Measuring the depth of invasion in vulvar squamous cell carcinoma: interobserver agreement and pitfalls

Aims The depth of invasion is an important prognostic factor for patients with vulvar squamous cell carcinoma (SCC). The threshold of 1 mm distinguishes between FIGO stages IA and >= IB disease and guides the need for groin surgery. Therefore, high interobserver agreement is crucial. The conventional and the alternative method are described to measure the depth of invasion. The aims of this study were to assess interobserver agreement for classifying the depth of invasion using both methods and to identify pitfalls. Methods and results Fifty slides of vulvar SCC with a depth of invasion approximately 1 mm were selected, digitally scanned and independently assessed by 10 pathologists working in a referral or oncology centre and four pathologists in training. The depth of invasion was measured using both the conventional and alternative method in each slide and categorised into <= 1 and >1 mm. The percentage of agreement and Light's kappa for multi-rater agreement were calculated, and 95% confidence intervals were calculated by bootstrapping (1000 runs). The agreement using the conventional method was moderate (kappa = 0.57, 95% confidence interval = 0.45-0.68). The percentage of agreement among the participating pathologists using the conventional method was 85.0% versus 89.4% using the alternative method. Six pitfalls were identified: disagreement concerning which invasive nest is deepest, recognition of invasive growth and where it starts, curved surface, carcinoma situated on the edge of the tissue block, ulceration and different measurement methods. Conclusions Pathologists reached only moderate agreement in determining the depth of invasion in vulvar SCC, without a notable difference between the two measurement methods.