- Email: [email protected]
Clinical outcomes from skin screening clinics within a community-based melanoma screening program
Clinical outcomes from skin screening clinics within a community-based melanoma screening program Joanne F. Aitken, PhD,a,b Monika Janda, PhD,a,c Mark Elwood, MD,d Philippa H. Youl, MPH,a Ian T. Ring, FAFPHM,e and John B. Lowe, DrPHf Brisbane, Queensland, Carlton, Victoria, and Wollongong, New South Wales, Australia; and Iowa City, Iowa Background: Within a randomized trial of population screening for melanoma, primary care physicians conducted whole-body skin examinations and referred all patients with suspect lesions to their own doctor for further treatment. Objective: Our aim was to describe characteristics of skin screening participants, clinical screening diagnoses, management following referral, and specificity and yield of screening examinations. Methods: Information collected from consent forms, referral forms, and histopathological reports of lesions that had been excised or undergone biopsy was analyzed by means of descriptive statistics. Results: A total of 16,383 whole-body skin examinations resulted in 2302 referrals (14.1% overall; 15.5% men, 18.2% $ 50 years of age) for 4129 suspect lesions (including 222 suspected melanoma, 1101 suspected basal cell carcinomas [BCCs], 265 suspected squamous cell carcinomas [SCCs]). Histopathologic results were available for 94.8% of 1417 lesions excised and confirmed 33 melanomas (23 in men; 24 in participants $ 50 years of age), 259 BCCs, and 97 SCCs. The probability of detecting skin cancer of any type within the program was 2.4%. The estimated specificity of whole-body skin examinations for melanoma was 86.1% (95% confidence interval = 85.6-86.6). The positive predictive value (number of confirmed/ number of lesions excised or biopsied 3 100) for melanoma was 2.5%, 19.3% for BCC, and 7.2% for SCC (overall positive predictive value for skin cancer, 28.9%). Limitations: Follow-up of participants with a negative screening examination has not been conducted for the present investigation. Conclusions: The rate of skin cancer detected per 100 patients screened was higher than previously reported and men and attendees older than 50 years more frequently received a referral and diagnosis of melanoma. The specificity for detection of melanoma through whole-body skin examination by a primary care physician was comparable to that of other screening tests, including mammography. ( J Am Acad Dermatol 2006;54:105-14.)
elanoma is the fourth most common cancer in Australia. Between 1990 and 2000, the incidence of melanoma in men and women increased per year by 2.4% and 1.5%, respectively.1 In 2001, in a population of approximately 19,400,000, approximately 8900 new cases of melanoma and just over 1000 deaths were recorded.1
M
Within the United States, the incidence of melanoma has also increased over past decades2 and for 2005, about 59,580 new cases of melanoma and 7770 deaths from melanoma have been estimated.3 The thickness of the tumor at diagnosis is the best predictor of survival from melanoma.4-6 Whole-body skin examination by a physician is a relatively rapid,
From Viertel Centre for Research in Cancer Control, Queensland Cancer Fund, Brisbanea; School of Population Health, University of Queensland, Brisbaneb; School of Public Health, Queensland University of Technology, Brisbanec; National Cancer Control Initiative, Carltond; Centre for Health Services Development, University of Wollongonge; and the Department of Community and Behavioral Health, College of Public Health, University of Iowa, Iowa City.f Funding source: Queensland Cancer Fund and Queensland Health.
Conflicts of interest: None identified. Accepted for publication August 30, 2005. Reprint requests: Joanne F. Aitken, PhD, Viertel Centre for Research in Cancer Control, Queensland Cancer Fund, PO Box 201, Spring Hill, QLD 4004, Australia. E-mail: [email protected]. Published online November 25, 2005. 0190-9622/$32.00 ª 2005 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2005.08.072
105
106 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
Abbreviations used: AAD: BCC: CI: KC: PPV: SCC: USPSTF:
American Academy of Dermatology basal cell carcinoma confidence interval keratinocyte carcinoma positive predictive value squamous cell carcinoma US Preventive Services Task Force
safe, and inexpensive method of screening for melanoma, which may improve early detection and reduce mortality rates associated with melanoma.7,8 Dermatologists and primary care physicians in several countries have initiated ‘‘skin cancer awareness days’’ offering skin screening to interested volunteers9-13; however, in the absence of evidence from randomized trials, there is insufficient objective scientific evidence for population screening for melanoma and most major medical organizations do not recommend this.14,15 For example, the US Preventive Services Taskforce (USPSTF) has stated that ‘‘. . . the evidence is insufficient to recommend for or against routine screening for skin cancer using a total-body skin examination for the early detection of cutaneous melanoma, basal cell cancer or squamous cell skin cancer’’. Nevertheless, the USPSTF described skin screening as the most promising strategy for reducing the excess burden of skin cancer in older persons.15 In 1998, the Queensland Cancer Fund began the first phase of a randomized trial of a population screening program for melanoma involving 18 communities, 9 of which were randomized to the intervention arm.16 The intervention program aimed to increase skin screening rates within the 9 intervention communities such that 60% of the adult population ( $ 30 years*) had at least one whole-body skin examination by a doctor in the 3-year intervention period (1998-2001). We have estimated that the uptake of skin screening (whole-body skin examination by a doctor within the past 3 years) within the 9 intervention communities was 47.5% (range, 33%61%). Detailed results are described elsewhere.17 The aim of the present study was to describe participation in the open-access skin screening clinics that were conducted within the intervention communities, the clinical and histopathological outcomes of the whole-body skin examinations performed at the clinics, the compliance with follow-up of suspect lesions detected during the screening examination and to estimate the specificity for melanoma of whole-body skin examination by primary *Although melanoma occurs in those persons younger than 30 years of age, it is relatively rare.
care physicians within a population screening program for melanoma.
METHODS The design of the community-based randomized controlled trial of screening for melanoma has been described in detail elsewhere.16 Eighteen communities (9 intervention and 9 control) were enrolled in the first phase of the trial. The 9 intervention communities received a 3-year intervention program, the aims of which were to increase community participation in skin screening by primary care physicians and to encourage whole-body skin self-examination with presentation of suspect lesions to a doctor. A detailed description of the background, development, implementation, and reach of the intervention program has been published elsewhere.18 In brief, the intervention program consisted of 3 components: a community education program, a support and education program for primary care physicians to encourage them to offer whole-body skin examination to their patients, and the provision of free openaccess skin screening clinics that were centrally organized and administered as part of the intervention program. Within the skin clinics, whole-body skin examinations were provided by primary care physicians, including physicians hired from outside the communities as well as local doctors, if they were available. Before holding the clinics, physicians participated in a 1-day training workshop that covered the background of the community-based, randomized trial of screening for melanoma and included sessions on skin cancer epidemiology, early diagnosis, management, and patient communication. Sessions also included clinical practice training with a dermatologist and a 1-day on-site clinical skills assessment by a physician with expertise in skin cancer. Dermatologists were not otherwise involved in the community intervention program. Clinics were held in workplaces, community venues, and local hospitals and included day and evening sessions. A personalized letter of invitation to attend a skin screening clinic was mailed to all residents of intervention communities, who were 30 to 79 years of age. Skin clinics were also advertised in local newspapers. Those attending the clinics gave written informed consent for the whole-body examination and for researchers to access the medical information resulting from any suspect lesion detected during the skin examination. Participants completed a short questionnaire before the examination to ascertain demographic information, history of skin screening, and skin cancer risk factors. Screening examinations involved visual examination by the doctor of the skin on the whole body,
Aitken et al 107
J AM ACAD DERMATOL VOLUME 54, NUMBER 1
excluding areas covered by underwear. Patients with suspect lesions discovered during the examination were referred back to their own doctor for diagnosis and management. SkinWatch clinic doctors were instructed to refer patients who reported a new lesion, an existing lesion that changed in size/shape or color, or any other lesions of concern. In addition, doctors were instructed to refer patients with any lesion that exhibited bleeding, itching, crusting, or pain and any lesion that the doctor viewed as being suspect for skin cancer based on the ABCDE criteria or a 7 point checklist provided to SkinWatch doctors. Patients who only had lesions that the skin clinic doctors considered to be solar or seborrheic keratoses did not require formal referral; however, the screening doctor advised the patients of their options for removal by their primary care physician. No treatment was conducted within the skin clinics. The referral form included a description of the site and provisional diagnosis (melanoma, benign nevus, dysplastic nevus, basal cell carcinoma [BCC], squamous cell carcinoma [SCC], seborrheic keratosis, solar keratosis, or other) of any lesion referred. Patients were provided with a copy of the referral form and were asked to visit their own doctor for diagnosis and management. A second copy of the referral form was sent to the patient’ s own doctor, who was asked to record the management of each referred lesion (no action necessary, monitor lesion, excision, biopsy, cryotherapy, referral, other). Copies of all referrals were also kept by the research team. For all patients who visited their own doctor, the management of the lesion and date of management were recorded. For all lesions excised or biopsied, the histopathological diagnosis was obtained from the treating doctor or directly from the pathology laboratory. Patients who did not attend their doctor within 2 months of referral were telephoned and sent a reminder letter advising them of the importance of seeing their doctor. This process was repeated 5 months after the skin screening examination for all participants who had not attended their doctor by that time. Skin clinic questionnaires. The questionnaires distributed at the clinics asked about the participant’ s age, sex, natural hair color, skin color, tendency to burn, number of moles on the body, current concern about a spot or mole, number of whole-body skin examinations within the past 3 years; history of removal of a skin lesion, history of melanoma, history of other skin cancer (Table I). During the intervention period slightly different versions of the questionnaire were used. All information collected, including skin clinic questionnaires, skin clinic referral forms, and
pathology reports if a lesion was excised or biopsied, was entered into a relational database for analysis. Data analysis Standard descriptive statistical analyses were used to summarize the characteristics of the skin clinic attendees and all those who received a clinical diagnosis for a suspect lesion. Chi-square tests were performed to determine differences between attendees who did or did not receive a referral for a suspect lesion by the skin clinic doctors, and attendees with a referral who did or did not visit their own doctor to establish a definite diagnosis for the suspect skin lesion(s). Estimation of specificity for melanoma diagnosis. We estimated the specificity of the screening examination as follows: true negative screens/(true-negative 1 false-positive screens) 3 100). A positive screen was defined as one that resulted in a referral for investigation of a skin lesion. Since we did not follow up screening participants who had negative screens, the exact number of trueand false-negative screens is unknown. However, reports from other screening programs for melanoma have found the probability of having no melanoma if the initial skin examination is negative as in the order of 99.8%.19 For the purposes of calculating specificity, we assumed the melanoma incidence rate for the Queensland population (51.0 per 100,000)20 as a reasonable upper limit of potentially missed melanomas in those with a negative screen. Adjusted for the age distribution within screening participants, this yielded an expected number of 49 potential false-negative cases for a 3-year screening program. Specificity was then estimated separately for participants older than 30 years and older than 50 years. Positive predictive value. We determined the positive predictive value (PPV) of the whole-body screening examination as the number of confirmed melanoma, BCC, SCC, and any of these 3 cancer diagnoses over all lesions subjected to excision or biopsy where a histopathological report was available. Ethics Ethical approval for this study was granted by the Behavioural and Social Sciences Ethical Review Committee of the University of Queensland.
