Screening for reducing morbidity and mortality in malignant melanoma
Screening for malignant melanoma has the potential to reduce morbidity and mortality from the disease through earlier detection, as prognosis is closely associated with the thickness of the lesion at the time of diagnosis. However, there are also potential harms from screening people without skin lesion concerns, such as overdiagnosis of lesions that would never have caused symptoms if they had remained undetected. Overdiagnosis results in harm through unnecessary treatment and the psychosocial consequences of being labelled with a cancer diagnosis. For any type of screening, the benefits must outweigh the harms. Screening for malignant melanoma is currently practised in many countries, and the incidence of the disease is rising sharply, while mortality remains largely unchanged.
To assess the effects on morbidity and mortality of screening for malignant melanoma in the general population.
We searched the following databases up to May 2018: the Cochrane Skin Specialised Register, CENTRAL, MEDLINE, Embase, and LILACS. We also searched five trials registries, checked the reference lists of included and other relevant studies for further references to randomised controlled trials (RCTs), used citation tracking (Web of Science) for key articles, and asked trialists about additional studies and study reports.
RCTs, including cluster‐randomised trials, of screening for malignant melanoma compared with no screening, regardless of screening modality or setting, in any type of population and in any age group where people were not suspected of having malignant melanoma. We excluded studies in people with a genetic disposition for malignant melanoma (e.g. familial atypical mole and melanoma syndrome) and studies performed exclusively in people with previous melanomas.
Data collection and analysis
We used standard methodological procedures expected by Cochrane. The primary outcomes of this review were total mortality, overdiagnosis of malignant melanoma, and quality of life/psychosocial consequences.
We included two studies with 64,391 participants. The first study was a randomised trial of an intervention developed to increase the rate of performance of thorough skin self‐examination. The intervention group received instructional materials, including cues and aids, a 14‐minute instruction video, and a brief counselling session, and at three weeks a brief follow‐up telephone call from a health educator, aimed at increasing performance of thorough skin self‐examination. The control group received a diet intervention with similar follow‐up. The trial included 1356 people, who were recruited from 11 primary care practices in the US between 2000 and 2001. Participant mean age was 53.2 years and 41.7% were men. This study did not report on any of our primary outcomes or the following secondary outcomes: mortality specific to malignant melanoma, false‐positive rates (skin biopsies/excisions with benign outcome), or false‐negative rates (malignant melanomas diagnosed between screening rounds and up to one year after the last round). All participants were asked to complete follow‐up telephone interviews at 2, 6, and 12 months after randomisation.
The second study was a pilot study for a cluster‐RCT of population‐based screening for malignant melanoma in Australia. This pilot trial included 63,035 adults aged over 30 years. The three‐year programme involved community education, an education and support component for medical practitioners, and the provision of free skin screening services. The mean age of people attending the skin screening clinics (which were held by primary care physicians in workplaces, community venues, and local hospitals, and included day and evening sessions) was 46.5 years, and 51.5% were men. The study included whole communities, targeting participants over 30 years of age, but information on age and gender of the whole study population was not reported. Study duration was three years (1998 to 2001), and outcomes were measured at the screening clinics during these three years. There was no further follow‐up for any outcomes. The control group received no programme. The ensuing, planned cluster randomised trial in 560,000 adults was never carried out due to lack of funding. At the time of this review, there are no published or unpublished data on our prespecified outcomes available, and no results for mortality outcomes from the pilot study are to be expected.
The risk of bias in these studies was high for performance bias (blinding study personnel and participants) and high or unclear for detection bias (blinding of outcome assessment). Risk of bias in the other domains was either unclear or low. We were unable to assess the certainty of the evidence for our primary outcomes as planned due to lack of data.
Adult general population screening for malignant melanoma is not supported or refuted by current evidence from RCTs. It therefore does not fulfil accepted criteria for implementation of population screening programmes. This review did not investigate the effects of screening people with a history of malignant melanoma or in people with a genetic disposition for malignant melanoma (e.g. familial atypical mole and melanoma syndrome). To determine the benefits and harms of screening for malignant melanoma, a rigorously conducted randomised trial is needed, which assesses overall mortality, overdiagnosis, psychosocial consequences, and resource use.
Minna Johansson, John Brodersen, Peter C Gøtzsche, Karsten Juhl Jørgensen
Plain language summary
Screening for malignant melanoma (a type of skin cancer)
We reviewed the evidence about the effect of screening for malignant melanoma (a type of skin cancer) in people who were not suspected of having this cancer i.e. people with no suspicious mole or lesion (an area of skin with an unusual appearance in comparison with the surrounding skin), compared with no screening. We included any type of screening (e.g. skin self‐examination, or by health professional) of any person not suspected of having malignant melanoma, irrelevant of age or gender. We included studies in people thought to have a high risk of developing malignant melanoma, but not those known to previously have had melanoma.
