Dermatology Online Journal
The relationship between biopsy technique and uncertainty in the histopathologic diagnosis of melanoma
- Author(s): Pariser, Robert J.
- Divers, Allison
- Nassar, Ahmed
- et al.
The relationship between biopsy technique and uncertainty in the histopathologic diagnosis of melanoma
Robert J. Pariser Allison Divers and Ahmed Nassar
Dermatology Online Journal 5(2):4
Eastern Virginia Medical School
Review of filed histopathology material of 525 cases of definite melanomas and other atypical pigmented lesions showed that diagnostic certainty was greatest for excisional or deep shave specimens. Shave and particularly punch biopsy specimens were associated with less certainty. The data of this study suggest that punch biopsy should be avoided for pigmented lesions and that a properly done deep shave biopsy is nearly equal to an excision in diagnostic content for such lesions.
Excisional biopsy is the recommended method for diagnosing suspected melanoma; however, less than complete lesion removal is commonly performed if the lesion is large or is located on or near important anatomic structures. If part of the lesion is to be sampled, incisional biopsy from normal skin to the most suspicious portion of the lesion is said to be preferred; punch biopsy through the thickest portion of the lesion is also said to be acceptable. Shave biopsy is discouraged because of the possibility of inadequate sampling, and because it tends to cut off the base of lesions, thus compromising the ability to determine lesion depth, a crucial element of patient management.[1-4]
The main purpose of this study was to tabulate the types of specimens of atypical pigmented lesions submitted to a dermatopathology laboratory and to determine the extent to which specimen type influenced the degree of certainty of the histopathologic diagnosis of melanoma. In addition, information was sought to evaluate the trends in incidence and tumor type seen over the survey period.
Filed histopathologic material from a small, private dermatopathology laboratory was reviewed. An attempt was made to find all cases diagnosed as definite or probable melanoma. A computer-based search was conducted for material received from mid 1988 through mid 1997. Records were scanned for the following terms: melanoma, maligna, atypical, melanocytic, and atypia. Non-melanocytic lesions found (e.g. "atypical fibroxanthoma") were eliminated. It should be noted that the term "dysplastic nevus" and "dysplasia," as regards melanocytic lesions, are not used in this laboratory. For each case a clinical record was sought. Data from these charts, from the specimen submission forms, and from the microscopic examination of the tissue itself were compiled. The following data elements were recorded:Type of specimen: Specimens were identified as excisions, incisional specimens, punch specimens, shave specimens, deep shave specimens, and curettings. For the purposes of this study, the following definitions were used:
- Excision removal of lesion by a scalpel yielding a fusiform specimen with edges perpendicular to the skin surface.
- Incision similar to an excision, except that only part of the lesion is removed.
- Punch removal of cylindrical specimen by skin punch.
- Shave removal of a thin disk of tissue by scalpel (generally a 15-blade) yielding a specimen generally limited to epidermis and upper dermis.
- Deep shave removal of a disk of tissue by a curved blade such as a Derma-Blade yielding a specimen which generally extends to at least mid dermis.
- Curettage removal of tissue fragments by skin curette.
- Unknown nature of specimen indeterminate.
- Entire lesion present on specimen.
- Less than entire lesion present on specimen, but missing portion was considered insignificant as to the determination of benignity vs malignancy.
- Less than entire lesion present on specimen, and missing portion was considered significant as to the determination of benignity vs malignancy.
- Completeness of specimen indeterminate from microscope slides available.
- Completeness of specimen indeterminate from pathology report and slides not available.
Lesion level: Clark's level recorded as Arabic numeral.
Degree of histologic certainty: The degree of certainty of the histopathologic diagnosis of melanoma was scored as:
- Completely certain.
- Fairly sure. The lesion was felt to be malignant, but with some element of doubt.
- Some suspicion. The pathologic features permit a diagnosis of melanoma, but with significant doubt.
- Don't know. Unable to distinguish benign from malignant.
- Likely to be benign on review.
- Specimen not evaluable.
In general the diagnosis of melanoma was based on the criteria of Ackerman. Degree of certainty of the histopathologic diagnosis was a subjective judgment made by one of the authors (RJP) for each case. There is little doubt that these judgments on degree of certainty would vary somewhat among pathologists; however, every attempt was made to maintain consistency throughout this study.Reason for uncertainty: The following reasons for uncertainty, which were not mutually exclusive, were delineated for each specimen:
- Inadequate specimen width.
- Inadequate specimen depth.
