- Email: [email protected]
Sampling of melanocytic nevi for research purposes: A prospective, pilot study to determine effect on diagnosis
DERMATOPATHOLOGY
Sampling of melanocytic nevi for research purposes: A prospective, pilot study to determine effect on diagnosis Scott R. Florell, MD,a,b Bruce R. Smoller, MD,d Kenneth M. Boucher, PhD,c Douglas Grossman, MD, PhD,a,b,c Ronald M. Harris, MD, MBA,a,b Glen M. Bowen, MD,a,b and Sancy A. Leachman, MD, PhDa,b Salt Lake City, Utah, and Little Rock, Arkansas Background: Research on melanocytic nevi predominantly utilizes formalin-fixed, paraffin-embedded tissue, largely limiting research to morphologic and immunohistochemical observations. Withholding portions of fresh nevus tissue for molecular studies could result in the loss of important diagnostic material. Objective: This study prospectively evaluated melanocytic nevi for histologic homogeneity to determine if using a portion for research would have affected diagnosis. Methods: Thirty-three subjects were enrolled in a prospective study in which pigmented lesions were chosen for biopsy on a clinical basis. Lesions were sectioned and each piece submitted in a separate block (mean, 2.7; range 2-5 blocks per lesion). Slides from nevi were examined in two phases. In phase I, sections of nevi were randomized and a diagnosis was rendered for each section of nevus. In phase II, the dermatopathologist reviewed all slides for each nevus as a case, similar to the original interpretation of the lesion provided to the clinician. Diagnoses from phases I and II were compared with the original diagnosis. Results: Case material included 51 melanocytic lesions from 31 subjects. The phase I diagnosis matched the original diagnosis for 99 of 121 slides that showed a melanocytic lesion (82%). The phase II diagnosis matched the original diagnosis for 45 of 51 specimens (88%). Limitations: The study was limited by: a small number of specimens; the clinician could have chosen clinically homogeneous nevi for biopsy; effect of interobserver and intraobserver variability on diagnosis. Conclusions: For the majority of melanocytic nevi in this study, the diagnostic information present in one section of a melanocytic nevus could be extrapolated to the remainder of the specimen without adverse consequences from a diagnostic or therapeutic perspective. ( J Am Acad Dermatol 2008;59:814-21.)
INTRODUCTION Melanocytic nevi are among the most common cutaneous lesions undergoing biopsy and submitted for histopathologic analysis, and the importance of molecular study of melanocytic nevi and melanoma cannot be overstated. The dermatopathology archive From The Melanoma Program, Huntsman Cancer Institutea and Departments of Dermatologyb and Oncological Sciences,c University of Utah, Salt Lake City; and the Department of Pathology, University of Arkansas for Medical Sciences, Little Rock.d Supported by grants from The Skin Cancer Foundation (to S. R. F.), the Dermatology Foundation Leaders Society Dermatologist Investigator Research Fellowship and Clinical Career Development Award (to S. R. F.); National Institutes of Health grants K23 RR17525-01 (to S. R. F.) and R01 AR50102 (to D. G.), Doris Duke Charitable Foundation (to S. A. L.), Fellowship-To-Faculty Transition Award from the University of Utah funded in part by the Howard Hughes Medical Institute (to S. A. L.), The Huntsman Cancer Foundation (to S. A. L., D. G.), the Tom C. Mathews Jr. Familial Melanoma Research Endowment, Huntsman General
814
is an important and widely available source of tissues for research purposes since the majority of the specimen is usually retained in paraffin after a diagnosis is rendered. However, research using formalinfixed, paraffin-embedded tissues is largely limited to morphologic and immunohistochemical studies Clinical Research Center Public Health Service grant (MO1 RR00064), National Cancer Institute (NCI) Cancer Center Support grants 5P30CA420-14, and the Utah Cancer Registry, funded by Contract # N01-PC-35141 from the NCI with additional support from the Utah Department of Health and the University of Utah. Conflicts of interest: None declared. Reprint requests: Scott R. Florell, MD, Department of Dermatology, University of Utah Health Sciences Center, 4B454 SOM, Salt Lake City, UT 84132. E-mail: [email protected]. 0190-9622/$34.00 ª 2008 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2008.07.020
J AM ACAD DERMATOL
Florell et al 815
VOLUME 59, NUMBER 5
because of difficulties in analyzing DNA and RNA due to cross-linking of nucleic acids with protein.1-3 RNA also degrades quickly at room temperature in the surgical pathology laboratory,4 and culturing cells is not an option after fixation. Consequently (for cellular and molecular studies), obtaining a portion of fresh or flash-frozen tissue is preferable, although this could result in the loss of potentially important diagnostic material. In the pathology laboratory, histologic diagnosis of specimens is made based on microscopic examination of a small percentage of the tissue. The assumptions made in the standard method in which surgical tissues are processed are (1) the small fraction of the tissue examined microscopically is representative of the entire specimen and (2) the diagnosis or recommendation made from examination of this ‘‘representative’’ sample would not change if more sections were examined. Similarly, when a portion of tissue is retained for research purposes, the implication is that the examined portion of the tissue is sufficient for diagnosis. Our melanoma program is engaged in cellular and molecular studies on melanocytic nevi, but we are concerned about the potential impact of utilizing portions of nevi for research on diagnosis and patient care. We present data from a prospective study in which biopsied melanocytic nevi were thoroughly sectioned to (1) determine the histologic homogeneity of nevi; (2) determine whether the diagnosis of one section of a nevus is comparable to the diagnosis on other sections of the same nevus; (3) correlate clinical eccentric foci of pigmentation with histologic findings; and (4) determine if retaining a portion of the tissue for research purposes would have had a deleterious effect on diagnosis in this sample set.
