Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 1.
Published in final edited form as: J Am Acad Dermatol. 2020 Jan 11;82(5):1158–1165. doi: 10.1016/j.jaad.2019.12.063

Meta-Analysis of Number Needed to Treat for Diagnosis of Melanoma by Clinical Setting

Amy J Petty 1, Bradley Ackerson 2, Reed Garza 3, Michael Peterson 4, Beiyu Liu 5, Cynthia Green 5, Michelle Pavlis 6
PMCID: PMC7167347  NIHMSID: NIHMS1549359  PMID: 31931085

Abstract

Background

The number needed to treat or excise (NNT) is commonly used to measure the number of skin biopsies performed to detect one melanoma. The NNT for melanoma varies widely between clinical settings.

Objective

To provide a formal statistical comparison of the efficacy of melanoma detection between different clinical settings.

Methods

A systematic review and meta-analysis of all relevant observational studies on NNT in relation to melanoma was performed on MEDLINE. We performed a random effects model meta-analysis and reported NNTs with 95% confidence intervals (CIs). The subgroup analysis was related to clinical setting.

Results

In all, 29 papers including a total of 398,549 biopsies/excisions were analyzed. The overall NNT was 9.71 (95% CI, 7.72-12.29): 22.62 (95% CI, 12.95-40.10) for primary care, 9.60 (95% CI, 6.97-13.41) for dermatologist, and 5.85 (95% CI, 4.24-8.27) for pigmented lesion specialists.

Limitations

There is heterogeneity in data reporting and possibility of missing studies. In addition, the incidence of melanoma varies among clinical settings and could affect NNT calculations.

Conclusion:

Specialists have the lowest NNT, followed by dermatologists, suggesting involving specialists in the diagnosis and treatment of pigmented skin lesions can likely improve patient outcomes.

Keywords: melanoma, melanoma in situ, pigmented lesions, dermatologic surgery, oncology, number needed to treat, number needed to excise

Capsule summary

  • The number needed to treat for melanoma varies widely across clinical settings.

  • Data from the present study highlight that pigmented lesion specialists have the lowest number needed to treat for melanoma, followed by dermatologists. Understanding the variability in melanoma detection across clinical settings allows for better cost comparison.

Introduction

Melanoma is a devastating cancer with high morbidity and mortality. Greater than 178,560 melanomas were diagnosed in 2018, and both the incidence and mortality have been steadily increasing for decades.1 Importantly, this increase is not due to methods of detection or changes in clinical or histologic diagnosis.2 Some studies have shown that the rise in incidence currently outpaces the rise in mortality3, leading to an increased number of high risk patients and therefore future diagnoses. While its incidence is increasing, melanoma remains relatively rare in the healthcare setting, making its quick and efficient diagnosis challenging.

An often-reported number for melanoma detection is the number needed to treat (NNT) or number needed to excise (NNE) which is the number of benign pigmented skin lesions excised compared to the number of confirmed melanomas. This metric varies widely depending on the clinical setting in which it is measured.4 A quantitative understanding of our ability to diagnose melanoma is important for many reasons. First, unnecessary biopsies increase the false positive rate, leading to emotional and psychological stress on patients. Second, the real economic impact of these biopsies can only be correctly evaluated with an understanding of the NNT across clinical settings. This allows for an accurate analysis of the increased cost of specialty care versus the increased cost of a higher rate of negative biopsies. Third, this understanding will establish standards of care and allow various clinical settings to measure progress and set goals for improvement. Previously reported NNT were analyzed solely based on practice setting or level of training.4–8 Many factors influence a provider’s decision to biopsy a suspicious lesion, including the patient’s age, personal and family history, and site of the lesion. It is important to acknowledge that both patient anxiety as well as an individual physician’s clinical experience influence the perceived need to biopsy.9, 10 Even physician compensation has been cited as a motivator.11 The aim of this meta-analysis is to analyze all published data on NNTs across different clinical settings and to report the difference in melanoma detection efficacy between them. This will provide a systematic comparison of published reports on NNT for melanoma.

Methods

This study was approved by the Duke University Institutional Review Board. This report was written in accordance to the Preferred Reporting Items for Systemic Reviews and Meta-Analyses (PRISMA) statement12 whenever possible.