- Pathologic features difficult to interpret. Lesions in this category were those prompting consideration of benign neoplasms which share features with melanoma, such as Spitz nevi, recurrent nevi, or those whose features were subtle or incompletely developed.
- No uncertainty.
These reasons were not mutually exclusive; some specimens were scored in more than one category.
Using the screening criteria noted above 525 cases were identified for study. They had an average Clark level of 1.5 and an average Breslow depth of 0.34 mm. The maximal Clark level seen was 4; the maximal Breslow depth was 5.80 mm. 341 lesions were in-situ. Three lesions were felt to be metastatic.Trends in Melanoma Levels
The incidence of lesions found in the screening increased notably over the years covered by this study. Excluding the two half-years (last half of 1988 and first half of 1997), the yearly incidence of cases rose from 25 in 1989 to 138 in 1996. The majority of this increase was accounted for by the earliest (in situ) lesions, which increased from 13 in 1989 to 105 in 1996. Invasive lesions increased from 11 in 1989 to 32 in 1996. If the analysis is limited to lesions felt to be certain melanomas histologically, the same pattern emerges: total lesions increased from 19 in 1989 to 74 in 1996; in situ lesions increased from 8 to 47; and invasive lesions increased from 10 to 27. These increases cannot be fully accounted for by an overall increase in specimens submitted to the laboratory, which rose from 4148 in 1989 to 7044 in 1996, a 69 0ncrease.Types of Specimens
Of the 525 specimens, 192 (37%) were excisions, 189 (36%) were punches, 48 (9%) were shaves, 19 (4%) were deep shaves, and 74 (14%) were of unknown type. One case was submitted as a shave followed by a punch, and two were curettage specimens. No specimens were identified as incisional.Overall Degree of Histologic Certainty
Of the 525 cases, 341 (65%) were felt to be "completely certain" melanomas. In 77 (15%) the diagnosis of melanoma was considered "fairly sure"; in 89 (17%) the lesions were felt to be melanoma, but with a significant element of doubt ("some suspicion"); in 10 (2%) the lesions were felt likely to be benign on review, and in 7 (1%) the lesion was not evaluable.
In the 184 cases in which there was some degree of uncertainty, the pathologic features of the case alone were felt to be the cause of the uncertainty in 132 cases (25%), inadequate width of specimen was felt to be the cause in 24 cases (5%), and a combination of inadequate width and inherent pathologic features was the cause in 21 cases (4%). 7 cases (1%) were not evaluable. In no case was inadequate depth felt to compromise histopathologic diagnosis so far as malignancy versus benignity was concerned.Specimen Quality and Histopathologic Certainty Related to Biopsy Type
Figure 1. Percentages of specimens of each type which removed entire lesion, part of lesion with insignificant missing portion, or part of lesion with significant missing portion.
Figure 2. Percentages of each specimen type resulting in a histopathologic diagnosis which was completely certain or fairly sure, or in which there was some suspicion. Shown in yellow are those cases in which there was little doubt of lesion malignancy (sum of completely certain and fairly sure).
19 specimens were identified as deep shaves. 15 of them (79%) contained the entire lesion and in 16 (84%) the diagnosis was felt to be certain. 4 deep shave specimens (21%) were missing "insignificant" portions. No deep shave specimens were missing "significant" portions.
Of the 48 shave specimens 25 (52%) contained the entire lesion and the same number yielded a completely certain diagnosis. In 19 (40%) of shave specimens an "insignificant" part of the lesion was missing; in 2 (4%) a "significant" portion was lacking.
Of the 189 punch specimens 46 (24%) contained the entire lesion and 98 (51%) yielded a completely certain diagnosis. In 92 (49%) of punch specimens an "insignificant" part of the lesion was missing; in 40 (21%) a "significant" portion of the lesion was missing.
As noted in Figures 1 and 2, excision and deep shave specimens were more likely to contain the complete lesion and to result in more histopathologic certainty than shave and particularly punch specimens. Specifically the proportion of specimens of each type resulting in "some suspicion," the lowest level of certainty, were 9%, 11%, 23% and 21 0.000000or excision, deep shave, punch, and shave, respectively. When certainty of histopathologic diagnosis is analyzed by depth of lesion a trend toward reduced certainty for earlier lesions emerges (Table). This trend is particularly notable for incisional biopsy specimens. For example for invasive lesions (Clark's levels II-V), excision produced 95ertainty and deep shave produced 82%. The respective numbers for in-situ lesions were 73% and 75%; this drop in certainty likely reflects the inherently less developed pathology of earlier lesions. By contrast, punch produced only 77ertainty and shave 67 0.000000or invasive lesions, while for in-situ lesions, the respective numbers were 44% and 42%.
