Pitfalls in the Diagnosis of Malignant Melanoma: Findings of a Risk Management Panel Study
The Doctors Company (TDC) is the largest physician-owned professional liability insurance company in the United States and insures more than 1,100 pathologists who practice in every state. Therefore, its claims should be representative of problems occurring in many pathology practices. In the process of reviewing pathology claims the author observed that the misdiagnosis of melanoma was a repetitive problem. This impression was reinforced by a computer search of all claims reported to TDC from 1990 through 2001, which revealed that 8.6% of pathology claims and 14.2% of dermatology claims involved the key words “skin cancer” and/or “melanoma”. This finding led to a detailed analysis of individual pathology claims reported from 1995 through 2001.
Materials and Methods:
The author reviewed 218 surgical pathology and fine needle aspiration biopsy claims reported to The Doctors Company from 1995 through 1997. 11% of these claims (#23) involved the misdiagnosis of melanoma; 70% were false negative diagnoses. Melanoma claims were second only to claims involving breast biopsy. Subsequently an additional 144 surgical pathology and cytology claims from 1998 through 2001 were reviewed. 16% of these claims (#23) involved misdiagnosed melanoma.
A Melanoma Risk Management Panel of expert dermatopathologists was convened in April 2002 at The Doctors Company home office in Napa, California. The purpose was to devise useful strategies that pathologists and dermatologists could use in their practices to reduce the risk of diagnostic error and/or patient mismanagement when dealing with melanocytic lesions – focusing on the recurrent “problem areas” identified by the author in the 23 melanoma claims reviewed from 1998 through 2001.
Ten recurrent problems were identified:
- Nodular melanoma misdiagnosed as melanocytic nevus without explanation
- Nodular melanoma with the low power architecture of melanocytic nevus, but the cytologic features of melanoma (Nevoid Melanoma)
- Claims involving partial biopsies (shave or punch biopsies)
- Superficial spreading melanoma misinterpreted as chronically inflamed nevus
- Melanoma misdiagnosed as Spitz nevi
- Unrecognized desmoplastic melanoma
- Melanoma presenting as a lymph node metastasis, misdiagnosed as lymphoma
- Melanoma misdiagnosed as “dysplastic nevus”
- Spindle cell melanoma misdiagnosed as spindle cell squamous carcinoma
- Patients presenting with metastatic melanoma without a known primary - and a history of having a skin lesion removed and presumably discarded without microscopic examination
The misdiagnosis of melanoma continues to be a major cause of malpractice claims filed against pathologists and dermatologists. While specific problem lesions were identified (nevoid melanoma, Spitz nevus, desmoplastic melanoma, and dysplastic nevus) many claims involve "classic" melanoma misdiagnosed as melanocytic nevus without an identifiable explanation. However, the reproducible diagnosis of melanoma appears to be a problem even for experienced dermatopathologists. A panel of 11 experts reviewing 37 classic melanocytic tumors agreed unanimously on the diagnosis of benign versus malignant on only 35% of cases1,2. When 6 university pathologists reviewed 94 melanoma slides, their agreement was good for Breslow thickness, moderate for growth phase, and fair for Clark level3. A discussion of the ten problem areas identified by the panel follows. The recommended guidelines are an attempt to identify strategies that may be helpful in reducing the risk of diagnostic error and enhancing patient management. They are not rules. Only the physician who is knowledgeable of the circumstances prevailing in each individual situation can make the ultimate decision regarding their appropriate application.
- Nodular melanomas misdiagnosed as melanocytic nevi without an obvious explanation (4 claims). Some managed care contracts restrict referrals and may not allow dermatologists to send specimens to their usual pathologist or dermatopathologist, or to read their own cases. Some plans mandate referral to large central laboratories or contractors offering deeply discounted prices. These labs may employ their own pathologists or they may subcontract with pathology groups or individual dermatopathologists. Pathologists employed in these labs may be reading too many slides too quickly due to the productivity standards required by the discounted charges and by the emphasis on rapid turn-around time. Multiple block levels and immunohistochemical stains may be sometimes discouraged due to cost. There may be minimal or no clinical information submitted with these specimens and the pathologist often doesn’t know the referring physician or can’t reach them or doesn’t have time to do so. This may lead to different levels of practice in the same practice setting depending upon whether a case is deeply discounted under contract or billed directly.