Search Strategy and Selection of Relevant Studies

All criteria for inclusion and exclusion of reports were determined before the literature search. To identify eligible studies, a comprehensive search strategy designed by a Master of Library and Information Science–trained librarian to identify all relevant studies of number needed to treat or number needed to excise in relation to melanoma in the electronic database MEDLINE. English language articles were included from January 1995 – December 2016. Terms related to melanoma, pigmented lesion, nevus/nevi, biopsy, number needed to treat, and number needed to excise were searched with all available synonyms.

Our search yielded a total of 790 articles which were analyzed for inclusion. There were five articles not included in the initial search that were included in this analysis.5, 6, 13–15 Titles and abstracts from the search results were assessed independently by two reviewers. Disagreements were resolved by two other reviewers. Subsequently, the full text and references of articles that met inclusion criteria were reviewed and the data was extracted.4–8, 10, 11, 13–33 One study34 found by our search included biopsies only from patients that were under longitudinal surveillance and at particularly high risk and was therefore excluded from statistical analysis.

Data Extraction

The selected studies data were abstracted by using a standardized data extraction form. Several articles we evaluated published NNTs for different clinical settings over different periods of time and were therefore separated into different studies in our analysis.4, 13, 23, 27, 33 General study characteristics (author, country of origin, type of study, clinic type, total number of biopsies, total number of melanomas) were recorded. Each selected study was determined to include NNT data from primary care physicians only (designated P), combined data from primary care physicians and primary physicians with some dermatologic training (designated PD), dermatologists only (designated D), combined data from dermatologists and dermatologists with some training in pigmented lesions (designated DS), or pigmented lesion specialists only (designated S). Pigmented lesion specialists are dermatologists with a subspecialty in pigmented lesions. Whenever possible, the number of reported nevi without seborrheic keratosis (SK) was used for our NNT calculations. For some included studies, this necessitated subtracting the reported number of SKs from total biopsies. Other studies did not report specific numbers, and therefore included both nevi and SK in the NNT calculation.5, 6, 8, 10, 11, 13–16, 19, 21, 22, 24, 27–30, 32, 33, 35, 36

Statistical Methods

The NNT with 95% confidence interval (CI) was calculated for all groups and according to specialty (dermatologist, pigmented lesion specialist, and primary care). For each meta-analysis conducted, we first computed the overall log odds of melanoma diagnosis and its CIs given that the log odds is approximately normal for large samples. The log odds of melanoma is equal to log(p/(1-p)) with p representing the proportion of melanoma in biopsy. We then transformed the overall log odds estimate and 95% CI back to the original NNT units. By using this strategy, we calculated the overall NNT and the NNT according to specialty (dermatologist, pigmented lesion specialist, and primary care). Heterogeneity between studies was assessed using Cochran’s Q and the I2 statistic. The studies were found not to share a common true effect, thus for each meta-analysis we used an inverse, variance-weighted, random-effects model. Funnel plots were used to determine the likelihood of publication bias.37 All analyses were performed using SAS Version 9.4 (SAS Institute Inc., Cary, NC).

Results

A flowchart of search results is shown in Fig. 1. After removal of duplicates, there was a total of 795 papers to review; then 748 were excluded based on the information in their title and abstract. Thus, 47 full-text original articles were evaluated. After reviewing these articles, we found that 29 studies fit our inclusion criteria and could be used in the meta-analysis. Table 1 summarizes data from all included studies.

Figure 1.

Figure 1.

Open in a new tab

Flow chart of search and study selection process.

