Proportion of changed benign melanocytic lesions by age group in patients undergoing short-term sequential digital dermoscopic imaging.
Menzies SW, Stevenson ML, Altamura D, Byth K. Variables Predicting Change in Benign Melanocytic Nevi Undergoing Short-term Dermoscopic Imaging. Arch Dermatol. 2011;147(6):655-659. doi:10.1001/archdermatol.2011.133
Copyright 2011 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2011
To determine whether certain patient demographics are associated with poorer specificity for the diagnosis of melanoma in nevi undergoing short-term sequential digital dermoscopic imaging.
Retrospective cohort study performed from April 1, 1998, through May 31, 2007.
Sydney Melanoma Diagnostic Centre, a tertiary referral institution.
A total of 2497 benign melanocytic lesions in 1765 patients undergoing short-term sequential digital dermoscopic imaging during 2.5 to 4.5 months (42.3% male; mean [SD] age, 40  years; age range, 1-86 years).
Main Outcome Measures
Proportion of changed nevi as a function of age, sex, lesion diameter, and anatomical site.
The only variable significantly associated with nevus change was age group (P = .002). When compared with the middle-aged (aged 36-50 years) group, the odds of change were significantly increased in the child and adolescent (aged 0-18 years: odds ratio, 2.60; 95% confidence interval, 1.30-5.22), young adult (aged 19-35 years: 1.50; 1.04-2.17), and elderly (> 65 years old: 2.04; 1.04-3.99) age groups. Within the changed benign lesions, a significant association was observed between histologic subtype and age group (P = .01). The proportion of changed lesions of the banal nevi type decreased and the proportion of the dysplastic nevi type increased with age. In the elderly group, 75.9% of changed lesions were of the dysplastic nevi type compared with 35.7% in the youngest group.
A poorer specificity is observed for the diagnosis of melanoma for nevi undergoing short-term sequential digital dermoscopic imaging in children and adolescents (75.7%) and elderly patients (77.9%) compared with other patients (84.6%).
Sequential digital dermoscopic imaging (SDDI) allows the detection of morphologic change of melanocytic lesions over time that permits the detection of early melanoma without specific dermoscopic features of malignant neoplasms (so-called featureless melanoma).1- 6 Sequential digital dermoscopic imaging is used in 2 general settings. Long-term SDDI is performed during standard surveillance periods (ie, 6-12 months) in patients with nevi that are not suspicious of melanoma. Long-term SDDI is usually performed in patients with multiple atypical nevi. In these patients, the observation of certain significant morphologic changes determines the need for excision.2,3 Such changes occur in 4% to 5% of monitored nevi.2,4,5 In contrast, short-term SDDI performed during a 3-month interval (range, 2.5-4.5 months) is used in patients with nevi that are more suspicious of melanoma.1,6 Short-term SDDI is usually performed in patients with mildly atypical nevi who have a history of change or moderately atypical nevi without specific dermoscopic features of melanoma and without a history of change. Unlike long-term SDDI, virtually any morphologic change seen with short-term SDDI leads to the excision of these nevi.1,3,6 Of importance, 99.2% of lesions remaining unchanged with short-term SDDI are benign. Nevertheless, 16% of benign nevi change during the short-term interval.6 In this study, we investigated whether certain patient characteristics predict change in benign nevi undergoing short-term SDDI. Such analysis helps to define whether certain patients are less suitable to undergo the technique because of a relatively poorer specificity for the diagnosis of melanoma.
All patients were examined in the tertiary referral center of the Sydney Melanoma Diagnostic Centre from April 1, 1998, through May 31, 2007, in a retrospective cohort, as previously described.6 In this clinic, SDDI was performed on selected patients using the SolarScan system (Polartechnics Ltd, Sydney, Australia), which allows precise color calibration between imaging sessions.1 In general, patients with lesions with a minor suspicion of melanoma, which are usually flat or slightly raised, who had no overt dermoscopic evidence of melanoma underwent short-term SDDI in 2 clinical scenarios.1,3,6 The first scenario was of patients with lesions with mild atypia defined as symmetrical or, more frequently, near symmetrical in pigmentation pattern and that had a history of change. The second scenario was of patients with lesions with moderate atypia defined as more asymmetrical with greater architectural disorder, sometimes with a single dermoscopic feature of melanoma (as seen in some dysplastic nevi) but without a history of change. In general, patients with lesions with the dermoscopic features of Spitz nevi or those with a rim of multiple peripheral brown globules did not undergo short-term SDDI because of the known natural history of rapid morphologic change in such lesions.
All consecutive patients undergoing short-term SDDI (ie, <4.5 months) were initially recruited into the study.6 All patients gave written consent for their clinical data to be used, and the study was approved by the ethics committee of the Sydney South West Area Health Service. Approximately half the patients underwent 6-week monitoring followed by 3-month (range, 2.5-4.5 months) monitoring if morphologic changes were not seen. The remainder of the patients underwent 3-month monitoring without 6-week SDDI. Any visual change during this time was considered significant to warrant excision, with the exception of an increase or decrease of milialike cysts or a uniform increase or decrease in pigmentation without architectural change consistent with surrounding skin tanning changes.1 All lesions subsequently excised at any future period and histologically confirmed as melanoma (invasive and in situ) were excluded. All lesions that were imaged beyond the field of view (ie, >24-mm diameter) or were less than 24 mm in diameter but had unclear lesion boundaries and, hence, no accurate measure of lesion diameter also were excluded.
