Nodular melanoma: five consecutive cases in a general practice with polarized and non-polarized dermatoscopy and dermatopathology

Background: The incidence of nodular melanoma (NM) has been consistently described as at least 10–15% of total melanomas for over 15 years despite advances in diagnostic algorithms and medical technology. NMs are strongly correlated with faster rates of growth and poorer prognosis and thus provide clinicians with a challenge for early recognition. Objective: To evaluate diagnostic clues of consecutive histopathologically proven NMs in one general practice with particular emphasis on dermatoscopic characteristics and compare this to the published literature. Method: A retrospective observational study was performed of five consecutive histologically proven NM, from a total of 212 consecutive melanomas from a general practice in Brisbane, Queensland, Australia. Dermatoscopic images, both polarized and non-polarized, which appears to be a unique resource, and dermatopathologic slides were available for all lesions. Results: All of the NMs in this series were pigmented although one was hypomelanotic. Two of them were symmetrical. The most highly sensitive clues to NM were gray or blue structures and polarizing-specific white lines. Limitations: Due to the small number of NMs in this report no statistical significance can be attributed to the observational findings. Conclusion: This small series supports what is already known: that a significant proportion of NMs may be dermatoscopically symmetrical but that known clues to melanoma are frequently present. Nodular lesions, pigmented or non-pigmented, should be excised to exclude NM if there is any clue to malignancy, regardless of symmetry, unless a confident specific benign diagnosis can be made.


Introduction
Nodular melanoma (NM), defined as a melanoma with any junctional component extending no more than three rete ridges beyond the invasive component, is the second largest melanoma subtype, comprising 10-15% of melanomas in Caucasians [1]. NMs have been shown to have a faster growth rate (GR) (median GR 0.49 mm/month) than lentigo maligna melanomas (median GR 0.13 mm/month) and superficial spreading melanomas (SSM) (median GR 0.12 mm/month) [2]. The GR of malignant melanomas (MM) has also been shown to be an independent prognostic indicator for the prediction of relapse after one year of follow up [3]. Studies have been performed in an attempt to determine diagnostic features, both clinical and dermatoscopic, which facilitate earlier diagnosis of NM, when Breslow thickness is less and prognosis therefore more favorable [4]. A general practice, with a special interest in skin cancer medicine and dermatoscopic photo-documentation of all treated lesions, provides a unique perspective on the evaluation of this condition.

Methods
A retrospective analysis was performed with respect to all NMs diagnosed between January 1, 2008, and June 30, 2013, in a general practice in Brisbane, Queensland, Australia. All lesions treated were prospectively recorded on the Skin Cancer Audit Research Database (SCARD) for both tracking and research purposes [5]. During

Cases
The five cases of histologically diagnosed NM are presented in Tables 1   and 2 with relevant clinical, dermatoscopic and dermatopathologic information. In estimation of melanoma growth rate (GR) we used the ratio MM thickness/duration of the MM visible growth as defined in a previously described assessment tool [3]. Clinical images are displayed in Figure   1, close-up images in Figure 2      The clinical ABCD method is the most widely known algorithmic method for the clinical diagnosis of melanoma and has been promoted both to healthcare professionals and patients [10]. One of the criteria for melanomas to be detected using this method is that they have a minimum diameter of 6 mm. It has been agreed among many authors that a significant proportion of NMs do not fulfil the ABCD criteria including the criterion of a minimum lesion size of 6mm [8,11]. In one series of eleven thin NMs (Breslow thickness 2 mm or less), 63.6% (n=7/11) had a diameter of less than 6 mm [4]. Similarly, in our series 60% (n=3/5) had a diameter less than 6mm and furthermore, each of these also had a Breslow thickness less than 2 mm.
One study of 1789 patients with melanoma found that NM was most frequently found in older men and most commonly on the lower limbs or head and neck [7]. In addition, it was shown to be more strongly correlated with actinic keratosis rather than high nevi counts [7]. This suggests that NMs have an association with sun-damaged skin. In our series 80% (n= 4/5) were male and 75% (n=3/4) of these males were over 50. With respect to body site, 60% (n=3/5) were on the leg (on sun-damaged skin) and 40% (n=2/5) were on the torso (on non-sun-damaged skin).
In a study involving 92 SSMs and 33 NMs, a higher proportion of NM was discovered by the patient (60.6%) compared to SSM (48.9%) [8], and in a study of 22 patients with NM, 61% were first detected by the patient and another 17% detected by another family member and the patient [9].
In our series 60% (n= 3/5) were reported by the patient and another was known of by the patient, but this information was not volunteered until after the lesion was discovered (Case 4). In one case where the lesion was only 3 mm in diameter and on the posterior torso, it was discovered by  10% of the lesion had 78.2% sensitivity for melanoma [14].
The clue of 'blue-black color' was present in 60% (n=3/5) of the NMs in our series.
The clues to malignancy of gray or blue structures and polarizing specific white lines (defined as perpendicularly orientated white lines visible only on polarized dermatoscopy) displayed the highest sensitivity for NM in our small consecutive series; each of them (n=5/5) had either one or the other clue and 60% (n=3/5) had both.
Polarizing-specific white lines were first named 'chrysalis structures' [15] and were attributed to the presence of increased collagen, which has birefringent properties causing rapid randomization of polarized light thus making the collagen more conspicuous. In a study by Balagula et al. it was found that in non-biopsied lesions these structures were most commonly found in dermatofibromas and scars, but in 265 biopsied lesions including 20 melanomas they were observed With respect to dermatoscopic examination there were some similarities between the melanomas in our series (see Table   2) and those in larger published studies.
It has been shown that in NM many of the classic dermatoscopic features of SSM are lacking, however, irregularity of color is usually present in those that do contain pigment [12]. All of the melanomas in our series contained melanin pigment, although in one (Case 3) 75% of the lesion was non-pigmented and this would categorize it as an amelanotic/ hypomelanotic melanoma (AHM) [13].
In our series symmetry was present in 40% (n=2/5; Cases 2 and 5) and if the accompanying nevus was ignored Case 1 was also symmetrical; all were pigmented. The asymmetrical melanomas in our series were asymmetrical in both structure and color. In a published series of 33 NMs, 80% were symmetrical and 60.7% were classified as amelanotic [8]. In another published study 64% (n=7/11) of thin NMs (Breslow 2 mm of less) were symmetrical and 18% (n=2/11) were classified as amelanotic [4].
A study of a series of 283 nodular pigmented lesions found that the presence of blue/black color covering at least We believe this supports the hypothesis that polarizingspecific white lines represent increased collagen production as vertically orientated bands, probably reflecting increased fibroblast activity related to the vertical growth phase of melanoma [16].  [16]. They were found to be more frequently observed in invasive melanomas than in-situ melanomas and their prevalence correlated to increased thickness of melanomas. In our series of 212 melanomas (23.2% n=49 invasive), both polarized and non-polarized dermatoscopy images were available in 142 (19.7% n=28 invasive). While 80% (n=4/5) of the nodular melanomas had polarizing-specific white lines only 7.2% (n=10/137) of the non-nodular melanomas displayed this feature and all but one of those were in-situ.

Conclusion
In Case 1 of our series the polarizing-specific white lines appear to correlate with white lines also seen with nonpolarized dermatoscopy and we speculate that they correlate with vertical bands of collagen seen in the dermatopathology [Copyright: ©2014 Rosendahl et al.]