Melanoma controversies

By Jan Bowers, contributing writer, September 1, 2021

At first glance, dermatologists would seem to be wildly successful in their decades-long effort to detect and remove cutaneous melanoma lesions before they can metastasize and become deadly. According to the American Cancer Society’s (ACS) Facts & Figures, the incidence rate of invasive melanoma has increased 2% every year from 2008 to 2017, and melanoma in situ has increased from 28,600 cases in 2000 to a predicted 101,280 in 2021. Despite the explosion in incidence, the mortality rate for melanoma remained remarkably steady until 2014, when, according to the ACS, advances in treatment led to a drop in the death rate of almost 7% per year between 2014 and 2018.

What factors explain the discrepancy between the incidence and mortality trends? Earlier this year, an article in the Sounding Board section of the New England Journal of Medicine (2021; 384(1): 73-9) reignited an ongoing controversy when the authors (one dermatologist, two non-dermatologist physicians) argued that “increased diagnostic scrutiny (i.e., the combined effect of more screening skin examinations, falling clinical thresholds to biopsy pigmented lesions, and falling pathological thresholds to label the morphologic changes as cancer) is the primary driver of the rapid rise in melanoma diagnoses.”

Three dermatologists weighed in on the issue in a recent DermWorld Insights and Inquiries article, “Epidemic of melanoma or epidemic of scrutiny?” One contributor, Sancy A. Leachman, MD, PhD, FAAD, professor and chair of the department of dermatology at Oregon Health & Science University, countered the NEJM authors’ perspective, insisting that “‘overdiagnosis,’ as used by Welch et al, is a term that should convey our current limitations in identifying a biologically lethal melanoma rather than suggest intentional over-calling of melanomas.”

DermWorld explores the possible reasons for the discrepancy; the value of screening and surveillance; what should and should not be labeled as melanoma; the potential harm to patients in misdiagnosing or overdiagnosing melanoma; and what a different approach to screening and diagnosis could look like. 

Early detection of dangerous malignancy or “overdiagnosis”?

The growing disparity between melanoma incidence and mortality was evident as early as 25 years ago, when data from the Surveillance, Epidemiology, and End Results (SEER) program caught the attention of Robert A. Swerlick, MD, FAAD, during a lunchroom conversation with a colleague. “I looked at that chart and said, ‘something doesn’t add up.’ I’m just a curious person, and started reading,” said Dr. Swerlick, who is professor and chair of dermatology at Emory University School of Medicine. His conclusion: the discovery and removal of more early-stage melanomas, a result of intensified screening, is not the underlying cause of the steady mortality rate. “The math is very implausible, because we’ve been able to run this scenario for probably close to 40 or 50 years. And as we increase the surveillance intensity, invariably the diagnosis rate goes up. The death rate remains absolutely level. We can’t surveil the entire population, so we would anticipate that there would be an increase in deaths associated with those who aren’t coming to us for surveillance. But that isn’t happening.” 

Concurring with Dr. Swerlick, Jason B. Lee, MD, FAAD, first allowed that “it makes plenty of sense that if we’re catching early melanomas, that is going to prevent the mortality from increasing.” But he pointed out that physicians screen only an estimated 5-7% of the population at increased risk of developing melanoma. “That means the more than 90% that we’re not seeing should eventually develop late-stage disease, and they’re not. So, whatever we’re diagnosing today, they’re the very early stages, in situ and stage one diseases.” Dr. Lee, who is professor of dermatology and director of the dermatopathology center at Jefferson University, attributes the discrepancy to “overdiagnosis, meaning that dermatopathologists are looking at lesions that, under the microscope, fulfill the criteria of malignancy, and therefore qualify as malignancies, but the vast majority don’t behave in a malignant manner; they are indolent.”

An early proponent of skin self-examination coupled with physician skin exams, Darrell S. Rigel, MD, FAAD, strongly disputed the notion that the rising incidence of melanoma can be attributed to aggressive screening. Dr. Rigel, who is clinical professor of dermatology at NYU Grossman School of Medicine and a former AAD president, compared melanoma screening with screening for prostate cancer using the PSA blood test. “As soon as they introduced the PSA test in the mid-1980s, the prostate cancer rate skyrocketed within two to three years; then, over time, it came back down to baseline.” In contrast, when melanoma screening was introduced around the same time, “there was no spike whatsoever. The cases haven’t come back down. The line has been constantly increasing.” In addition, during the period 2006-2016, “both the absolute number of deaths from melanoma and the survival percentage were increasing. The only way that could mathematically be true is the real incidence has to be rising.”

What screening reveals

While there are few if any randomized controlled trials testing the effectiveness of screening in reducing melanoma morbidity and mortality, researchers have examined the methods and results of screening to evaluate its value. One study analyzed the results of a skin cancer screening initiative conducted at the University of Pittsburgh Medical Center (JAMA Oncol. 2017;3(8):1112-15) in which 333,788 eligible patients were offered melanoma screening during an office visit with a PCP. The melanoma incidence rate among the screened patients was more than double that of the unscreened patients, and the melanomas detected in screened patients were thinner than those in unscreened patients. The authors called the greater preponderance of thin melanomas among the screened patients “a clinically significant finding.” 

