Copyright © 2011 John Wiley & Sons A/S

J Cutan Pathol 2012: 39: 17–20 doi: 10.1111/j.1600-0560.2011.01848.x John Wiley & Sons. Printed in Singapore

Journal of Cutaneous Pathology

Perspectives in Dermatopathology

The melanoma ‘epidemic’: lessons from prostate cancer The rise in malignant melanoma incidence has been termed ‘‘epidemic’’. Closer scrutiny of epidemiologic data suggests overdiagnosis as the true cause of the dramatic rise in melanoma incidence. In epidemiologic terms, ‘‘overdiagnosis’’ describes lesions that are histologically malignant but biologically benign. Overdiagnosis is not unique to melanoma screening but is prevalent in screening for cancers of other organs, including the thyroid and prostate glands. Keywords: epidemic, dysplastic nevus, melanoma, pathology Glusac EJ. The melanoma ‘epidemic’: lessons from prostate cancer. J Cutan Pathol 2012; 39: 17–20. © 2011 John Wiley & Sons A/S.

In this issue, Wolfgang Weyers offers thoughtful insights regarding the melanoma ‘epidemic’ in his article The epidemic of melanoma between under and overdiagnosis.1 This work is, in part, a response to an editorial that I wrote.2 Dr Weyers and I concur that ‘the incredibly increasing incidence of melanoma is incredible’.1 We agree that ‘the rising number of melanomas diagnosed has resulted in increased diagnostic scrutiny, more pigmented lesions being biopsied and more melanomas recognized thus enhancing the ‘‘epidemic’’ in self-perpetuating fashion’.1 We agree that the ‘epidemic, once initiated, held its ground in self-perpetuating fashion’ and that ‘the problem of overdiagnosis is real, and it needs to be addressed’.1 However, Weyers suggests that the melanoma ‘epidemic’ is primarily the result of previous underdiagnosis rather than current overdiagnosis of melanoma. He posits that criteria for melanoma were developed largely in the 1970s, with significant contribution by Ackerman and colleagues, and that this resulted in the correct identification of lesions that are melanoma and had been previously unrecognized.1 Histopathologic history While there are significant publications regarding the histopathology of malignant melanoma in the 1930s (and in every subsequent decade), it is fair to conclude

Earl J. Glusac Department of Pathology and Dermatology, Yale University School of Medicine, New Haven, CT 06510-3206, USA Earl J. Glusac, MD, Department of Pathology and Dermatology, Yale University School of Medicine, New Haven, CT +06510-3206, USA Tel: +203 785 4094 Fax: +203 785 6869 e-mail: [email protected] Accepted for publication November 30, 2011

that criteria for morphologic diagnosis of melanoma (and of other disorders) are a work in progress. Particularly strong progress was made in the 1970s. With decades of experience, criteria tend to be pared, added and weighted. Dr Weyers and I agree that diagnoses are ‘. . . based on the degree of . . . changes and the importance attached to them’,1 Herein lies the rub. The criteria for melanoma have been known for some decades. The issue is how to apply them. Virtually every criterion for malignant melanoma can episodically (and sometimes commonly) be seen in melanocytic nevi. It is clusters of criteria that result in diagnoses, and how one ‘clusters’ them results in accuracy or inaccuracy. It is reassuring to believe that the ‘startling increase in the diagnosis of melanoma in the 1980s and 90s was clearly not caused primarily by overdiagnosis of melanoma but by underdiagnosis previously’.1 How accurate is that assessment? It is likely impossible to know. There is certainly an element of previous underdiagnosis and current overdiagnosis, and I believe Dr Weyers makes a solid point that melanoma was underdiagnosed in the early and mid decades of the 20th century. However, returning to Welch et al.3 upon which I based The melanoma ‘epidemic’ a dermatopathologist’s perspective, it is reasonable to say that the histopathologic criteria for melanoma had been

