Epidemic of melanoma or epidemic of scrutiny?

By Jason B. Lee, MD
May 19, 2021
Vol. 3, No. 20

The goal of population screening is to detect cancers at their early stages to prevent the associated morbidity and mortality of the late stages. The efforts to detect melanoma in its early stages went into full gear over 30 years ago. In the early 80s, the late A. B. Ackerman pronounced that “Malignant melanoma can be diagnosed clinically and histologically when it is small, flat, and confined to the epidermis.” (1) There was well-deliberated promotion of skin cancer screenings and public campaigns to increase melanoma awareness. The widespread adoption of dermatoscopy, a diagnostic technique promoted to detect incipient and incognito melanomas, further contributed to the detection of even earlier melanomas. (2) Cases of melanoma in situ have dramatically increased from 28,600 in 2000 to 101,280 in 2021. (3) All the efforts to detect early melanomas appear to have been highly successful as Stage I disease makes up the majority of the melanomas and the mortality has been stable for the last 40 years.

From an epidemiologic perspective, the sharp rise in the incidence of melanoma in the face of stable mortality for the past 40 years signify the epidemiologic signature of overdiagnosis. (4) For the current trends in melanoma to be valid, there are two possible prevailing explanations. (5) The first explanation is that our specialty was able to identify and cure all aggressive forms of melanoma in their early stages, preventing any rise in mortality in the backdrop of seeing less than 10% of the population at risk each year. This would take incredible luck, winning the lottery kind of luck. The second explanation is that the sharp rise in the incidence of early-stage melanomas represent indolent or biologically benign forms, which are not the obligate precursors to the deadly forms of melanoma, and the stable mortality reflects the stable incidence of the deadly forms of melanoma that have occurred all along, a trend that fulfills the epidemiologic phenomenon of overdiagnosis. (4, 6)

According to Welch and coworkers, the rapid rise in the incidence of melanoma is not due to a true rise in incidence, but it is the byproduct of increased scrutiny, which they refer to as the “epidemic of inspection, surveillance, and biopsy of pigmented skin lesions.” (6) They assert that the rapid rise in the incidence of melanoma is due to the increasing skin cancer screening activities, low threshold to biopsy, and low threshold to diagnose melanoma by dermatopathologists fueled by heightened public awareness of melanoma, financial incentive, and fear of missing melanoma that have resulted in what they refer to as the “cycle of melanoma overdiagnosis.” (6) In order to break this cycle, the authors have broad sweeping recommendations that include curtailing self-referral of skin biopsy specimens, increasing the threshold to biopsy, particularly small pigmented lesions, increasing the histopathological threshold in the diagnosis of melanomas, and, most of all, ceasing all population-based skin cancer screening activities. (6)


These recommendations counter the current teachings and practices of dermatologists and dermatopathologists. The perceived benefits of early detection of skin cancers are ingrained in the minds of dermatologists and dermatopathologists and it is difficult for us to imagine any significant harms that come from it. Some will argue that the benefits of screening outweigh the inherent expected harms of cancer screening. It must be acknowledged, however, that there are no population-level or randomized controlled trials that demonstrate the benefits of melanoma screening. There was initial excitement about the preliminary population data that indicated a decreased melanoma-specific mortality in northern Germany, but the benefit was short-lived and longer follow-up and the subsequent nationwide population screening had no impact on melanoma-specific mortality. (7) Screening usually identifies slow growing indolent melanomas since their early stages are stretched out much longer in time. Screening for melanomas appears to have tapped into a large reservoir of indolent melanomas in the population that are not obligate precursors to the deadly forms of melanoma. In the latest study out of Australia, Guitera and coworkers reported the detection of early melanomas, mostly in situ stage, in a structured surveillance program of high-risk patients. (8) Lacking any data on its impact on mortality, Ferris, in her editorial, questions whether we should continue the path of such labor-intensive surveillance in the battle against melanoma. (9) Over 40 years of incidence and mortality data indicate a broad overdiagnosis of melanoma, a trend that our specialty can no longer overlook.