RESULTS Overall, during the 3-year intervention period, 16,383 whole-body skin examinations were performed. Men and women were approximately
108 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
Table I. Characteristics of skin clinic attendees referred (n = 2303) and not referred (n = 14,081) for a suspected lesion Trial communities (n = 66,576)
Sex Men Women Age (y) \20 20-29 30-39 40-49 50-59 60-69 $70 No. of visits First Second Third or fourth visit Skin cancer risk factors Hair color Brown/black Blond Red Not recorded Skin color Brown/olive/Asian Medium Fair Not recorded Tendency to burn after 1 h of unprotected sun exposure Tan a lot Tan slightly without burning Burn, then tan Burn, no tan Not recorded No. of moles on the body as presented on a pictogram None \10 10-50 [50 Not recorded Current concern about a spot or mole No/don’ t know Yes Not recorded Had a whole-body skin examination by a doctor during the past 3 y No/don’ t know Yes Not recorded
No.
%
33,932 32,644
51.0 49.0
20,528 10,998 10,895 8407 6081 4785 4882
30.8 16.5 16.4 12.6 9.1 7.2 7.3
Clinic attendees not referred (n = 14,081) No.
Referred for suspect lesion (n = 2302)
%
No.
%
6715 7366
47.7 52.3
1230 1072
53.4 46.6
845 1492 2842 3034 2876 1855 1137
6.0 10.6 20.2 21.5 20.4 13.2 8.1
62 146 316 470 534 435 339
2.7 6.3 13.7 20.4 23.2 18.9 14.7
13,264 772 45
94.2 5.5 0.3
2079 212 11
90.3 9.2 0.5
7105 1463 507 5006
78.3 16.1 5.6
1185 245 96 776
77.7 16.1 6.3
1122 2136 5878 4945
12.3 23.4 64.3
167 325 1032 778
11.0 21.3 67.7
528 1930 4384 2251 498
5.8 13.7 31.1 16.0
79 323 726 387 787
5.2 21.3 47.9 25.5
1264 5264 2007 378 5165
14.2 59.0 22.5 4.2
166 835 366 103 1470
11.3 56.3 24.9 4.5
6812 6005 1264
53.1 46.9
785 1403 104
36.2 63.8
10,457 2509 1115
80.6 19.4
1679 540 83
75.7 24.3
J AM ACAD DERMATOL
Aitken et al 109
VOLUME 54, NUMBER 1
equally represented at the clinics (Table I). The mean age of those attending the clinics was 46.5 years (standard deviation = 16.4). The screening program was targeted to people aged 30 years and over and most of those attending (84.5%) were in this age range compared with 52.6% of all residents within trial communities who were older than 30 years (Table I). A total of 907 (5.5%) of those attending the clinics were under the age of 20 years, and 1638 (10%) were aged 20 to 29 years. In total, 47.7% of all those attending reported at least one common risk factor for skin cancer, such as fair skin, a tendency to burn after 1 hour of unprotected sun exposure, or more than 10 moles on the body. Most participants (15,343; 93.7%) visited the skin clinics only once during the 3-year intervention period, whereas 1040 participants (6.3%) attended two or more times. There was no difference between men and women in the number of times they visited the clinics (x 2 = 0.046; P = .83). Participants 50 years of age and older were more likely to visit the clinics more than once compared with younger participants (8.8% vs 4.4%; x2 = 128.4; P \.01). Among the total 16,383 skin screening examinations performed at the clinics, 2302 participants (14.1%) were referred for investigation of 4129 suspect lesions. Men (15.5%) were more likely to receive a referral for a suspect lesion than were women (12.7%) (x 2 = 26.1; P\.01). The probability of referral increased with the age of the patients. Of the 2545 people attending who were younger than 30 years of age, 208 (8.2%) were referred compared with 786 (11.8%) of those 30 to 49 years of age and 1308 (18.2%) of those 50 years old or older (x 2 = 204.4; P \ .01). Those referred were also more likely to have indicated they were concerned about a spot or mole (63.8% vs 46.9%; x2 = 223.0; P \ .01), to have 10 or more moles on the body (39.4% vs 26.7%; x 2 = 20.9; P \ .01) and to have had a whole-body skin examination within the past 3 years (24.3% vs 19.4%; x2 = 29.6; P \ .01) compared with other clinic attendees (Table I). Among the 4129 referred lesions, 1588 were suspected by the doctor at the skin clinic to be skin cancer: 222 melanoma (5.4%), 1101 BCC (26.7%), and 265 SCC (6.4%). Overall, 1.3 lesions per 100 examinations were suspected to be melanoma, and 8.3 lesions per 100 examinations were suspected to be BCC or SCC. The remaining 2541 referred lesions (61.5%) were suspected to be Hutchinson’s melanotic freckle, benign or dysplastic nevi, seborrheic keratosis, solar keratosis, or other nonmalignant lesions (Fig 1). Of the total 2302 patients who received a referral for a suspect lesion, 60.1% visited their doctor
without a reminder, and an additional 14.4% and 4.7% visited their doctor after the first and second reminders, respectively. Overall, 480 (20.8%) did not visit their doctor or could not be contacted for follow-up. Compared with other referred patients, these 480 patients were more likely to be younger than 40 years of age (20.3% vs 32.1%; x 2 = 67.4; P \ .01), more likely to have received the referral from a workplace clinic (16.8% vs 27.5%; x 2 = 28.8; P \.01), and less likely to have received a clinical diagnosis of melanoma (9.9% vs 5.0%; x2 = 11.7; P \.01). There was no difference between these patients and those who filled their referral with respect to sex, skin cancer risk factors, current concern about a spot or mole, skin screening history, and personal or family history of skin cancer (data not shown). Among the 1822 patients who visited a doctor with their skin clinic referral, some form of lesion management was recorded for 2982 lesions (Fig 1). Among these, 1417 lesions were excised or underwent biopsy, and histopathology was subsequently available for 1343 (94.8%) of these lesions (Table II). Patients who were subjected to excision or biopsy did not differ significantly in their characteristics from those who received a referral. Fig 1 presents a summary of the suspected diagnosis and pathologic outcomes for the 1343 lesions. The suspected diagnosis exactly matched the histopathologic diagnosis for 427 lesions (31.8%, 95% confidence interval [CI] = 0.29-0.34). Within all lesions that were excised or underwent biopsy, the PPV was 2.5% (number of confirmed melanomas/number of lesions excised or undergoing biopsy 3 100 = 33/1343 3 100) for melanoma, 7.2% (number of confirmed SCC/ number of lesions excised or undergoing biopsy 3 100 = 97/1343 3 100) for SCC, and 19.3% (number of confirmed BCC/number of lesions excised or undergoing biopsy 3 100 = 259/1343 3 100) for BCC, with an overall PPV (number of confirmed skin cancers /number of lesions excised or undergoing biopsy 3 100) of 29.0%. The overall probability of detecting skin cancer within the program (389 confirmed melanomas, BCCs, and SCCs out of 16,383 screening examinations) was 2.4%. Assuming a 3-year screening interval and a maximum number of false-negative screens for melanoma of 49, the specificity of melanoma diagnosis by whole-body skin examination was estimated to be approximately 14,032/16,301 3 100 = (truenegative screens/(true-negative 1 false-positive screens) 3 100 = 86.1 (95% CI = 85.6-86.6). Even after doubling the potential rate of false-negative screens, this estimate of specificity changed very little. Specificity changed slightly with the age of the patients; therefore, assuming a 3-year screening
110 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
Fig 1. Clinical outcomes from 16,383 examinations within a population screening program for melanoma. HMF, Hutchinson’s melanotic freckle.