One paper from Queensland, Australia, published in the Journal of Investigative Dermatology noted that “more people actually die from melanomas thinner than 1 mm than from those thicker than 4 mm” due to the higher incidence of thin melanomas in the population (2015;135(4):1190-1193). Utilizing data from the Queensland Cancer Registry (QCR), the authors calculated age-standardized incidence and mortality rates for each year for all melanomas and by thickness of the first primary, and found that thin melanomas (≤1 mm) accounted for 19% of melanoma deaths overall (68% of all melanomas), but increased from 14% in 1990–1994 to 23% in 2005-2009.

The authors of a Cutis review of five observational studies (2015;96:175-82), citing a lack of consensus regarding screening guidelines, sought to clarify the value and impact of physician skin examinations on tumor thickness at diagnosis and melanoma mortality. Overall, the authors said, “Our findings suggest that physician skin exams are associated with a decline in melanoma tumor thickness and melanoma-specific mortality.” One of the studies included in their analysis described a population-based screening project in northern Germany (known as SCREEN) that enrolled 360,288 patients (19% of eligible citizens) (J Am Acad Dermatol. 2012 Feb;66(2):201-11). The Cutis authors noted that the SCREEN program was associated with a 47% and 49% reduction in melanoma mortality for men and women, respectively, five years after the program was implemented in 2003, while mortality rates in surrounding areas remained steady. 

While the Cutis authors point to the SCREEN study as providing “the greatest evidence for population-based screening programs,” the strength of that evidence has come into question in the ensuing years, due to factors like the lack of a randomized control group and the use of regional data for the evaluation of melanoma mortality. According to a study published in BMJ Open, melanoma mortality rates returned to levels observed before screening initiation from 2009 to 2013. In addition, the expansion of the program to the national level has not yielded a reduction in mortality commensurate with that of the regional program. 

While population screening is often discussed in the context of lighter-skinned individuals, who are at increased risk of developing melanoma, Dr. Rigel pointed out that self-examination and screening may be lifesaving measures for people of color. “Melanoma has the largest racial disparity in terms of prognosis and survival versus any other cancer,” he remarked. “If you’re Black, your chance of getting melanoma is 1/25th that of a Caucasian, so the index of suspicion is not high in this group. Thus, by the time it’s diagnosed, it’s often more advanced.”

What is a potentially lethal melanoma?

There’s broad agreement that intensive screening and surveillance allow physicians to detect small, thin melanomas. But what constitutes a melanoma, and whether the smallest in situ melanomas truly pose a threat to patients, are questions still being argued. “Epidemiologists have been generous in thinking that what we as dermpaths view under the microscope all fulfill the diagnosis of melanoma,” said Dr. Lee. “What’s not clear is how many of these are actually misdiagnosis; I think that percentage is high. But nobody knows because no one reviews anyone’s slides.” 

Dr. Lee cited a study (BMJ. 2017;357:j2813) that sought to quantify the accuracy and reproducibility of pathologists’ diagnoses of melanocytic skin lesions; the study authors concluded that “diagnoses of melanoma in situ and early stage invasive melanoma, together are more common than all other stages of melanoma combined, were neither reproducible nor accurate.” Dr. Lee said he was “shocked by that article because it basically said that what we’re doing for these thin lesions with our diagnosis and our method is invalid.”

In the NEJM article, Welch et al noted that the incidence of melanoma in situ (MIS) “is now 50 times as high as it was in 1975,” adding that the increasing detection and treatment of MIS has failed to reduce the incidence of invasive melanoma, which more than tripled in the same period. In Dr. Lee’s opinion, “these small (less than one centimeter) lesions that are very discrete, I don’t think they’re a cause of patients having a bad outcome.” He alluded to the preponderance of successful melanoma lawsuits and remarked that “we don’t know the contexts and situations where we’re missing that melanoma, but looking at the epidemiology data, it’s not because we’re not biopsying and removing these small, flat, thin melanomas. I think these lesions stay at the in situ stage for a long time, and epidemiologic data tells us that the vast majority don’t really progress. Figuring out which of these will progress should be one of our priorities.” 

Dr. Leachman recommends removing all MIS lesions because “we don’t know which are slow-growing and indolent. There are multiple genetic mutations that can enable metastasis in some in situ melanomas, rare as it might be. If you catch an in situ melanoma, it will probably never attain the capacity to invade; but there’s a small indistinguishable subset that will. Unfortunately, the tools we currently have, like immunohistochemistry, aren’t good enough to distinguish them. No one can know whether in situ melanoma cells have the ability to metastasize or not, but the patient deserves to know what we know, along with the probabilities.” Dr. Rigel agreed that there’s no such thing as a “safe melanoma. If I had an in situ melanoma on me, I’m not going to leave it on and say ‘yeah, it’s probably going to be okay, I’ll leave it.’ There’s no data to support that.”

