The objective of this study was to determine the accuracy of image-guided f
ine needle aspiration cytology (FNAC) in the diagnosis of pulmonary lesions
. A retrospective study was undertaken of 286 patients with 288 lesions, wh
o underwent a total of 302 procedures. The FNAC diagnoses were reported as
malignant, suspicious, atypical, benign or non-diagnostic. Subsequently the
FNAC diagnoses were correlated with either the histological or clinical di
agnoses. Of the 288 lesions, 64.6% were reported on FNAC as malignant, 2.1%
suspicious, 2.4% atypical, 20.8% benign and 10.1% non-diagnostic. On revie
w of the suspicious, atypical, selected benign cases and non-diagnostic FNA
C by an independent pathologist there was agreement with the original FNAC
diagnosis in all cases. All of 186 malignant FNAC diagnoses were confirmed
malignant either clinically or on subsequent histology. Four of the six sus
picious FNAC diagnoses had a malignant outcome, one patient had organising
pneumonia on excision biopsy and one was lost to follow up. Six of the seve
n atypical FNAC diagnoses were confirmed on histology as malignant, while o
ne lesion resolved spontaneously. Fifty-two of 60 benign FNAC diagnoses wer
e confirmed benign either clinically or on histology. Seven of the lesions
diagnosed as benign on FNAC were proven to be malignant. One patient with a
benign FNAC diagnosis was lost to follow-up. Ten of the 29 non-diagnostic
FNAC group were later shown on clinical or histological follow up to be mal
ignant. This study shows that image guided FNAC for the diagnosis of malign
ant pulmonary lesions has a sensitivity of at least 92% and a specificity o
f at least 96%. It is a reliable diagnostic test although its accuracy is l
imited by technical difficulties in obtaining an adequate sample.