Aitken et al 111
J AM ACAD DERMATOL VOLUME 54, NUMBER 1
Table II. Histopathological diagnosis for lesions excised or biopsied* Histopathological diagnosis Benign Dysplastic Seborrheic Solar Total Melanoma BCC SCC HMF nevus nevus Lentigo keratosis keratosis Other Suspected diagnosis (n = 1343) (n = 33) (n = 259) (n = 97) (n = 1) (n = 433) (n = 96) (n = 103) (n = 85) (n = 87) (n = 149)
Melanoma BCC SCC HMF Benign nevus Dysplastic nevus Lentigo Seborrheic keratosis Solar keratosis Other
161 371 87 5 171 368 11 52
15 1 1 0 0 10 1 3y
7 200 21 0 4 8 0 2
3 51 28 0 0 2 0 1
1 0 0 0 0 0 0 0
68 19 1 1 101 211 4 19
18 2 0 0 21 51 0 2
19 5 1 2 23 39 4 5
12 12 8 1 4 28 0 11
4 28 16 1 3 4 0 3
14 53 11 0 15 15 2 6
47 70
0 2z
13 4
6 6
0 0
2 7
0 2
0 5
3 6
17 11
6 27
BCC, Basal cell carcinoma; HMF, Hutchinson’s melanotic freckle; SCC, squamous cell carcinoma. *No pathology report was noted for 74 lesions. y All 3 lesions were suspected to be seborrheic keratosis; however, the skin clinic doctor also noted some additional features, making the lesions atypical enough to warrant excision to rule out melanoma. z Not further specified.
interval and only including patients 30 years of age or older, specificity was 85.0 (95% CI = 84.4-85.6). For patients 50 years of age or older, the specificity was 81.9 (95% CI = 81.0-82.8). Melanomas were confirmed in 23 men and 10 women. Fifteen lesions of these 33 patients had been suspected by the skin screening doctor to be melanoma, 10 melanomas were found among the 368 suspected dysplastic nevi, 3 melanomas were found among the 52 lesions suspected to be seborrheic keratosis, and one melanoma among the lesions suspected to be lentigo (Table II). Twenty-four of these 33 melanoma patients were 50 years of age or older, 22 reported that they were currently concerned about a spot or mole at the time of the screening examination, and 26 had not had a wholebody skin examination by a doctor in the past 3 years. Patients who received the diagnosis of melanoma were significantly younger than patients with a diagnosis of BCC or SCC (x 2 = 20.1; P = .01). There were no other significant differences in self-reported skin cancer risk factors and skin screening behaviors between patients diagnosed with melanoma and those diagnosed with BCC or SCC (data not shown). Of the 33 confirmed melanomas, 13 were in-situ, 18 were less than 1.00 mm thick, and 2 were 1.00 mm thick or more.
DISCUSSION To our knowledge, this is the first study to report the clinical outcomes of whole-body skin
examinations within the context of a population screening program for melanoma by primary care practitioners. Unlike other skin screening programs which attracted women at a much higher rate,12,13,21-23 men and women were equally represented at the clinics, probably reflecting the fact that the community intervention program included personal letters of invitation to men (and women) to attend the clinics.18 More than half of the total 16,383 skin examinations were performed on people younger than 50 years of age. Overall, 14% of all screening examinations resulted in a referral for a suspect lesion, a referral rate comparable to that reported elsewhere as summarized for the USPSTF.24 Men and those persons 50 years of age or older were disproportionately represented within the group of referred patients, consistent with the higher risk for skin cancer within these groups.25 Only 15 of 33 melanomas were among the lesions in which the skin clinic doctor suspected this diagnosis; 10 were suspected to be dysplastic nevi, 3 were suspected to be seborrheic keratosis, and one suspected of being lentigo. This implies that further restriction of the numbers of patients referred for assessment on the basis of the clinical diagnosis would lead to missing some melanomas. At present, patients with dysplastic nevi are not assiduously followed up in all screening programs. Doing so would likely increase the yield of the screening program, but at increased cost and decreased specificity. There was one large group that appears to have fairly benign outcomes, the lesions with a
112 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
Table III. Skin screening program characteristics
Hefland et al24 (2001); review Swetter et al30 (2003) Bolognia et al22 (1990) Carli et al13 (2003) Present study
Referred/ 100 screened
Biopsied/ 100 screened
Suspected KC/100 screened
2-34
4-31
2-20
54 12
Confirmed KC/100 screened
0.0004-0.05 30%-58%
13.4 11
Among suspected KC, how many confirmed KC/100 screened
13.9
0.03
4
3.7
4*
0.8
14
8.6
8.3
2.2
Suspected melanoma/ 100 screened
Confirmed melanoma/ 100 screened
Among suspected melanoma, how many confirmed melanoma/ 100 screened
0-0.9
0-0.4
6%-58%
31-62% for BCC, 4 33-43% for SCC 43% for BCC, 0.8 14% for SCC 23% overall* Not specified 53% for BCC, 1.3 32% for SCC
7%-12.5% 0
0
0.2 0.2
9.3%
KC, Keratinocyte carcinoma. *KC and melanoma.
clinical diagnosis of benign nevus, which included 4 BCCs but no other malignancy. Of those melanomas found through the screening program, about 39% were in situ lesions, 55% were thin invasive lesions less than 1 mm thick, and 6% were 1 mm thick or greater. Within the population of Queensland during the period from 1999 through 2002, the corresponding percentages are as follows: 36%, in-situ melanomas; 48%, invasive melanomas less than 1 mm thick; 16%, invasive melanomas 1 mm thick or more,26 indicating, as has been found elsewhere,12 that melanomas found through screening tend to be less advanced than those detected symptomatically. Of course, it is not possible to know what proportion of these in-situ and thin invasive lesions would have progressed over time had they not been detected and treated. Further follow-up of this cohort and comparison with the characteristics of melanomas within the unscreened community may provide some insight into the possible long-term benefit of this screening program. Most referred patients (80%) presented for additional advice or treatment from their doctor. Although 60% of all referred participants visited their doctor without a reminder, the two follow-up telephone calls and letters motivated an additional 19% to fill their referral, thus significantly improving the overall complete follow-up rate achieved. This compliance rate of approximately 80% is similar to those rates reported from the American Academy of Dermatology’ s (AAD) Skin Cancer Screening Program,10,12 a skin cancer screening day in British Columbia,27 within a mobile screening program in
the Netherlands,28 and a community-based skin screening program in San Diego.29 Although patients who apparently ignored their referral were more likely to be younger than 40 years of age and less likely to be referred for a suspected melanoma, we did not detect any association between compliance and the perceived personal risk of developing skin cancer or commonly noted skin cancer risk factors. In an evaluation of the AAD screening program, compliance with a recommended biopsy was higher in persons who received a diagnosis of ‘‘suspected melanoma’’ compared with ‘‘rule out melanoma’’.10 Younger individuals, those referred for more than one lesion, and individuals without a family history of skin cancer were reported to be less compliant with the recommended follow-up of a screen-detected lesion in the San Diego screening program,29 whereas serious illness, concern about melanoma but not about other skin cancers, concern about a different lesion, and denial were among the most prominent reasons reported by screenees who did not follow a referral after skin screening in Connecticut.22 In summary, these findings suggest that future screening programs need to specifically educate younger screening participants with a diagnosis of suspected skin cancer about the importance of seeing their doctor to receive appropriate management. Approximately 8.3 per 100 patients screened within the present screening program received a clinical diagnosis of keratinocyte carcinoma (KC) (BCC and SCC). This is at the upper end of the range summarized for the USPSTF of up to 9 suspected KCs
J AM ACAD DERMATOL
Aitken et al 113
VOLUME 54, NUMBER 1