As described previously,6 47% of patients with unchanged lesions undergoing SDDI had long-term data indicating the benign nature of the lesions. Three categories of recorded follow-up were considered evidence of a benign lesion: patients undergoing 6-month to 8-month SDDI from baseline that revealed no morphologic change; patients undergoing longer-than-8-month SDDI from baseline that revealed none of the significant morphologic changes of size, shape, color, evidence of regression, or appearance of known dermoscopic features of melanoma, according to the criteria of Kittler et al2; and patients who had a routine skin examination later than 12 months after baseline imaging without detecting melanoma.
For the included lesions, the maximum lesion diameter at the baseline image (using a calibrated measurement tool in SolarScan), patient age, patient sex, and lesion anatomical site were recorded. The statistical software S-PLUS, version 8 (TIBCO Software Inc, Palo Alto, California), was used to analyze the data. Generalized linear mixed models with patient identifier as a random effect were used to assess the degree of association between possible risk factors and the presence of change. These models take into account the correlation between nevi from the same patient and use odds ratios and their 95% confidence intervals to quantify associations.
The 1765 study patients (42.3% male; mean [SD] age, 40  years; range, 1-86 years) with benign melanocytic lesions undergoing short-term SDDI during 2.5 to 4.5 months had 2497 lesions. Of these lesions, 2093 (83.8%) had no morphologic change but 404 (16.2%) showed change and were excised and confirmed by histopathologic examination to be benign.
A significant association (P = .002) was found between age group and the presence of change in benign melanocytic lesions in patients undergoing short-term SDDI (Table 1). The younger (0-18 and 19-35 years) and elderly (> 65 years) age groups showed greater change than the middle-aged (36-50 and 51-65 years) groups (Figure). Poorer specificity was observed for the diagnosis of melanoma for nevi undergoing short-term SDDI in children and adolescents (75.7%) and elderly patients (77.9%) compared with other adults (84.6%) (Table 1).The odds of change were significantly higher in the younger age groups and in the elderly group when compared with the middle-age group aged 36 to 50 years (Table 2). Within the changed benign lesions, a significant association was observed between histologic subtype and age group (P = .01) (Table 3). The proportion of changed benign lesions of the banal nevi type decreased and the proportion of the dysplastic nevi type increased with age. In this regard, 35.7% of changed lesions were dysplastic nevi in the child and adolescent group, but this proportion more than doubled (75.9%) in elderly patients.
No significant association was observed between the presence of benign melanocytic lesions undergoing change in males (16.7%) compared with females (15.8%) (P = .56). A borderline but nonsignificant association was found between the presence of change in benign melanocytic lesions and anatomical site (P = .07) (Table 4). This association disappeared when head and neck lesions (excluding the lip and scalp) were removed (P = .51). A borderline but nonsignificant association (P = .055) was found between the presence of change and maximum lesion diameter at baseline. The mean (SD) diameter of changed lesions was 6.8 (2.7) mm compared with 7.1 (3.0) mm for unchanged lesions. No significant association (P = .28) was found between the presence of change and grouped maximum diameter (≤3.0, 3.1-6.0, 6.1-10.0, and 10.1-24.0 mm).
A generalized linear mixed model for the presence of change with main effects of age (ie, 0-18 years, 19-65 years, or > 65 years) and site (ie, head and neck including the scalp and lips, trunk, or limbs) together with their interaction was fitted to the data. No statistically significant interaction was found between the age and site effects (P = .89).
Finally, we examined the morphologic changes found in the lesions of elderly patients. In the lesions that had changed, 37.9% showed some evidence of involution (ie, any loss of structure or pigmentation within the lesion), 27.6% showed an increase in size (of which 62.5% showed an asymmetric size increase), 89.7% showed some architectural change, 62.1% showed a focal increase in pigmentation, 41.4% showed a focal decrease in pigmentation, and 10.3% appeared to have changed color.
To our knowledge, ours is the first study to examine patient lesion characteristics that might predict change in suspicious benign melanocytic lesions in patients undergoing short-term SDDI. These lesions undergo monitoring when a minor clinical suspicion of melanoma (dermoscopic or historical) is present without overt dermoscopic evidence thereof. Our results show that the only significant factor predicting change among age group, sex, anatomical site, and lesion diameter was patient age. Surprisingly, this association was U-shaped, with the elderly patients having a similar proportion of changed lesions to that of children and adolescents.
