The diagnosis of melanoma can be challenging for many physicians. In a study reviewing >2000 pathology reports, it was shown that 15% of cases were misdiagnosed by pathologists for both in-situ melanoma and invasive malignant melanoma.¹ This discrepancy suggests there are limitations with the current diagnosis methodology.

The best approach to diagnosing melanoma is to use a clinicodermoscopic approach. To learn about the specific features of difficult to diagnose melanomas (MM), Puig et al. reviewed the clinical and dermatoscopic features of 93 difficult-to-diagnose melanoma (DDM) and found how summarized in Table 1.²

However, even when using a dermatoscope, the validity and reliability of dermoscopic criteria is not standardized and does not always lead to melanoma detection. When differentiating nevi from melanoma, most dermoscopic criteria had poor to fair interobserver agreement.³ The criteria that did reach moderate amounts of agreement are listed in Table 2.³

Deep learning convolutional neural networks (CNN) may facilitate melanoma detection. For the first time Haenssle et al. compared a CNN’s diagnostic performance with a large international group of 58 dermatologists, including 30 experts.⁴ In this article, the CNN performed consistently better than dermatologists when identifying melanocytic lesions. However, in most of those cases, the information was taken out of context. As melanoma is an atypical mole syndrome, the entire clinical context is necessary to ascertain whether a mole is melanotic or not.

In tandem with dermoscopic criteria, patient-related factors (age, skin type, history of melanoma, ultraviolet exposure, pregnancy, and growth dynamics) are essential to determine the diagnosis and management of pigmentated melanocytic nevi as any or all of them can influence the diagnosis.⁵ For example, brown bands with irregular color lines, width, and spacing on the nail plate are not indicative of melanoma in children but are very worrisome in adults.⁶ The consensus panel on melanonychia nail plate dermoscopy agreed that any decision to excise should be based on established clinical criteria (history and physical exam) and not on nail plate dermoscopy patterns further qualifying the need for clinical parameters and not sole reliance on dermoscopy.⁶

Another important consideration is the comparison of lesions in patients with multiple nevi. When taken individually, many nevi may appear abnormal. However, when taken as a whole, that is reviewing all the nevi in one patient and comparing the morphology to each other, it allows dermatologists to make a more informed decision as to which nevi may be malignant.⁷ This approach has been referred to as the “Ugly Duckling Sign” and it is a major factor in melanoma detection.⁸

In a study of 80 patients, including 7 with melanoma, all nevi were presented to dermatologists during 2 separate experiments.⁸ Findings demonstrated that all melanomas were labeled as ugly duckling nevi when dermatologists were given access to intrapatient comparative analysis.⁸ As a result, it reduced the biopsy of nevi by a factor of 6.9 showcasing the improved effectiveness of intrapatient comparative analysis on the diagnosis of melanoma.⁸

Dermoscopy and even the ugly duckling approach are not enough in some instances to detect melanoma. In a study using 236 baseline dermatoscopic images (59 quartets from 59 patients, each including one melanoma detected during follow-up and three nevi), 26 dermatologists were asked to assess the grade of dermoscopic atypia on a numerical scale and to identify the MM.⁹ On average, dermatologist only identified 24 of the 59 MM (40%, range 11-37).⁹ This suggests that a significant proportion of MM detected during follow-up cannot be differentiated at baseline. This necessitates the additional inclusion of less atypical lesions for monitoring.

Digital dermoscopy has been very useful in monitoring patients over time to determine any changes in nevi that deserve additional attention.¹⁰ It can be used as a total body digital scan as well as a digital review of individual lesions, both of which have proven useful in long-term monitoring.^11,12

The addition of reflectance confocal microscopy (RCM) analysis to dermoscopy reduces unnecessary excisions with a high diagnostic accuracy and could be a means for reducing the economic impact associated with the management of skin cancer.^13,14 However, not all dermatologists have access to RCM.