Pathologists are encouraged to personally contact referring physicians in order to discuss the importance of providing clinical information (age, sex, site, working diagnosis, whether this is a new or recurrent lesion, and lesional diameter if a shave or punch biopsy). In many offices, the critical contact is with the nurse or other member of the office staff who fills out the pathology requisition. The value of complete biopsies vs. partial biopsies that incompletely sample a lesion and the importance of completely re-excising a partially biopsied lesion when recommended should be stressed.
Pathologists should also focus on those factors which are under their control, ie, thorough sampling of a lesion, looking at multiple levels through the block, insisting on good histologic sections, and not doing frozen sections on melanocytic lesions. Adequate sampling is important since diagnostic features may be seen in only one or two sections taken from one of multiple levels. A recommended approach is to serially section the entire biopsy and examine multiple sections from each of three levels. The intervening unstained sections are saved and can be subsequently stained if needed or used for immunohistochemical studies.
- Nodular melanomas with the low power architecture of melanocytic nevus, but the cytologic features of melanoma: Nevoid Melanoma (3 claims) – suggesting that these lesions were examined only under low power and not assessed cytologically at higher magnification. Pathologists should routinely examine all melanocytic lesions under high power – even if they appear to be typical nevi under low power. Look for mitoses (frequent and deep), cytologic atypia, coalescence of nests (loss of nesting pattern), melanin in deep tumor cells, and lymphohistiocytic infiltrates. Relying on low power symmetry can be misleading.
- Claims involving partial biopsies (shave or punch biopsies): Of the 23 pathology claims reviewed from 1998 through 2001 involving misdiagnosed melanoma, 30% were shave biopsies (7 claims); 26% were punch biopsies (6 claims); 26% were melanomas that recurred locally (6 claims) and had therefore been incompletely excised (type of biopsy unknown); and only 17% (4 claims) involved melanomas adequately excised. Overall 56% of claims (13 claims) involved punch or shave biopsies misdiagnosed as benign nevi and incompletely excised (many were superficial spreading melanomas) and 83% of all claims involved melanomas that were incompletely excised.
Excisional biopsy is recommended for all melanocytic lesions because a partial biopsy may result in a partial diagnosis, which may be a misdiagnosis. Nevertheless, despite admonitions to completely excise melanocytic lesions, primary care clinicians and dermatologists continue to perform punch and shave biopsies - and this practice appears to be increasing under “managed care”. These “partial biopsies” may sample non-diagnostic areas of a melanoma or an associated benign nevus; they don't include the deepest portion of the lesion or the diagnostically helpful radial growth phase; and unlike excisional biopsy specimens there is a residual lesion that may recur and/or metastasize.
Punch and shave biopsies are often driven by cost considerations: scalpels cost more than razor blades, sutures are expensive, and a return visit is often required if a lesion is completely removed and sutured. However, these “savings” are trivial compared with the cost of a malpractice claim. If referring physicians continue to send partial biopsies, it is important for the pathologist to know the lesional diameter on the patient since this allows an estimate of the extent to which the biopsy sampled the lesion. Pathologists should encourage referring physicians to provide this information on all partial biopsies.
It is important to comment on the presence of nevocytes at margins – even for benign lesions AND to ask for re-excision of “problem” lesions (lesions showing some but not all the features required for a diagnosis of melanoma): a 0.5cm margin for superficial lesions and a 1.0cm margin for deep lesions is generally appropriate. A partial skin biopsy is analogous to a colonoscopic biopsy of a dysplastic adenomatous polyp, ie, if the biopsy is not excisional, the entire lesion is always removed. Suggested comments to add to the pathology report to encourage re-excision of partial biopsies include:
- When there is little concern:
“one might consider re-excision”
“if this biopsy is part of a larger clinically suspicious lesion, complete excision is advised”
- When there is major concern:
“re-excision of this lesion is mandatory”
“re-excision of this lesion is warranted”
“this is a partial biopsy of a larger lesion, which should prompt complete excision of the entire lesion”
Such comments will often precipitate a phone call from the referring physician or a request for consultation, which is apt to assure appropriate follow-up.
