Interobserver reproducibility of the diagnosis of differentiated exophytic vulvar intraepithelial lesion (DEVIL) and the distinction from its mimics.

Abstract

Most vulvar squamous cell carcinomas are human papillomavirus (HPV)-associated or TP53-mutant. A third category of HPV-independent TP53-wild-type lesions is uncommon and not fully understood. Differentiated exophytic vulvar intraepithelial lesion (DEVIL) has been characterised as a precursor of this latter category. The reproducibility of the diagnosis of DEVIL and its distinction from lesions with overlapping morphology has not been studied. Our aim was to establish the interobserver agreement in the diagnosis of DEVIL and its distinction from neoplastic and reactive conditions of the vulva on haematoxylin and eosin evaluation. A set of 35 slides was evaluated by eight reviewers (two trainees and six practising gynaecological pathologists). The set included DEVIL, condyloma, established vulvar precursors [high-grade squamous intraepithelial lesion (HSIL) and differentiated vulvar intraepithelial neoplasia (dVIN)] with superimposed acanthosis or verruciform growth, lichen simplex chronicus (LSC), and psoriasis. Kappa (κ) values were calculated. Overall, interobserver agreement was moderate (κ = 0.56), improving to substantial (κ = 0.7) when evaluation was performed by practising pathologists. Agreement was strong for the diagnosis of HSIL (κ = 0.88), and substantial for the diagnosis of DEVIL (κ = 0.61), condyloma (κ = 0.79), and LSC (κ = 0.72). Agreement was moderate for the diagnosis of dVIN (κ = 0.59) and psoriasis (κ = 0.53). Perfect agreement (6/6) among practising pathologists was observed in 43% of cases, and majority agreement (5/6 or 4/6) was observed in 48% of cases. Reproducibility in the diagnosis of verruciform vulvar lesions, including the novel DEVIL, is acceptable overall. Reproducibility is higher for well-known lesions such as HSIL and condyloma than for more challenging diagnoses such as DEVIL, dVIN, and psoriasis. Agreement is higher among practising gynaecological pathologists, suggesting that training and experience improve reproducibility. Our findings support the inclusion of DEVIL as a diagnostic entity in the classification of vulvar squamous lesions.