An overdiagnosis of melanoma in-situ in a 28-year-old has lifelong grave consequences, harms that are highly significant for the patient. Unfortunately, though overdiagnosis is easily discernable at the population level, it is not easily discernible at the patient level as the clinician has no way of knowing which lesion will or will not progress. What can clinicians do to minimize the overdiagnosis of melanoma? What is clear is that less scrutiny and less biopsies are in order, particularly the small flat melanocytic lesions. Drawing from the epidemiology data, clinicians can be reassured that whether the designation is severe atypia or melanoma in situ, the small epidermal lesions do not progress. Furthermore, the concordance rate among pathologists is poor for a range of melanocytic lesions that include dysplastic nevi and atypical nevi of any grade, Spitz nevus, and melanoma in-situ (10) and the histologic diagnosis of small epidermal melanocytic lesions is no longer predictive of the biologic behavior. Another area that deserves less scrutiny is the surveillance of Clark/dysplastic nevi and their removal, for they no longer have a precursor status and they are a large source of overdiagosis. (11) More strategies need to be developed to improve the recognition of the aggressive forms of melanoma in their early stages while minimizing practices that detect only the indolent forms of the disease.

For all the good intentions of screening and early detection of melanoma, the unintended consequence of overdiagnosis is now apparent with an unknown magnitude of harm. Over 25 years ago, Swerlick and Chen cautioned the dermatology community that an increase in scrutiny was responsible for the epidemic of melanoma that had no impact on mortality and, over a decade ago, Glusac cautioned about the swinging of the pendulum toward histopathologic overdiagnosis of melanoma. (11, 12) Their forewarnings, however, have gone unheeded. In this new decade, Welch and coworkers sounded the alarm about the same issues of melanoma diagnosis to a broader audience. For the benefit of our patients, it is now time to pivot away from our current approach to melanoma diagnosis and advance the strategies that will begin to close the gap on overdiagnosis.

Points to Remember: Deadly forms of melanoma are real. Overdiagnosis of melanoma is real. A shift in strategy in melanoma detection is needed to close the gap on overdiagnosis.

Our expert’s viewpoint

Sancy A. Leachman, MD, PhD
John D. Gray Endowed Chair in Melanoma Research
Professor & Chair, Department of Dermatology
Oregon Health & Science University
Chair, AAD Melanoma/Skin Cancer Community Programs Committee

Inconclusive epidemiologic data — and passion for patient protection — has sparked a controversy on overdiagnosis of melanoma, precipitated by a recent paper by H.G. Welch et al. in the New England Journal of Medicine (NEJM). (6) This is good to see. It is a sign of health in the medical profession, despite the strong biases that exist at both ends of the debate. The beauty of the debate is that it tickles our emotions, creates a “call to arms,” and inspires us to speak up. The danger of a polarizing debate is that the cause can become an end in itself, fueled by the adrenalin produced in the process, and often leading to unconscious bias as well as a tendency to over-interpret any data that might be used to support one’s own argument. I would like to reframe the debate in a way that openly acknowledges that there is much we do not know and balances the over-interpretation of non-definitive data.

Let’s begin with the terms “overdiagnosis” and “diagnostic scrutiny.” Both have become charged words and are used in the NEJM paper with a negative connotation that suggests harm to patients. I do not like these terms or the tactic. “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. They suggest a core problem: that we are imposing too much “diagnostic scrutiny,” leading to overdiagnosis. They support this contention by showing that the rising rate of melanoma has occurred in the context of stable mortality. They warn that technological advances, such as molecular tests and artificial intelligence, are “worrisome” and likely to worsen the problem. They suggest that the answer to the overdiagnosis problem is to stop doing skin cancer screening examinations, to encourage pathologists to practice “linguistic de-escalation,” and to stop biopsying pigmented lesions smaller than 6 mm in diameter. Opponents to this point of view claim that stable mortality is a positive consequence of the increased detection of earlier lesions, suggesting that improved “diagnostic scrutiny” has saved lives. These opponents make the point that an excised indolent melanoma does not cause the same harm to an individual as an unnecessary prostatectomy, and that this reality has not been appropriately included in the overall discussion. And neither group acknowledges the necessity of the “pendulum effect” that occurs in all advancing areas of medicine, a trial-and-error method by which we work our way toward the ideal process by learning as we go, incorporating new data, and iteratively improving our course.