- Superficial spreading melanomas misinterpreted as nevi showing chronic inflammation (3 claims). These cases suggest that melanocytic lesions containing numerous lymphocytes should be more carefully examined (especially under high power) for other features suggestive of melanoma - and completely excised. Among dermatopathologists, there is an adage that “lymphocytes are better than pathologists in recognizing melanomas”
- Melanomas misdiagnosed as Spitz nevi (4 claims). The diagnosis of Spitz nevus in an adult is a “high risk” and difficult diagnosis for most pathologists. Several areas of particular risk involving the diagnosis of adult Spitz nevus include:
- lesions composed predominantly of epithelioid melanocytes
- large lesions (over 6mm in diameter)
- lesions on sun damaged
- broadly transected lesions
- lesions in which more than one population of melanocytes is present
When considering this diagnosis in an adult it is advisable and prudent to obtain consultation from an expert dermatopathologist. Spitz nevus is a high risk/low frequency diagnosis – analogous to soft tissue tumors or bone tumors. If a pathologist is not seeing Spitz nevi on a regular basis, and the patient is > 20 years old, the case should be sent to an expert. If Spitz nevi are being seen on a regular basis and the patient is > 20 years old, unless all of the typical diagnostic criteria are present send the case to an expert. All Spitz nevi should be completely excised – even if the patient is < 20 years old. Even experts may disagree on problematic or atypical Spitzoid lesions; when this occurs some referral centers may perform sentinel lymph node biopsy and in one published report metastatic cells were found in the sentinel node in 5 of 10 cases4.
- Unrecognized desmoplastic melanoma (3 claims); over half of these were shave biopsies (often misdiagnosed microscopically as dermatofibroma). Most desmoplastic melanomas are associated with lentigo maligna or atypical lentiginous proliferations. At times, a pigmented actinic keratosis can be difficult to differentiate from lentigo maligna. In the presence of an actinic keratosis it is always prudent to assess whether lentigo maligna may coexist. In sun-damaged skin there is often an increased number of single basally located enlarged melanocytes that may be mildly atypical. In the absence of substantial cytologic atypia, formation of junctional nests and extension down the hair follicles, this is insufficient to justify a diagnosis of lentigo maligna. If the microscopic appearances are less than clear cut, suggest in the report that the clinician review the clinical aspects of the case to exclude lentigo maligna and undertake additional sampling if indicated.
The diagnosis of lentigo maligna is indicated when, in atrophic skin with solar elastosis, there is an increased frequency of atypical melanocytes, singly and in nests in the basal layer of the interfollicular epidermis with extension of the melanocytic proliferation into the skin appendages. While the melanocytes are largely present in the basal layer, some may be crowded into the suprabasal epidermis; full thickness migration of single cells and nests is not seen. The papillary dermis is usually widened by fibrosis and lymphohistiocytic infiltrates, often with some pigment incontinence. Atypical melanocytes often extend beyond the clinically visable lesion and surgical margins often transect skin that contains atypical melanocytes. Pathologists need to carefully evaluate margins, which often requires use of immunohistochemical stains for MART-1/Melan-A.
While invasive melanoma arising from lentigo maligna most often takes the form of a vertical growth phase “nodule”, some are desmoplastic melanomas. These are comprised of spindle-shaped cells that resemble myofibroblasts and are associated with abundant fibrous stroma that may resemble a scar. Their appearance is often deceptively subtle and the only protection against missing them is to routinely look for desmoplastic melanoma in all biopsies of lentigo maligna. A useful rule is: Actinic damaged skin + Scar or Desmoplasia = rule out Desmoplastic Melanoma. This often requires immunohistochemical stains for S100 protein; HMB45 has a low sensitivity for desmoplastic melanoma and is not useful.
Desmoplastic melanoma is often present in and around cutaneous nerves (neurotropism) and can extend for considerable distances peripherally via this route. Surgical margin evaluation requires margin-parallel “en-face” blocks, inked to determine the true margin, using S100 stains to highlight nerves present at the margin and any neurotropic cells associated with them.
