Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
104884 | Pathology | 2012 | 7 Pages |
SummaryAimTo assess concordance between the histopathological reports of referring pathologists and those of pathologists reviewing the cases for the Western Australia Melanoma Advisory Service.MethodsA retrospective review of 721 pathology reports from 2000 to 2009 was conducted. Histological features including Breslow thickness, Clark level, tumour type and clinico-pathological staging [American Joint Committee on Cancer (AJCC)] were compared. Further analysis was undertaken for 169 cases to compare mitotic rate, excision margins, regression, growth phase, vascular invasion, neurotropism, tumour infiltrating lymphocytes, microsatellites and predominant cell type.ResultsReferring pathologists consistently reported Breslow thickness, Clark level and excision margins. Reporting of other parameters including ulceration, mitotic rate and vascular invasion, however, was variable. There was almost perfect concordance (kappa = 0.81-1.00) for tumour thickness, ulceration, microsatellites and growth phase; substantial concordance (κ = 0.61-0.80) for Clark level, mitotic rate, completeness of excision and neurotropism; moderate concordance (κ = 0.41-0.60) for vascular invasion, regression, predominant cell type and histological type; and only slight concordance (κ = 0-0.2) for tumour infiltrating lymphocytes. There was a high level of agreement for diagnosis of lesions as melanoma versus benign (97.3%). Overall concordance for pathological tumour staging was substantial (81.9%, κ = 0.79). Lowest concordance was found for stage 1b (91.3%, κ = 0.62).ConclusionOverall concordance in clinicopathological stage was high due to consistency of reporting of tumour thickness and ulceration. Lower concordance was found for pathological substages due to discrepancies in Clark level, highlighting its limited reliability as a prognostic indicator and supporting the revision of its use in the latest AJCC melanoma staging protocol.