Automated quantification of Ki67/MART1 stains may prevent false-negative melanoma diagnoses.

BACKGROUND

Inability to distinguish melanomas from benign nevi is the most frequent reason for malpractice lawsuits in surgical pathology. Reliable diagnostic tools to support hematoxylin and eosin (H&E) stains and induce diagnostic vigilance are thus highly needed. Because high diagnostic performance recently was showed using automated image analysis, the immunohistochemical proliferation marker Ki67 seems a potential candidate. This study aimed to investigate if this previously presented automated algorithm could have prevented 10 false-negative melanoma diagnoses. In addition, diagnostic utility of another, but narrower, immunohistochemical proliferation marker, phosphohistone H3 (PHH3), was explored.

METHODS

A total of 10 formalin-fixed paraffin-embedded melanocytic tumors, initially classified as benign or dysplastic but revised as melanomas at metastatic debut, were dual-stained for Ki67/MART1 and PHH3/MART1. A Ki67 index was automatically calculated in epidermis, dermis, a combination of such, and a dermal hot spot. Dermal PHH3/MART1 scores were established semi-automatically.

RESULTS

The dermal Ki67 index identified all 10 melanomas, the hot-spot index 8 and the epidermal and combined indices only 2 and 5, respectively. Nine melanomas were PHH3 positive and scores correlated with Ki67.

CONCLUSIONS

PHH3 added limited information, but supplemental automated Ki67 assessment could possibly have prevented the misdiagnosis of most melanomas had the algorithm been available at the time of diagnosis.