Early detection of lung cancer is a highly unmet medical need. Even though low-dose computed tomography (LDCT) has been shown to improve lung cancer mortality in high-risk individuals , the rate of clinical adoption remains low at 1.9 % [2, 3]. Cell-free DNA (cfDNA) testing might substitute LDCT as a screening tool, according to preliminary results of the Circulating Cell-free Genome Atlas (CCGA) Study presented at the ASCO Congress .
CCGA is a prospective longitudinal cohort study designed for early cancer detection. Approximately 15,000 participants will be enrolled, 70 % of whom have cancer, while 30 % do not. In the non-cancer participants, benign comorbid conditions are not excluded, making it a real-world population. The cancer patients have been diagnosed with any malignancy. Non-cancer and cancer participants are enrolled from the same institutions to control for preanalytical variability due to geographic distribution. Blood samples and clinical data are collected from all participants; tissue samples are obtained from the cancer patients.
Characterisation of cancer-specific cfDNA signals
Genome-based screening calls for a broad approach. The researchers apply extensive sequencing of blood samples, including targeted and whole-genome sequencing of cfDNA and white blood cells (WBCs), targeted and whole-genome bisulfite sequencing of cfDNA, and whole transcriptome sequencing of cell-free RNA. This allows for characterisation of all major somatic and epigenetic cfDNA features. Correction for WBC variants is performed, as these are a major source of interference. Tumour tissue undergoes whole genome sequencing. During the 5-year follow-up, patients with cancer are followed up with regard to treatment, recurrence and mortality; in the non-cancer group, any new cancer diagnosis is assessed, as well as treatment and mortality.
More than 12,000 participants have been enrolled at 142 sites in the USA and Canada to date. At the ASCO Congress, the results obtained in a pre-specified case-control sub-study comprising 2,800 participants were presented. This population was divided into a training set (n = 1,406; 118 patients with lung cancer) and an independent test set (n = 834; 46 patients with lung cancer) intended for verification of the findings. Indeed, comprehensive sequencing of plasma cfDNA was shown to generate high-quality data across the spectrum of genomic features, permitting non-invasive cancer detection. The assays used detected lung cancer across stages, histologies, and populations. Importantly, WBC-derived mutations and copy number variations proved a major source of potential false positives and must be accounted for to achieve high specificity. Together, these early results support the promise of using cfDNA-based sequencing to develop an early cancer detection test with high specificity. Further assay and clinical development in large-scale clinical studies is ongoing.