METHODS Subjects This study was approved by the Institutional Review Board at the University of Utah (IRB No. 00-7916). During the period from March 2002 to April 2004, 33 subjects were enrolled in a prospective study in which pigmented lesions were chosen for biopsy on clinical grounds (ie, lesion falls out of the range of normal for that patient, including ABCDE criteria). The entire clinical lesion was removed by either saucerization shave or punch technique and the tissues were fixed in 10% buffered formalin. Examination and processing of tissue A dermatopathologist (S. R. F.) examined the fixed tissue. Punch biopsy specimens were bisected and each half was submitted in a separate cassette. Shave specimens were sectioned at 2-mm intervals,
perpendicular to the long axis, and each of the pieces of tissue were submitted in separate cassettes. If a pigmented lesion had a discrete pigmentary change clinically identified as a ‘‘dark dot,’’ the specimen was carefully inked along the deep margin beneath the ‘‘dark dot’’ as well as the epidermis overlying the area prior to sectioning and processing. The tissues were processed in the usual fashion. Four-micron sections from at least two levels of each block were applied to a glass microscopic slide and stained with hematoxylin and eosin. One of two board-certified dermatopathologists reviewed all of the slides for each specimen as part of their daily sign-out and prepared a diagnostic report. The clinician was notified of the results as usual. This report is referred to as the ‘‘original diagnosis.’’ The slides were subsequently examined in two phases (Fig 1). Phase I The purpose of phase I of the study was to determine whether the pathologic findings in a small part of the nevus were representative of the entire lesion. The diagnosis of the specimen was established from multiple blinded portions of the lesion and compared for consistency. As such, each specimen was at least bisected and each piece was submitted in a single block, resulting in a one slide for each piece of tissue. All of the slides were randomized and a diagnosis was rendered for each slide. The diagnostic categories were as follows: 1, benign ordinary intradermal or compound melanocytic nevus; 2, melanocytic nevus with mild or moderate architectural disorder and/or mild or moderate cytologic atypia; 3, melanocytic nevus with severe architectural disorder and/or severe cytologic atypia; 4, malignant melanoma (in-situ or invasive); 5, other melanocytic nevus (eg, Spitz nevus, blue nevus); and 6, no melanocytic lesion present. The dermatopathologist also rated each slide as benign (categories 1, 2, 5, and 6 above), atypical (category 3), or malignant (category 4). The dermatopathologist indicated whether a re-excision should be done (yes or no), usually for category 3 or 4 melanocytic lesions. The diagnosis for each slide, representing different portions of the same lesion, was compared with the original clinical dermatopathology report. Phase II In phase II, the dermatopathologist was given the complete, fully representational set of slides for each specimen that was reviewed in a similar fashion to the original diagnosis utilizing the same diagnostic categories. The dermatopathologist also indicated whether a re-excision should be done (yes or no). The dermatopathologist (blinded to the results of
816 Florell et al
J AM ACAD DERMATOL NOVEMBER 2008
Fig 1. Study design. Melanocytic nevi were sectioned into multiple pieces and were examined by one of two dermatopathologists as part of the regular workload for the day. A diagnostic report, the ‘‘original diagnosis,’’ was generated for each melanocytic lesion and the report was provided to the clinician as usual. After all subjects were enrolled in the study, the melanocytic lesions were examined in two phases by a single dermatopathologist and the phase I and II diagnoses were compared with the original diagnosis.
phase I) examined the slides 6 weeks after completing phase I. In this phase, the purpose was to prospectively evaluate whether removing a portion of the biopsy (eg, for research purposes) would have negatively impacted patient care. Case material A total of 33 subjects were enrolled from March 2002 through April 2004. Fifty-six specimens were collected (20 subjects had one specimen, 8 had two specimens, 1 had 3 specimens, 3 had 4 specimens, and 1 had 6 specimens). The entire clinical lesion was removed by shave saucerization (47 specimens), punch (7 specimens), or excision (1). One specimen was a re-excision.
RESULTS Fifty-six specimens were prospectively collected. Sixteen (29%) of the specimens were benign ordinary melanocytic nevi (category 1); 27 (48%) were melanocytic nevi with mild to moderate architectural disorder and/or focal mild or moderate cytologic atypia (category 2), and one melanocytic nevus (2%) was diagnosed as a melanocytic nevus with severe atypia (category 3). There was one excision for a highly suspect lesion that demonstrated invasive
melanoma arising in association with a melanocytic nevus (Clark level III, Breslow thickness 0.68 mm). The area was re-excised with 1-cm margins with no residual melanocytic lesion. There was one melanoma-in situ. Five lesions were classified as ‘‘other’’ melanocytic lesions (category 5): 3 Spitz nevi, 1 blue nevus, and 1 combined ordinary and blue nevus. One specimen showed features of regression without a residual melanocytic lesion. Three lesions were seborrheic keratoses. Only melanocytic nevi and malignant melanomas were included in the study; the re-excision specimen (of the invasive melanoma), 3 seborrheic keratoses, and lesion showing only regression were excluded. The case material used in this study included 31 subjects with 51 melanocytic lesions and a total of 140 glass microscopic slides. The mean (standard deviation) number of blocks per specimen was 2.7 (0.7), with a range of 2-5 (case material summarized in Table I). Of the 51 lesions, 42 were shave specimens (average number of levels per slide, 3.4; range, 2-5), 8 were punch specimens (average number of levels per slide, 3.6; range, 3-4), and one was an excision (each slide with 2 levels). Two dermatopathologists examined the slides initially and generated a diagnostic report; one read
Florell et al 817
J AM ACAD DERMATOL VOLUME 59, NUMBER 5
Table I. Summary of case material* Total No. No. in (%) study (%)
Subjects enrolled 33 Specimens collected 56 Benign melanocytic nevus, 16 (29) category 1 Melanocytic nevus with 27 (48) mild-moderate atypia, category 2 Melanocytic nevus with 1 (2) severe atypia, category 3 Spitz nevus, category 5 3 (6) Blue nevus, category 5 1 (2) Combined nevus (ordinary 1 (2) benign 1 blue), category 5 Invasive malignant melanoma, 1 (2) category 4 Re-excision of invasive melanoma 1 (2) Melanoma in situ, category 4 1 (2) Features of regression , no residual 1 (2) melanocytic lesion Seborrheic keratoses 3 (6) Total pigmented lesions 56 (100) Total melanocytic nevi included in study
31 (94) 51 (91) 16 (31) 27 (53) 1 (2) 3 (6) 1 (2) 1 (2) 1 (2)
1 (2)
51 (100)
*Only melanocytic nevi were included in the study.
