Background:

Melanoma is the leading cause of skin cancer mortality. Visual skin examination for skin cancer screening could impact disease incidence and mortality in U.S. adults and adolescents.

Purpose:

We conducted a systematic evidence review of visual skin examination for skin cancer screening in primary care settings to support the U.S. Preventive Services Task Force (USPSTF) in updating its previous recommendation. Our review addressed five key questions in adults and adolescents age 15 years and older without a prior diagnosis of skin cancer: 1) What is the direct evidence that visual skin cancer screening by a primary care provider or dermatologist reduces skin cancer morbidity and mortality and all-cause-mortality? 2) What are the harms of skin cancer screening and diagnostic followup? 3) What are the test characteristics of visual skin cancer screening when performed by primary care providers or dermatologists? 4) Does visual skin cancer screening lead to earlier detection of skin cancer compared to usual care? and 5) What is the association between earlier detection of skin cancer and skin cancer morbidity and mortality and all-cause mortality?

Data Sources:

We searched MEDLINE, PubMed, and the Cochrane Central Register of Controlled Trials for studies published from January 1, 1995 through June 1, 2015. We supplemented searches by examining bibliographies from previous systematic reviews, and retrieved articles and studies included in the previous USPSTF review for potential inclusion. We searched federal agency trial registries for ongoing and unpublished trials.

Study Selection:

We conducted dual independent review of 12,514 abstracts. We reviewed 453 full-text articles, which two reviewers independently evaluated against well-defined inclusion/exclusion criteria and quality rated. Discrepancies were discussed with a third reviewer and resolved by consensus.

Data Extraction and Analysis:

Four investigators abstracted data from 13 studies and 15 articles into evidence tables and a second reviewer checked these data. We qualitatively summarized the evidence for each key question, since data were insufficient in quantity or consistency for meta-analysis.

Results:

Key question 1. What is the direct evidence that visual skin cancer screening by a primary care provider or dermatologist reduces skin cancer morbidity and mortality and all-cause mortality? One fair-quality ecologic study addressed the impact of physician visual skin cancer examination on melanoma mortality. The Skin Cancer Research to Provide Evidence for Effectiveness of Screening in Northern Germany (SCREEN study), conducted in the Schleswig-Holstein region of Germany, involved a multicomponent intervention including the following: 1) training nondermatologists and dermatologists in skin cancer screening; 2) a media campaign to encourage skin cancer screening in adults age 20 years and older; and 3) a followup dermatology referral protocol for nondermatologists to refer adults with either suspicious lesions or multiple risk factors for skin cancer. During the 1-year intervention period (2003 to 2004), nearly 361,000 adults (19% of the age-eligible adults) were screened with a visual skin cancer examination, mainly by nondermatologists. The majority of those screened were women (73.6%) and the mean age was 49.7 years (standard deviation, 16.2 years). Using a pre-post design comparing melanoma mortality in the population in 1998 to 1999 and 2008 to 2009, the SCREEN study demonstrated a 48 percent reduction in melanoma mortality in the Schleswig-Holstein (intervention) region but no reductions in melanoma mortality were observed in the four neighboring (control) regions without an active skin cancer screening program or in Germany as a whole. The reduction in absolute mortality was a decline of 0.8 deaths due to melanoma per 100,000 persons in the intervention region. As an ecologic study, the results do not provide individual-level data about risk reduction associated with screening, and it is not possible to directly compare changes in mortality among those exposed versus not exposed to skin cancer screening and account for confounding.

Key question 2. What are the harms of skin cancer screening and diagnostic followup? Two fair-quality studies evaluated the harms of skin cancer screening by assessing biopsy yield and patient satisfaction with shave biopsy results. We found no studies that evaluated harms due to overdiagnosis, procedure-related adverse events, or psychosocial harms. The SCREEN study demonstrated variation by age in the number of skin excisions needed to detect one melanoma, squamous cell carcinoma, or basal cell carcinoma. For all cancers detected, fewer excisions were needed to detect one case in older adults age 65 years or older compared to younger adults. For melanoma, detecting one case in women age 65 years or older required 22 excisions compared to 41 excisions in women ages 20 to 34 years. Similar patterns were observed in men and for other skin cancer types. In a case series of 45 men and women who participated in skin cancer screening and underwent shave biopsy for suspected nonmelanoma skin cancer, 7.1 percent of patients expressed poor satisfaction with the cosmetic results from shave biopsy after 6 months compared to 16.1 percent of physicians rating the same site as poor.

