Subtyping of Non-Small Cell Lung Carcinoma (NSCLC): Comparison of Cytology and Small Biopsy Specimens
CS Sigel, MA Friedlander, MF Zakowski, AL Moreira, N Rekhtman. Memorial Sloan-Kettering Cancer Center, New York, NY
Background: The accurate classification of NSCLC has gained importance due to the emergence of histology-based treatment options. The majority of NSCLCs are unresectable at presentation and treatment is usually determined by small biopsy (Bx) or cytology (Cyto) specimens. While Cyto has been established to be superior to Bx for the diagnosis (Dx) of small cell lung carcinoma (SCLC), the relative efficacy of these modalities in subtyping of NSCLC has not been established.
Design: A review of the departmental database was conducted to identify all primary lung carcinomas with concurrent Cyto and Bx specimens including cases with subsequent resection during a two-year period (9/1/2006-9/1/2008). Metastases, SCLCs and carcinoids were excluded. In our clinical practice, Cyto and Bx specimens are reviewed independently. 102 paired specimens were identified (M:F ratio 1:1.9, average age 68, age range 37-91) including fine needle aspirates with core Bx (n=66), and bronchial wash/brush/lavage specimens with transbronchial Bx (n=36). 20 cases had subsequent resection.
Results: Of 102 cases, Cyto Dx was definitive (adenocarcinoma or squamous cell carcinoma) vs favored vs unclassified in 70% vs 19% vs 11%, whereas the distribution for Bx was 72% vs 22% vs 6%, respectively. The unclassified rate was reduced to 4% when the two modalities were considered together. Overall, Cyto was more definitive in 13% of cases, and Bx in 17%. All Cyto Dx except 2 were rendered purely on morphology, whereas immunohistochemistry (IHC) (typical panel TTF-1, p63, 34βE12) aided in the classification of 22% of Bx. Tumor type was concordant in Cyto and Bx for 93%, whereas discordant Dx were reached for 7% of cases (n=7). For these cases, resection and/or additional IHC revealed the correct Dx was rendered by Cyto in 3 cases and Bx in 4 cases. Tumor type in all concordant Cyto/Bx specimens was supported by subsequent resection.
Conclusions: We find that in our routine clinical practice both Cyto and Bx achieve high rate of NSCLC subtyping with comparable accuracy. Combining the modalities reduces the rate of unclassified NSCLC to 4%. Although the use of IHC may be withheld in Cyto with a known paired Bx, morphological features alone are sufficient to subtype the majority of cases. Furthermore, similar to SCLC, Cyto is superior to Bx in classification of NSCLC in a subset of cases. Our data confirms the suitability of both small Bx and Cyto specimens, particularly when combined, for histology-based treatment paradigms.
Wednesday, March 24, 2010 9:30 AM
Poster Session V # 248, Wednesday Morning