Performance of Cytologic Specimens in EGFR and KRAS Molecular Testing with a Focus on Minimal Cellularity Requirements.
Suzanne M Brandt, Laura J Tafe, Maria E Arcila, Andre L Moreira, Maureen F Zakowski, Marc Ladanyi, Natasha Rekhtman. Weill Cornell Medical College, New York, NY; Memorial Sloan-Kettering Cancer Center (MSKCC), New York, NY
Background: Testing for EGFR and KRAS mutations is used to guide molecular-targeted therapy for lung adenocarcinoma. We evaluated the performance of cytologic specimens in detecting EGFR/KRAS mutations in clinical practice with a focus on minimal cellularity requirements.
Design: Cytologic specimens submitted for EGFR/KRAS molecular testing at MSKCC during a 1-year period (n=128) were reviewed. Most specimens were paraffin-embedded cell blocks, with a mean of 14 sections (5um-thick) submitted for testing after a pathologist's triage as “adequately cellular”. Included were FNAs (n=96), pleural fluids (n=29), and bronchial washings/brushings (n=3). We analyzed the test failure rate, DNA yield, concordance with mutations in other specimens, and clinicopathologic correlates. Manual cell counts were performed to determine the minimal cellularity sufficient to detect the mutations.
Results: Of 128 cytologic specimens, 125 (97.6%) were sufficient for complete EGFR/KRAS testing, 1 (0.8%) was sufficient for partial analysis (EGFR only), and 2 (1.6%) were not analyzable due to PCR failure. Of analyzable samples, 31 (25%) had EGFR mutations and 25 (20%) had KRAS mutations. The mean DNA yield was 1.13ug (range 0.02-18.14ug). Ten samples with the lowest DNA amounts (range 0.08-0.47ug) which yielded EGFR/KRAS mutations had a median tumor cell count of 773 (range 112-1687) per representative section, with tumor cells representing on average 59% of the cellularity (range 14-90%). Twenty-one (16%) of 128 patients had an additional surgical or cytologic specimen that had undergone molecular testing. All mutations identified in multiple samples were concordant, except for specimens from patients with multiple primary tumors (n=4). EGFR mutations identified in cytologic specimens were strongly associated with never-smoker status (p<0.001) and were more common in females (p=0.28), whereas KRAS mutations were associated with smoking (p<0.001).
Conclusions: Various types of cytologic specimens are suitable for EGFR/KRAS testing. Cell blocks triaged by pathologists as “adequately cellular” yield sufficient DNA for molecular testing, with only rare exceptions. Mutations can be identified in cell blocks containing only several hundred tumor cells. Concordance with mutations from other specimens and clinicopathologic correlates similar to those previously established for surgical specimens further support the validity of mutation analysis in cytologic specimens.
Tuesday, March 1, 2011 1:00 PM
Poster Session IV # 283, Tuesday Afternoon