17

Glusac known by 1986 (the first year of data from that study). In the studied population, the diagnosis of melanoma (mostly early melanoma) increased 2.4fold between 1986 and 2001. The sampling rate increased 2.5-fold during that time period. While it is possible that this increase came about largely because of correct application of newly discovered criteria, this postulate seems to me unlikely. Given the lack of commensurate increase in mortality, those more skilled than we in epidemiologic analysis suggest that overdiagnosis was the cause. Not our fault This does not imply that the melanoma ‘epidemic’ is the fault of histopathologists. Here, clarification of epidemiologic terms is in order. I believe the vast majority of lesions that have contributed to the melanoma ‘epidemic’ are benign lesions (as demonstrated epidemiologically) that most expert pathologists would agree show microscopic features diagnostic of melanoma. This is what epidemiologists term ‘overdiagnosis’. By contrast, we histopathologists often use ‘overdiagnosis’ to describe lesions that are inappropriately called malignant. Epidemiologists employ the frank term ‘misdiagnosis’ for that. Dr Weyers suggests that we confuse ‘diagnosis’ with ‘prognosis’.1 If one believes, as Dr Weyers implies, that any lesion that is morphologically diagnostic of melanoma is melanoma, that argument can be made. That argument assumes, as Dr Weyers states, that morphology is the true gold standard. Perhaps it is fairer to say that morphology is currently the best diagnostic tool we have, but, because of the number of microscopic variables involved and subjective nature of their interpretation, it falls short of gold standard reliability. The prostate monologues What lends even greater credence to ‘overdiagnosis’ of melanoma is the presence of this phenomenon in screening for cancer in several other organs, including the prostate, breast and thyroid. We might take some lessons from prostate cancer. In 2008 an estimated 186,000 men in the United States were diagnosed with prostate cancer, making it the secondmost common cancer diagnosis in males (excluding non-melanoma skin cancer).4 The lifetime risk of death from prostate cancer is approximately 3%. The lifetime risk for a diagnosis of prostate cancer is 16%. However, many more than 16% of men appear to have prostate cancer but do not know it. Studies have showed that the likelihood of identifying prostate cancer rises in direct proportion to how often and how thoroughly the prostate is sampled. Studies

18

show that six screening biopsies resulted in 20–25% rate of identifying carcinoma, while 13 screening cores uncovered a 40% cancer rate.5,6 Another study examined the prostate glands of 525 men who had died in accidents and who had no history of prostate cancer. Even among young men in their 20s, 10% were found to have prostate cancer. Among men in their 70s, more than 80% were found to have prostate cancer.7 Given the 3% lifetime risk of dying of prostate cancer and up to 80% prevalence of prostate carcinoma (defined microscopically) in the elderly, it has been reasonably asked whether most of these lesions are true malignancies. In other words, is what looks like cancer under the microscope really cancer? Studies from thyroid cancer also lend perspective. Like melanoma, the diagnosis of thyroid cancer has risen dramatically as a result of increased sampling. (Thyroid cancer has a low and steady mortality rate. It causes approximately 1600 deaths annually in the United States.)4 An autopsy study from Finland which examined the thyroid glands of 101 consecutive autopsies of older patients concluded that, if we look hard enough, virtually all elderly individuals would be found to have thyroid cancer.8 Like melanoma and the skin, the thyroid gland appears to harbor a large reservoir of lesions that are typically diagnosed as malignant (mostly small papillary carcinomas) but are not biologically malignant. As a result some authorities have recommended against thyroid screening. Seek and ye shall find The problem is even more complicated regarding melanoma, because lesions which can episodically resemble melanoma, namely melanocytic nevi, are arguably the commonest neoplasms in man.9 And while on the order of 99% of nevi are easily diagnosed as nevi, this leaves a large number of lesions that could be overdiagnosed as melanoma (given the commonness of nevi). It is interesting to speculate about what might happen if we subjected our Caucasian population to the (brutal) removal of all pigmented lesions. What incidence of melanoma might be uncovered? I suspect it would be impossibly high. What cancer screening has showed, in the skin and in several other organ systems, is that ‘cancer’ is found in direct proportion to how closely it is sought. The pressures of patient anxieties, medicallegal concerns and financial incentives are aligned in the same direction and appear to push the tide for increased sampling, propelling the melanoma ‘epidemic’. At the same time, ‘underdiagnoses’ are severely punished, and there is no corresponding penalty for ‘overdiagnosis’.10

The melanoma ‘epidemic’ Dr Weyers and I agree that ‘the ultimate proof of diagnosis is recurrence, metastasis and death. In other words, in equivocal cases diagnosis of a nevus can never be proven right, and a diagnosis of melanoma never wrong. This alone is a strong incentive to err on the malignant side, rather than the benign one’.1 This is an important point, and Dr Weyers has stated it better than I could have. It implies an element of fortitude in establishing a benign diagnosis, if one believes that a benign diagnosis is indeed correct. (It also implies that biologic behavior is the gold standard.) Partial biopsies Dr Weyers states that ‘overdiagnoses caused by incomplete biopsies that have been performed with ever increasing frequency since the 1970s may have been an important contributing factor in the so-called ‘‘melanoma epidemic’’’.1 I agree with Dr Weyers that incomplete biopsy is a pervasive and extraordinarily difficult problem. In my experience partial biopsies are more likely to result in ‘underdiagnosis’ than ‘overdiagnosis’, and I regret that partial biopsies have contributed to underdiagnoses I have made and seen others make. The problem It is reasonable to conclude that the apparent melanoma epidemic is a result of previous underdiagnosis and current overdiagnosis of melanoma. While the previous underdiagnosis of melanoma was not a problem, the current overdiagnosis of it is. In A dermatopathologist’s perspective I stated ‘I believe the overdiagnosis of melanoma is arguably the most difficult problem we face in dermatopathology today’.2 Today I differ somewhat with that opinion in that I no longer believe it is arguable. In A dermatopathologist’s perspective, I highlighted the story of patient Chrys Uicker, who received an inaccurate diagnosis of melanoma, which resulted in symptoms of depression, untoward changes in behavior, many minor surgeries and affected plans regarding having additional children. I regret I have personally encountered similar patients who have experienced unnecessary anxiety and difficulty obtaining medical insurance and life insurance, among other problems as a result of their diagnoses.