Pathologists should be alert for the presence of scars in the dermis of all melanocytic lesions. A scar often indicates that there was a previous attempt at biopsy or therapy. Since the clinician frequently does not provide this important information, it is important for pathologists to become familiar with the histologic features of a scar. These include:
- loss of normal rete ridges
- loss of adnexal structures
- horizontally orientated fibroblasts and collagen bundles
- vertically or diagonally orientated, thin straight vessels
Benign melanocytic nevi may recur following partial biopsy and produce a “pseudomelanoma” in situ pattern. Typically, one sees small individual melanocytes scattered in a haphazard fashion throughout the suprabasal epidermis overlying a dermal scar; junctional nests of melanocytes may also be present. This simulates Pagetoid melanocytosis and, if the pathologist is unaware of the prior biopsy history, may be misdiagnosed as in situ melanoma. An important diagnostic clue is that the melanocytes do not extend within the epidermis beyond the margins of the dermal scar.
If a scar is present in a melanocytic lesion, contact the referring physician’s office to see if they have a record of a previous biopsy. It is often necessary to have them ask the patient. Many times they will simply check their records, and if the biopsy was not performed in their office, answer “no”. Review of any previous biopsies is essential in providing comprehensive evaluation of a recurrent melanocytic lesion. This may prevent both under and over diagnosis.
- Melanoma presenting as a lymph node metastasis and misdiagnosed as large cell lymphoma (2 claims). Since an undifferentiated tumor composed of large malignant cells with prominent nucleoli may be a large cell lymphoma or an undifferentiated carcinoma or a melanoma, a panel of immunohistochemical stains (CD45, CK AE1/AE3, S100) is required to make a definitive diagnosis. If ordering immunohistochemical stains requires pre-authorization by an insurer and this is denied, document the “informed refusal” in the report.
- Melanomas misdiagnosed as “dysplastic nevi involving margins” (2 claims) - one was a punch biopsy and in the other a requested re-excision was not performed. Claims involving “Dysplastic or Atypical” nevi appear to result in part from miscommunication between pathologists and dermatologists. A pathologist may use the term “atypical” or “dysplastic” in a generic sense – meaning that cytologically disturbing cells are present in a lesion lacking diagnostic criteria for melanoma. Complete excision is often requested, but the dermatologist may be reluctant to do so since to them these terms have a more specific meaning and connote a clinical syndrome or benign clinical entity.
For lesions falling in the diagnostic “grey” area between melanocytic nevus and melanoma, lesions for which the biologic potential is unpredictable, pathologists should use the term “Atypical Melanocytic Neoplasm” and request complete excision. Pathologists should avoid the terms “Atypical Nevus” or “Dysplastic Nevus” for these lesions, since these terms connote a clinical syndrome or a benign clinical entity and dermatologists may not perform the indicated re-excision. This reluctance to re-excise a lesion may be unconsciously reinforced by the economic disincentives of capitation.
Nevi that show the architectural and cytologic features characteristic of Dysplastic or Atypical Nevus should be completely excised. Dysplastic or Atypical nevi can “blend” with and be difficult to differentiate from in situ melanoma. They are referred to by some experts as “melanocytic lesions of unknown medicolegal potential”.
- Spindle cell melanoma misdiagnosed as spindle cell squamous carcinoma (2 claims). Immunohistochemical stains were not performed on either of these cases. Whenever an amelanotic spindle cell skin lesion is seen, a panel of immunohistochemical stains (S100, MART-1/Melan-A and CK AE1/AE3) is required to differentiate squamous carcinoma from desmoplastic/spindle cell melanoma and atypical fibroxanthoma. HMB45 has a low sensitivity for desmoplastic and spindle cell melanomas and is not useful in this differential diagnosis.
- Patients presenting with metastatic melanoma without a known primary (2 dermatology claims). Each patient gave a history of having skin lesions (including “moles”) removed in a physicians office. Since there were no pathology reports, the lesions were presumably discarded and not sent to a pathologist.
Finally, pathologists should always issue written reports to document verbal consultations – particularly with dermatologists, since the melanocytic lesions they want an opinion on are apt to be from lesions they have clinical concern about. When we are asked to look at a slide and give an opinion, we are giving a consultation for which we can be held liable. Without a written report, the only record is the clinicians recollection of the conversation or a handwritten note made in the office record or chart, which is often incomplete or inaccurate.