Table 1 –

Summary of selected studies (n = 36).4–8, 10, 11, 13–33

Study Year Specialty NMR NNT Melanoma No. (%) Total Biop/Exci No. Melanoma in situ (%) Lesions used to calculate
Ahnlide16 2014 D 5.54 6.81 252 (15) 1717 49.6 Nevi, SK
Argenziano17 2008 S 3.42 4.42 12 (23) 54 50 Nevi
Argenziano (Study 1)4 2012 P 28.49 29.49 7263 (3) 214122 36.8 Nevi
Argenziano (Study 2)4 2012 S 7.69 8.69 9910 (12) 86093 14.1 Nevi
Baade14 2008 P 10.82 19.59 152 (5) 2977 36.2 Nevi, SK
Bauer18 2005 S 15.50 16.50 2 (6) 33 100 Nevi
Carli19 2003 S 5.56 6.75 16 (15) 108 25 Nevi, SK
Carli (BJD)20 2004 S 8.51 9.57 319 (10) 3053 46.4 Nevi
Carli10 2004 S 4.00 5.33 15 (19) 80 NR Nevi, SK
Chia21 2008 D NR 3.52 195 (28) 686 NR Pigmented lesions
English22 2003 P 18.96 29.03 295 (3) 8563 39 Nevi, SK
English11 2004 P 19.53 29.37 160 (3) 4699 38.8 Nevi, SK
Esdaile (Study 1)23 2014 D 2.46 3.46 188 (29) 650 23.9 Nevi
Esdaile (Study 2)23 2014 S 1.74 2.74 266 (36) 730 37.2 Nevi
Green5 2004 D 26.14 26.14 156 (4) 4078 NR Nevi, SK
Haenssle24 2006 D 12.02 12.02 53 (8) 637 52.8 Melanocytic
Hansen6 2009 P 22.25 30.49 348 (3) 10612 38.5 Nevi, SK
Kittler26 2006 DS 4.48 5.48 91 (18) 499 58.2 Nevi
Kittler25 2000 S 8.38 9.38 8 (11) 75 62.5 Nevi
Marks (Study 1)27 1997 PD 10.77 15.64 707 (6) 11055 33.8 Nevi, SK
Marks (Study 2)27 1997 PD 7.97 12.53 1099 (8) 13766 41.1 Nevi, SK
Menzies28 2001 S 7.14 8.57 7 (12) 60 71.4 Nevi, SK
Nault15 2015 D NR 21.39 23 (5) 492 NR Pigmented lesions
Rolfe29 2012 D 6.18 11.47 55(9) 631 56.0 Nevi, SK
Rosendahl30 2012 P NR 9.25 2367 (11) 21900 NR Pigmented lesions
Sidhu31 2012 D 5.25 6.25 750 (16) 4691 NR Nevi
Soares7 2009 D 9.20 10.51 147 (10) 1545 49.7 Nevi
Soltani-Arabshani32 2015 DS 10.82 14.56 165 (7) 2402 46.7 Nevi, SK
Terushkin (Study 1)13 2010 D 12.17 13.92 12 (7) 167 NR Nevi, SK
Terushkin (Study 2)13 2010 D 12.55 14.09 11 (7) 155 NR Nevi, SK
Terushkin (Study 3)13 2010 S 2.54 3.77 13 (27) 49 NR Nevi, SK
Terushkin (Study 4)13 2010 S 5.83 7.67 6 (13) 46 NR Nevi, SK
Tromme (Study 1)33 2012 D 8.86 9.86 93 (10) 917 20.4 Nevi
Tromme (Study 2)33 2012 D 7.11 8.11 74 (12) 600 36.5 Nevi
Tromme (Study 3)33 2012 S 2.09 3.09 64 (32) 198 37.5 Nevi
Wilson8 2012 D 7.67 14.64 28 (7) 410 NR Nevi, SK
Open in a new tab

Abbreviations: P, primary care physicians; D, dermatologists; S, specialists; DS, dermatologists with specialized training; PD, primary care physician with dermatologic training; NMR, nevi-melanoma ratio; NNT, number needed to treat; SK, seborrheic keratosis; NR, not reported; Biop/Exci, biopsies/excisions.

Overall, data from 29 published reports, representing 36 individual studies and a total of 398,549 biopsies/excisions was analyzed. NNT for melanoma in 36 individual studies grouped by the specialty are plotted in Figure 2. The Q statistic for the log odds of all melanoma diagnoses was statistically significant (Q=10182.2; p < 0.001). The random effects model (REM) was then used to estimate the mean log odds with 95% CI and then transformed back to NNT, which was 9.71 (95% CI, 7.72-12.29; I2 = 99.7%) (Fig. 3). The funnel plot for this analysis showed a uniform distribution, indicating a low publication bias (Supplementary Fig. 1)

Figure 2.

Figure 2.

Open in a new tab

Bar graph of NNT for melanoma in 29 published articles (36 individual studies). Dash line represents the overall NNT for all studies evaluated. P, primary care physicians (blue); D, dermatologist (red); S, specialists (green); PD, primary care physician with dermatologic training (brown); DS, dermatologist with some specialized training in pigmented lesions (purple); NNT, number needed to treat.

Figure 3.

Figure 3.

Open in a new tab

Forest plot of NNT for melanoma in 29 published articles (36 individual studies). D, dermatologists; P, primary care physicians; S, specialists; DS, dermatologist with some specialized training in pigmented lesions; PD, primary care physician with dermatologic training; NNT, number needed to treat.

Next, the NNT was calculated by specialty.

