P. Guitera 1 , A. Waddell2, G. Carlos3, B.M. Gouveia4, C. Sinz5, T. Ahmed6,7, S. Lo7, R. Rawson8, R. Scolyer8
Lentigo maligna (LM) represents the most prevalent form of melanoma in situ in Australia with an incidence rising rapidly. The presence of (micro-) invasion impacts man- agement and prognosis.
We prospectively evaluated the performance of a single- reader in accurately detecting invasion at the time of diag- nosis and evaluated retrospectively RCM features to find the most valuable ones assisting in differentiating LM Mel- anoma (LMM) from LM.
Histopathology verified cases of LM and LMM cases were retrieved from the database of two melanoma centres (Sydney Melanoma Diagnostic Centre; Melanoma Institute Australia) between 2005 and 2014. Confocal features were summarised through proportions and compared by chi- squares test or Fisher exact test as appropriate. The associ- ation of each factors with LMM was assessed using uni- variate logistic regression model.
On the 19 LMM viewed in our units between 2005 and
2014, 17 were suspected to be invasive by a single reader (PG) while 2 were not (false negative); On the 210 LM imaged, invasion was suspected in 20 (false positive) and not in 190; Giving a sensitivity of 89% and the specificity of
90% for the diagnosis of invasion.
We analysed retrospectively 9 categories of RCM features and grade them. Comparing the 19 LMM and 60 controls (LM), we found the confocal features associated with LMM. Their OR and p value with the interobserver agree- ment will be updated.
In conclusion, RCM is a valuable tool to help diagnosing invasion on suspicious macule of the face.