*Figures listed are numbers of cases in each category. Figures in parentheses are percentages. For example, for in situ lesions, there were 116 excisions, of which 85 were certain and 28 were less than certain ("fairly sure"+"some suspicion"). Not shown are cases scored as "don't know," "likely to be benign on reevaluation," and "not evaluable."
The loss of histopathologic certainty when changing from complete or nearly complete lesion removal to partial removal was more pronounced for in situ lesions than for invasive lesions. The average percentage of certainty for completely or nearly completely removed in situ lesions (excision + deep shave) was 74%; for incisional biopsy specimens (punch + shave) it was 43%, a loss of 31 percentage points. For invasive lesions, the corresponding figures are 88% and 72%, a loss of 16 percentage points.Causes of Uncertainty
The diagnosis of melanoma was felt to be less than certain in 184 cases; in 78 cases the degree of uncertainty was small ("fairly sure") and in 89 it was greater ("some suspicion"). Possible causes for histopathologic uncertainty were the inherent pathologic features, lack of specimen width (breadth), lack of specimen depth, or combinations of these factors.
In the "fairly sure" category the numbers and proportions of cases accounted for by inherent pathology, specimen width, and specimen depth were 56 (72%), 13 (17%), and 0 (0%), respectively. A combination of width and inherent pathology was felt responsible for the uncertainty in 8 of these cases (10%). For one case in this category, microscope slides were not available for review. In the "some suspicion" category the numbers and proportions of cases were 67 (76%), 10 (11%), and 0(0%), respectively. In this group a combination of width and inherent pathology accounted for 12 cases (13%). In no case was lack of specimen depth a contributor to uncertainty; in 260f cases, however, lack of specimen width alone or in combination with inherent pathology was a contributor to uncertainty.
Current trends in the epidemiology and management of cutaneous melanoma have bearing on the histopathologic diagnosis of melanoma as addressed by this study. The incidence of melanoma in the United States and worldwide continues to rise. In addition heightened awareness of melanoma by physicians and lay individuals, and increased opportunities for screening of pigmented lesions have resulted in more pathologic specimens of earlier melanocytic lesions. The result is an increase in lesions presenting "tougher calls" for the general pathologist and the dermatopathologist. In this survey there was a notable increase in in-situ melanomas relative to invasive lesions, and there was an increase in less-than-certain melanomas, over the period screened.
The trend toward more conservative surgical management of melanomas is also relevant to this study. It is clear that traditional wide surgical margins offer little or no survival advantage to individuals with melanoma. While it is generally accepted that lesion depth has prognostic importance, this factor is of decreasing overall importance in actual patient management, as lesions on average become thinner and surgical margins more conservative. The truly thin melanoma will likely be adequately biopsied, so far as depth is concerned, by all but the skimpiest of shave specimens. For the truly thick melanomas, such as those over several millimeters in thickness, there is little loss of clinically useful data even if the base of the lesion is cut off, so long as the specimen is at least several millimeters thick. For lesions of intermediate thickness, some loss of clinically useful histopathologic data may certainly occur for specimens which are too thin, but the clinician can generally avoid this pitfall by careful observation and palpation of the lesion. These considerations should tend to partly allay the fear that cutting off the base of a melanoma during biopsy will compromise patient management. It is well accepted that cutting through a melanoma has no bearing on prognosis.[9,10,11] It should be emphasized that the primary goal of a biopsy of a pigmented lesion is to establish benignity or malignancy. Knowledge of lesion depth is of little importance if this distinction cannot be made.