Key question 3. What are the test characteristics of visual skin cancer screening when performed by primary care providers versus dermatologists? Two fair-quality observational studies reported test characteristics among screening-eligible populations. In the first study, primary care physicians conducted screenings in 16,383 adults in Queensland, Australia. Cancer outcomes were determined by pathology or biopsy reports. False-negative rates were estimated using published literature and population melanoma rates. Within 36 months of the first screening examination, sensitivity for melanoma detection was 40.2 percent (calculated) and specificity was 86.1 percent (95% confidence interval [CI], 85.6 to 86.6). The positive predictive value for melanoma was 1.4 percent. The second study evaluated the performance of volunteer dermatologists and plastic surgeons who conducted screening in Western Australia among 7,436 adult men and women. At 24 months, sensitivity for melanoma detection was 49.0 percent (95% CI, 34.4 to 63.7) and specificity was 97.6 percent (95% CI, 97.2 to 97.9), with an overall recall rate of 2.7 percent. The positive predictive value was 11.9 percent (95% CI, 7.8 to 17.2%). Different followup times for cancer outcomes prohibits direct comparison of screening accuracy between the two physician types.

Key question 4. Does visual skin cancer screening lead to earlier detection of skin cancer compared to usual care? One fair-quality case-control study from Queensland, Australia measured the association between whole-body skin examination by a physician in the previous 3 years among men and women ages 20 to 75 years with or without melanoma. Cases (n=3,762) were diagnosed with first primary melanoma between 2000 and 2003; controls (n=3,824) were randomly selected from electoral rolls according to 5-year age categories and the sex distribution of the cases. Among controls, 28.3 percent reported receiving a whole-body skin examination by a physician within the previous 3 years compared to 35.3 percent of melanoma cases. In multivariate-adjusted models, cases diagnosed with thin melanoma (≤0.75 mm) had a 38 percent higher odds (odds ratio [OR], 1.38 [95% CI, 1.22 to 1.56]) of receiving physician whole-body skin examination in the previous 3 years compared to controls. Further, cases diagnosed with thicker lesions (>0.75 mm) had a 14 percent reduced odds (OR, 0.86 [95% CI, 0.75 to 0.98]) of receiving physician skin examination compared to controls. The thickest melanoma lesion cases (≥3.00 mm) had a 40 percent reduced odds of recent physician skin examination compared to controls (OR, 0.60 [95% CI, 0.43 to 0.83]). These results should be confirmed using a prospective study design.

Key question 5. What is the association between earlier detection of skin cancer and skin cancer morbidity and mortality and all-cause mortality? Eight fair- to good-quality studies evaluated the association between lesion thickness or stage at diagnosis and melanoma mortality and all-cause mortality. Four of the studies were conducted in U.S. populations, one in Germany, and three in Australia. All eight studies demonstrated a consistent statistically significant relationship between the degree of disease involvement at diagnosis and melanoma mortality, regardless of the characterization of the stage or lesion thickness. Thicker lesions (>4.0 mm) were associated with a 3.1- to 32.6-fold increased risk of melanoma mortality compared to thinner lesions. Similarly, advanced-stage melanomas (stage III or above) were associated with a 9.9- to 27.1-fold increased risk of melanoma mortality compared to early-stage melanomas. Stage at melanoma detection was associated with a statistically significant increase in all-cause mortality among melanoma cases identified from California SEER registries; compared to stage I disease at detection, the adjusted hazard ratio of all-cause mortality was 2.26 times higher for stage II disease (95% CI, 2.14 to 2.39), 4.27 for stage III disease (95% CI, 3.90 to 4.67), and 10.39 for stage IV disease (95% CI, 8.96 to 12.00).

Limitations:

Very few screening studies met our inclusion criteria, and few were conducted in U.S. settings or with clear relevance to U.S. primary care.

Conclusions:

On a population level, with limited evidence on skin cancer screening, a clear statement cannot be made about the benefit of skin cancer screening for melanoma mortality and all-cause mortality or association with thinner lesions. With few studies to confirm these results, the applicability for widespread skin cancer screening could be limited. Later stage at diagnosis of melanoma is associated with strong effect on melanoma mortality within 5 years of diagnosis. Future research on skin cancer screening should focus on evaluating the effectiveness of targeted screening in persons considered to be at higher risk for skin cancer.