Malignant melanoma is a skin tumour that can cause death by spreading to other parts of the body; the number of tumours is rising, while in many countries the risk of dying from the disease has not increased in a similar way. Screening for malignant melanoma is performed by visual self‐examination of the skin, or visual inspection by a doctor or other health professional. Screening has the potential to reduce deaths from melanoma. However, there are also potential harms from screening people without symptoms of melanoma, such as finding melanomas that would never have caused symptoms if they had remained undetected (i.e. overdiagnosis), unnecessary surgery, and possible psychological stress. It is important to establish the evidence base for screening.
Two studies met our inclusion criteria. The first study, based in the US, aimed to investigate how to increase the frequency people undertake skin self‐examinations. All 1356 participants were asked to complete follow‐up telephone interviews at 2, 6, and 12 months after randomisation. The average age of participants was 53.2 years; 41.7% were men.
The second study included 18 communities in Australia (63,035 adults) that were assigned to either have a three‐year community‐based melanoma screening programme or not. The study did not report information on the mean age or proportion of men and women in the whole study population, but the average age of those attending the skin screening clinics was 46.5 years and 51.5% were men. The study lasted three years; outcomes were measured at the screening clinics during this time. There was no further follow‐up. The purpose of the study was to investigate whether it was possible to conduct a larger trial, which was stopped by lack of funding.
The first study was funded by the National Cancer Institute (US); the second, by Queensland Cancer Fund and Queensland Health (Australia).
There was no information from either study on the effects of screening on total deaths, overdiagnosis from screening, or participant quality of life. The following outcomes were also not reported: deaths from skin cancer and false‐positive/‐negative rates (i.e. diagnosing a skin lesion as a melanoma when it is not present/not recognising a melanoma when it is present). Thus, we do not know whether screening for malignant melanoma results in any benefit, or whether such a possible benefit would be outweighed by harms of screening. General adult population screening for malignant melanoma is not supported or refuted by evidence from well‐designed trials up to May 2018 and therefore does not fulfil accepted criteria for implementing screening programmes.
Reliability of the evidence
We could not assess the reliability of the evidence for our primary outcomes as they were not assessed.
Minna Johansson, John Brodersen, Peter C Gøtzsche, Karsten Juhl Jørgensen
Implications for practice
Adult general population screening for malignant melanoma is not supported or refuted by current evidence from randomised controlled trials. The intervention therefore does not fulfil current criteria for implementation of population screening programmes (UKNSC 2015; WHO 2008).
We do not have sufficient evidence to determine the effects on morbidity and mortality of screening for malignant melanoma in the general population.
This review did not investigate the effects of screening people with a history of malignant melanoma or those who have a familial predisposition.
Implications for research
To determine the benefits and harms of screening for malignant melanoma, a rigorously conducted randomised trial is needed. As screening effects (both benefits and harms) are generally small at the population level, effects on total and disease‐specific mortality are more likely to be created or erased by bias in a trial than what is commonly the case in trials of medical interventions. A trial would therefore have to be very large and rigorously conducted to allow an assessment of overall mortality, which is the only outcome that incorporates both the possible reduction of disease‐specific mortality and the possible increased mortality arising from harmful effects of screening. Such a trial may not be feasible.
An alternative approach may be to conduct trials of, for example, old, light‐skinned men or people with light skin living in countries with high sun exposure, because these selected population are at higher risk than other populations for developing melanoma.
Since opportunistic screening is already widespread in many countries, a challenge to any trial would be to make sure that the control group is not subject to such screening (i.e. to avoid contamination), since this may dilute both potential benefits and potential harms of screening picked up in the trial. Apart from a potential effect on mortality, as discussed above, other important outcomes to consider in future trials include overdiagnosis, psychosocial consequences, and resource use.
Future trials must ensure they follow the CONSORT guideline for clinical trials, to improve the quality of research, reducing risk of bias, and guide decision making (Moher 2010).
Before implementation of population‐based screening for cancer in asymptomatic citizens, high‐quality evidence from randomised trials showing that benefits outweigh harms is a specified requirement (UKNSC 2015; WHO 2008). The case of screening for malignant melanoma reinforces the importance of this requirement. First, as is apparent from the SCREEN study, non‐randomised studies may lead to seriously misleading results. Second, screening has important harms, such as overdiagnosis and overtreatment of malignant melanomas, and robust trials would need to be performed to quantify them and weigh them against the benefit. Third, the majority of people who take part in the screening programmes cannot benefit from screening as they will never develop the disease. Fourth, screening programmes have a high potential for opportunity costs (Harris 2014). Fifth, when offering screening, healthcare systems invite asymptomatic people to an intervention that they have not asked for, which leads to ethical considerations that differ from those in regular health care (Sackett 2002).