To the Editor: Since its introduction, the diagnostic utility of PReferentially expressed Antigen in Melanoma (PRAME) has been investigated with variable results.1–5 In general, PRAME shows strong and diffuse expression in most malignant melanocytic tumors, demonstrating high sensitivity and specificity for primary and metastatic melanomas. In some subsets, such as in the evaluation of lentigo maligna (LM), PRAME immunoreactivity is a sensitive and specific method to determine the malignant nature of the lesion and to assess tumor margins.6 However, in sun-exposed skin of the elderly, pigmented actinic keratosis (PAK), a variant of actinic keratosis in which melanin may be present in the keratinocytes of the lower epidermal layers, in melanocytes, or in dermal melanophages, represents a challenging differential diagnosis that has crucial therapeutic implications, as it can be treated conservatively, while in situ melanoma would require a potentially disfiguring re-excision, if located in the head and neck region. Although noninvasive diagnostic methods, such as dermoscopy and reflectance confocal microscopy, have been shown to improve the discrimination between in situ melanoma and PAK,7,8 differential diagnosis remains challenging. Thus, a biopsy could be required in equivocal cases along with ancillary histological techniques. The use of immunohistochemistry has yielded conflicting results. Melan-A tends to overestimate the number of melanocytes in sun-exposed skin,9 while MITF-1 and SOX10 are reliable markers for melanocytic quantification.10 In this study, we describe our experience with PRAME on 2 different pigmented lesions in the sun-exposed skin of the same patient. An 86-year-old man presented for the evaluation of a flat, irregularly pigmented macule on the cheek (Fig. 1A, arrows) and a large, ill-defined, similar patch on the vertex (Fig. 1B, arrows), for which he underwent 2 punch biopsies. Histological examination revealed severe actinic dermal damage in both cheek (Fig. 2A) and scalp (Fig. 2B) specimens, with thick elastosis and vascular ectasias. In the cheek, atypical basal keratinocytes were present, with large nuclei, visible nucleoli, and altered maturation. Some pigmented melanophages were also seen in the superficial dermis (Fig. 2A). Keratinocyte stratification was regular in the scalp, but some small hyperchromatic cells were unevenly lined at the dermal–epidermal junction (Fig. 2B).FIGURE 1.: A, An irregular, brownish macule on a sun-damaged cheek. B, A large, ill-defined pigmented patch on the scalp. Margins are macroscopically unpredictable.FIGURE 2.: A, Some atypical keratinocytes are present in the basal layers, with large nuclei, visible nucleolus, and altered maturation; few pigmented melanophages can be seen in the superficial dermis. B, Keratinocyte stratification is regular, but some small hyperchromatic cells are unevenly lined at the dermal–epidermal junction. The actinic dermal damage is severe both in the cheek and in the scalp. C and D, Melan-A staining is superimposable in the cheek (C) and in the scalp (D). E and F, PRAME stains few elements in the cheek (E) and is diffusely positive in the scalp (F).Although Melan-A showed similar results in both lesions (Figs. 2C, D), PRAME decorated very few elements in the cheek (Fig. 2E) and was diffusely positive in the scalp (Fig. 2G). Based on these morphological and immunohistochemical results, the macule on the cheek was diagnosed as PAK, while the lesion on the scalp was called a cell-poor early in situ melanoma (MIS) with an LM growth pattern. It may be difficult to distinguish early MIS with an LM pattern from other lesions occurring in chronically sun-exposed skin, such as solar lentigos, nevi, and actinic keratosis, or even from normal sun-exposed skin that carries an increased number of melanocytes per se. The LM growth pattern is characterized by continuous single-cell proliferation, with no or very little nesting, little pigmentation, and no pagetoid spread along the epidermis. Tumor cells are cytologically bland and are usually located along the dermo–epidermal junction. There is a typical loss of mutual spatial relations and crowding in the epidermal basal layer; however, these features can be extremely deceptive, especially in the early phases. Sometimes, these inconspicuous findings contradict the worrisome clinical picture, recalling the important issue of clinicopathological correlation when evaluating bladder melanocytic lesions. Similarly, actinic keratosis can show mild keratinocyte dysplasia involving only the basal epidermal layers, thus being either underdiagnosed as reactive changes or misdiagnosed as early MIS, especially if keratinocytes are pigmented, as in the present case. PRAME has recently been found to be a useful marker for distinguishing benign from malignant lesions in re-excision specimens with abundant reactive melanocytic hyperplasia, as in the head and neck region. In our patient, PRAME helped in the differential diagnosis with clinical implications. The PAK was treated with cryotherapy with a complete clinical response, while the LM is currently undertreatment with imiquimod 5% cream because of patient comorbidities.