TY - JOUR
T1 - Atypical cellular blue nevi (cellular blue nevi with atypical features)
T2 - Lack of consensus for diagnosis and distinction from cellular blue nevi and malignant melanoma ("malignant blue nevus")
AU - Barnhill, Raymond L.
AU - Argenyi, Zsolt
AU - Berwick, Marianne
AU - Duray, Paul H.
AU - Erickson, Lori
AU - Guitart, Joan
AU - Horenstein, Marcello G.
AU - Lowe, Lori
AU - Messina, Jane
AU - Paine, Susan
AU - Piepkorn, Michael W.
AU - Prieto, Victor
AU - Rabkin, Michael S.
AU - Schmidt, Birgitta
AU - Selim, Angelica
AU - Shea, Chris R.
AU - Trotter, Martin J.
PY - 2008/1
Y1 - 2008/1
N2 - The distinction of cellular blue nevi (CBN) with atypical features ["atypical" CBN (ACBN)] from conventional CBN and malignant melanomas related to or derived from CBN remains a difficult problem. Here, we report on the diagnosis of various cellular blue melanocytic neoplasms by 14 dermatopathologists who routinely examine melanocytic lesions. Three parameters were assessed: (1) for between rater analyses, we calculated interobserver agreement by the κ statistic (regardless of whether the diagnosis was correct). (2) For each individual lesion, we reported whether a majority agreement (>50%) was reached and, if so, whether the majority agreed with the gold standard diagnosis, derived from standardized histopathologic criteria for melanoma, definitive outcome such as metastatic event or death of disease, or disease-free follow-up for ≥4 years. (3) For the individual pathologists, we calculated sensitivity and specificity for each type of lesion. The study set included 26 melanocytic lesions: (1) 6 malignant melanomas developing in or with attributes of CBN; (2) 11 CBN with atypical features and indeterminate biologic potential (ACBN); (3) 8 conventional CBN; and (4) 1 common BN. The κ values for interrater agreement varied from 0.52 (95% confidence interval 0.45, 0.58) for melanoma to 0.02 (0.05, 0.08) for ACBN and 0.20 (0.13, 0.28) for CBN. The κ for all lesions was 0.25 (0.22, 0.28). The pathologists' sensitivities were 68.6% (61.0%, 76.1%) for melanoma, 33.1% (21.0%, 45.2%) for ACBN, and 44.6% (29.0%, 60.3%) for CBN. The specificities were 65.7% (55.8%, 75.6%) for melanoma, 84.7% (77.3%, 92.2%) for ACBN, and 89.9% (82.7%, 97.1%) for CBN. Overall, greater than 50% of the pathologists agreed and were correct in their diagnosis 38.5% (10 lesions) of the time. There was a majority agreement, but with an incorrect diagnosis, another 26.9% (7 lesions) of the time. Six of the 7 majority agreements with an incorrect diagnosis were for ACBN lesions. In summary, the results of our study indicate that there is substantial confusion and disagreement among experienced histopathologists about the definitions and biologic nature of cellular blue melanocytic neoplasms particularly those thought to have atypical features ("atypical" CBN).
AB - The distinction of cellular blue nevi (CBN) with atypical features ["atypical" CBN (ACBN)] from conventional CBN and malignant melanomas related to or derived from CBN remains a difficult problem. Here, we report on the diagnosis of various cellular blue melanocytic neoplasms by 14 dermatopathologists who routinely examine melanocytic lesions. Three parameters were assessed: (1) for between rater analyses, we calculated interobserver agreement by the κ statistic (regardless of whether the diagnosis was correct). (2) For each individual lesion, we reported whether a majority agreement (>50%) was reached and, if so, whether the majority agreed with the gold standard diagnosis, derived from standardized histopathologic criteria for melanoma, definitive outcome such as metastatic event or death of disease, or disease-free follow-up for ≥4 years. (3) For the individual pathologists, we calculated sensitivity and specificity for each type of lesion. The study set included 26 melanocytic lesions: (1) 6 malignant melanomas developing in or with attributes of CBN; (2) 11 CBN with atypical features and indeterminate biologic potential (ACBN); (3) 8 conventional CBN; and (4) 1 common BN. The κ values for interrater agreement varied from 0.52 (95% confidence interval 0.45, 0.58) for melanoma to 0.02 (0.05, 0.08) for ACBN and 0.20 (0.13, 0.28) for CBN. The κ for all lesions was 0.25 (0.22, 0.28). The pathologists' sensitivities were 68.6% (61.0%, 76.1%) for melanoma, 33.1% (21.0%, 45.2%) for ACBN, and 44.6% (29.0%, 60.3%) for CBN. The specificities were 65.7% (55.8%, 75.6%) for melanoma, 84.7% (77.3%, 92.2%) for ACBN, and 89.9% (82.7%, 97.1%) for CBN. Overall, greater than 50% of the pathologists agreed and were correct in their diagnosis 38.5% (10 lesions) of the time. There was a majority agreement, but with an incorrect diagnosis, another 26.9% (7 lesions) of the time. Six of the 7 majority agreements with an incorrect diagnosis were for ACBN lesions. In summary, the results of our study indicate that there is substantial confusion and disagreement among experienced histopathologists about the definitions and biologic nature of cellular blue melanocytic neoplasms particularly those thought to have atypical features ("atypical" CBN).
KW - Atypical cellular blue nevus
KW - Cellular blue nevus
KW - Malignant blue nevus
KW - Melanoma
UR - http://www.scopus.com/inward/record.url?scp=37549002132&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=37549002132&partnerID=8YFLogxK
U2 - 10.1097/PAS.0b013e3181573aaf
DO - 10.1097/PAS.0b013e3181573aaf
M3 - Article
C2 - 18162768
AN - SCOPUS:37549002132
SN - 0147-5185
VL - 32
SP - 36
EP - 44
JO - American Journal of Surgical Pathology
JF - American Journal of Surgical Pathology
IS - 1
ER -