The rationale for the punch biopsy for pigmented lesions formerly was that the diagnosis of melanoma could generally be made from a punch specimen, and, if the thickest portion of the lesion was sampled, lesion depth could be determined. As more lesions which are intraepidermal or only in the superficial dermis are sampled, however, the breadth of the specimen, not the depth, becomes the limiting factor for a good biopsy specimen. Pathologic factors such as size, symmetry, and margination are compromised by a specimen which does not include the full width of the lesion. In addition, such specimens compromise the ability to assess factors which are focal and must be sought over as broad a front as possible, such as spacing between epidermal melanocytes, size and shape of melanocyte nests, tendency to upward migration of melanocytes, and cytologic details. A particular risk of erroneous diagnosis may occur in the fairly common situation of a melanoma arising in association with a preexisting nevus. Punch biopsy through the thickest portion of the lesion can easily sample the nevus only and completely miss the melanoma (Figure 3). The data of this study suggest that a wide, thin specimen (namely a well done deep shave biopsy) will produce a more satisfactory specimen than a narrow, deep one (punch biopsy) for the majority of melanomas seen today. We acknowledge that the actual number of deep shave specimens in this study is relatively small, which may somewhat limit the strength of this assertion. A similar conclusion without supporting data has been expressed by other authors, who suggest that the use of well-done saucerization biopsies could result in more early melanomas being diagnosed promptly, particularly in the case of lesions which are only slightly atypical clinically and for which the clinician might be concerned that excision would be deemed "overkill" if the lesion turned out to be benign .[12,13] Other reports have documented apparent popularity of such saucerization specimens for diagnosis of pigmented lesions.[14,15]
Figure 3. This pigmented lesion presents a particular risk of misdiagnosis if a punch biopsy were taken from the thickest portion of the lesion. The thick portion is a dermal nevus. The flat, intraepidermal portion, which is a melanoma arising in association with this nevus, could easily be missed by such an incisional biopsy. This error is more likely to occur in nevus/melanoma combinations in which the clinical distinction between the two is not so clear as in this case.
We suggest that it is time to revise the oft-repeated advice on how to biopsy suspicious pigmented lesions with which this article began. Excision remains the "gold standard," but a well done deep shave (saucerization) specimen is nearly as good. Traditional shave specimens and particularly punch specimens are to be avoided because they yield specimens that may compromise histopathologic interpretation. The guiding principle should always be to deliver the entire lesion, or as much of it as possible, to the histopathologist. For non-excisional specimens the shape of the specimen should ideally match that of the lesion. A broad, flat lesion typical of the early melanomas seen so commonly today can be well biopsied by a broad, flat specimen; a deeper, more advanced lesion requires more specimen depth.
References1. Elder DE, Murphy GF. Melanocytic tumors of the skin. In: Atlas of tumor pathology. Washington, DC: Armed Forces Institute of Pathology; 1991:111.
2. NIH consensus conference. Diagnosis and treatment of early melanoma. JAMA 1992;268:1314-19.
3. Balch CM, Houghton A, Milton G, et al. Biopsy of melanoma. In: Cutaneous melanoma: clinical management and treatment results worldwide. 2nd ed. Philadelphia: JB Lippincott Co;1991:264-268.
4. Roses DF. Principles of biopsy of lesions suspected of being malignant melanoma. In: Harris MN, Roses DF, Ackerman AB, eds. Diagnosis and management of cutaneous malignant melanoma. Philadelphia: WB Saunders Co;1983:105-112.
5. Ackerman AB, Cerroni L, Kerl H. Pitfalls in Histopathologic Diagnosis of Malignant Melanoma. Philadelphia: Lea and Febiger; 1994.
6. Rigel DS, Friedman RJ, Kopf AW. The incidence of malignant melanoma in the United States: issues as we approach the 21st century. J Am Acad Dermatol 1996;34:838-47.
7. Merlino LA, Sullivan KJ, Whitaker DL. The independent pathology laboratory as a reporting source for cutaneous melanoma incidence in Iowa. J Am Acad Dermatol 1997;37:578-85.
8. Zitelli JA, Brown CD, Hanusa BH. Surgical margins for excision of primary cutaneous melanoma. J Am Acad Dermatol 1977;37:422-9.
9. Lees VC, Briggs JC. Effect of initial biopsy procedure on prognosis in stage I invasive cutaneous malignant melanoma: a review of 1986 patients. Br J Surg 1991;78:1108-10.
10. Lederman JS, Sober AJ. Does biopsy type influence survival in clinical stage I cutaneous melanoma? J Am Acad Dermatol 1985;13;983-7.
11. Ho VC, Sober AJ. Therapy for cutaneous melanoma: an update. J Am Acad Dermatol 1990;22:159-76.
12. Salache SJ, Grabski WJ. Transverse sectioning of a pigmented lesion. Dermatol Surg 1997;23:578-82.
13. Piepkorn M, Odland PB. Quality of care in the diagnosis of melanoma and related melanocytic lesions. Arch Dermatol 1997;133:1393-96.
14. Salopek TG, Slade J, Marghoob AA, et al. Management of cutaneous malignant melanoma by dermatologists of the American Academy of Dermatology, I: survey of biopsy practices of pigmented lesion suspeted as melanoma. J Am Acad Dermatol 1995;33:441-50.
15. Drake LA, Ceilley RI, Cornelison LR, et al. Guidelines of care for malignant melanoma. J Am Acad Dermatol 1993;28:638-41.
© 1999 Dermatology Online Journal