Using 6 studies with NNT diagnosed by primary care (P), the Q statistic was statistically significant (Q=2557.3; p < 0.001). The overall NNT diagnosed by P was estimated as 22.62 (95% CI, 12.95-40.10; I2= 99.8%; REM) (Fig. 3). Adding two studies with primary care/dermatologist (PD) designation, the combined NNT is 20.02 (95% CI, 13.07-30.99; I2=99.8%; REM) (Supplementary Fig. 2).

For dermatologists (D), 14 studies were included. the Q statistic for the log odds was again statistically significant (Q=663.3; p < 0.001 and the overall NNT for dermatologist was estimated as 9.60 (95% CI, 6.97-13.41; I2 = 98.0%; REM) (Fig. 3).

For the 12 studies with lesion specialists (S), the Q statistic was statistically significant (Q=461.1; p < 0.001) and NNT by specialists was calculated to be 5.85 (95% CI, 4.24-8.27; I2 = 97.6%; REM) (Fig. 3). Adding two additional studies with dermatologist/specialist (DS) designation, the combined NNT is 6.23 (95% CI, 4.72-8.36; I2=97.6%; REM) (Supplementary Fig. 3).

A general linear mixed model was created to compare NNT between two types of physicians (D vs P, D vs S, P vs S). The NNT of the primary care physicians (P) was found to be 2.52 times greater than that of the dermatologists (D) (95% CI, 1.31-4.85], p=0.008). The NNT for the dermatologists (D) was 1.77 times greater than that of the specialists (S) (95% CI, 1.01-3.09], p=0.045), while the NNT of the primary care physicians (P) was 4.50 times greater than that of the specialists (S) (95% CI, 2.43-8.34], p < 0.001).

Discussion

Understanding how the level of training and practice setting of physicians treating melanoma affect the ability to accurately diagnose and treat melanoma is essential. This study is the first to our knowledge to compile all current information on NNT across various practice settings and perform a systematic statistical comparison.

We showed that pigmented lesion specialists have the lowest NNT, followed by dermatologists. While many factors are at play, more specialized training and experience likely provide them with better intuition as to which lesions to biopsy. Additionally, the frequency of high-risk patients encountered by specialists is likely variable. Another important consideration is the role of referrals. Paine et al found that the more suspicious a general practitioner is of malignancy, the more likely they are to refer the patient to see a specialist.38 This could decrease the NNT for specialists, but also increase the NNT for primary care providers who see a much lower frequency of melanomas in their practice.27 Since there is no published study that examined the potential effects of referral bias on NNT calculations across clinical settings, a prospective study is likely needed in the future to further evaluate this.

While we did not compare physicians and advanced practice providers (APPs), existing studies show a similar relationship between level of training and NNT. Nault et al found a significantly higher NNT for APPs, primarily nurse practitioners, compared with physicians.15 Anderson et al also found a significantly higher NNT for physician assistants (PAs) compared with physicians (39.4 vs 25.4).39 They did note that PAs were less likely to see patients with significant risk factors such as a personal history of melanoma.

The use of dermoscopy in the detection of melanoma has been shown to directly impact NNT. Kittler et al found that diagnostic accuracy with the use of dermoscopy was significantly higher. However, this difference was only observed with its use by specialists, and its use by untrained or less experienced physicians showed no improvement to clinical inspection alone.40 Lorentzen et al compared the use of dermoscopy between “experts” and “non-experts” in the detection of melanoma, and found a sensitivity of 0.83 and 0.69 respectively (p=0.04). Its use in the expert group doubled the positive likelihood ratios.41 As positive predictive value directly correlates with prevalence, this may account for some of the variation seen between specialty clinics and “non-expert” settings. Binder et al found that “non-expert” use of dermoscopy led to a decrease in sensitivity.42 Others have shown that the use of dermoscopy, while not significantly improving melanoma detection, does lead to a decrease in the number of lesions biopsied.43, 44 Unfortunately, the limitations of this meta-analysis did not allow for us to directly compare NNT with or without dermoscopy due to unavailability of data or inconsistencies in reporting. Further studies are indicated to more formally analyze how its use affects the NNT in different clinical settings.