per 100 screened, whereas the rate of 2.2 confirmed KCs per 100 screened is higher than the 0.0004 to 0.8 reported in other populations13,22,24 (Table III). Residents of Queensland are at high risk for skin cancer and melanoma because of the combination of a susceptible, mostly fair-skinned population and high exposure to solar radiation in proximity to the equator. The high yield of KCs through screening within the present program reflects the high incidence of such lesions in Queensland.1 Diagnosis of KC is not the aim of a skin screening program, given the slow progression of such lesions and the small effect early detection is likely to have on mortality rates within the screened population. However, referral for, and diagnosis and management of, KC is likely to be responsible for a large proportion of the cost of such a program and needs to be considered should population-based screening for melanoma be introduced. On the other hand, receiving the diagnosis of a KC could increase the participants’ awareness of skin cancer in general and could improve their tendency to participate in future screening for melanoma. Approximately 1.3 per 100 patients screened were referred with a suspected diagnosis of melanoma and 0.2 per 100 had a melanoma histologically confirmed on referral, comparable to the rates reported within other screening populations,13,22,24 some of which specifically targeted high- risk populations.29 Our result is somewhat higher than the 0.15 per 100 among 242,374 screenees of the AAD screening program12 (Table III), which again may reflect the higher prevalence of melanoma in Queensland. Further, only between 38% (during screening period 1991-1994) and 53% (during the screening period 1997-1999) of patients screened within the AAD screening program received a whole-body skin examination. Our estimates of specificity for melanoma (86%) compare favorably with those of other established population screening procedures for cancer, such as mammography for breast cancer (specificity 94%97%), fecal occult blood test for colorectal cancer (specificity 96%-98%) or prostate-specific antigen test for prostate cancer (81%-98%),15 and reflect the skill of doctors in Queensland for diagnosing melanoma. Specificity decreased slightly with older age, most likely because of the higher prevalence of suspect lesions in older populations. The PPV (number of confirmed melanomas/ number of lesions excised or biopsied 3 100) or clinical accuracy for diagnosing melanoma of 2.5% based on the excised lesions within the present program is within the range of 0%-12.5% reported in the literature, mostly from screening programs
facilitated by dermatologists.24 The PPV (number of confirmed SCC/number of lesions excised or biopsied 3 100) of 7.2% within the present study for SCC was similar to those reported by Jonna et al29 (range, 4.2%-12.5%) and within a targeted skin screening program in northern California.30 Green, Leslie, and Weedon31 in a screening program in which dermatologists only excised lesions they suspected to be skin cancer reported a PPV for SCC of 39%. The PPV for BCC (number of confirmed BCCs/number of lesions excised or biopsied 3 100) within the present sample was 19.2%, which is lower than the 31% reported by Swetter et al30 and the 59% reported by Green, Leslie, and Weedon,31 for studies in which dermatologists performed the examinations. As has been discussed earlier,32 the PPV of a screening test is determined by its sensitivity and the prevalence of the preclinical disease in the population. It is therefore difficult to compare the present findings with those reported in the literature because of the much higher prevalence of skin cancer in the Queensland population33 and the provision of skin examinations by primary care physicians rather than dermatologists. In conclusion, to our knowledge this study is the first to document the clinical outcomes and specificity of whole-body skin examination within a population screening program for melanoma conducted by primary care physicians. Reflecting the high incidence of skin cancer in Queensland, the program detected a higher number of skin cancers per 100 patients screened than has been reported previously. Referral and diagnosis of melanoma were significantly more common for men and attendees older than 50 years of age. The specificity for melanoma of whole-body skin examination by a primary care physician was comparable to that of other accepted screening tests for cancer, including mammography, indicating that this is a potentially useful screening test. Whether or not skin screening would be effective in reducing mortality rates of melanoma remains unproven. REFERENCES 1. Australian Institutes of Health and Welfare (AIHW) & Australasian Association of Cancer Registries (AACR). Cancer in Australia 2001. AIHW Cat. No. CAN 23. Canberra: AIHW (Cancer Series no. 28). 2004. 2. Howe HL, Wingo PA, Thun MJ, Ries LA, Rosenberg HM, Feigal EG, et al. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 2001;93:824-42. 3. Jemal A, Murray T, Ward E, Samuels A, Tiwari TC, Ghafoor A, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55:10-30. 4. Balch CM, Buzaid AC, Soong SJ, Atkins MB, Cascinelli N, Coit DG, et al. Prognostic factors analysis of 17,600 melanoma patients: validation of the American Joint Committee on Cancer melanoma staging system. J Clin Oncol 2001;19:3622-34.
114 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
5. Balch CM, Buzaid AC, Soong SJ, Atkins MB, Cascinelli N, Coit DG, et al. New TNM melanoma staging system: linking biology and natural history to clinical outcomes. Semin Surg Oncol 2003;21:43-52. 6. Rigel DS, Carucci JA. Malignant melanoma: prevention, early detection, and treatment in the 21st century. CA Cancer J Clin 2000;50:215-36. 7. Friedman LC, Bruce S, Webb JA, Weinberg AD, Cooper HP. Skin self-examination in a population at increased risk for skin cancer. Am J Prev Med 1993;9:359-64. 8. Oliveria SA, Christos PJ, Marghoob AA, Halpern AC. Skin cancer screening and prevention in the primary care setting: national ambulatory medical care survey 1997. J Gen Intern Med 2001;16:297-301. 9. Koh HK, Caruso A, Gage I, Geller AC, Prout MN, White H, et al. Evaluation of melanoma/skin cancer screening in Massachusetts. Cancer 1990;65:375-9. 10. Koh HK, Norton LA, Geller AC, Sun T, Rigel DS, Miller DR, et al. Evaluation of the American Academy of dermatology’s national skin cancer early detection and screening program. J Am Acad Dermatol 1996;34:971-8. 11. Pehamberger H, Binder M, Knollmayer S, Wolff K. Immediate effects of a public education campaign on prognostic features of melanoma. J Am Acad Dermatol 1993;29:106-9. 12. Geller AC, Zhang Z, Sober AJ, Halpern AC, Weinstock MA, Daniels S, et al. The first 15 years of the American Academy of Dermatology Skin Cancer Screening Programs: 1985-1999. J Am Acad Dermatol 2003;48:34-41. 13. Carli P, De Giorgi V, Giannotti B, Seidenart S, Pellacani G, Peris K, et al. Skin cancer day in Italy: method of referral to open access clinics and tumor prevalence in the examined population. Eur J Dermatol 2003;13:76-9. 14. The Cancer Council of Australia (2004). National Cancer Prevention Policy 2004-2006. Available at: http://www.cancer. org.au/documents/NatCancerPreventPol_04-06.pdf. Accessed June 06, 2005. 15. US Preventive Services Task Force (USPSTF)* Guide to clinical preventive services. 3rd ed. Recommendations. Available at: http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=hstat3.part.1. Accessed June 06, 2005. 16. Aitken JF, Elwood M, Lowe JB, Firman DW, Balanda KP, Ring IT. A randomised trial of population screening for melanoma. J Med Screen 2002;9:33-8. 17. Aitken JF, Youl PH, Janda M, Lowe JB, Ring IT, Elwood M. Increase in skin cancer screening during a community-based randomised intervention trial. Int J Cancer 2005. Available at: http://www3.interscience.wiley.com/cgi-bin/jissue/76502439. Published online ahead of print September 8, 2005. 18. Lowe JB, Ball J, Lynch BM, Baldwin L, Janda M, Stanton WR, Aitken JF. Acceptability and feasibility of a community-based
19.
20.
21. 22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32. 33.
screening programme for melanoma in Australia. Health Promot Int 2004;19:437-44. Rampen FH, Casparie-van Velsen JI, van Huystee BE, Kiemeney LA, Schouten LJ. False-negative findings in skin cancer and melanoma screening. J Am Acad Dermatol 1995;33:59-63. QCR and QCF (2004). Cancer in Queensland. Incidence and mortality 1982 to 2002. Brisbane, Queensland Cancer Registry and Queensland Cancer Fund Available at: http://www.health. qld.gov.au/publications/infocirc/12860.pdf. Accessed November 2, 2005. McGee R, Elwood M, Williams S, Lowry F. Who comes to skin checks? N Z Med J 1994;107:58-60. Bolognia JL, Berwick M, Fine JA. Complete follow-up and evaluation of skin cancer screening in Connecticut. J Am Acad Dematol 1990;23:1098-106. De Rooij MJ, Rampen FH, Schouten LJ, Neumann HA. Factors influencing participation among melanoma screening attenders. Acta Derm Venereol 1997;77:467-70. Hefland M, Mahon SS, Eden KB, Frame PS, Orleans CT. Screening for skin cancer: a summary of the evidence. Am J Prev Med 2001;20:47-58. Hanrahan PF, Hersey P, D’Este CA. Factors involved in presentation of older people with thick melanoma. Med J Aust 1998;169:410-4. Coory M, Baade P, Aitken JF, Smithers M, McLead GRC, Ring IT. Trends for in-situ and invasive melanoma in Queensland, Australia 1982 to 2002. Cancer Causes Control. In press. Engelberg D, Gallagher RP, Rivers JK. Follow-up and evaluation of skin cancer screening in British Columbia. J Am Acad Dermatol 1999;41:37-42. Krol S, Kreijser LMT, Van Der Rhee HJ, Welvaart K. Screening for skin cancer in the Netherlands. Acta Derm Venereol (Stockh) 1991;71:317-21. Jonna BP, Delfino RJ, Newman WG, Tope WD. Positive predictive value for presumptive diagnoses of skin cancer and compliance with follow-up among patients attending a community screening program. Prev Med 1998;27:611-6. Swetter SM, Waddell BL, Vazquez MD, Khosravi VS. Increased effectiveness of targeted skin cancer screening in the Veterans Affairs population of Northern California. Prev Med 2003;36: 164-71. Green A, Leslie D, Weedon D. Diagnosis of skin cancer in the general population: clinical accuracy in the Nambour study. Med J Aust 1988;148:447-50. Hennekens CH, Bruing JE. Epidemiology in medicine. Philadelphia: Lippincott Williams & Wilkins; 1987. Aitken J, Elwood M. Population screening for melanoma: current evidence and a population-based randomised trial. In: Thompson JF, Morton DL, Kroon BBR, editors. Textbook of melanoma. London: Martin Dunitz; 2004. pp. 100-14.