32 of the specimens, the other read 19 of the specimens. Six of the original diagnostic reports recommended a re-excision of the melanocytic lesion (spindle cell nevus, n = 1; compound dysplastic nevus with mild or moderate atypia, n = 2; markedly atypical junctional lesion on sun-damaged skin, n = 1; melanoma in situ, n = 1; invasive melanoma, n = 1). Phase I In phase I, the diagnosis matched the original diagnosis for 99 of 140 slides (71%). Of the 41 discrepancies, 19 were identified as ‘‘no melanocytic lesion present;’’ these slides generally represented normal skin at the edges of the specimen. If these 19 slides are excluded, then the phase I diagnosis matched the original diagnosis for 99 of 121 slides (82%) that showed a pigmented lesion. The diagnosis was upgraded to a more atypical lesion in 17 slides (benign nevus [category 1] to nevus with mild or moderate atypia [category 2], n = 10; nevus with mild or moderate atypia [category 2] to severe atypia [category 3], n = 6; and nevus with severe atypia [category 3] to melanoma in situ [category 4], n = 1) and downgraded to a more benign diagnosis for 5 slides (nevus with mild-moderate atypia (category 2) to benign (category 1), n = 3; nevus with severe atypia (category 3) to nevus with mild-moderate atypia (category 2), n = 1; malignant melanoma
(category 4) to nevus with mild-moderate atypia (category 2), n = 1). When 3 diagnostic categories were considered and slides were diagnosed as benign (including nevi with architectural disorder or dysplastic features and mild to moderate atypia), atypical (nevi with architectural disorder or dysplastic features and severe atypia) or malignant, the diagnosis correlated with the original pathology report in 111 of 121 slides (92%). Of the 10 discrepancies, the diagnosis was upgraded for 7 of the slides (benign to atypical, n = 6, and atypical to melanoma in situ, n = 1, specimen 39) and the diagnosis was downgraded for 3 slides (atypical to benign, n = 2 and malignant to benign, n = 1, specimen #28, representing melanocytic nevus at the edge of an invasive melanoma, Fig 2). Phase I findings are summarized in Table II. A recommendation for re-excision was made for 14 slides involving 6 specimens which correlated well with the original pathology report with the exception of two slides upgraded to ‘‘severe atypia’’ (category 3) in phase I. Phase II In phase II, in which the slides were organized into cases, the phase II diagnosis matched the original diagnosis for 45 of 51 specimens (88%; k statistic 0.81, 95% confidence interval [CI] 0.66-0.96, good correlation5,6); the phase II diagnosis was upgraded to a more atypical lesion in 5 cases (benign nevus to nevus with mild or moderate atypia, n = 4; nevus with mild-moderate atypia to severe atypia, n = 1) and downgraded in 1 case (nevus with mildmoderate atypia to benign nevus). When 3 diagnostic categories were considered (benign, atypical, or malignant), the phase II diagnosis matched the original diagnosis for 50 of 51 specimens (98%; k statistic 0.85, 95% CI = 0.56-1.00, good correlation5,6) reflecting a category upgrade for the nevus with mild-moderate atypia (benign) to severe atypia (atypical). There were no significant differences in recommendation for re-excision. These findings are summarized in Table III. Variability of histology within specimens Of the 51 specimens, #28 (55%) demonstrated homogeneous histopathologic findings throughout and all of the sections were given the same diagnostic code. For example, if a nevus was divided into 3 pieces, each of the 3 pieces was diagnosed identically in phase I (Fig 3). Twelve of 51 specimens (24%) showed homogeneous diagnostic codes except for slides that did not show a melanocytic process. For example, if a nevus was divided into 3 pieces, two of the pieces were given the same diagnostic code and the third was ‘‘no melanocytic
818 Florell et al
J AM ACAD DERMATOL NOVEMBER 2008
Fig 2. Invasive malignant melanoma arising in association with a benign nevus. The lesion was found on the left upper back of a 73-year-old man. The clinical impression was ‘‘atypical nevus, rule-out melanoma.’’ Phase I diagnoses were not homogeneous: two sections showed invasive melanoma, one section showed benign nevus (category 2), and one section showed normal skin without a melanocytic lesion. This type of lesion should not be sampled for research purposes. Numbers in parentheses indicate diagnostic category.
lesion.’’ In 11 of 51 specimens (21%), the phase I diagnosis was not uniform. For example, one excision specimen (specimen #28) was sectioned into 4 pieces. When the slides of the 4 sections were randomized in phase I, two of the sections were diagnosed as invasive malignant melanoma (category 4), one section as a category 2 nevus, and one section with no melanocytic lesion (category 6). The phase II diagnosis for this lesion was invasive malignant melanoma (category 4) and this correlated with the original pathologic diagnosis (see Fig 2). Lesions with foci of eccentric pigmentation Twelve of the 51 pigmented lesions (24%) had a discrete pigmentary change clinically identified as a ‘‘dark dot.’’ Six of these foci of eccentric pigmentation corresponded histologically to keratinocyte pigmentation and discrete collections of melanophages within the superficial dermis. Two specimens demonstrated host response manifested by papillary dermal expansion, fibrosis, and melanophage deposition. One specimen each demonstrated pigmented junctional melanocytic nests, congenital pattern nevus with a deep penetrating component, follicular rupture, and superficial dermal melanocytic theques with prominent cytoplasmic pigmentation.