We can only speculate regarding the biological reasons why benign melanocytic lesions in elderly patients are more likely to change than in other adults. It may relate to the observation of a significantly increased proportion of dysplastic nevi in the changed lesions of elderly patients compared with younger patients, with the hypothesis that dysplastic nevi are more likely to change than nondysplastic lesions. However, in childhood, changed nevi clearly are more likely to be of the nondysplastic banal nevi type than is the case in adults. We have excluded any effect of differences in anatomical site as a cause of change in elderly patients.
This study is limited by the uncertainty of the criterion standard histologic definition of early in situ melanoma. Because melanoma is more prevalent in elderly patients, it is possible that the number of histologically reported false-negative melanomas is increased in that group. However, a bias may exist in the form of an increase in histologically reported false-positive results in elderly patients, when, as in this study, the physician is not masked to age (ie, pathologists are more likely to diagnose melanoma in elderly patients than in young patients).
Previous studies have examined age changes of nevi in long-term SDDI. Kittler et al7 showed a high proportion of childhood nevi enlarging during a median of 11 months compared with that in older age groups. The proportion that enlarged then decreased until reaching a minimum in adults older than 40 years. The authors did not stratify age beyond older than 50 years. Nevertheless, change in their data set corresponded strongly to the presence of nevi with multiple peripheral brown globules observed via dermoscopic examination. However, because these lesions were known to increase in size, they are excluded from undergoing short-term SDDI in our cohort. Bauer et al8 also showed an inverse relationship with growth of clinically atypical nevi and age during long-term SDDI during a median of 25 months. However, the oldest stratified group was older than 47 years. Hence, as in the study by Kittler et al,7 the increased changes observed in elderly patients in our study could not be evaluated in the former studies.
Spitz nevi are known to undergo rapid morphologic change.9 As a result, we do not routinely monitor these lesions. Nevertheless, a small number of them were excised after changes in short-term SDDI in our study (Table 3). However, when data from these patients were removed from the age analysis, the significant association between age and change remained. Also, it is known that age-related differences exist in the dermoscopic global patterns of nevi, with a globular pattern predominating in childhood in contrast to the reticular and/or homogeneous pattern that occurs increasingly with older age.10 It is possible that such global pattern differences also may be involved in differences in the proportion of changing nevi undergoing short-term SDDI.
We found a borderline association that failed to reach significance of decreased change in benign melanocytic lesions undergoing short-term SDDI on the head and neck (excluding the lips and scalp) compared with other anatomical sites. Kittler et al7 found no significant relationship with body site and changes in nevi in patients undergoing long-term SDDI. In contrast, Bauer et al8 found a significant increase in growth of nevi in patients undergoing long-term SDDI on the trunk vs on the extremities. Head and neck lesions were not analyzed in their study. As we also observed, no association was found between the proportion changed and sex.
Short-term SDDI is used to monitor melanocytic lesions. However, because a lentigo or ephelis occasionally can be included within the differential diagnosis of lentigo maligna, a small number of patients with lentigo (n = 10) and ephelis (n = 1) underwent biopsy and were included in the analysis. However, a wide age distribution and small sample size of patients with these lesions (0-18 years, n = 0; 19-35 years, n = 5; 36-50 years, n = 2; 51-65 years, n = 3; > 65 years, n = 1) were observed, indicating a negligible effect on the analysis.
This study has some limitations. It is a retrospective study of patients undergoing short-term SDDI in the clinic. Although the criteria for selecting lesions for monitoring are well defined, clearly some selection bias could exist regarding lesions monitored within the different age groups. Hence, the results would be different if randomly selected nevi would have been chosen in a prospective setting. Nevertheless, the aim of the study is to determine the variables predicting change of benign nevi within the well-defined clinical setting used for short-term SDDI and not to describe changes of all nevi within a population. Also, the lesions were histopathologically reported within the standard clinical practice of including patient demographic data, including age, rather than masking the investigators to the clinical criteria. It remains possible that a bias of histopathologic diagnosis could occur within this setting.
One of us has previously shown that a patient history of change of the nevus at initial presentation did not correlate with subsequent changes detected via short-term SDDI.1 Hence, the outcome of short-term SDDI of benign melanocytic lesions is independent of patient sex, baseline history of change in lesion, lesion diameter, and anatomical site. Only age influences the outcome, with greater change (ie, poorer specificity for the diagnosis of melanoma) occurring in childhood and adolescence (0-18 years, specificity of 75.7%) and old age (> 65 years, specificity of 77.9%).
Correspondence: Scott W. Menzies, MB, BS, PhD, Sydney Melanoma Diagnostic Centre, Royal Prince Alfred Hospital, Camperdown, NSW, 2050 Australia (firstname.lastname@example.org).
Accepted for Publication: November 24, 2010.
Author Contributions: All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Menzies. Acquisition of data: Altamura and Stevenson. Analysis and interpretation of data: Menzies, Byth, Stevenson, and Altamura. Drafting of the manuscript: Menzies. Critical revision of the manuscript for important intellectual content: Menzies, Byth, Stevenson, and Altamura. Statistical analysis: Byth, Menzies, and Stevenson.
Financial Disclosure: None reported.