Selection of Samples for Epidermal Growth Factor Receptor (EGFR) Mutation Analysis in Non-Squamous Non-Small Cell Lung Carcinoma
Carolyn J Shiau, Jesse Babwah, Gilda da Cunha Santos, Scott L Boerner, William R Geddie, Suzanne Kamel-Reid, Cuihong Wei, David M Hwang, Ming S Tsao. University Health Network, Toronto, ON, Canada
Background: Tyrosine kinase (TK) domain mutations in the EGFR gene are well documented in non-small cell lung cancer, with molecular testing established as a critical factor to initiate first-line treatment with EGFR kinase inhibitors. We reviewed data from the first 18 months of EGFR mutation testing in a province-wide population to identify best practice principles for the selection of tumor samples.
Design: Data from EGFR mutation analysis were collected from 1853 consecutive cases starting from March 2010. Samples were reviewed by pathologists to evaluate material for testing, including tumor classification, sample type, tumor cellularity, and immunohistochemical profile. Mutation analysis was performed by PCR fragment analysis for exon 19 deletions and Sau961 restriction enzyme digest for exon 21 L858R mutation. The significance of rates of mutation and test success were analyzed by chi-squared comparison testing.
Results: 1305 histology and 368 cytology samples were evaluated for testing. 176 cases (9.0%) were excluded at the time of review, most frequently due to inadequate cellular material (n=136). Analysis performed on 1677 samples revealed an EGFR mutation rate of 21.2% with approximately equal prevalence for exon 19 and L858R mutation (11.0% and 10.2%). EGFR mutation rate was higher in cytology (26.7%) versus histology (19.6%) samples, with the greatest difference in mediastinal lymph node samples (40.0% in 43 cytology, 10.1% in 125 histology, p=0.0015). The overall rate of inconclusive test results was 6.5% (n=109) with no significant difference between histology and cytology samples. The most significant factor for test success in histology samples is high percentage of tumour cellularity, regardless of sample type. TTF-1 status was noted in 1156 histology samples (929 positive, 227 negative). 13 of 212 (6.1%) of EGFR mutatant cases were TTF-1 negative, with a negative predictive value of 93.7%.
Conclusions: Histo/cytological review of samples prior to testing will remove cases with inadequate cellular material for testing. Ideal tumor cellularity should be over 30% to improve test success rates. Tumors negative for TTF-1 by immunohistochemistry have a low likelihood of EGFR mutation. Cytology samples are equivalent to histology samples for test success rate and may have advantages for detecting EGFR mutation.
Monday, March 19, 2012 9:30 AM
Poster Session I Stowell-Orbison/Surgical Pathology/Autopsy Awards Poster Session # 299, Monday Morning