DISCUSSION In this study, we sought to determine whether retaining a portion of a melanocytic nevus for research purposes could have a deleterious effect on
histologic diagnosis and consequently, patient care and outcome. To our knowledge, published studies addressing this question are not available. Although there are vast libraries of archived tissue in private and academic dermatopathology practices, the potential experimental utility is limited since many contemporary investigative techniques are not optimal (or possible) on formalin-fixed tissue. In phase I, we found that one section of nevus was generally representative of the remainder of the specimen. Specifically, the diagnosis of a single portion of a nevus matched the original diagnosis in more than 80% of cases and more than 90% when only 3 diagnostic categories were considered (benign, atypical, or malignant). As such, most of the nevi in this study demonstrated homogeneous histologic features throughout the lesion. However, it is important to note that the diagnosis for 7 of the slides (5%) was upgraded to a more atypical diagnosis, and in two specimens additional re-excision was recommended in phase I that was not recommended in the original dermatopathology report. One specimen was upgraded from a severely atypical (category 3) lesion to melanoma in situ, though a recommendation for complete excision was made for each of the random slides for this lesion and this correlated with the original dermatopathology report. More problematic, however, was the invasive melanoma (see Fig 2), in which 2 of the 4 microscopic slides showed benign nevus and normal skin. Had the more atypical area been selected for research, an invasive
Florell et al 819
J AM ACAD DERMATOL VOLUME 59, NUMBER 5
Table II. Summary of phase I data*
Table III. Summary of phase II data No. (%)
Total slides Exact match with original diagnosis Slides without pigmented lesion (edges of specimen)y Slides with a pigmented lesion Exact match with original diagnosis Diagnosis upgraded to more atypical lesion Diagnosis downgraded to more benign lesion Three diagnostic categories: Benign, atypical, malignant Correlate with original diagnosis Diagnosis upgraded to more atypical lesion Diagnosis downgraded to more benign lesion
140 99/140 (71) 19 121 99/121 (82) 17/121 (14)
No. (%)
Total cases Exact match with original diagnosis Diagnosis upgraded to more atypical lesion Diagnosis downgraded to more benign lesion Three diagnostic categories: Benign, atypical, malignant Correlate with original diagnosis Diagnosis upgraded to more atypical lesion
51 45 (88) 5 (10) 1 (2)
50 (98) 1 (2)
5/121 (4)
111/121 (92) 7/121 (6) 3/121 (2)
*When 3 diagnostic categories are considered, melanocytic lesions are diagnosed as benign (category 1, 2, and 5 nevi), atypical (category 3 nevi), and malignant (category 4 in situ or invasive malignant melanoma). y These are slides that did not demonstrate a melanocytic lesion and generally represented normal skin at the edges of the specimen.
melanoma could have been missed. In sum, these results suggest that retaining a small portion of a melanocytic nevus for research purposes does not compromise diagnosis. In phase II, all of the slides for each case were reviewed together, similar to the way the lesion was diagnosed originally. The purpose of phase II was to determine the level of concordance with the way the lesion was diagnosed originally. If the concordance was poor, the phase I results would accordingly be less reliable. The results indicate good concordance5,6 with the original diagnosis. A clinical focus of eccentric hyperpigmentation was seen in 12 of the 49 specimens, all of which were associated with benign pathologic findings, most commonly keratinocyte pigmentation and melanophage deposition in the superficial dermis. The results are similar to those published by Bolognia, Lin, and Shapiro,7 although in their series of 59 melanocytic lesions with an eccentric focus of hyperpigmentation, 3 contained foci of melanoma arising within a melanocytic nevus. The relatively small number of biopsy specimens limits this study and multi-institutional expansion of the study would permit greater assurance of the safety of utilizing a portion of nevi for research. In addition, it is possible that the clinician chose clinically homogeneous nevi for biopsy rather than those with clinically atypical features, which could
skew the data in favor of histologic homogeneity; however, the clinicians submitting melanocytic lesions did not attempt to grade clinical homogeneity. Some of the diagnostic discrepancy between the original diagnosis and phase I and II diagnoses could be related to interobserver and intraobserver variability. This was likely minimized when fewer diagnostic categories were considered (ie, benign, atypical, and malignant). Because the study was not designed to be an interobserver concordance study and the known lack of interobserver concordance among expert pathologists in diagnosis of melanocytic lesions,8-12 only one dermatopathologist evaluated slides in phase I and phase II, similar to our experience that research collaborations involving sampling of melanocytic lesions usually enlist the assistance of one dermatopathologist. In summary, we found that, for the majority of specimens in this study, the diagnostic information present in one section of a melanocytic nevus could be extrapolated to the remainder of the specimen without adverse consequences to the patient from a diagnostic or therapeutic perspective. Broader issues are raised by the findings in this study. The first is that collaboration with colleagues in pathology and dermatopathology is critical for studies in which nevi are sampled for research purposes. A pathologist should be available to assist the clinician in the clinical evaluation of the patient and the lesion(s) to be sampled, then gross the specimen to adequately sample any areas that the pathologist deems important to evaluate. The second is whether histologic evaluation of one part of the lesion can be correlated with molecular findings in other parts of the specimen. According to our results, the histologic findings are usually homogeneous, suggesting that the molecular findings would also be similar, but this was not specifically studied here. The third issue is that in a small number of biopsies, the histologic findings were not homogeneous throughout the lesion and sampling part of a nevus in this setting could potentially affect patient care. There is a small,
820 Florell et al
J AM ACAD DERMATOL NOVEMBER 2008
Fig 3. Compound nevus with congenital features demonstrating homogeneous pathologic findings in all 4 sections. The lesion was found on the left side of the upper part of the neck of a 42-year-old woman. The clinical impression was ‘‘atypical nevus.’’ Phase I diagnoses were uniform and correlated with the phase II and original diagnosis.
