SSAT - PERFORMANCE CHARACTERISTICS OF A NEW FLUORESCENCE IN SITU HYBRIDIZATION (FISH) PROBE SET ON EUS FNA AS A STAND ALONE OR REFLEX APPROACH FOR THE CYTOLOGIC DIAGNOSIS OF PANCREATIC DUCTAL ADENOCARCINOMA

Back to 2018 Posters

PERFORMANCE CHARACTERISTICS OF A NEW FLUORESCENCE IN SITU HYBRIDIZATION (FISH) PROBE SET ON EUS FNA AS A STAND ALONE OR REFLEX APPROACH FOR THE CYTOLOGIC DIAGNOSIS OF PANCREATIC DUCTAL ADENOCARCINOMA
Michael J. Levy^*, Amber Schneider, Emily Barr Fritcher, Benjamin R. Kipp, Sarah Kerr, Ferga C. Gleeson
Mayo Clinic, Rochester, MN

BACKGROUND
Conventional cytology (CC) analysis of EUS-FNA pancreatic ductal adenocarcinoma (PDAC) specimens provides a diagnostic sensitivity and specificity of 60-95% and 99-100%, respectively. A clinical pancreatobiliary FISH probe set has been validated for enhancing diagnostic sensitivity in biliary brushings, but not FNA specimens. FISH utilizes fluorescently labeled DNA probes to detect cells that have copy number abnormalities indicative of malignancy. Aneuploidy is not present in all malignancies. Premalignant disease states may also produce aneuploid cell populations. Furthermore, not all FNAs from malignant tumors obtain malignant cells. In the context of these diagnostic challenges, our aim was to determine the performance characteristics of the pancreatobiliary FISH probe set in PDAC when used alone or in combination with CC.

METHODS
We retrospectively selected a diagnostically challenging cohort that provided a spectrum of CC interpretations: positive (n=25), suspicious (n=36), atypical (n=18), and negative (n=21) for malignancy from 100 patients (age 64.5±12.4 years; male 68%) who underwent EUS FNA for presumed PDAC. A smear containing the most atypical cells was de-stained and hybridized with 4 locus specific probes 1q21, 7p12, 8q24, and 9p21. FISH was scored by scanning for abnormal cells and determining signal copy numbers. A FISH positive result was defined by ≥ 3 copies of ≥ 2 probes in 4 or more cells, excluding tetrasomy. Single locus gain, deletions, and tetrasomy were deemed negative.

RESULTS
A final diagnosis of malignancy (n=62) or benign disease (n=38) was established, including: PDAC (n=62), chronic pancreatitis, (n=13), AIP (n=5), IPMN (n=4), acute pancreatitis (n=3), lipoma (n=1), and indeterminate benign disease (n=12). Mean slide age for FISH analysis was 5.5 years (range, 1-10 years). CC and FISH provided a sensitivity, specificity, and accuracy of 40.3% vs. 37.1%, 100% vs. 89.5%, and 63% vs. 57%, respectively. FISH failed due to lack of adequate hybridization signals in 11 patients; attributing to the lower sensitivity. When using FISH as a reflex technique, PDAC was diagnosed in 9 additional patients, but 4 false positive results occurred.

CONCLUSIONS
Our data indicate poor performance of the pancreaticobiliary FISH probe set when used as a stand-alone test. While FISH/CC combination testing enhanced the sensitivity by 14.5%, the specificity reduced to a clinically unacceptable rate of 89.5%. Given that the analysis was conducted in an intentionally challenging cohort with a low CC sensitivity and FISH was retrospectively performed, additional data are needed to confirm our findings.

Test Result Sensitivity Specificity Accuracy
Routine Cytology (Only Pos = Pos) 25/62 (40.3%) 38/38 (100%) 63/100 (63%)
FISH (All patients) 23/62 (37.1%) 34/38 (89.5%) 57/100 (57%)
FISH (exclude inadequate samples) 23/51 (45.1%) 34/38 (89.5%) 57/89 (64%)
Composite Testing (RC or FISH Pos) 34/62 (54.8%) 34/38 (89.5%) 68/100 (68%)

Back to 2018 Posters