elanoma is the fourth most common cancer in Australia. Between 1990 and 2000, the incidence of melanoma in men and women increased per year by 2.4% and 1.5%, respectively.1 In 2001, in a population of approximately 19,400,000, approximately 8900 new cases of melanoma and just over 1000 deaths were recorded.1
M
Within the United States, the incidence of melanoma has also increased over past decades2 and for 2005, about 59,580 new cases of melanoma and 7770 deaths from melanoma have been estimated.3 The thickness of the tumor at diagnosis is the best predictor of survival from melanoma.4-6 Whole-body skin examination by a physician is a relatively rapid,
From Viertel Centre for Research in Cancer Control, Queensland Cancer Fund, Brisbanea; School of Population Health, University of Queensland, Brisbaneb; School of Public Health, Queensland University of Technology, Brisbanec; National Cancer Control Initiative, Carltond; Centre for Health Services Development, University of Wollongonge; and the Department of Community and Behavioral Health, College of Public Health, University of Iowa, Iowa City.f Funding source: Queensland Cancer Fund and Queensland Health.
Conflicts of interest: None identified. Accepted for publication August 30, 2005. Reprint requests: Joanne F. Aitken, PhD, Viertel Centre for Research in Cancer Control, Queensland Cancer Fund, PO Box 201, Spring Hill, QLD 4004, Australia. E-mail: [email protected]. Published online November 25, 2005. 0190-9622/$32.00 ª 2005 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2005.08.072
105
106 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
Abbreviations used: AAD: BCC: CI: KC: PPV: SCC: USPSTF:
American Academy of Dermatology basal cell carcinoma confidence interval keratinocyte carcinoma positive predictive value squamous cell carcinoma US Preventive Services Task Force
safe, and inexpensive method of screening for melanoma, which may improve early detection and reduce mortality rates associated with melanoma.7,8 Dermatologists and primary care physicians in several countries have initiated ‘‘skin cancer awareness days’’ offering skin screening to interested volunteers9-13; however, in the absence of evidence from randomized trials, there is insufficient objective scientific evidence for population screening for melanoma and most major medical organizations do not recommend this.14,15 For example, the US Preventive Services Taskforce (USPSTF) has stated that ‘‘. . . the evidence is insufficient to recommend for or against routine screening for skin cancer using a total-body skin examination for the early detection of cutaneous melanoma, basal cell cancer or squamous cell skin cancer’’. Nevertheless, the USPSTF described skin screening as the most promising strategy for reducing the excess burden of skin cancer in older persons.15 In 1998, the Queensland Cancer Fund began the first phase of a randomized trial of a population screening program for melanoma involving 18 communities, 9 of which were randomized to the intervention arm.16 The intervention program aimed to increase skin screening rates within the 9 intervention communities such that 60% of the adult population ( $ 30 years*) had at least one whole-body skin examination by a doctor in the 3-year intervention period (1998-2001). We have estimated that the uptake of skin screening (whole-body skin examination by a doctor within the past 3 years) within the 9 intervention communities was 47.5% (range, 33%61%). Detailed results are described elsewhere.17 The aim of the present study was to describe participation in the open-access skin screening clinics that were conducted within the intervention communities, the clinical and histopathological outcomes of the whole-body skin examinations performed at the clinics, the compliance with follow-up of suspect lesions detected during the screening examination and to estimate the specificity for melanoma of whole-body skin examination by primary *Although melanoma occurs in those persons younger than 30 years of age, it is relatively rare.
care physicians within a population screening program for melanoma.
METHODS The design of the community-based randomized controlled trial of screening for melanoma has been described in detail elsewhere.16 Eighteen communities (9 intervention and 9 control) were enrolled in the first phase of the trial. The 9 intervention communities received a 3-year intervention program, the aims of which were to increase community participation in skin screening by primary care physicians and to encourage whole-body skin self-examination with presentation of suspect lesions to a doctor. A detailed description of the background, development, implementation, and reach of the intervention program has been published elsewhere.18 In brief, the intervention program consisted of 3 components: a community education program, a support and education program for primary care physicians to encourage them to offer whole-body skin examination to their patients, and the provision of free openaccess skin screening clinics that were centrally organized and administered as part of the intervention program. Within the skin clinics, whole-body skin examinations were provided by primary care physicians, including physicians hired from outside the communities as well as local doctors, if they were available. Before holding the clinics, physicians participated in a 1-day training workshop that covered the background of the community-based, randomized trial of screening for melanoma and included sessions on skin cancer epidemiology, early diagnosis, management, and patient communication. Sessions also included clinical practice training with a dermatologist and a 1-day on-site clinical skills assessment by a physician with expertise in skin cancer. Dermatologists were not otherwise involved in the community intervention program. Clinics were held in workplaces, community venues, and local hospitals and included day and evening sessions. A personalized letter of invitation to attend a skin screening clinic was mailed to all residents of intervention communities, who were 30 to 79 years of age. Skin clinics were also advertised in local newspapers. Those attending the clinics gave written informed consent for the whole-body examination and for researchers to access the medical information resulting from any suspect lesion detected during the skin examination. Participants completed a short questionnaire before the examination to ascertain demographic information, history of skin screening, and skin cancer risk factors. Screening examinations involved visual examination by the doctor of the skin on the whole body,
Aitken et al 107
J AM ACAD DERMATOL VOLUME 54, NUMBER 1
excluding areas covered by underwear. Patients with suspect lesions discovered during the examination were referred back to their own doctor for diagnosis and management. SkinWatch clinic doctors were instructed to refer patients who reported a new lesion, an existing lesion that changed in size/shape or color, or any other lesions of concern. In addition, doctors were instructed to refer patients with any lesion that exhibited bleeding, itching, crusting, or pain and any lesion that the doctor viewed as being suspect for skin cancer based on the ABCDE criteria or a 7 point checklist provided to SkinWatch doctors. Patients who only had lesions that the skin clinic doctors considered to be solar or seborrheic keratoses did not require formal referral; however, the screening doctor advised the patients of their options for removal by their primary care physician. No treatment was conducted within the skin clinics. The referral form included a description of the site and provisional diagnosis (melanoma, benign nevus, dysplastic nevus, basal cell carcinoma [BCC], squamous cell carcinoma [SCC], seborrheic keratosis, solar keratosis, or other) of any lesion referred. Patients were provided with a copy of the referral form and were asked to visit their own doctor for diagnosis and management. A second copy of the referral form was sent to the patient’ s own doctor, who was asked to record the management of each referred lesion (no action necessary, monitor lesion, excision, biopsy, cryotherapy, referral, other). Copies of all referrals were also kept by the research team. For all patients who visited their own doctor, the management of the lesion and date of management were recorded. For all lesions excised or biopsied, the histopathological diagnosis was obtained from the treating doctor or directly from the pathology laboratory. Patients who did not attend their doctor within 2 months of referral were telephoned and sent a reminder letter advising them of the importance of seeing their doctor. This process was repeated 5 months after the skin screening examination for all participants who had not attended their doctor by that time. Skin clinic questionnaires. The questionnaires distributed at the clinics asked about the participant’ s age, sex, natural hair color, skin color, tendency to burn, number of moles on the body, current concern about a spot or mole, number of whole-body skin examinations within the past 3 years; history of removal of a skin lesion, history of melanoma, history of other skin cancer (Table I). During the intervention period slightly different versions of the questionnaire were used. All information collected, including skin clinic questionnaires, skin clinic referral forms, and
pathology reports if a lesion was excised or biopsied, was entered into a relational database for analysis. Data analysis Standard descriptive statistical analyses were used to summarize the characteristics of the skin clinic attendees and all those who received a clinical diagnosis for a suspect lesion. Chi-square tests were performed to determine differences between attendees who did or did not receive a referral for a suspect lesion by the skin clinic doctors, and attendees with a referral who did or did not visit their own doctor to establish a definite diagnosis for the suspect skin lesion(s). Estimation of specificity for melanoma diagnosis. We estimated the specificity of the screening examination as follows: true negative screens/(true-negative 1 false-positive screens) 3 100). A positive screen was defined as one that resulted in a referral for investigation of a skin lesion. Since we did not follow up screening participants who had negative screens, the exact number of trueand false-negative screens is unknown. However, reports from other screening programs for melanoma have found the probability of having no melanoma if the initial skin examination is negative as in the order of 99.8%.19 For the purposes of calculating specificity, we assumed the melanoma incidence rate for the Queensland population (51.0 per 100,000)20 as a reasonable upper limit of potentially missed melanomas in those with a negative screen. Adjusted for the age distribution within screening participants, this yielded an expected number of 49 potential false-negative cases for a 3-year screening program. Specificity was then estimated separately for participants older than 30 years and older than 50 years. Positive predictive value. We determined the positive predictive value (PPV) of the whole-body screening examination as the number of confirmed melanoma, BCC, SCC, and any of these 3 cancer diagnoses over all lesions subjected to excision or biopsy where a histopathological report was available. Ethics Ethical approval for this study was granted by the Behavioural and Social Sciences Ethical Review Committee of the University of Queensland.