but definite, risk of missing significant pathology when sampling melanocytic lesions for research purposes because such tissues cannot be used (in most cases) for additional sectioning and histologic examination. This risk can be minimized by carefully selecting lesions for biopsy and by rapid histopathologic examination of the tissue by the dermatopathologist at the time of biopsy. On the basis of the results of this study, our melanoma program has developed guidelines for selection, sampling, and processing of melanocytic nevi for research purposes (after obtaining Institutional Review Board [IRB] approval): 1. Select homogeneous lesions without significant asymmetry in border, color, or foci of eccentric pigmentation 2. Biopsy of ‘‘typical’’ atypical nevus—biopsy of a clinically atypical nevus is permissible when other similar-appearing nevi are present (ie, signature nevus) 3. When a clinically atypical nevus undergoes biopsy and there is return of pigment or the presence of a persistent pink macule or papule within the scar, a repeat biopsy is performed. 4. Clinically atypical lesions suspected of being malignant melanoma are submitted in their entirely for histopathologic analysis due to variation in histologic features unless express IRB approval has been obtained for collection for these lesions. Specimen #28 (Fig 2) was clinically
suspected of being malignant melanoma and thus would not have been a candidate for research sampling. In our program, following these guidelines, portions of clinically atypical nevi are held for potential research use until cleared by the dermatopathologist (with express patient consent). There are various methods for preserving tissue while awaiting histologic diagnosis.4 Usually, the remainder of the specimen is retained in tissue culture medium and/or RNA preservative. Then next morning, the routine hematoxylin-eosin profiles are reviewed by a dermatopathologist and a diagnostic report is issued. If there are any suspect histologic features, the retained portions of the nevus are submitted entirely for hematoxylin-eosin processing and not used for research. If the lesion demonstrates benign histologic features, the researcher may then use the remaining specimen for research purposes. We gratefully acknowledge the willing participation of all individuals in this study.
REFERENCES 1. Feldman MY. Reactions of nucleic acids and nucleoproteins with formaldehyde. Prog Nucleic Acid Res Mol Biol 1973;13: 1-49. 2. Moller K, Rinke J, Ross A, Buddle G, Brimacombe R. The use of formaldehyde in RNA-protein cross-linking studies with ribosomal subunits from Escherichia coli. Eur J Biochem 1977;76: 175-87.
Florell et al 821
J AM ACAD DERMATOL VOLUME 59, NUMBER 5
3. Werner M, Chott A, Fabiano A, Battifora H. Effect of formalin tissue fixation and processing on immunohistochemistry. Am J Surg Pathol 2000;24:1016-9. 4. Florell SR, Coffin CM, Holden JA, Zimmermann JW, Gerwels JW, Summers BK, et al. Preservation of RNA for functional genomic studies: a multidisciplinary tumor bank protocol. Mod Pathol 2001;14:116-28. 5. Kendall BS, Ronnett BM, Isacson C, Cho KR, Hedrick L, Diener-West M, Kurman RJ. Reproducibility of the diagnosis of endometrial hyperplasia, atypical hyperplasia, and welldifferentiated carcinoma. Am J Surg Pathol 1998;22:1012-9. 6. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74. 7. Bolognia JL, Lin A, Shapiro PE. The significance of eccentric foci of hyperpigmentation (’small dark dots’) within melanocytic nevi. Analysis of 59 cases. Arch Dermatol 1994;130:1013-7. 8. Corona R, Mele A, Amini M, De Rosa G, Coppola G, Piccardi P, et al. Interobserver variability on the histopathologic
9.
10.
11.
12.
diagnosis of cutaneous melanoma and other pigmented skin lesions. J Clin Oncol 1996;14:1218-23. Duncan LM, Berwick M, Bruijn JA, Byers HR, Mihm MC, Barnhill RL. Histopathologic recognition and grading of dysplastic melanocytic nevi: an interobserver agreement study. J Invest Dermatol 1993;100:318S-21S. Farmer ER, Gonin R, Hanna MP. Discordance in the histopathologic diagnosis of melanoma and melanocytic nevi between expert pathologists. Hum Pathol 1996;27:528-31. Florell SR, Boucher KM, Leachman SA, Azmi F, Harris RM, Malone JC, et al. Histopathologic recognition of involved margins of lentigo maligna excised by staged excision: an interobserver comparison study. Arch Dermatol 2003;139: 595-604. Wechsler J, Bastuji-Garin S, Spatz A, Bailly C, Cribier B, AndracMeyer L, et al. Reliability of the histopathologic diagnosis of malignant melanoma in childhood. Arch Dermatol 2002;138: 625-8.
Sampling of melanocytic nevi for research purposes: A prospective, pilot study to determine effect on diagnosis Scott R. Florell, MD,a,b Bruce R. Smoller, MD,d Kenneth M. Boucher, PhD,c Douglas Grossman, MD, PhD,a,b,c Ronald M. Harris, MD, MBA,a,b Glen M. Bowen, MD,a,b and Sancy A. Leachman, MD, PhDa,b Salt Lake City, Utah, and Little Rock, Arkansas Background: Research on melanocytic nevi predominantly utilizes formalin-fixed, paraffin-embedded tissue, largely limiting research to morphologic and immunohistochemical observations. Withholding portions of fresh nevus tissue for molecular studies could result in the loss of important diagnostic material. Objective: This study prospectively evaluated melanocytic nevi for histologic homogeneity to determine if using a portion for research would have affected diagnosis. Methods: Thirty-three subjects were enrolled in a prospective study in which pigmented lesions were chosen for biopsy on a clinical basis. Lesions were sectioned and each piece submitted in a separate block (mean, 2.7; range 2-5 blocks per lesion). Slides from nevi were examined in two phases. In phase I, sections of nevi were randomized and a diagnosis was rendered for each section of nevus. In phase II, the dermatopathologist reviewed all slides for each nevus as a case, similar to the original interpretation of the lesion provided to the clinician. Diagnoses from phases I and II were compared with the original diagnosis. Results: Case material included 51 melanocytic lesions from 31 subjects. The phase I diagnosis matched the original diagnosis for 99 of 121 slides that showed a melanocytic lesion (82%). The phase II diagnosis matched the original diagnosis for 45 of 51 specimens (88%). Limitations: The study was limited by: a small number of specimens; the clinician could have chosen clinically homogeneous nevi for biopsy; effect of interobserver and intraobserver variability on diagnosis. Conclusions: For the majority of melanocytic nevi in this study, the diagnostic information present in one section of a melanocytic nevus could be extrapolated to the remainder of the specimen without adverse consequences from a diagnostic or therapeutic perspective. ( J Am Acad Dermatol 2008;59:814-21.)
