Interobserver variability in ultrasound (US) guided fine needle aspiration cytology (FNAC) of sentinel nodes (SN): Experience in 1,000 melanoma patients.

Abstract

9097 Background: FNAC is used to determine nature of suspected lesions that are either clinically apparent (palpable) or discovered by imaging (non-palpable) with the same high sensitivity (Voit et al.JNCI 2011). However, there seems to be interobserver variability. Methods: Morphology criteria (Voit et al. JCO 2010) were tested in 3 pre-SLNB-US examiners. US criteria, FNAC results were collected in a prospective database.Positive cytology was confirmed by histopathology. Examiner #1 had a long US-FNAC experience, 2 others were learners, #2 with a long US, but not FNAC experience, #3 new to both. Results: In 1,000 consecutive prospective patients mean/median Breslow thickness was 2.58/1.57mm. Mean/median follow-up was 56 / 53 months (1 – 132). SN positivity rate was 21%. Overall US-FNAC sensitivity was 71%(US only) and 51%(US-FNAC). Sensitivity of US-FNAC was highest for T4 (76%) and ulcerated melanomas (63%). #1 performed 724(72%), #2 202(20%), #3 74(7%). US validation of #1 was 431/724(59.9%) benign, in 227/724(31.4%) suspicious and in 66/724(9.1%) malignant. #2 estimated in 144/202(71.3%) benign, in 50/202(24,7%) suspicious and in 8/202(4%) malignant; #3 in 57(77%) benign, in 8/75 (10.8%) suspicious and in 9/74 (12.2%) malignant (p<0.001). #1 detects early sign peripheral perfusion(PP) in 32%, #2 in 21% and #3 in 22%. #1 performed FNAC in 291(40%), #2 in 55(25%) and #3 in17(15%)(p<0.001). #1 had sensitivity of 86/157(55%), spec of 99%, PPV of 99% and NPV of 89% (p<0.001). #2 had 34%,100%,100% and 88%,#3 56%, 100%,100% and 89% respectively(p<0.001). Detected sentinel node Tuburden <1mm was successfully FNACed in 45% by #1 and in 20% by #2 and#3 respectively. Conclusions: This largest database covering FNACs in SN of melanoma patients showed no difference between ultrasonographists with regard to performing a FNAC if a late sign, baloonshape (BS) or loss of central echoes (LCE), was seen. An examiner experienced in both US and FNAC sees more PP, performs more FNACs and has more positive FNACs. Examiners not used to FNAC tend to not perform FNACS. Interobserver difference is due to a significantly lower detection rate of PP.