Application of ESCAT and OncoKB scales in Liquid biopsy (LB) in Advanced NSCLC patients (pts): Is it feasible and reliable?

Date 28 September 2019
Event ESMO 2019 Congress
Session Poster Display session 1
Topics Non-Small Cell Lung Cancer
Pathology/Molecular Biology
Presenter Michael McCusker
Citation Annals of Oncology (2019) 30 (suppl_5): v797-v815. 10.1093/annonc/mdz269
Authors M.G. McCusker, A. Russo, J. Chen, K.A. Scilla, R. Mehra, C.D. Rolfo
  • Thoracic Oncology & Early Clincial Trials, Medicine Department, University of Maryland Marlene and Stewart Greenebaum Comprehensive Cancer Center, MD 21201 - Baltimore/US



Next generation sequencing (NGS) of cfDNA in NSCLC is entering in clinical practice. Several commercially available platforms using NGS report wide variety of somatic aberrations. Vendors also include therapeutic suggestions to guide oncologists. In addition, new levels of evidence tools are now available for tissue results, but not still used in LB.


Advanced NSCLC pts underwent commercial 73-gene cfDNA NGS analysis. ESMO Scale for Clinical Actionability (ESCAT) and OncoKB were used to grade levels of evidence for categorize aberrations and compared with variant allele frequency (VAF), treatment decisions, and vendor suggestions.


77 samples from 73 advanced NSCLC pts (73% adenocarcinoma, 27% squamous cell carcinoma) at the time of diagnosis (49%) or during disease course (51%) were analyzed. Median turnaround time was 8 days (range 5-17), with no genotyping failures. There were 323 unique somatic alternations identified: Sequence mutations (254), amplifications (43), synonymous mutations (23), and fusions (3). Median cfDNA % was 0.7 (range 0.05-49.5); 7 samples had no detectable genetic alterations. Variants of unknown significance were 33% of point mutations. We detected 87 and 88 potentially actionable genetic alterations according to ESCAT (IA-IV) and OncoKB (1-4), respectively, and 26% received a matched targeted drug. Discrepancies between these two tools and vendor suggestions were reported in 4 cases. We performed a subset analysis of 64 samples: Median VAF was 4.37 % (range 0.16-43.05) and a VAF < 1% was reported in 16 samples; 45% of alterations (excluding amplifications) were clonal events (cfDNA% divided by VAF > 0.5). Among EGFR and ALK positive samples median VAF was 7.98% (range 0.29-41.2); 3/17 samples presented a VAF below 1%, with no detrimental effect on treatment response.


The application of ESCAT and OncoKB is feasible in LB. Driver mutations with low VAF are amenable to receive treatment. VAF could be included as complementary tool to grading systems to better understand the significance of aberrations.Discrepancies between vendor therapeutic suggestions and evidence-based grading systems requests caution in the use of information outside the molecular tumor board.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.


Has not received any funding.


C.D. Rolfo: Honoraria (self), personal fees : Novartis; Honoraria (self), personal fees : MSD; Non-remunerated activity/ies, non-financial support : OncoDNA; Honoraria (self), personal fees and non-financial support : GuardantHealth. All other authors have declared no conflicts of interest.