A Retrospective Study of Diameter and Breslow Thickness in Invasive Melanomas

In cutaneous melanoma, greater Breslow histological thickness is associated with an upstaging and worse-predicted prognosis [1]. A proposed approach was to apply the threshold used in the past of the diameter of a pencil eraser (>6 mm) as a threshold to biopsy suspicious pigmented lesions, with the rationale that larger melanomas are thicker and more aggressive tumors [2].

Nevertheless, the association of the diameter of melanoma with the Breslow thickness has not been systematically investigated. We are aware of five studies investigating this topic [3‒7], that reported weak or not significant correlation of Breslow thickness with melanoma diameter. Most of these studies did not adjust for factors possibly confounding the association of Breslow thickness with the melanoma diameter, such as age, sex, or the body location of melanoma. Only the study of Moreno-Ramirez et al. [4] considered the effect of confounding factors. This study reported in an analysis by body location that T2 to T4 melanomas were significantly more likely (vs. in situ and T1 grouped together) to have diameter larger than 5 mm (vs. ≤5 mm).

These studies have included in situ together with invasive melanomas. In situ melanomas have a Breslow thickness of 0 mm that will bias the summary results of the Breslow thickness toward smaller values. In addition, it is important to disentangle in situ from invasive small melanomas as only the latter have the potential to affect prognosis. We performed a single-center, university hospital, retrospective study with the aim to investigate the association of diameter with Breslow thickness in invasive cutaneous melanomas.

We retrospectively studied the Greek patients included in the European Association of Dermato-Oncology (EADO) database, with invasive melanomas of the nodular melanoma (NM) or superficial spreading melanoma (SSM) subtype, diagnosed from 2006 to 2015, as reported previously [8]. In brief, eligible cases included patients more than 16 years old with a diagnosis of primary cutaneous melanoma of NM or SSM subtype. Melanomas in situ were excluded. There were complete data for age, sex, Breslow thickness, and subtype. In the current study, variables of interest at initial diagnosis included patient age, sex assigned at birth, Breslow thickness, melanoma subtype, the primary melanoma body site, and the greatest diameter mentioned in the macroscopic description of the histopathological report. Institutional Review Board approval was obtained. The study is reported according to the STROBE checklist for observational studies [9].

In statistical analysis, categorical variables were investigated with χ² test. The Mann-Whitney test was used for the comparison of medians. The quartiles of the diameter (lower, median, upper) were studied in non-parametric quantile regression model using bootstrap standard errors. Multivariable analysis was performed, adjusting for age, sex, localization of melanoma. In post hoc analysis, a multivariable analysis was performed, adjusting for age, sex, localization of melanoma, and melanoma subtype. Statistical tests were two-sided, and a p value <0.05 was considered statistically significant. Analyses were carried out using STATA, version 12 (StataCorp. (2011). Stata Statistical Software: Release 12. College Station, TX, USA: StataCorp LP).

In total, 537 cases of invasive melanomas were included. Median age was 53 years and 48% were male. Most melanomas (60%) had Breslow thickness ≤1.0 mm (T1 tumors according to AJCC8 staging), 26.6% were T2 tumors (Breslow 1.1–2.0 mm), 8.6% were T3 (Breslow 2.1–4.0 mm), and 4.8% were T4 (Breslow >4.0 mm). The median diameter of melanoma was 13 mm (interquartile range: 9, 20), while 57 melanomas (10.6%) had a diameter of ≤6 mm.

Regarding the melanoma subtype, there were 429 SSM (79.9%) and 108 NM (20.1%). There was not a statistically significant difference in the median diameter between SSM and NM (13 mm, IQR 8, 20) vs. 15 mm, IQR 10, 21, respectively, p = 0.71). There was a significant difference in thickness between SSM and NM. Most SSM (72.7%) were thin with Breslow thickness ≤1 mm, while only 10 of NMs (9.3%) were thin (p < 0.001). In these lines, there was a significantly greater median Breslow thickness in NMs compared to SSMs (2.7 mm vs. 0.7 mm, p < 0.0001). These results underscore that, although significantly thicker, NMs were not associated with larger diameter compared to SSMs.