But dermatologists need to consider the consequences of labeling a small lesion as melanoma, Dr. Swerlick said. “When you tell people they have melanoma, the first thing that comes to their mind is, ‘am I going to die?’ So, if what you’re telling them may or may not be true, that’s not a good thing. If it initiates a cascade of interventions which affect them for the rest of their lives, where there’s little or no data that they derive any benefit whatsoever, that’s not good.”

The way forward

In a perfect world, dermatologists would screen and surveil the right patients at the right intervals, remove only those lesions with lethal potential, and never get sued for malpractice. But given the world in which they practice, with a lack of hard data to guide many of their decisions, the dermatologists described their vision for a rational approach to detecting and diagnosing melanoma.

Drs. Leachman and Lee advocate stratified screening rather than broad-based population screening. “We should acknowledge that it’s probably a waste of precious resources to recommend that every person in the general population get an exam by a dermatologist. That’s physically impossible anyway,” said Dr. Leachman. “In Oregon, we recommend that the general population do self-screening, and we provide materials and education and facilitate self-exams. People with high-risk features probably do need an annual exam by a physician. But if you don’t have any history of melanoma, haven’t had a lot of sun overexposure, you probably don’t need a dermatologist to do that. However, if you’ve had a melanoma, or a strong family history, then you definitely want to have a dermatologist or a melanoma expert looking over you regularly.” 

Dr. Leachman and several co-authors published a set of data-driven screening guidelines (Melanoma Manag. 2017;4(1):13-37) for adults aged 35-75 that lays out specific risk characteristics for personal history, family history, physical features and UVR overexposure. Dr. Lee also urged a focus on high-risk people, which he defined as patients who are “immunosuppressed, have white skin, are 60 and over, and those who are socioeconomically disadvantaged. They don’t seek help, and they’re the ones who come in with late-stage melanomas.” The NEJM authors recommend the elimination of most screening, reserving it “for only those groups at highest risk.” In Dr. Leachman’s view, this approach “is too prohibitive and algorithmic, taking away the physician’s ability to personalize screening on an individualized basis. There will be some important unintended consequences if we choose to follow their recommendations, namely a shift from the role of the physician as an individual patient advocate and medical innovator, to a gatekeeper restricted by the need for societal cost reduction. Insurance companies, payers, and plenty of government agencies have societal cost:benefits in mind, but patients only have their doctors to advocate for them. These recommendations teeter on an incredibly slippery slope that could alter the doctor-patient relationship; it’s worth deep contemplation.” Self- and partner-examination, “somebody looking at your back and the back of your legs, is probably more important than you coming in every year,” said Dr. Lee. “Maybe too many people will come in with things that are benign, I don’t know. But I do promote that.” Dr. Rigel also emphasized the importance of self-examination, and said he urges his patients to “learn your own skin. If you see something that’s changing, you’re the first person to know. It’s a team effort — I’m there to help, but you see your skin every day.”

Rethinking the diagnostic criteria and nomenclature around small, thin melanomas might reduce overdiagnosis and too-aggressive treatment, said Dr. Lee. “Let’s have a little bit higher threshold to biopsy and call things melanoma, particularly these thin, very small lesions. And let’s strive for better communication between the dermatologist and dermatopathologist to explain to them the context of the lesion.” Dermatopathologists, he added, “need to own up, if the gold standard is histopathology. There’s just too much hedging and too much overcall. They can’t call everything dysplastic and atypical, and call things melanoma. When it comes to melanocytic lesions, it’s easy to call things melanoma; it’s hard to call things benign.” Dr. Swerlick agreed that because there is ambiguity among pathologists when it comes to diagnosing melanoma, “it is my thought to reclassify [thin, small lesions] as not necessarily disease, but as risk factors for bad outcomes. That reframes the discussion, which then becomes: how likely is that to happen? You can probably discern that likelihood from outcomes data from early melanoma, which would suggest that if it’s in situ, it virtually never goes on to develop metastasis.”

Dr. Leachman sounded a note of caution about moving in this direction. “How do we change the nomenclature to better reflect the low probability of death from the disease without risking development of a false sense of security that could lead to failure to follow up on a suspicious change? The key is for providers to take the time to clearly explain what the risk is and what it means to the patient, which can be difficult in fast-paced practices. Changing the nomenclature won’t address the underlying issue of patients’ misunderstanding of risk levels.” Physicians need to acknowledge that “the ability to essentially separate the world of cancer or not cancer is an impossible task,” said Dr. Swerlick. “We like very simple bright lines and binaries, but the world is shades of gray, and this is a perfect example of that. We ultimately need to adapt to a world where there’s greater uncertainty — not just physicians, but our patients as well.”

Additional DermWorld Resources