INTRODUCTION Melanocytic nevi are among the most common cutaneous lesions undergoing biopsy and submitted for histopathologic analysis, and the importance of molecular study of melanocytic nevi and melanoma cannot be overstated. The dermatopathology archive From The Melanoma Program, Huntsman Cancer Institutea and Departments of Dermatologyb and Oncological Sciences,c University of Utah, Salt Lake City; and the Department of Pathology, University of Arkansas for Medical Sciences, Little Rock.d Supported by grants from The Skin Cancer Foundation (to S. R. F.), the Dermatology Foundation Leaders Society Dermatologist Investigator Research Fellowship and Clinical Career Development Award (to S. R. F.); National Institutes of Health grants K23 RR17525-01 (to S. R. F.) and R01 AR50102 (to D. G.), Doris Duke Charitable Foundation (to S. A. L.), Fellowship-To-Faculty Transition Award from the University of Utah funded in part by the Howard Hughes Medical Institute (to S. A. L.), The Huntsman Cancer Foundation (to S. A. L., D. G.), the Tom C. Mathews Jr. Familial Melanoma Research Endowment, Huntsman General
814
is an important and widely available source of tissues for research purposes since the majority of the specimen is usually retained in paraffin after a diagnosis is rendered. However, research using formalinfixed, paraffin-embedded tissues is largely limited to morphologic and immunohistochemical studies Clinical Research Center Public Health Service grant (MO1 RR00064), National Cancer Institute (NCI) Cancer Center Support grants 5P30CA420-14, and the Utah Cancer Registry, funded by Contract # N01-PC-35141 from the NCI with additional support from the Utah Department of Health and the University of Utah. Conflicts of interest: None declared. Reprint requests: Scott R. Florell, MD, Department of Dermatology, University of Utah Health Sciences Center, 4B454 SOM, Salt Lake City, UT 84132. E-mail: [email protected]. 0190-9622/$34.00 ª 2008 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2008.07.020
J AM ACAD DERMATOL
Florell et al 815
VOLUME 59, NUMBER 5
because of difficulties in analyzing DNA and RNA due to cross-linking of nucleic acids with protein.1-3 RNA also degrades quickly at room temperature in the surgical pathology laboratory,4 and culturing cells is not an option after fixation. Consequently (for cellular and molecular studies), obtaining a portion of fresh or flash-frozen tissue is preferable, although this could result in the loss of potentially important diagnostic material. In the pathology laboratory, histologic diagnosis of specimens is made based on microscopic examination of a small percentage of the tissue. The assumptions made in the standard method in which surgical tissues are processed are (1) the small fraction of the tissue examined microscopically is representative of the entire specimen and (2) the diagnosis or recommendation made from examination of this ‘‘representative’’ sample would not change if more sections were examined. Similarly, when a portion of tissue is retained for research purposes, the implication is that the examined portion of the tissue is sufficient for diagnosis. Our melanoma program is engaged in cellular and molecular studies on melanocytic nevi, but we are concerned about the potential impact of utilizing portions of nevi for research on diagnosis and patient care. We present data from a prospective study in which biopsied melanocytic nevi were thoroughly sectioned to (1) determine the histologic homogeneity of nevi; (2) determine whether the diagnosis of one section of a nevus is comparable to the diagnosis on other sections of the same nevus; (3) correlate clinical eccentric foci of pigmentation with histologic findings; and (4) determine if retaining a portion of the tissue for research purposes would have had a deleterious effect on diagnosis in this sample set.
