Low-dose computed tomography (LDCT) remains the only validated modality for lung cancer screening, yet participation is suboptimal. Blood-based assays could eventually complement LDCT, but their performance for detecting early-stage disease in screening populations is uncertain. This study evaluated how well a cell-free DNA blood-based screening test can detect early-stage lung cancer in patients with LDCT screen-detected or incidental lung nodules. We recruited patients via email with upcoming LDCT appointments whose results were classified under the Lung-RADS categories of 1–4. Guardant Health LUNAR® test collection kits were mailed to the participant. Participants provided a blood sample on the day of their LDCT visit, or within 30 days of the scan. Sensitivity and specificity were computed using the tumor registry as the gold standard and compared to assay results from the blood draw date closest to the date of diagnosis of cancer, but prior to receipt of any treatment. Of the 765 participants, 18 lung cancers were diagnosed. Most of the study population had a previous lung cancer screen. Eleven cases were diagnosed with stage 1 disease (68.8%). The sensitivity was 43.8% (95% CI:19.4–68.1%) and the specificity was 84.1% (95% CI: 81.5–86.7%), which resulted in a false positive rate of 15.9%. The negative predictive value was 98.6% and the positive predictive value was 5.6%. Exploratory stage-adjusted sensitivity was 75.4% (95% CI:63.6–87.2%). The cfDNA assay demonstrated moderate specificity but limited sensitivity. Given the false positive rate of 15.9%, our results suggest that cfDNA-based assays for lung cancer may play a role as tools in combination with existing screening strategies pending continued validation. This proof-of-concept study is an important step in determining the feasibility of blood-based testing for early lung cancer detection.
Zepp et al. (Wed,) studied this question.