RESULTS Overall, during the 3-year intervention period, 16,383 whole-body skin examinations were performed. Men and women were approximately
108 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
Table I. Characteristics of skin clinic attendees referred (n = 2303) and not referred (n = 14,081) for a suspected lesion Trial communities (n = 66,576)
Sex Men Women Age (y) \20 20-29 30-39 40-49 50-59 60-69 $70 No. of visits First Second Third or fourth visit Skin cancer risk factors Hair color Brown/black Blond Red Not recorded Skin color Brown/olive/Asian Medium Fair Not recorded Tendency to burn after 1 h of unprotected sun exposure Tan a lot Tan slightly without burning Burn, then tan Burn, no tan Not recorded No. of moles on the body as presented on a pictogram None \10 10-50 [50 Not recorded Current concern about a spot or mole No/don’ t know Yes Not recorded Had a whole-body skin examination by a doctor during the past 3 y No/don’ t know Yes Not recorded
No.
%
33,932 32,644
51.0 49.0
20,528 10,998 10,895 8407 6081 4785 4882
30.8 16.5 16.4 12.6 9.1 7.2 7.3
Clinic attendees not referred (n = 14,081) No.
Referred for suspect lesion (n = 2302)
%
No.
%
6715 7366
47.7 52.3
1230 1072
53.4 46.6
845 1492 2842 3034 2876 1855 1137
6.0 10.6 20.2 21.5 20.4 13.2 8.1
62 146 316 470 534 435 339
2.7 6.3 13.7 20.4 23.2 18.9 14.7
13,264 772 45
94.2 5.5 0.3
2079 212 11
90.3 9.2 0.5
7105 1463 507 5006
78.3 16.1 5.6
1185 245 96 776
77.7 16.1 6.3
1122 2136 5878 4945
12.3 23.4 64.3
167 325 1032 778
11.0 21.3 67.7
528 1930 4384 2251 498
5.8 13.7 31.1 16.0
79 323 726 387 787
5.2 21.3 47.9 25.5
1264 5264 2007 378 5165
14.2 59.0 22.5 4.2
166 835 366 103 1470
11.3 56.3 24.9 4.5
6812 6005 1264
53.1 46.9
785 1403 104
36.2 63.8
10,457 2509 1115
80.6 19.4
1679 540 83
75.7 24.3
J AM ACAD DERMATOL
Aitken et al 109
VOLUME 54, NUMBER 1
equally represented at the clinics (Table I). The mean age of those attending the clinics was 46.5 years (standard deviation = 16.4). The screening program was targeted to people aged 30 years and over and most of those attending (84.5%) were in this age range compared with 52.6% of all residents within trial communities who were older than 30 years (Table I). A total of 907 (5.5%) of those attending the clinics were under the age of 20 years, and 1638 (10%) were aged 20 to 29 years. In total, 47.7% of all those attending reported at least one common risk factor for skin cancer, such as fair skin, a tendency to burn after 1 hour of unprotected sun exposure, or more than 10 moles on the body. Most participants (15,343; 93.7%) visited the skin clinics only once during the 3-year intervention period, whereas 1040 participants (6.3%) attended two or more times. There was no difference between men and women in the number of times they visited the clinics (x 2 = 0.046; P = .83). Participants 50 years of age and older were more likely to visit the clinics more than once compared with younger participants (8.8% vs 4.4%; x2 = 128.4; P \.01). Among the total 16,383 skin screening examinations performed at the clinics, 2302 participants (14.1%) were referred for investigation of 4129 suspect lesions. Men (15.5%) were more likely to receive a referral for a suspect lesion than were women (12.7%) (x 2 = 26.1; P\.01). The probability of referral increased with the age of the patients. Of the 2545 people attending who were younger than 30 years of age, 208 (8.2%) were referred compared with 786 (11.8%) of those 30 to 49 years of age and 1308 (18.2%) of those 50 years old or older (x 2 = 204.4; P \ .01). Those referred were also more likely to have indicated they were concerned about a spot or mole (63.8% vs 46.9%; x2 = 223.0; P \ .01), to have 10 or more moles on the body (39.4% vs 26.7%; x 2 = 20.9; P \ .01) and to have had a whole-body skin examination within the past 3 years (24.3% vs 19.4%; x2 = 29.6; P \ .01) compared with other clinic attendees (Table I). Among the 4129 referred lesions, 1588 were suspected by the doctor at the skin clinic to be skin cancer: 222 melanoma (5.4%), 1101 BCC (26.7%), and 265 SCC (6.4%). Overall, 1.3 lesions per 100 examinations were suspected to be melanoma, and 8.3 lesions per 100 examinations were suspected to be BCC or SCC. The remaining 2541 referred lesions (61.5%) were suspected to be Hutchinson’s melanotic freckle, benign or dysplastic nevi, seborrheic keratosis, solar keratosis, or other nonmalignant lesions (Fig 1). Of the total 2302 patients who received a referral for a suspect lesion, 60.1% visited their doctor
without a reminder, and an additional 14.4% and 4.7% visited their doctor after the first and second reminders, respectively. Overall, 480 (20.8%) did not visit their doctor or could not be contacted for follow-up. Compared with other referred patients, these 480 patients were more likely to be younger than 40 years of age (20.3% vs 32.1%; x 2 = 67.4; P \ .01), more likely to have received the referral from a workplace clinic (16.8% vs 27.5%; x 2 = 28.8; P \.01), and less likely to have received a clinical diagnosis of melanoma (9.9% vs 5.0%; x2 = 11.7; P \.01). There was no difference between these patients and those who filled their referral with respect to sex, skin cancer risk factors, current concern about a spot or mole, skin screening history, and personal or family history of skin cancer (data not shown). Among the 1822 patients who visited a doctor with their skin clinic referral, some form of lesion management was recorded for 2982 lesions (Fig 1). Among these, 1417 lesions were excised or underwent biopsy, and histopathology was subsequently available for 1343 (94.8%) of these lesions (Table II). Patients who were subjected to excision or biopsy did not differ significantly in their characteristics from those who received a referral. Fig 1 presents a summary of the suspected diagnosis and pathologic outcomes for the 1343 lesions. The suspected diagnosis exactly matched the histopathologic diagnosis for 427 lesions (31.8%, 95% confidence interval [CI] = 0.29-0.34). Within all lesions that were excised or underwent biopsy, the PPV was 2.5% (number of confirmed melanomas/number of lesions excised or undergoing biopsy 3 100 = 33/1343 3 100) for melanoma, 7.2% (number of confirmed SCC/ number of lesions excised or undergoing biopsy 3 100 = 97/1343 3 100) for SCC, and 19.3% (number of confirmed BCC/number of lesions excised or undergoing biopsy 3 100 = 259/1343 3 100) for BCC, with an overall PPV (number of confirmed skin cancers /number of lesions excised or undergoing biopsy 3 100) of 29.0%. The overall probability of detecting skin cancer within the program (389 confirmed melanomas, BCCs, and SCCs out of 16,383 screening examinations) was 2.4%. Assuming a 3-year screening interval and a maximum number of false-negative screens for melanoma of 49, the specificity of melanoma diagnosis by whole-body skin examination was estimated to be approximately 14,032/16,301 3 100 = (truenegative screens/(true-negative 1 false-positive screens) 3 100 = 86.1 (95% CI = 85.6-86.6). Even after doubling the potential rate of false-negative screens, this estimate of specificity changed very little. Specificity changed slightly with the age of the patients; therefore, assuming a 3-year screening
110 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
Fig 1. Clinical outcomes from 16,383 examinations within a population screening program for melanoma. HMF, Hutchinson’s melanotic freckle.