Having a melanoma with Breslow thickness ≤1.0 mm was not significantly associated with the lower, median, and upper quartile of the diameter (p values: 0.149, 0.296, and 0.875, respectively) after adjustment for age and sex (online suppl. Table 1; Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000536151). The significance of the effects of age and gender differed across the quartiles of diameter. Each additional year of older age was significantly associated with a small increase in the diameter, e.g., the median diameter was greater by 0.13 mm for each additional year of age. Compared to females, males had a significantly greater median diameter by 2.75 mm (p = 0.014) and a significantly greater diameter by 5.13 mm for the upper quartile (p = 0.023), after adjustment for age and Breslow thickness (online suppl. Table 1).

The results were similar when a multivariable analysis was performed, adjusting for age, sex and melanoma location. Breslow thickness ≤1.0 mm was not significantly associated with the lower, median, and upper quartile of the diameter (p values: 0.157, 0.103, and 0.928, respectively) (online suppl. Table 2).

Next, we explored the association of diameter with Breslow thickness in multivariable quantile regression analysis, adjusting for age, sex, melanoma location, and the melanoma subtype. The results were similar, and Breslow thickness ≤1.0 mm was not significantly associated with the lower, median, and upper quartile of the diameter (p values: 0.063, 0.083, and 0.791, respectively) (Table 1).

The criterion of diameter is included in the ABCDE mnemonic (A: asymmetry, B: border irregularity, C: color irregularity, D: diameter larger than that of a pencil eraser >6 mm, E: elevation), used as a public health message and a self-detection tool for melanoma. Our study showed that the diameter of cutaneous melanoma was not associated with Breslow thickness, adding evidence to support the limitations of using the “Diameter” criterion [10‒13].

To our knowledge, there are 5 studies in the literature, including more than 20 small invasive cases, reporting the frequency of small-diameter (≤6 mm) invasive melanomas, and these studies did not investigate the association of Breslow thickness with diameter [14‒18]. Three studies reported that small-diameter melanomas were detected at a frequency of 10.6% [14], 14.3% [15], or 14.7% [16] of all invasive melanomas, while the frequency of small-diameter invasive melanomas was higher in the Australian studies, e.g., 31% in the study by Shaw et al. [17] in 1992 and 34% in the study of Regio Pereira et al. [18] in 2022. Australia focused on extensive campaigns and strategies for the early detection of melanoma, that could have contributed to their higher rate of detecting small invasive melanomas [17]. In addition, metastatic small melanomas have been reported, underscoring their potential to be aggressive tumors [14, 17]. The American Academy of Dermatology website mentions for “D,” traditionally defined as diameter greater than 6 mm in the ABCDE for melanoma, that “while melanomas are usually greater than 6 mm or about the size of a pencil eraser when diagnosed, they can be smaller” [19], suggesting a shift of the “D” definition toward “any diameter.” Apart from the possibility of a melanoma appearing as a small lesion, what matters most for the patients is whether a small melanoma could be a thick lesion, given that greater Breslow thickness is an established marker of worse prognosis.

In our study, Breslow thickness ≤1.0 mm (thin melanomas) was not independently associated with the diameter in invasive cutaneous melanomas. Also, we report that the NM (vs. SSM) subtype was not independently associated with diameter and that males had larger melanomas compared to females, after multivariate adjustment. Gender may influence the self-detection of melanoma, and men tend to engage less in health practices than women [20]. There are scarce studies on the association of Breslow thickness with the diameter of melanoma; we are aware of only five studies summarized in online supplementary Table 3 [3‒7]. Fernandez et al. [3] reported smaller trimmed mean Breslow thickness in melanomas with diameter ≤6 mm compared to those >6 mm (0.35 vs. 0.55 mm). The melanoma subtypes were not reported. Crocetti et al. [5] reported a weak correlation between log-Breslow and log-diameter in melanomas including invasive and in situ tumors. Jimenez Balcells et al. [6] reported no significant correlation between Breslow thickness and diameter in invasive and in situ melanomas, including the NM, SSM, lentigo maligna, lentigo maligna melanoma (LMM), and acral lentiginous melanoma (ALM) subtypes. Salijuma et al. [7], in a study including NM and SSM as well as in situ melanomas, reported weak correlation between Breslow thickness and diameter and a significantly larger median diameter for T2, T3, T4 versus T1 and in situ melanomas. Importantly, these studies did not perform a multivariate analysis to consider confounding factors that may affect the results. The multicenter study of Moreno-Ramirez et al. [4] reported results by body location and melanomas in situ, and T1 had a statistically significant smaller mean size than melanomas T2-T4 (1.08 vs. 1.57 cm, respectively). The melanoma subtypes were not reported. On the other hand, higher ABCDE scores were not associated with thickness (less or more than 0.8 mm) in another retrospective study of 272 in situ and invasive melanomas [21].