Both sides should consider the following: Epidemiological data rarely proves causality. Epidemiological data can be powerful, especially when it is accurate and complete, but it is not possible to know for sure why the trends seen in melanoma have been observed. Good epidemiologic data permits the generation of testable hypotheses. If one hypothesis is unequivocally proven, there is no room for debate. Both sides of the screening controversy have reasonable hypotheses, and there are others, as well, particularly for the increasing rates of melanoma in the absence of increased mortality. For example, the Arizona Cancer Registry reported a doubling of cases in 2012–13 following a campaign to increase reporting by dermatologists in the state. Melanoma deaths were more accurately reported by the registry all along, leading to an apparent disproportionate increase in incidence over mortality. Could some of the rise in incidence over mortality reported nationwide be due to systematic improvement of the reporting infrastructure through the SEER Registry system? Perhaps even more importantly, the hypothesis that overdiagnosis is the only cause of the effect being observed is over-simplified; it is highly unlikely that the cause of the incidence-mortality disparity has a single explanation. As Grossman et al. reported in a rebuttal to the Welch paper, there is clear data from the Utah Cancer Registry showing that some part of the increase in incidence cannot be attributable to overdiagnosis alone. (13) Indeed, nearly every trend observed in large human populations is more complex than it seems, so isn’t it most likely that the overall picture we are seeing is a combination of several effects? In sum, both sides are probably partially right and partially wrong; our job is to keep an open mind and tease out the closest version of the ground truth.

Given this lack of complete understanding, it is concerning that the solution proposed by Welch et al. is to return to the “old ways,” rather than embrace new science. They suggest that we revert to not biopsying lesions less than 6 mm, rather than making an effort to better understand the biology (and clinical signs) that makes some 2 mm lesions deadly. Shouldn’t our call-to-action be to improve, rather than to decrease, our diagnostic scrutiny? Wouldn’t it be better to utilize the COVID “experiment of nature” to evaluate the true (data-based) impact of decreased melanoma screening, rather than keep all screening programs closed (as suggested by Welch et al.)? Should the SPOT Skin Cancer™ screening program be halted completely, as they suggest, or should we purposefully stratify risk and screen those with highest need? In Oregon, we are supporting a statewide risk-stratification research project, the WarOnMelanoma™, that is designed to test the hypothesis that melanoma screening will reduce death due to melanoma. This will soon provide prospective, controlled data for the discussion.

Finally, physicians have a professional obligation not only to care for the individual patient, but also to advocate for progress. Approximately 1–2% of thin melanomas are lethal. Is it appropriate to decline to biopsy something that might kill the patient, particularly if monitoring is not an option? Shouldn’t the provider at least offer to follow the lesion carefully if a biopsy is not done, or teach the patient to follow it themselves? Shouldn’t the provider utilize everything in their diagnostic armamentarium (like specialized imaging or molecular testing) to improve the objectivity of the diagnosis? To fail to recognize the need for continuous progress dooms us to remaining where we are: unable to accurately diagnose all lethal melanomas. What we are experiencing now is a swing of the screening pendulum, produced by the recognition that we may be over-screening those at low risk; we shouldn’t wildly overshoot the center, though, by halting screening for those that will benefit. We have a social responsibility both to prevent harm to our patients and promote improvements and advances that will help the next generation. Balancing the risks of potential harms from an action (e.g., screening) with risks of harms from lost opportunity (e.g., no improved screening) is an important part of our professional contract with society.

Thanks to everyone who has engaged in this debate. Good things are coming out of it. As a field, we are talking about when and how to limit self-referral of pathology, about which individuals are at sufficient risk to warrant regular screening, and about the inherent subjectivity of a pathologic diagnosis. We are also discussing the difference between a physician’s obligation to the individual and to society. We need to keep up the good work; it will result in better outcomes for our patients over time.

Editor’s note: Dr. Leachman serves as chair of the AAD’s Melanoma/Skin Cancer Community Programs Committee, but the opinions expressed herein are her own.

Our editor’s viewpoint

Warren R. Heymann, MD

Drs. Lee and Leachman raise many important considerations regarding trends in over-diagnosing early melanomas.

My viewpoint on this topic is based on my experience as a melanoma patient, a dermatologist, and a dermatopathologist.

My melanoma was thin (0.3 mm Breslow thickness). I also have some atypical nevi. I really do not wish to have biopsies performed on these lesions unless my dermatologist is worried.

As a dermatologist, I am happy to follow patients clinically (with photography and dermoscopy), biopsying only those lesions that I find disconcerting, or those lesions that patients are convinced have changed significantly. Once patients have been diagnosed with melanoma, however, even for in situ lesions, many feel very differently; some people scrutinize and obsess over every pigmented lesion, insisting that those lesions be removed. Even when I tell patients that I had melanoma myself and reiterate my personal approach, some are very insistent on having those lesions excised. Honestly, sometimes it is just easier to acquiesce and just do the procedure. (I’m sure that there are those who do this for financial gain, but I hope that is the proverbial “bad apple” group that is found in any occupation.)