METHODS Subjects This study was approved by the Institutional Review Board at the University of Utah (IRB No. 00-7916). During the period from March 2002 to April 2004, 33 subjects were enrolled in a prospective study in which pigmented lesions were chosen for biopsy on clinical grounds (ie, lesion falls out of the range of normal for that patient, including ABCDE criteria). The entire clinical lesion was removed by either saucerization shave or punch technique and the tissues were fixed in 10% buffered formalin. Examination and processing of tissue A dermatopathologist (S. R. F.) examined the fixed tissue. Punch biopsy specimens were bisected and each half was submitted in a separate cassette. Shave specimens were sectioned at 2-mm intervals,
perpendicular to the long axis, and each of the pieces of tissue were submitted in separate cassettes. If a pigmented lesion had a discrete pigmentary change clinically identified as a ‘‘dark dot,’’ the specimen was carefully inked along the deep margin beneath the ‘‘dark dot’’ as well as the epidermis overlying the area prior to sectioning and processing. The tissues were processed in the usual fashion. Four-micron sections from at least two levels of each block were applied to a glass microscopic slide and stained with hematoxylin and eosin. One of two board-certified dermatopathologists reviewed all of the slides for each specimen as part of their daily sign-out and prepared a diagnostic report. The clinician was notified of the results as usual. This report is referred to as the ‘‘original diagnosis.’’ The slides were subsequently examined in two phases (Fig 1). Phase I The purpose of phase I of the study was to determine whether the pathologic findings in a small part of the nevus were representative of the entire lesion. The diagnosis of the specimen was established from multiple blinded portions of the lesion and compared for consistency. As such, each specimen was at least bisected and each piece was submitted in a single block, resulting in a one slide for each piece of tissue. All of the slides were randomized and a diagnosis was rendered for each slide. The diagnostic categories were as follows: 1, benign ordinary intradermal or compound melanocytic nevus; 2, melanocytic nevus with mild or moderate architectural disorder and/or mild or moderate cytologic atypia; 3, melanocytic nevus with severe architectural disorder and/or severe cytologic atypia; 4, malignant melanoma (in-situ or invasive); 5, other melanocytic nevus (eg, Spitz nevus, blue nevus); and 6, no melanocytic lesion present. The dermatopathologist also rated each slide as benign (categories 1, 2, 5, and 6 above), atypical (category 3), or malignant (category 4). The dermatopathologist indicated whether a re-excision should be done (yes or no), usually for category 3 or 4 melanocytic lesions. The diagnosis for each slide, representing different portions of the same lesion, was compared with the original clinical dermatopathology report. Phase II In phase II, the dermatopathologist was given the complete, fully representational set of slides for each specimen that was reviewed in a similar fashion to the original diagnosis utilizing the same diagnostic categories. The dermatopathologist also indicated whether a re-excision should be done (yes or no). The dermatopathologist (blinded to the results of
816 Florell et al
J AM ACAD DERMATOL NOVEMBER 2008
Fig 1. Study design. Melanocytic nevi were sectioned into multiple pieces and were examined by one of two dermatopathologists as part of the regular workload for the day. A diagnostic report, the ‘‘original diagnosis,’’ was generated for each melanocytic lesion and the report was provided to the clinician as usual. After all subjects were enrolled in the study, the melanocytic lesions were examined in two phases by a single dermatopathologist and the phase I and II diagnoses were compared with the original diagnosis.
phase I) examined the slides 6 weeks after completing phase I. In this phase, the purpose was to prospectively evaluate whether removing a portion of the biopsy (eg, for research purposes) would have negatively impacted patient care. Case material A total of 33 subjects were enrolled from March 2002 through April 2004. Fifty-six specimens were collected (20 subjects had one specimen, 8 had two specimens, 1 had 3 specimens, 3 had 4 specimens, and 1 had 6 specimens). The entire clinical lesion was removed by shave saucerization (47 specimens), punch (7 specimens), or excision (1). One specimen was a re-excision.
RESULTS Fifty-six specimens were prospectively collected. Sixteen (29%) of the specimens were benign ordinary melanocytic nevi (category 1); 27 (48%) were melanocytic nevi with mild to moderate architectural disorder and/or focal mild or moderate cytologic atypia (category 2), and one melanocytic nevus (2%) was diagnosed as a melanocytic nevus with severe atypia (category 3). There was one excision for a highly suspect lesion that demonstrated invasive
melanoma arising in association with a melanocytic nevus (Clark level III, Breslow thickness 0.68 mm). The area was re-excised with 1-cm margins with no residual melanocytic lesion. There was one melanoma-in situ. Five lesions were classified as ‘‘other’’ melanocytic lesions (category 5): 3 Spitz nevi, 1 blue nevus, and 1 combined ordinary and blue nevus. One specimen showed features of regression without a residual melanocytic lesion. Three lesions were seborrheic keratoses. Only melanocytic nevi and malignant melanomas were included in the study; the re-excision specimen (of the invasive melanoma), 3 seborrheic keratoses, and lesion showing only regression were excluded. The case material used in this study included 31 subjects with 51 melanocytic lesions and a total of 140 glass microscopic slides. The mean (standard deviation) number of blocks per specimen was 2.7 (0.7), with a range of 2-5 (case material summarized in Table I). Of the 51 lesions, 42 were shave specimens (average number of levels per slide, 3.4; range, 2-5), 8 were punch specimens (average number of levels per slide, 3.6; range, 3-4), and one was an excision (each slide with 2 levels). Two dermatopathologists examined the slides initially and generated a diagnostic report; one read
Florell et al 817
J AM ACAD DERMATOL VOLUME 59, NUMBER 5
Table I. Summary of case material* Total No. No. in (%) study (%)
Subjects enrolled 33 Specimens collected 56 Benign melanocytic nevus, 16 (29) category 1 Melanocytic nevus with 27 (48) mild-moderate atypia, category 2 Melanocytic nevus with 1 (2) severe atypia, category 3 Spitz nevus, category 5 3 (6) Blue nevus, category 5 1 (2) Combined nevus (ordinary 1 (2) benign 1 blue), category 5 Invasive malignant melanoma, 1 (2) category 4 Re-excision of invasive melanoma 1 (2) Melanoma in situ, category 4 1 (2) Features of regression , no residual 1 (2) melanocytic lesion Seborrheic keratoses 3 (6) Total pigmented lesions 56 (100) Total melanocytic nevi included in study
31 (94) 51 (91) 16 (31) 27 (53) 1 (2) 3 (6) 1 (2) 1 (2) 1 (2)
1 (2)
51 (100)
*Only melanocytic nevi were included in the study.
