Diagnostic Accuracy of Fine-Needle Aspiration Biopsy of the Parapharyngeal Space
AT Arnason, RD Hart, JG Nasser, JR Trites, SM Taylor, MJ Bullock. Dalhousie University, Halifax, NS, Canada
Background: Fine-needle aspiration biopsy (FNA) of the parapharyngeal space (PPS) is a diagnostic challenge. Sampling is difficult (often trans-oral FNA), primary PPS tumors are rare, and there is a broad differential diagnosis, including salivary gland, neurogenic, carotid body, lymphoreticular, and soft tissue tumors. PPS FNA has rarely been studied, with only four series of more than 20 cases, all from major referral centers, reporting high rates of diagnostic accuracy (87-100% accuracy, sample size 24 to 63 FNAs). It is unclear whether this high rate of diagnostic accuracy can be achieved in a smaller tertiary care centre with fewer CT-guided FNAs and a lower volume of routine head and neck cytology.
Design: Pathology records from our institution (an 1100 bed Canadian academic tertiary care centre) were searched to identify all patients who underwent PPS FNA from Sept. 1991 to Aug. 2009. Records were reviewed for all cases. The FNA diagnosis was compared to the gold standard of subsequent histopathology or long term clinical follow up.
Results: 27 patients (12 male, 15 female, mean age 53 years) had 36 FNAs (9 patients had repeat FNAs). 33 FNAs were office based and 3 were done under image guidance. 11/36 (31%) FNAs were non-diagnostic. In the 25 diagnostic FNAs, there was sensitivity 89%, specificity 94%, PPV 89%, NPV 94%, and accuracy 92% for the diagnosis of positive or negative for malignancy. A correct specific diagnosis (e.g. “schwannoma”) was made in 9/25 (36%) cases. The non-diagnostic rate was significantly higher (p<0.025) in FNAs prepared as conventional smear cytology (9/17 = 53%) versus liquid based ThinPrep cytology (2/19 = 11%). A specific diagnosis was made significantly more often (p<0.05) with ThinPrep (8/19 = 43%) versus conventional (1/17 = 5.9%). Comments about specimen quality were less frequent with ThinPrep (2/19 = 11%) than with conventional (8/17 = 47%), where obscuring blood and degeneration were frequently cited by pathologists. One minor complication from FNA occurred, temporary facial nerve palsy secondary to xylocaine.
Conclusions: Our institution has a high rate of diagnostic accuracy for classifying PPS FNAs as benign or malignant, but a lower rate of reporting a specific diagnosis. Non-diagnostic FNAs are frequent and occur more often with conventionally prepared smears than with ThinPrep. Improved specimen quality with ThinPrep seems to be a factor, but other confounding variables not studied here may contribute to this difference.
Wednesday, March 24, 2010 1:00 PM
Poster Session VI # 57, Wednesday Afternoon