There appear to be geographic differences that may contribute to NNT, even within consistent practice settings. Comparing a few examples of numbers reported from dermatologists for example, Green et al calculated a NNT of 26 in Miami5, compared to 15 for Wilson et al in North Carolina8, 9 for Soares et al in Arizona7, 13 for Marks et al in Australia27, and around 3 for Esdaile et al in the United Kingdom.23 We included 14 studies reporting data from dermatologists only and found an NNT of 9.6. However, these 14 studies ranged from an NNT of 3.5 to 26.1. While geographic differences play a role, we found that analyzing all published data by clinical setting, irrespective of geography, gives the best estimation of NNT. More studies from consistent regions are needed to formally analyze geographic variations.

As our meta-analysis covers cases that range over more than 20 years, it is quite possible that the NNT was not stable throughout that period. A multicenter survey of over 300,000 cases found that between 1998 and 2007 there was an improvement in NNT for skin cancer specialists but not for non-specialists.4 Conversely, Wang et al showed that between 2000 and 2015 there was an increase in per capita skin biopsies in the Medicaid population without a corresponding increase in excision rates, suggesting an increase in NNT over time.45

The current meta-analysis has limitations. First, there was the inconsistency in data reports, resulting in imperfect and potentially incomplete comparisons across studies. For example, many studies included seborrheic keratoses or pigmented basal cell carcinomas in their number of biopsies and NNT calculations. It was not possible to mitigate these inconsistencies when specific numbers of seborrheic keratoses were not reported. Second, we pooled data from different studies despite high heterogeneity. Additionally, the definition of pigmented lesion specialists varied geographically, creating difficulties in classifying data in the right category. Lastly, it is unclear how referral bias would affect the calculation of NNTs across clinical settings and it is likely that baseline patient characteristic differences at primary care clinics versus pigmented lesion specialty clinics could skew the calculations of NNTs. Criteria for referral to a pigmented lesion specialty clinic may include patients with high number of nevi, personal history of previous melanoma, or family history of melanoma. However, there is data to support that higher risk patients are more likely to receive their initial care from a primary care physician and have their melanomas detected during a routine skin check.46

The treatment of melanoma and other skin cancer is associated with significant cost to patients and health care systems.14 As the disease stage progresses, the cost of treatment increases rapidly. One study estimated the 5 year cost of treating MMIS at $4,648.48 compared with $159,808.17 for stage IV melanoma.47 Another study found that the cost savings from early diagnosis of a single melanoma justify 170 benign biopsies.48 However, in an Australian “open access skin cancer clinic” staffed by family practitioners providing consultations solely for diagnosing and treating skin cancers and suspicious skin lesions, NNT for melanoma was calculated to be as high as 287.49 This suggests that there is great variation in the NNT between different practitioner groups and cost effectiveness must be properly studied and considered. With the large economic burden of healthcare, particularly in the United States, it necessitates that we improve the efficacy of melanoma detection.

In conclusion, in this meta-analysis we found that pigmented lesion specialists have the lowest NNT, followed by dermatologists, suggesting involving specialists and/or dermatologists in the care of patients with many nevi or at high risk of melanoma can likely lead to improved clinical outcome.

Supplementary Material

Supplemental files

Supplementary Figure 1. Funnel plot assessing publication bias of 29 publications (36 individual studies) included in this meta-analysis.

Supplementary Figure 2. Forest plots for sub-analysis of melanoma NNT for primary care physician and primary care physician/dermatologist in 8 published reports. P, primary care physicians; PD, primary care physician with dermatologic training; NNT, number needed to treat.

Supplementary Figure 3. Forest plots for sub-analysis of melanoma NNT for 458 pigmented lesion specialist and dermatologist/specialist in 14 published reports. S; pigmented lesion specialist; DS, dermatologist with some specialized training in pigmented lesions; NNT, number needed to treat.

Acknowledgments

Funding sources: Research reported in this publication was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number UL1TR002553. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflicts of interests: None declared.

IRB approval status: Reviewed and approved by Duke University IRB: approval #0007909.

