Reproducibility of histopathologic diagnosis and classification of non-melanocytic skin cancer: A panel exercise in the framework of the multicenter Southern European study HELIOS

Citation
J. Wechsler et al., Reproducibility of histopathologic diagnosis and classification of non-melanocytic skin cancer: A panel exercise in the framework of the multicenter Southern European study HELIOS, TUMORI, 87(2), 2001, pp. 95-100
Citations number
30
Categorie Soggetti
Onconogenesis & Cancer Research
Journal title
TUMORI
ISSN journal
03008916 → ACNP
Volume
87
Issue
2
Year of publication
2001
Pages
95 - 100
Database
ISI
SICI code
0300-8916(200103/04)87:2<95:ROHDAC>2.0.ZU;2-N
Abstract
Aims and background: The reproducibility of histologic diagnoses, and in pa rticular of the distinction between basal and squamous cell forms, has been evaluated in the framework of a multicenter case-control study tin Italy, France, Spain and Switzerland) aimed to assess the causes of non-melanocyti c skin cancers. Methods: A panel composed of 10 pathologists from the collaborating centers was appointed. A total of 1774 slides of routine diagnoses were blindly re viewed by a second panelist; discordant diagnoses underwent a third examina tion. Controversial and interesting cases were discussed during general ses sions. Results: The results showed a high degree of concordance (99.5%), with a Co hen's kappa of 0.85 (95% CI, 0.77-0.94) in the assessment of malignancy of lesions. The concordance in the distinction between large morphologic group s, basal cell carcinoma and squamous cell carcinoma was also high (Cohen's kappa = 0.85; 95% CI, 0.82-0.89). The assessment of histologic subtypes, de gree of invasion and differentiation showed a lower degree of concordance, presumably as a consequence of a weaker consistency in the relevant classif ications. Conclusions: The reproducibility study has therefore validated the epidemio logic study and in particular allowed to correct some misclassifications th at could have lessened the analysis of the case-control study. In general, because of its characteristics (number of pathologists and variety of their origins, the large number of cases examined, blind examination of histolog ic slides), the conclusions of the study may show a certain degree of gener alization, at least with regard to the countries represented. Routine histo logic diagnoses of cutaneous carcinoma showed a high degree of reliability with reference to the assessment of malignancy and the distinction between basal and squamous cell carcinoma morphotypes.