The inclusion of exclusively invasive melanomas (and not in situ tumors) in our study may, at least in part, account for differences in results as melanomas in situ have a Breslow thickness of 0 mm that would bias the summary results of Breslow thickness toward smaller values. Also, it is clinically relevant due to the impact on survival depending on Breslow thickness. In situ melanomas (confined to the epidermis) have a different biological behavior from invasive melanomas and have the same life expectancy as the general population, irrespective of their diameter [22]. Another strength of our study is the quantile analysis with adjustments for age, sex, melanoma location, and subtype.

There are some limitations to our study. The measurement of maximal diameter from the macroscopic description of the histopathology report was used. A shrinkage of the tumor of up to 15% has been reported after formalin tissue fixation [4, 23]. This would lead to an underestimation of up to 1.5 mm for a tumor diameter of 10 mm. Also, our results cannot be extrapolated to other melanoma subtypes such as LMM or ALM which were not included in this study. However, LMM has a distinct clinical profile, is typically characterized by a larger diameter, location on the head and neck, slow progression in older patients, and could bias the diameter toward greater values. Notably, it has been reported that the size of LMM was poorly associated with the presence of invasion (invasive or in situ) [24]. ALM is a relatively rare melanoma subtype in white-skinned populations, accounting for up to 5% of all cutaneous melanomas [25].

Regarding the melanoma subtype, we showed that, although significantly thicker, NMs were not associated with larger diameter compared to SSMs. Similarly, Pinero-Madrona et al. [26] showed that a higher Breslow:diameter ratio (that may correspond to thick lesions with small diameter) was significantly more frequent in the NM (vs. SSM) subtype. Population-based studies and large retrospective studies have shown that the NM subtype has the poorest survival among melanoma subtypes and that NM was independently associated with worse melanoma-specific survival in thin melanomas [8, 27, 28]. Collectively, these findings suggest that NMs may be smaller and aggressive lesions, highlighting the importance of their timely detection.

These analyses need to be interpreted in context. It would be wrong to infer that the ABCDE criteria are not useful for the differentiation of melanocytic nevi from cutaneous melanomas. Several studies have shown the usefulness of the ABCDE mnemonic in this regard [29]. On the other hand, there are limitations in the “Diameter larger than 6 mm” criterion. Expecting that melanomas will be larger than 6 mm may hamper the diagnosis of small invasive melanomas. Research focusing on small invasive melanomas reported these may be thick (by reporting weak correlation of Breslow thickness with diameter [3‒7]) or aggressive lesions (by reporting follow-up with metastasis and death) [14, 17]. In conclusion, in our study, including invasive melanomas of the NM or SSM subtype, diameter was not associated with Breslow thickness, indicating the potential of smaller melanomas to be thicker tumors.

The Andreas Sygros Hospital Institutional Review Board approval was obtained. Approval number 3957. Patient consent is not applicable; this was a retrospective study of de-identified information from the medical files.

Dr. Stratigos reports personal fees and/or research support from Novartis, Roche, BMS, AbbVie, Sanofi, Regeneron, and Genesis Pharma, outside the submitted work. Clio Dessinioti, Mihaella Plaka, Angelliki Befon, Dorothea Polydorou, Irene Stefanaki, Katerina Kypreou, and Varvara Theologi report no conflict of interest.

No funding was received.

Clio Dessinioti contributed to the conception and design of the study, collection of the data, performed the statistical analysis, contributed to writing the article, and approved the final version. Mihaella Plaka, Angeliki Befon, Dorothea Polydorou, Irene Stefanaki, Katerina Kypreou, and Varvara Theologi contributed to the collection of the data and writing the article and approved the final version. Alexander Stratigos contributed to the conception and design of the study, writing the article, critical revision of the article, and approved the final version.

Data generated or analyzed during this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding author.