References

  • 1.American Cancer Society. Cancer Facts and Figures 2018. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annual-cancer-facts-and-figures/2018/cancer-facts-and-figures-2018.pdf. Published 2018 Accessed March 8th, 2018
  • 2.Rigel DS. Trends in dermatology: melanoma incidence. Arch Dermatol 2010;146:318. [DOI] [PubMed] [Google Scholar]
  • 3.Leiter U, Eigentler T , Garbe C. Epidemiology of skin cancer. Adv Exp Med Biol 2014;810:120–40. [DOI] [PubMed] [Google Scholar]
  • 4.Argenziano G, Cerroni L, Zalaudek I, Staibano S, Hofmann-Wellenhof R, Arpaia N et al. Accuracy in melanoma detection: a 10-year multicenter survey. J Am Acad Dermatol 2012;67:54–9. [DOI] [PubMed] [Google Scholar]
  • 5.Green AR, Elgart GW, Ma F, Federman DG , Kirsner RS. Documenting dermatology practice: ratio of cutaneous tumors biopsied that are malignant. Dermatol Surg 2004;30:1208–9. [DOI] [PubMed] [Google Scholar]
  • 6.Hansen C, Wilkinson D, Hansen M, Argenziano G. How good are skin cancer clinics at melanoma detection? Number needed to treat variability across a national clinic group in Australia. J Am Acad Dermatol 2009;61:599–604. [DOI] [PubMed] [Google Scholar]
  • 7.Soares TF, Laman SD, Yiannias JA, Connolly SM, Lim KK, Wu Q et al. Factors leading to the biopsy of 1547 pigmented lesions at Mayo Clinic, Scottsdale, Arizona, in 2005. Int J Dermatol 2009;48:1053–6. [DOI] [PubMed] [Google Scholar]
  • 8.Wilson RL, Yentzer BA, Isom SP, Feldman SR , Fleischer AB Jr. How good are US dermatologists at discriminating skin cancers? A number-needed-to-treat analysis. J Dermatol Treat 2012;23:65–9. [DOI] [PubMed] [Google Scholar]
  • 9.Morton CA , Mackie RM. Clinical accuracy of the diagnosis of cutaneous malignant melanoma. Br J Dermatol 1998;138:283–7. [DOI] [PubMed] [Google Scholar]
  • 10.Carli P, Nardini P, Crocetti E, De Giorgi V , Giannotti B. Frequency and characteristics of melanomas missed at a pigmented lesion clinic: a registry-based study. Melanoma Res 2004;14:403–7. [DOI] [PubMed] [Google Scholar]
  • 11.English DR, Del Mar C , Burton RC. Factors influencing the number needed to excise: excision rates of pigmented lesions by general practitioners. Med J Aust 2004;180:16–9. [DOI] [PubMed] [Google Scholar]
  • 12.Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015;4:1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Terushkin V, Warycha M, Levy M, Kopf AW, Cohen DE , Polsky D. Analysis of the benign to malignant ratio of lesions biopsied by a general dermatologist before and after the adoption of dermoscopy. Arch Dermatol 2010;146:343–4. [DOI] [PubMed] [Google Scholar]
  • 14.Baade PD, Youl PH, Janda M, Whiteman DC, Del Mar CB , Aitken JF. Factors associated with the number of lesions excised for each skin cancer: a study of primary care physicians in Queensland, Australia. Arch Dermatol 2008;144:1468–76. [DOI] [PubMed] [Google Scholar]
  • 15.Nault A, Zhang C, Kim K, Saha S, Bennett DD , Xu YG. Biopsy Use in Skin Cancer Diagnosis: Comparing Dermatology Physicians and Advanced Practice Professionals. JAMA Dermatol 2015;151:899–902. [DOI] [PubMed] [Google Scholar]
  • 16.Ahnlide I, Nielsen K , Bjellerup M. Diagnosis of pigmented skin tumours in a dermatological setting: different aspects of the number needed to excise as a measure of efficiency. Acta Derm Venereol 2014;94:683–6. [DOI] [PubMed] [Google Scholar]
  • 17.Argenziano G, Mordente I, Ferrara G, Sgambato A, Annese P , Zalaudek I. Dermoscopic monitoring of melanocytic skin lesions: clinical outcome and patient compliance vary according to follow-up protocols. Br J Dermatol 2008;159:331–6. [DOI] [PubMed] [Google Scholar]
  • 18.Bauer J, Blum A, Strohhacker U , Garbe C. Surveillance of patients at high risk for cutaneous malignant melanoma using digital dermoscopy. Br J Dermatol 2005;152:87–92. [DOI] [PubMed] [Google Scholar]
