Objective. For patients with a primary cutaneous head and neck melanoma the pathologic
parameters are crucial. The parameters are essential for staging, to determine the appropriate
treatment and are the strongest predictors of outcome. Until now, the pathologic parameters
remain dependent on the expertise of the individual pathologist. Therefore, inter-observer
variation between pathologists is almost inevitable. Pathologists working at the Netherlands
Cancer Institute (NKI) routinely review melanoma patients referred for treatment.
The purpose of this study was to analyze the inter-observer variation of histopathological
reports between the reports from the pathologist at the NKI versus reports from non-NKI
general pathologist. This analysis is done of primary cutaneous head and neck melanomas.
Study design. A retrospective observational patient file study of case series was conducted. A
total of 483 cases of pathological features of primary cutaneous head and neck melanoma
over a time period of 11 years (2005-2016) were compared. These cases were collected by the
PALGA system (Pathologisch Anatomisch Landelijk Geautomatiseerd Archief), a national
database for pathological biopsies.
Methods. Two hundred ninety-six patients who underwent excision of a primary cutaneous
head and neck melanoma were included. Patients were excluded due to a recurrent cutaneous
head-neck melanoma, referring hospital outside the Netherlands, only a request for second
opinion or due to >2 missing histopathologic parameters in NKI or non-NKI reports. Reports
were collected from both referring general pathologists and the pathologists at NKI. All the
reports were analyzed for patients demographics, tumor characteristics and histopathologic
features in both the referring and NKI reports. The histopathologic features include the
melanoma sub-type, Breslow thickness (BT), tumor mitotic rate (TMR) and ulceration. For
these features the intraclass correlation coefficient and kappa score calculated the interobserver
Results. In 53.4% of the patients the histopathologic parameters were disconcordant. The
kappa scores were 0.648 for sub-type, 0.775 for BT, 0.472 for TMR and 0.802 for ulceration.
The intraclass correlation coefficient was 0.981 for BT. These results indicated a perfect interobserver
agreement among the pathologist for the measurement of BT, a substantial
agreement for subtype and ulceration and a moderate agreement for TMR.
After NKI review in 38 patients (12.8%) recommended treatment policies were changed. The
indication changed for proposed surgical excision margins in 26 cases (8.8%) and for
scheduling sentinel lymph node biopsy (SLNB) in 14 reports (4.8%) after review by an NKI
Conclusion. Review by an expert NKI pathologist of histopathologic parameters of primary
cutaneous head and neck melanoma leads to significant changes, with in some cases serious
consequences for treatment planning. Due to continuous changes over the years in
classification and staging system these histopathologic disconcordances are from great
© 2003-2007 RUG : De Rijksuniversiteit Groningen heeft de rechten van deze repository. Alle rechten voorbehouden.Powered by WildFire