When algorithms falter: a case report of a very small melanoma excised due to the dermatoscopic “ugly duckling” sign

We present a case report of a 3.5 mm diameter superficial spreading melanoma on the upper back of a 27-year-old woman, signed out as Clark level 2, Breslow thickness 0.2 mm with regression to 0.45 mm. The patient, with Fitzpatrick type 1 skin and minimal actinic damage, had presented for a routine skin check with no previous history of skin cancers. At the age of 17 she had received chemotherapy and radiotherapy for Ewing’s sarcoma of the right hip with pulmonary metastases. The skin lesion was assessed as dermatoscopically symmetrical and was not predicted as a melanoma by any algorithmic method. The provisional diagnosis of melanoma was made on the basis that this lesion was completely different in dermatoscopic pattern to her other nevi, a dermatoscopic “ugly duckling” lesion. We draw attention to the recently established link between defects in the STAG2 gene and Ewing’s sarcoma, glioblastoma and melanoma.


Introduction
Small melanomas (diameter < 4 mm) present diagnostic difficulties, as the clinical and dermatoscopic characteristics of such lesions have been reported only rarely [1]. Another recent case report by Pellizzari et al suggests that in very small lesions chaos (asymmetry of structure or color) may not be unequivocally present [2]. This particular melanoma had the diagnostic clue of lines radial, but as with the smallest published invasive melanoma [2], the radial lines were distributed circumferentially and any asymmetry was equivocal.

Case report
A 27-year-old woman presented for a routine skin cancer check in a primary care skin cancer clinic in Melbourne, Aus-We present a case report of a 3.5 mm diameter superficial spreading melanoma on the upper back of a 27-year-old woman, signed out as Clark level 2, Breslow thickness 0.2 mm with regression to 0.45 mm. The patient, with Fitzpatrick type 1 skin and minimal actinic damage, had presented for a routine skin check with no previous history of skin cancers. At the age of 17 she had received chemotherapy and radiotherapy for Ewing's sarcoma of the right hip with pulmonary metastases. The skin lesion was assessed as dermatoscopically symmetrical and was not predicted as a melanoma by any algorithmic method. The provisional diagnosis of melanoma was made on the basis that this lesion was completely different in dermatoscopic pattern to her other nevi, a dermatoscopic "ugly duckling" lesion. We draw attention to the recently established link between defects in the STAG2 gene and Ewing's sarcoma, glioblastoma and melanoma.
Of the seven presumed melanocytic, pigmented lesions, there were six which all had similar clinical and dermatoscopic features. They were each lightly pigmented with a structureless brown pattern and with a vascular pattern of curved vessels. These six lesions were assessed as being consistent with dermal nevi and were designated as "signature nevi" [3] for this patient. The seventh lesion was located on the right para-thoracic location, and it was slightly different to the other six lesions on naked eye examination, being Only seven other pigmented skin lesions (presumed melanocytic) were discovered on her skin. All lesions were examined   with darker lesions being discovered, while paler, hypomelanotic lesions may not be suspected at the stage where they are very small [2].
The clinician excised this lesion not because malignancy was predicted by any algorithm, but because in his opinion it was a dermatoscopic "ugly duckling." The "ugly duckling" sign was proposed as a useful clinical tool to increase specificity in contrast to the clinical ABCD algorithm [13]. One small study evaluated the "ugly duckling" sign as a dermatoscopic clue and found that in most patients, 80% or more of their nevi could be grouped into one, two or three dermatoscopic patterns [14], although there were no melanomas included in the studied lesions.
The melanoma described in this report occurred on type 1 skin but there was no history to suggest excessive exposure to UV radiation. UV exposure is a causative factor for melanoma [15], but it is also known that exposure to electromagnetic radiation in cancer therapy can increase the risk of cutaneous malignancy, including melanoma [16] This was arguably a symmetrical pattern, taking into account that perfect symmetry is rare in biological material, and although radial lines are regarded as a clue to malignancy by a number of published algorithms [4,5,6,7], this lesion failed to reach the threshold for excision with respect to any algorithm which had a clearly defined flowchart method of predicting malignancy [5,6,7,8,9,10]. The treating clinician made the decision to excise this lesion because of its having a distinctly different dermatoscopic pattern to the signature nevi; he regarded it as a dermatoscopic "ugly duckling."

Discussion
A review published in 2004 found that small melanomas (<6 mm so not conforming to the ABCD criteria where D stands for D = diameter of 6 mm or more) make up less than 1 to 38% of all invasive melanomas [11]. It has previously been reported that small melanomas may not be asymmetrical [1,2]. It has also been reported that hyperpigmentation was the defining feature in all of 13 small melanomas (<4 mm diameter) in one series of 95 melanomas [12], but Pellizzari et al pointed out that this may in fact be due to selection bias,  tistically [17]. Radiotherapy may be an etiological factor in this case with a history of the patient having been treated for Ewing's sarcoma with pulmonary metastases ten years earlier. It is recommended that survivors of childhood cancer have systematic ongoing surveillance for sequelae of their therapy, including their risk of melanoma [18].
Although this lesion was signed out as an invasive melanoma, the reporting pathologist conceded that invasive status was equivocal. In such circumstances it is appropriate to render a report of invasive status, as this will determine surgical margins, which will default to the more significant option.
There is a recently established genetic link between Ewing's sarcoma, melanoma and glioblastoma. Researchers at Georgetown School of Medicine, Washington, USA, have found defective copies of STAG2 gene in 21% of Ewing's sarcoma, 19% of melanoma and 19% of glioblastoma [19]. The possibility of a genetic link between the preceding Ewing's sarcoma and the subsequent melanoma in this case is speculative but mentioned.

Conclusion
Small melanomas (<6 mm diameter) will not be predicted by the clinical ABCD rule and they may not be predicted by currently published dermatoscopic algorithms. This particular lesion was not diagnosable by any of the published algorithms for pigmented skin malignancy. A correct provisional diagnosis was made purely by application of the "ugly duckling" sign applied on the basis of dermatoscopic pattern; i.e.,