  • 19.Carli P, De Giorgi V, Betti R, Vergani R, Catricala C, Mariani G et al. Relationship between cause of referral and diagnostic outcome in pigmented lesion clinics: a multicentre survey of the Italian Multidisciplinary Group on Melanoma (GIPMe). Melanoma Res 2003;13:207–11. [DOI] [PubMed] [Google Scholar]
  • 20.Carli P, De Giorgi V, Crocetti E, Mannone F, Massi D, Chiarugi A et al. Improvement of malignant/benign ratio in excised melanocytic lesions in the ‘dermoscopy era’: a retrospective study 1997-2001. Br J Dermatol 2004;150:687–92. [DOI] [PubMed] [Google Scholar]
  • 21.Chia AL, Simonova G, Dutta B, Lim A , Shumack S. Melanoma diagnosis: Australian dermatologists’ number needed to treat. Australas J Dermatol 2008;49:12–5. [DOI] [PubMed] [Google Scholar]
  • 22.English DR, Burton RC, del Mar CB, Donovan RJ, Ireland PD , Emery G. Evaluation of aid to diagnosis of pigmented skin lesions in general practice: controlled trial randomised by practice. BMJ 2003;327:375. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Esdaile B, Mahmud I, Palmer A , Bowling J. Diagnosing melanoma: how do we assess how good we are? Clin Exp Dermatol 2014;39:129–34. [DOI] [PubMed] [Google Scholar]
  • 24.Haenssle HA, Krueger U, Vente C, Thoms KM, Bertsch HP, Zutt M et al. Results from an observational trial: digital epiluminescence microscopy follow-up of atypical nevi increases the sensitivity and the chance of success of conventional dermoscopy in detecting melanoma. J Investig Dermatol 2006;126:980–5. [DOI] [PubMed] [Google Scholar]
  • 25.Kittler H, Pehamberger H, Wolff K , Binder M. Follow-up of melanocytic skin lesions with digital epiluminescence microscopy: patterns of modifications observed in early melanoma, atypical nevi, and common nevi. J Am Acad Dermatol 2000;43:467–76. [DOI] [PubMed] [Google Scholar]
  • 26.Kittler H, Guitera P, Riedl E, Avramidis M, Teban L, Fiebiger M et al. Identification of clinically featureless incipient melanoma using sequential dermoscopy imaging. Arch Dermatol 2006;142:1113–9. [DOI] [PubMed] [Google Scholar]
  • 27.Marks R, Jolley D, McCormack C , Dorevitch AP. Who removes pigmented skin lesions? J Am Acad Dermatol 1997;36:721–6. [DOI] [PubMed] [Google Scholar]
  • 28.Menzies SW, Gutenev A, Avramidis M, Batrac A , McCarthy WH. Short-term digital surface microscopic monitoring of atypical or changing melanocytic lesions. Arch Dermatol 2001;137:1583–9. [DOI] [PubMed] [Google Scholar]
  • 29.Rolfe HM. Accuracy in skin cancer diagnosis: a retrospective study of an Australian public hospital dermatology department. Australas J Dermatol 2012;53:112–7. [DOI] [PubMed] [Google Scholar]
  • 30.Rosendahl C, Williams G, Eley D, Wilson T, Canning G, Keir J et al. The impact of subspecialization and dermatoscopy use on accuracy of melanoma diagnosis among primary care doctors in Australia. J Am Acad Dermatol 2012;67:846–52. [DOI] [PubMed] [Google Scholar]
  • 31.Sidhu S, Bodger O, Williams N , Roberts DL. The number of benign moles excised for each malignant melanoma: the number needed to treat. Clin Exp Dermatol 2012;37:6–9. [DOI] [PubMed] [Google Scholar]
  • 32.Soltani-Arabshahi R, Sweeney C, Jones B, Florell SR, Hu N , Grossman D. Predictive value of biopsy specimens suspicious for melanoma: support for 6-mm criterion in the ABCD rule. J Am Acad Dermatol 2015;72:412–8. [DOI] [PubMed] [Google Scholar]
  • 33.Tromme I, Sacre L, Hammouch F, Legrand C, Marot L, Vereecken P et al. Availability of digital dermoscopy in daily practice dramatically reduces the number of excised melanocytic lesions: results from an observational study. Br J Dermatol 2012;167:778–86. [DOI] [PubMed] [Google Scholar]
  • 34.Robinson JK , Nickoloff BJ. Digital epiluminescence microscopy monitoring of high-risk patients. Arch Dermatol 2004;140:49–56. [DOI] [PubMed] [Google Scholar]
  • 35.Wilkinson D, Askew DA , Dixon A. Skin cancer clinics in Australia: workload profile and performance indicators from an analysis of billing data. Med J Aust 2006;184:162–4. [DOI] [PubMed] [Google Scholar]