Aitken et al 111
J AM ACAD DERMATOL VOLUME 54, NUMBER 1
Table II. Histopathological diagnosis for lesions excised or biopsied* Histopathological diagnosis Benign Dysplastic Seborrheic Solar Total Melanoma BCC SCC HMF nevus nevus Lentigo keratosis keratosis Other Suspected diagnosis (n = 1343) (n = 33) (n = 259) (n = 97) (n = 1) (n = 433) (n = 96) (n = 103) (n = 85) (n = 87) (n = 149)
Melanoma BCC SCC HMF Benign nevus Dysplastic nevus Lentigo Seborrheic keratosis Solar keratosis Other
161 371 87 5 171 368 11 52
15 1 1 0 0 10 1 3y
7 200 21 0 4 8 0 2
3 51 28 0 0 2 0 1
1 0 0 0 0 0 0 0
68 19 1 1 101 211 4 19
18 2 0 0 21 51 0 2
19 5 1 2 23 39 4 5
12 12 8 1 4 28 0 11
4 28 16 1 3 4 0 3
14 53 11 0 15 15 2 6
47 70
0 2z
13 4
6 6
0 0
2 7
0 2
0 5
3 6
17 11
6 27
BCC, Basal cell carcinoma; HMF, Hutchinson’s melanotic freckle; SCC, squamous cell carcinoma. *No pathology report was noted for 74 lesions. y All 3 lesions were suspected to be seborrheic keratosis; however, the skin clinic doctor also noted some additional features, making the lesions atypical enough to warrant excision to rule out melanoma. z Not further specified.
interval and only including patients 30 years of age or older, specificity was 85.0 (95% CI = 84.4-85.6). For patients 50 years of age or older, the specificity was 81.9 (95% CI = 81.0-82.8). Melanomas were confirmed in 23 men and 10 women. Fifteen lesions of these 33 patients had been suspected by the skin screening doctor to be melanoma, 10 melanomas were found among the 368 suspected dysplastic nevi, 3 melanomas were found among the 52 lesions suspected to be seborrheic keratosis, and one melanoma among the lesions suspected to be lentigo (Table II). Twenty-four of these 33 melanoma patients were 50 years of age or older, 22 reported that they were currently concerned about a spot or mole at the time of the screening examination, and 26 had not had a wholebody skin examination by a doctor in the past 3 years. Patients who received the diagnosis of melanoma were significantly younger than patients with a diagnosis of BCC or SCC (x 2 = 20.1; P = .01). There were no other significant differences in self-reported skin cancer risk factors and skin screening behaviors between patients diagnosed with melanoma and those diagnosed with BCC or SCC (data not shown). Of the 33 confirmed melanomas, 13 were in-situ, 18 were less than 1.00 mm thick, and 2 were 1.00 mm thick or more.
DISCUSSION To our knowledge, this is the first study to report the clinical outcomes of whole-body skin
examinations within the context of a population screening program for melanoma by primary care practitioners. Unlike other skin screening programs which attracted women at a much higher rate,12,13,21-23 men and women were equally represented at the clinics, probably reflecting the fact that the community intervention program included personal letters of invitation to men (and women) to attend the clinics.18 More than half of the total 16,383 skin examinations were performed on people younger than 50 years of age. Overall, 14% of all screening examinations resulted in a referral for a suspect lesion, a referral rate comparable to that reported elsewhere as summarized for the USPSTF.24 Men and those persons 50 years of age or older were disproportionately represented within the group of referred patients, consistent with the higher risk for skin cancer within these groups.25 Only 15 of 33 melanomas were among the lesions in which the skin clinic doctor suspected this diagnosis; 10 were suspected to be dysplastic nevi, 3 were suspected to be seborrheic keratosis, and one suspected of being lentigo. This implies that further restriction of the numbers of patients referred for assessment on the basis of the clinical diagnosis would lead to missing some melanomas. At present, patients with dysplastic nevi are not assiduously followed up in all screening programs. Doing so would likely increase the yield of the screening program, but at increased cost and decreased specificity. There was one large group that appears to have fairly benign outcomes, the lesions with a
112 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
Table III. Skin screening program characteristics
Hefland et al24 (2001); review Swetter et al30 (2003) Bolognia et al22 (1990) Carli et al13 (2003) Present study
Referred/ 100 screened
Biopsied/ 100 screened
Suspected KC/100 screened
2-34
4-31
2-20
54 12
Confirmed KC/100 screened
0.0004-0.05 30%-58%
13.4 11
Among suspected KC, how many confirmed KC/100 screened
13.9
0.03
4
3.7
4*
0.8
14
8.6
8.3
2.2
Suspected melanoma/ 100 screened
Confirmed melanoma/ 100 screened
Among suspected melanoma, how many confirmed melanoma/ 100 screened
0-0.9
0-0.4
6%-58%
31-62% for BCC, 4 33-43% for SCC 43% for BCC, 0.8 14% for SCC 23% overall* Not specified 53% for BCC, 1.3 32% for SCC
7%-12.5% 0
0
0.2 0.2
9.3%
KC, Keratinocyte carcinoma. *KC and melanoma.
clinical diagnosis of benign nevus, which included 4 BCCs but no other malignancy. Of those melanomas found through the screening program, about 39% were in situ lesions, 55% were thin invasive lesions less than 1 mm thick, and 6% were 1 mm thick or greater. Within the population of Queensland during the period from 1999 through 2002, the corresponding percentages are as follows: 36%, in-situ melanomas; 48%, invasive melanomas less than 1 mm thick; 16%, invasive melanomas 1 mm thick or more,26 indicating, as has been found elsewhere,12 that melanomas found through screening tend to be less advanced than those detected symptomatically. Of course, it is not possible to know what proportion of these in-situ and thin invasive lesions would have progressed over time had they not been detected and treated. Further follow-up of this cohort and comparison with the characteristics of melanomas within the unscreened community may provide some insight into the possible long-term benefit of this screening program. Most referred patients (80%) presented for additional advice or treatment from their doctor. Although 60% of all referred participants visited their doctor without a reminder, the two follow-up telephone calls and letters motivated an additional 19% to fill their referral, thus significantly improving the overall complete follow-up rate achieved. This compliance rate of approximately 80% is similar to those rates reported from the American Academy of Dermatology’ s (AAD) Skin Cancer Screening Program,10,12 a skin cancer screening day in British Columbia,27 within a mobile screening program in
the Netherlands,28 and a community-based skin screening program in San Diego.29 Although patients who apparently ignored their referral were more likely to be younger than 40 years of age and less likely to be referred for a suspected melanoma, we did not detect any association between compliance and the perceived personal risk of developing skin cancer or commonly noted skin cancer risk factors. In an evaluation of the AAD screening program, compliance with a recommended biopsy was higher in persons who received a diagnosis of ‘‘suspected melanoma’’ compared with ‘‘rule out melanoma’’.10 Younger individuals, those referred for more than one lesion, and individuals without a family history of skin cancer were reported to be less compliant with the recommended follow-up of a screen-detected lesion in the San Diego screening program,29 whereas serious illness, concern about melanoma but not about other skin cancers, concern about a different lesion, and denial were among the most prominent reasons reported by screenees who did not follow a referral after skin screening in Connecticut.22 In summary, these findings suggest that future screening programs need to specifically educate younger screening participants with a diagnosis of suspected skin cancer about the importance of seeing their doctor to receive appropriate management. Approximately 8.3 per 100 patients screened within the present screening program received a clinical diagnosis of keratinocyte carcinoma (KC) (BCC and SCC). This is at the upper end of the range summarized for the USPSTF of up to 9 suspected KCs
J AM ACAD DERMATOL
Aitken et al 113
VOLUME 54, NUMBER 1
