1226P - Predictive value of circulating tumor DNA (ctDNA) before and shortly after curative treatment in early stage non-small cell lung cancer (NSCLC), and exploration of (pre-)analytical factors

Background

Data suggest ctDNA detection in the weeks following curative treatment is predictive of outcome in early-stage NSCLC, potentially guiding adjuvant treatment. Clinically, it would be convenient if samples collected before or shortly after resection were predictive, even when sample is limited and ctDNA levels are low.

Methods

We investigated the value of plasma collected pre- and 1-3 days after treatment (d1-3), in patients from the LEMA and LUCID studies, tested using RaDaR®. Tumor exome sequencing (WES) guided assay design.

Results

In LEMA, 99% of surgical resection and 77% of diagnostic biopsies yielded sufficient DNA for WES and assay design. ctDNA was detected pre-treatment in 48% of 87 patients, with median eVAF 0.02%. Detection rose with stage; 19%, 64% and 92% in I, II and III, as did median eVAF; 0.01%, 0.02% and 0.06%. These data were comparable to LUCID (Gale et al, 2022), with ctDNA detected in 51% patients, including 24%, 77% and 87% with stage I, II and III. Considering both cohorts, pre-treatment ctDNA associated with reduced recurrence free- (RFS; HR=2.5; 95% CI 1.5-4.2, p<0.001) and overall- survival (OS; 2.04; 1.1-3.7, p=0.02). Sensitivity, specificity, positive- (PPV) and negative- predictive value (NPV) were 66%, 60%, 49% and 75%. Pre-treatment ctDNA was detected with eVAF as low as 0.0008%. Underlining the value of detection to such levels, we observed worse RFS (2.5; 1.3-5.0, p=0.01) and OS (2.3; 1.1-4.9, p=0.03) in patients with pre-treatment ctDNA detected at <0.01% vs. ctDNA negative patients. Detection <0.01% was possible with as little as 0.7mL plasma and ∼4ng DNA input. Compared to pre-treatment, d1-3 ctDNA was detected as low as 0.00002% eVAF, in 19% of 80 samples, and was associated with worse RFS (8.6; 3.8-19.6, p>0.001) and OS (11.2; 4.2-30.2, <0.001). Sensitivity, specificity, PPV and NPV were 44%, 89%, 53% and 85%.

Conclusions

We show sensitive tumor-informed detection of ctDNA is possible even from ‘real-world’ samples of suboptimal quantity, and is prognostic in samples collected before and soon after treatment. The predictive value of ctDNA may be useful to guide (neo)adjuvant treatment. An expanded analysis will be presented.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

NeoGenomics.

Disclosure

C.G. Smith: Other, Institutional, Full or part-time Employment: NeoGenomics. R. Rintoul: Financial Interests, Institutional, Speaker, Consultant, Advisor: Olympus Medical; Non-Financial Interests, Institutional, Leadership Role: UK Lung Cancer Coalition; Non-Financial Interests, Institutional, Research Grant, Project grant funding to institution: Asthma and Lung UK, UKRI Medical Research Council, Cancer Research UK; Non-Financial Interests, Institutional, Funding, Salary support: NIHR Cambridge Biomedical Research Centre, Cancer Research UK. K. Monkhorst: Non-Financial Interests, Institutional, Speaker, Consultant, Advisor, Consulting fees. To my institution: Lily, Bayer, Amgen. N. Rosenfeld: Non-Financial Interests, Institutional, Funding, Institution: Cancer Research UK, University of Cambridge, Cambridge University Hospitals, Royal Papworth Hospital; Non-Financial Interests, Institutional, Funding, In-kind contribution: NeoGenomics Ltd, Inivata Ltd; Non-Financial Interests, Institutional, Research Grant, Grants to institution: Cancer Research UK, UKRI, AstraZeneca; Financial Interests, Personal and Institutional, Royalties: Inivata Ltd, NeoGenomics Ltd; Non-Financial Interests, Personal, Speaker, Consultant, Advisor, Consulting fees: Inivata Ltd; Financial Interests, Personal, Stocks/Shares: NeoGenomics Ltd, Inivata Ltd. All other authors have declared no conflicts of interest.