  • 36.Youl PH, Baade PD, Janda M, Del Mar CB, Whiteman DC , Aitken JF. Diagnosing skin cancer in primary care: how do mainstream general practitioners compare with primary care skin cancer clinic doctors? Med J Aust 2007;187:215–20. [DOI] [PubMed] [Google Scholar]
  • 37.Song F, Khan KS, Dinnes J , Sutton AJ. Asymmetric funnel plots and publication bias in meta-analyses of diagnostic accuracy. Int J Epidemiol 2002;31:88–95. [DOI] [PubMed] [Google Scholar]
  • 38.Paine SL, Cockburn J, Noy SM , Marks R. Early detection of skin cancer. Knowledge, perceptions and practices of general practitioners in Victoria. Med J Aust 1994;161:188–9, 92–5. [DOI] [PubMed] [Google Scholar]
  • 39.Anderson AM, Matsumoto M, Saul MI, Secrest AM , Ferris LK. Accuracy of Skin Cancer Diagnosis by Physician Assistants Compared With Dermatologists in a Large Health Care System. JAMA Dermatol 2018;154:569–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Kittler H, Pehamberger H, Wolff K , Binder M. Diagnostic accuracy of dermoscopy. Lancet Oncol 2002;3:159–65. [DOI] [PubMed] [Google Scholar]
  • 41.Lorentzen H, Weismann K, Petersen CS, Larsen FG, Secher L , Skodt V. Clinical and dermatoscopic diagnosis of malignant melanoma. Assessed by expert and non-expert groups. Acta Derm Venereol 1999;79:301–4. [DOI] [PubMed] [Google Scholar]
  • 42.Binder M, Schwarz M, Winkler A, Steiner A, Kaider A, Wolff K et al. Epiluminescence microscopy. A useful tool for the diagnosis of pigmented skin lesions for formally trained dermatologists. Arch Dermatol 1995;131:286–91. [DOI] [PubMed] [Google Scholar]
  • 43.van der Rhee JI, Bergman W , Kukutsch NA. The impact of dermoscopy on the management of pigmented lesions in everyday clinical practice of general dermatologists: a prospective study. Br J Dermatol 2010;162:563–7. [DOI] [PubMed] [Google Scholar]
  • 44.Nachbar F, Stolz W, Merkle T, Cognetta AB, Vogt T, Landthaler M et al. The ABCD rule of dermatoscopy. High prospective value in the diagnosis of doubtful melanocytic skin lesions. J Am Acad Dermatol 1994;30:551–9. [DOI] [PubMed] [Google Scholar]
  • 45.Wang DM, Morgan FC, Besaw RJ , Schmults CD. An ecological study of skin biopsies and skin cancer treatment procedures in the United States Medicare population, 2000 to 2015. J Am Acad Dermatol 2018;78:47–53. [DOI] [PubMed] [Google Scholar]
  • 46.Watts CG, Madronio CM, Morton RL, Goumas C, Armstrong BK, Curtin A et al. Diagnosis and clinical management of melanoma patients at higher risk of a new primary melanoma: A population-based study in New South Wales, Australia. Australas J Dermatol 2017;58:278–85. [DOI] [PubMed] [Google Scholar]
  • 47.Alexandrescu DT. Melanoma costs: a dynamic model comparing estimated overall costs of various clinical stages. Dermatol Online J 2009; 15:1. [PubMed] [Google Scholar]
  • 48.Aires DJ, Wick J, Shaath TS, Rajpara AN, Patel V, Badawi AH et al. Economic Costs Avoided by Diagnosing Melanoma Six Months Earlier Justify >100 Benign Biopsies. J Drugs Dermatol 2016;15:527–32. [PubMed] [Google Scholar]
  • 49.Moffatt CR, Green AC , Whiteman DC. Diagnostic accuracy in skin cancer clinics: the Australian experience. Int J Dermatol 2006;45:656–60. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental files

Supplementary Figure 1. Funnel plot assessing publication bias of 29 publications (36 individual studies) included in this meta-analysis.

Supplementary Figure 2. Forest plots for sub-analysis of melanoma NNT for primary care physician and primary care physician/dermatologist in 8 published reports. P, primary care physicians; PD, primary care physician with dermatologic training; NNT, number needed to treat.

Supplementary Figure 3. Forest plots for sub-analysis of melanoma NNT for 458 pigmented lesion specialist and dermatologist/specialist in 14 published reports. S; pigmented lesion specialist; DS, dermatologist with some specialized training in pigmented lesions; NNT, number needed to treat.

NIHMS1549359-supplement-Supplemental_files.docx (14.3MB, docx)

RESOURCES