Abstract

AIM: To determine the interobserver variation in scoring presence and grade of vulvar intraepithelial neoplasia (VIN) in haematoxylin/eosin (H/E) slides, MIB 1 slides, and the combined use of H/E and MIB 1 slides. METHODS: 10 slides were stained with H/E and MIB 1 with each of the following diagnoses: normal vulvar skin, VIN 1, VIN 2, and VIN 3. Six observers first scored the H/E slides separately from the MIB 1 slides and second the combined H/E and MIB 1 slides. RESULTS: Unweighted group kappa for MIB 1 was 0.62 and the weighted group kappa was 0.91. This was significantly better than the unweighted group kappa for H/E slides (0.47, p = 0.023) as well as the weighted group kappa for H/E slides (0.82, p = 0.014). There was no improvement by the combined use of H/E and MIB 1 slides. VIN 2 is far less confused with VIN 3 in the combined use of H/E and MIB 1 slides (9%) than in H/E slides (38%) (p = 0.007). There is a tendency to grade VIN in a two tailed grading system rather than a three tailed grading system, which became more apparent with the combined use of H/E and MIB 1 slides. CONCLUSIONS: The interobserver variation with sole use of MIB 1 is better than with the use of H/E stain in VIN. The use of MIB 1 in grading VIN diminishes confusion between VIN 2 and VIN 3 fourfold. A two tailed grading system for VIN seems already to work in daily practice.