In melanoma, the distinction between ‘overdiagnosis’ and ‘misdiagnosis’ is not so well-defined, as evidenced by many cases upon which expert histopathologists disagree.11 Data from epidemiologic studies of prevalence and survival of melanoma appear to correlate with trends that experienced dermatopathologists see currently in the histopathologic diagnosis. In A dermatopathologist’s perspective, I stated that an informal poll of trusted colleagues found that they are seeing more ‘overcalls’ than ‘undercalls’ of melanomas by ratios ranging from 3:2 to 4:1.2 (I should add that the vast majority of melanoma cases that they see and that I see are correctly diagnosed.) I did not add (but will now) that the ratio of overcalls to undercalls that I am currently seeing in misdiagnosed cases is ≥10:1. I believe most of these ‘overcalls’ are ‘Clark’s/dysplastic’ nevi that are diagnosed as thin melanomas. It is because personal experience coincides with epidemiologic data that I believe that there is substance to overdiagnosis. What to do? Given the epidemiologic data, it would be reasonable for each of us to look at how we cluster criteria and consider which types of cases, if any, we diagnose as melanoma, but which may not be. Dr Weyers appears to take issue with one of my conclusions: I had stated, ‘A first step toward addressing the apparent false melanoma epidemic involves acknowledging the problem. I do not bring it up because I have the answer to the problem or because I claim not to be a part of it’.2 Alternatively, Dr Weyers advocates ‘establishment of well-defined diagnostic categories, efforts to establish reliable criteria for recognition of those categories, better clinicopathologic correlation, postponement of biopsy of pigmented lesions in the case of irritation and excisional rather than incisional biopsies’.1 These recommendations are excellent. They are not novel. Regrettably, previous calls for them appear largely to have been ignored. What has not been exhaustively attempted is study of the epidemiologic data, comparing it with our practical experience and considering the profound if disheartening conclusion that melanoma is being overdiagnosed. Perhaps, amidst that, means and motivation for effecting change might be uncovered.

References 1. Weyers W. The epidemic of melanoma between under- and over-diagnosis. J Cutan Pathol 2011. 2. Glusac EJ. The melanoma ‘epidemic’, a dermatopathologist’s perspective. J Cutan Pathol 2011; 38: 264.

3. Welch HG, Woloshin S, Schwartz LM. Skin biopsy rates and incidence of melanoma: population based ecological study. BMJ 2005; 331: 481. 4. SEER. URL http://seer.cancer.gov/ [accessed on 23 August 2011]

5. Eskew LA, Bare RL, McCullough DL. Systematic 5 region prostate biopsy is superior to sextant method for diagnosing carcinoma of the prostate. J Urol 1997; 157: 199. 6. Welch HG, Schwartz L, Woloshin S. We look harder for prostate cancer: how screening

19

Glusac made it clear that overdiagnosis exists in cancer. In Welch HG, Schwartz L, Woloshin S, eds. Overdiagnosed: making people sick in the pursuit of health. Boston: Beacon Press, 2011; 45. 7. Sakr WA, Grignon DJ, Haas GP, Heilbrun LK, Pontes JE, Crissman JD. Age and racial distribution of prostatic intraepithelial neoplasia. Eur Urol 1996; 30: 138.

20

8. Harach HR, Franssila KO, Wasenius VM. Occult papillary carcinoma of the thyroid. A ‘‘normal’’ finding in Finland. A systematic autopsy study. Cancer 1985; 56: 531. 9. Glusac EJ. Under the microscope: doctors, lawyers, and melanocytic neoplasms. J Cutan Pathol 2003; 30: 287. 10. Welch HG, Schwartz L, Woloshin S. We look harder for other cancers. In Welch HG,

Schwartz L, Woloshin S, eds. Overdiagnosed: making people sick in the pursuit of health. Boston: Beacon Press, 2011; 61. 11. Farmer ER, Gonin R, Hanna MP. Discordance in the histopathologic diagnosis of melanoma and melanocytic nevi between expert pathologists. Hum Pathol 1996; 27: 528.