per 100 screened, whereas the rate of 2.2 confirmed KCs per 100 screened is higher than the 0.0004 to 0.8 reported in other populations13,22,24 (Table III). Residents of Queensland are at high risk for skin cancer and melanoma because of the combination of a susceptible, mostly fair-skinned population and high exposure to solar radiation in proximity to the equator. The high yield of KCs through screening within the present program reflects the high incidence of such lesions in Queensland.1 Diagnosis of KC is not the aim of a skin screening program, given the slow progression of such lesions and the small effect early detection is likely to have on mortality rates within the screened population. However, referral for, and diagnosis and management of, KC is likely to be responsible for a large proportion of the cost of such a program and needs to be considered should population-based screening for melanoma be introduced. On the other hand, receiving the diagnosis of a KC could increase the participants’ awareness of skin cancer in general and could improve their tendency to participate in future screening for melanoma. Approximately 1.3 per 100 patients screened were referred with a suspected diagnosis of melanoma and 0.2 per 100 had a melanoma histologically confirmed on referral, comparable to the rates reported within other screening populations,13,22,24 some of which specifically targeted high- risk populations.29 Our result is somewhat higher than the 0.15 per 100 among 242,374 screenees of the AAD screening program12 (Table III), which again may reflect the higher prevalence of melanoma in Queensland. Further, only between 38% (during screening period 1991-1994) and 53% (during the screening period 1997-1999) of patients screened within the AAD screening program received a whole-body skin examination. Our estimates of specificity for melanoma (86%) compare favorably with those of other established population screening procedures for cancer, such as mammography for breast cancer (specificity 94%97%), fecal occult blood test for colorectal cancer (specificity 96%-98%) or prostate-specific antigen test for prostate cancer (81%-98%),15 and reflect the skill of doctors in Queensland for diagnosing melanoma. Specificity decreased slightly with older age, most likely because of the higher prevalence of suspect lesions in older populations. The PPV (number of confirmed melanomas/ number of lesions excised or biopsied 3 100) or clinical accuracy for diagnosing melanoma of 2.5% based on the excised lesions within the present program is within the range of 0%-12.5% reported in the literature, mostly from screening programs
facilitated by dermatologists.24 The PPV (number of confirmed SCC/number of lesions excised or biopsied 3 100) of 7.2% within the present study for SCC was similar to those reported by Jonna et al29 (range, 4.2%-12.5%) and within a targeted skin screening program in northern California.30 Green, Leslie, and Weedon31 in a screening program in which dermatologists only excised lesions they suspected to be skin cancer reported a PPV for SCC of 39%. The PPV for BCC (number of confirmed BCCs/number of lesions excised or biopsied 3 100) within the present sample was 19.2%, which is lower than the 31% reported by Swetter et al30 and the 59% reported by Green, Leslie, and Weedon,31 for studies in which dermatologists performed the examinations. As has been discussed earlier,32 the PPV of a screening test is determined by its sensitivity and the prevalence of the preclinical disease in the population. It is therefore difficult to compare the present findings with those reported in the literature because of the much higher prevalence of skin cancer in the Queensland population33 and the provision of skin examinations by primary care physicians rather than dermatologists. In conclusion, to our knowledge this study is the first to document the clinical outcomes and specificity of whole-body skin examination within a population screening program for melanoma conducted by primary care physicians. Reflecting the high incidence of skin cancer in Queensland, the program detected a higher number of skin cancers per 100 patients screened than has been reported previously. Referral and diagnosis of melanoma were significantly more common for men and attendees older than 50 years of age. The specificity for melanoma of whole-body skin examination by a primary care physician was comparable to that of other accepted screening tests for cancer, including mammography, indicating that this is a potentially useful screening test. Whether or not skin screening would be effective in reducing mortality rates of melanoma remains unproven. REFERENCES 1. Australian Institutes of Health and Welfare (AIHW) & Australasian Association of Cancer Registries (AACR). Cancer in Australia 2001. AIHW Cat. No. CAN 23. Canberra: AIHW (Cancer Series no. 28). 2004. 2. Howe HL, Wingo PA, Thun MJ, Ries LA, Rosenberg HM, Feigal EG, et al. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 2001;93:824-42. 3. Jemal A, Murray T, Ward E, Samuels A, Tiwari TC, Ghafoor A, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55:10-30. 4. Balch CM, Buzaid AC, Soong SJ, Atkins MB, Cascinelli N, Coit DG, et al. Prognostic factors analysis of 17,600 melanoma patients: validation of the American Joint Committee on Cancer melanoma staging system. J Clin Oncol 2001;19:3622-34.
114 Aitken et al
J AM ACAD DERMATOL JANUARY 2006
5. Balch CM, Buzaid AC, Soong SJ, Atkins MB, Cascinelli N, Coit DG, et al. New TNM melanoma staging system: linking biology and natural history to clinical outcomes. Semin Surg Oncol 2003;21:43-52. 6. Rigel DS, Carucci JA. Malignant melanoma: prevention, early detection, and treatment in the 21st century. CA Cancer J Clin 2000;50:215-36. 7. Friedman LC, Bruce S, Webb JA, Weinberg AD, Cooper HP. Skin self-examination in a population at increased risk for skin cancer. Am J Prev Med 1993;9:359-64. 8. Oliveria SA, Christos PJ, Marghoob AA, Halpern AC. Skin cancer screening and prevention in the primary care setting: national ambulatory medical care survey 1997. J Gen Intern Med 2001;16:297-301. 9. Koh HK, Caruso A, Gage I, Geller AC, Prout MN, White H, et al. Evaluation of melanoma/skin cancer screening in Massachusetts. Cancer 1990;65:375-9. 10. Koh HK, Norton LA, Geller AC, Sun T, Rigel DS, Miller DR, et al. Evaluation of the American Academy of dermatology’s national skin cancer early detection and screening program. J Am Acad Dermatol 1996;34:971-8. 11. Pehamberger H, Binder M, Knollmayer S, Wolff K. Immediate effects of a public education campaign on prognostic features of melanoma. J Am Acad Dermatol 1993;29:106-9. 12. Geller AC, Zhang Z, Sober AJ, Halpern AC, Weinstock MA, Daniels S, et al. The first 15 years of the American Academy of Dermatology Skin Cancer Screening Programs: 1985-1999. J Am Acad Dermatol 2003;48:34-41. 13. Carli P, De Giorgi V, Giannotti B, Seidenart S, Pellacani G, Peris K, et al. Skin cancer day in Italy: method of referral to open access clinics and tumor prevalence in the examined population. Eur J Dermatol 2003;13:76-9. 14. The Cancer Council of Australia (2004). National Cancer Prevention Policy 2004-2006. Available at: http://www.cancer. org.au/documents/NatCancerPreventPol_04-06.pdf. Accessed June 06, 2005. 15. US Preventive Services Task Force (USPSTF)* Guide to clinical preventive services. 3rd ed. Recommendations. Available at: http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=hstat3.part.1. Accessed June 06, 2005. 16. Aitken JF, Elwood M, Lowe JB, Firman DW, Balanda KP, Ring IT. A randomised trial of population screening for melanoma. J Med Screen 2002;9:33-8. 17. Aitken JF, Youl PH, Janda M, Lowe JB, Ring IT, Elwood M. Increase in skin cancer screening during a community-based randomised intervention trial. Int J Cancer 2005. Available at: http://www3.interscience.wiley.com/cgi-bin/jissue/76502439. Published online ahead of print September 8, 2005. 18. Lowe JB, Ball J, Lynch BM, Baldwin L, Janda M, Stanton WR, Aitken JF. Acceptability and feasibility of a community-based
19.
20.
21. 22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32. 33.
screening programme for melanoma in Australia. Health Promot Int 2004;19:437-44. Rampen FH, Casparie-van Velsen JI, van Huystee BE, Kiemeney LA, Schouten LJ. False-negative findings in skin cancer and melanoma screening. J Am Acad Dermatol 1995;33:59-63. QCR and QCF (2004). Cancer in Queensland. Incidence and mortality 1982 to 2002. Brisbane, Queensland Cancer Registry and Queensland Cancer Fund Available at: http://www.health. qld.gov.au/publications/infocirc/12860.pdf. Accessed November 2, 2005. McGee R, Elwood M, Williams S, Lowry F. Who comes to skin checks? N Z Med J 1994;107:58-60. Bolognia JL, Berwick M, Fine JA. Complete follow-up and evaluation of skin cancer screening in Connecticut. J Am Acad Dematol 1990;23:1098-106. De Rooij MJ, Rampen FH, Schouten LJ, Neumann HA. Factors influencing participation among melanoma screening attenders. Acta Derm Venereol 1997;77:467-70. Hefland M, Mahon SS, Eden KB, Frame PS, Orleans CT. Screening for skin cancer: a summary of the evidence. Am J Prev Med 2001;20:47-58. Hanrahan PF, Hersey P, D’Este CA. Factors involved in presentation of older people with thick melanoma. Med J Aust 1998;169:410-4. Coory M, Baade P, Aitken JF, Smithers M, McLead GRC, Ring IT. Trends for in-situ and invasive melanoma in Queensland, Australia 1982 to 2002. Cancer Causes Control. In press. Engelberg D, Gallagher RP, Rivers JK. Follow-up and evaluation of skin cancer screening in British Columbia. J Am Acad Dermatol 1999;41:37-42. Krol S, Kreijser LMT, Van Der Rhee HJ, Welvaart K. Screening for skin cancer in the Netherlands. Acta Derm Venereol (Stockh) 1991;71:317-21. Jonna BP, Delfino RJ, Newman WG, Tope WD. Positive predictive value for presumptive diagnoses of skin cancer and compliance with follow-up among patients attending a community screening program. Prev Med 1998;27:611-6. Swetter SM, Waddell BL, Vazquez MD, Khosravi VS. Increased effectiveness of targeted skin cancer screening in the Veterans Affairs population of Northern California. Prev Med 2003;36: 164-71. Green A, Leslie D, Weedon D. Diagnosis of skin cancer in the general population: clinical accuracy in the Nambour study. Med J Aust 1988;148:447-50. Hennekens CH, Bruing JE. Epidemiology in medicine. Philadelphia: Lippincott Williams & Wilkins; 1987. Aitken J, Elwood M. Population screening for melanoma: current evidence and a population-based randomised trial. In: Thompson JF, Morton DL, Kroon BBR, editors. Textbook of melanoma. London: Martin Dunitz; 2004. pp. 100-14.