiment Video
Updated: Jun 30, 2026
MicroRNA Based Liquid Biopsy: The Experience of the Plasma miRNA Signature Classifier MSC for Lung Cancer Screening
Published on: October 26, 2017
Budget Impact of the LungFlag™ Predictive Risk Model for Lung Cancer Screening
Carolina Heuser1, Michael K Gould2, Eran Choman3
1F. Hoffmann-La Roche Ltd, Grenzacherstrasse 124, Building 002, 4070, Basel, Switzerland. carolina.heuser@roche.com.
Background/Objective:
Lung cancer screening has historically been associated with false-positive results that can lead to complications and increased costs. LungFlag is a machine-learning risk prediction model using individual-level data to identify people at high risk of developing non-small cell lung cancer (NSCLC), prompting healthcare professionals to consider appropriate actions such as recommended screening with low-dose computed tomography. This study evaluated the budget impact of using the LungFlag risk prediction model to identify candidates for lung cancer screening from a US payer perspective.
Methods:
This budget impact model estimated total annual costs with or without LungFlag for a hypothetical 1 million-member US commercial health plan. Incremental costs were evaluated over a 5-year period with a cycle length of 1 year; no discounting was applied. All costs were adjusted to 2021 US dollars. The population included screening-naïve individuals aged 50-80 years with a ≥20-pack-years smoking history who were either current smokers or former smokers who had quit smoking in the previous 15 years (US Preventive Services Task Force 2021 criteria). LungFlag performance was measured 9-12 months before clinical diagnosis. Sensitivity and scenario analyses explored the robustness of base-case assumptions.
Results:
The budget impact model showed that LungFlag was associated with a total cumulative cost savings of US$2.8 million over 5 years, largely attributable to reduced advanced NSCLC treatment costs (-$4.4 million) as more patients were identified with early stage I rather than stage III/IV disease. Using LungFlag, an estimated 17 additional NSCLC diagnoses were identified, 22 fewer NSCLC-related deaths occurred over 5 years, 50 more patients had stage I NSCLC at diagnosis, and 33 fewer patients had stage III or IV NSCLC at diagnosis.
Conclusions:
Using LungFlag to identify candidates for lung cancer screening was estimated to save US$2.8 million over 5 years from a US commercial health plan perspective, largely attributable to reduced costs of treating advanced NSCLC.