As a dermatopathologist, if there is some atypia in nevi, of course I mention this on the biopsy report. Despite the multiple recent papers concluding that re-excisions of minimal or moderately atypical nevi that extend to the margins may be followed clinically, there are clinicians who will routinely re-excise these lesions.

At the root of this cycle is fear. Patients hear the word melanoma and think it is a death sentence. Clinicians fear malpractice litigation. Until there is improved education regarding overdiagnosis of early melanoma in concert with tort reform, it is hard to envision this relentless cycle of worry, biopsy, tendency to overcall melanoma histologically, and re-excision abating. Perhaps molecular studies and artificial intelligence will lead the way.

It must also be recognized that health care discrepancies abound in melanoma — in socioeconomically stressed populations of color, melanoma is underrecognized and has a worse prognosis. (14) As Cortez et al state: “Disparities in melanoma care exist in the United States. Disparities in provider type, patient demographics, place of residence, insurance status, socioeconomic status (SES), race/ethnicity, and age impact melanoma outcomes. Melanomas detected by dermatologists are thinner, at an earlier stage, and have better survival outcomes, compared with detection by primary care providers or patients. Lower SES, race/ethnicity, and place of residence are associated with decreased access to and/or utilization of dermatologists, and more advanced melanomas at diagnosis. Additionally, uninsured and publicly insured individuals are more likely to present with late-stage melanomas, resulting in worse outcomes.” (15)

As a society, when it comes to melanoma screening and improving outcomes, we have a lot of work to do.

References:

1. Ackerman AB. Macular and patch lesions of malignant melanoma: malignant melanoma in situ. J Dermatol Surg Oncol 1983;9:615-8.

2. Kittler H, Guitera P, Riedl E, et al. Identification of clinically featureless incipient melanoma using sequential dermoscopy imaging. Arch Dermatol. 2006;142(9):1113-1119.

3. American Cancer Society. Cancer facts and figures. Accessed April 3, 2021. https://www.cancer.org/research/cancer-facts-statistics/all-cancer-facts-figures.html

4. Welch HG, Kramer BS, Black WC. Epidemiologic Signatures in Cancer. N Engl J Med. 2019;381(14):1378-1386.

5. McHargue C. The melanoma epidemic: Real or Imaginary? Accessed April 3, 2021. https://www.youtube.com/watch?v=Y8teF9SWspA&t=16s

6. Welch HG, Mazer BL, Adamson AS. The Rapid Rise in Cutaneous Melanoma Diagnoses. N Engl J Med. 2021;384(1):72-79.

7. Boniol M, Autier P, Gandini S. Melanoma mortality following skin cancer screening in Germany. BMJ Open. 2015;5(9):e008158.

8. Elmore JG, Barnhill RL, Elder DE, et al. Pathologists' diagnosis of invasive melanoma and melanocytic proliferations: observer accuracy and reproducibility study. BMJ. 2017;357:j2813.

9. Guitera P, Menzies SW, Coates E, et al. Efficiency of Detecting New Primary Melanoma Among Individuals Treated in a High-risk Clinic for Skin Surveillance [published online ahead of print, 2021 Mar 17]. JAMA Dermatol. 2021;10.1001/jamadermatol.2020.5651.

10. Ferris LK. Early Detection of Melanoma: Rethinking the Outcomes That Matter [published online ahead of print, 2021 Mar 17]. JAMA Dermatol. 2021;10.1001/jamadermatol.2020.5650.

11. Glusac EJ. The melanoma 'epidemic', a dermatopathologist's perspective. J Cutan Pathol. 2011;38(3):264-267.

12. Swerlick RA, Chen S. The melanoma epidemic. Is increased surveillance the solution or the problem?. Arch Dermatol. 1996;132(8):881-884.

13. Grossman D, Sweeney C, Doherty JA. The Rapid Rise in Cutaneous Melanoma Diagnoses. N Engl J Med. 2021 Apr 8;384(14):e54. doi: 10.1056/NEJMc2101980.

14. Heymann WR: Melanoma Healthcare and Outcome Disparities: You Can Never Be Too Rich or Too Thin. J Am Acad Dermatol 2021 Apr 3;S0190-9622(21)00679-4. doi: 10.1016/j.jaad.2021.03.107. Online ahead of print.

15. Cortez JL, Vasquez J, Wei ML. The Impact of Demographics, Socioeconomics, and Healthcare Access on Melanoma Outcomes. J Am Acad Dermatol. 2020 Aug 9:S0190-9622(20)32405-1. doi: 10.1016/j.jaad.2020.07.125. Epub ahead of print. PMID: 32783908.

---