32 of the specimens, the other read 19 of the specimens. Six of the original diagnostic reports recommended a re-excision of the melanocytic lesion (spindle cell nevus, n = 1; compound dysplastic nevus with mild or moderate atypia, n = 2; markedly atypical junctional lesion on sun-damaged skin, n = 1; melanoma in situ, n = 1; invasive melanoma, n = 1). Phase I In phase I, the diagnosis matched the original diagnosis for 99 of 140 slides (71%). Of the 41 discrepancies, 19 were identified as ‘‘no melanocytic lesion present;’’ these slides generally represented normal skin at the edges of the specimen. If these 19 slides are excluded, then the phase I diagnosis matched the original diagnosis for 99 of 121 slides (82%) that showed a pigmented lesion. The diagnosis was upgraded to a more atypical lesion in 17 slides (benign nevus [category 1] to nevus with mild or moderate atypia [category 2], n = 10; nevus with mild or moderate atypia [category 2] to severe atypia [category 3], n = 6; and nevus with severe atypia [category 3] to melanoma in situ [category 4], n = 1) and downgraded to a more benign diagnosis for 5 slides (nevus with mild-moderate atypia (category 2) to benign (category 1), n = 3; nevus with severe atypia (category 3) to nevus with mild-moderate atypia (category 2), n = 1; malignant melanoma
(category 4) to nevus with mild-moderate atypia (category 2), n = 1). When 3 diagnostic categories were considered and slides were diagnosed as benign (including nevi with architectural disorder or dysplastic features and mild to moderate atypia), atypical (nevi with architectural disorder or dysplastic features and severe atypia) or malignant, the diagnosis correlated with the original pathology report in 111 of 121 slides (92%). Of the 10 discrepancies, the diagnosis was upgraded for 7 of the slides (benign to atypical, n = 6, and atypical to melanoma in situ, n = 1, specimen 39) and the diagnosis was downgraded for 3 slides (atypical to benign, n = 2 and malignant to benign, n = 1, specimen #28, representing melanocytic nevus at the edge of an invasive melanoma, Fig 2). Phase I findings are summarized in Table II. A recommendation for re-excision was made for 14 slides involving 6 specimens which correlated well with the original pathology report with the exception of two slides upgraded to ‘‘severe atypia’’ (category 3) in phase I. Phase II In phase II, in which the slides were organized into cases, the phase II diagnosis matched the original diagnosis for 45 of 51 specimens (88%; k statistic 0.81, 95% confidence interval [CI] 0.66-0.96, good correlation5,6); the phase II diagnosis was upgraded to a more atypical lesion in 5 cases (benign nevus to nevus with mild or moderate atypia, n = 4; nevus with mild-moderate atypia to severe atypia, n = 1) and downgraded in 1 case (nevus with mildmoderate atypia to benign nevus). When 3 diagnostic categories were considered (benign, atypical, or malignant), the phase II diagnosis matched the original diagnosis for 50 of 51 specimens (98%; k statistic 0.85, 95% CI = 0.56-1.00, good correlation5,6) reflecting a category upgrade for the nevus with mild-moderate atypia (benign) to severe atypia (atypical). There were no significant differences in recommendation for re-excision. These findings are summarized in Table III. Variability of histology within specimens Of the 51 specimens, #28 (55%) demonstrated homogeneous histopathologic findings throughout and all of the sections were given the same diagnostic code. For example, if a nevus was divided into 3 pieces, each of the 3 pieces was diagnosed identically in phase I (Fig 3). Twelve of 51 specimens (24%) showed homogeneous diagnostic codes except for slides that did not show a melanocytic process. For example, if a nevus was divided into 3 pieces, two of the pieces were given the same diagnostic code and the third was ‘‘no melanocytic
818 Florell et al
J AM ACAD DERMATOL NOVEMBER 2008
Fig 2. Invasive malignant melanoma arising in association with a benign nevus. The lesion was found on the left upper back of a 73-year-old man. The clinical impression was ‘‘atypical nevus, rule-out melanoma.’’ Phase I diagnoses were not homogeneous: two sections showed invasive melanoma, one section showed benign nevus (category 2), and one section showed normal skin without a melanocytic lesion. This type of lesion should not be sampled for research purposes. Numbers in parentheses indicate diagnostic category.
lesion.’’ In 11 of 51 specimens (21%), the phase I diagnosis was not uniform. For example, one excision specimen (specimen #28) was sectioned into 4 pieces. When the slides of the 4 sections were randomized in phase I, two of the sections were diagnosed as invasive malignant melanoma (category 4), one section as a category 2 nevus, and one section with no melanocytic lesion (category 6). The phase II diagnosis for this lesion was invasive malignant melanoma (category 4) and this correlated with the original pathologic diagnosis (see Fig 2). Lesions with foci of eccentric pigmentation Twelve of the 51 pigmented lesions (24%) had a discrete pigmentary change clinically identified as a ‘‘dark dot.’’ Six of these foci of eccentric pigmentation corresponded histologically to keratinocyte pigmentation and discrete collections of melanophages within the superficial dermis. Two specimens demonstrated host response manifested by papillary dermal expansion, fibrosis, and melanophage deposition. One specimen each demonstrated pigmented junctional melanocytic nests, congenital pattern nevus with a deep penetrating component, follicular rupture, and superficial dermal melanocytic theques with prominent cytoplasmic pigmentation.
DISCUSSION In this study, we sought to determine whether retaining a portion of a melanocytic nevus for research purposes could have a deleterious effect on
histologic diagnosis and consequently, patient care and outcome. To our knowledge, published studies addressing this question are not available. Although there are vast libraries of archived tissue in private and academic dermatopathology practices, the potential experimental utility is limited since many contemporary investigative techniques are not optimal (or possible) on formalin-fixed tissue. In phase I, we found that one section of nevus was generally representative of the remainder of the specimen. Specifically, the diagnosis of a single portion of a nevus matched the original diagnosis in more than 80% of cases and more than 90% when only 3 diagnostic categories were considered (benign, atypical, or malignant). As such, most of the nevi in this study demonstrated homogeneous histologic features throughout the lesion. However, it is important to note that the diagnosis for 7 of the slides (5%) was upgraded to a more atypical diagnosis, and in two specimens additional re-excision was recommended in phase I that was not recommended in the original dermatopathology report. One specimen was upgraded from a severely atypical (category 3) lesion to melanoma in situ, though a recommendation for complete excision was made for each of the random slides for this lesion and this correlated with the original dermatopathology report. More problematic, however, was the invasive melanoma (see Fig 2), in which 2 of the 4 microscopic slides showed benign nevus and normal skin. Had the more atypical area been selected for research, an invasive
Florell et al 819
J AM ACAD DERMATOL VOLUME 59, NUMBER 5
Table II. Summary of phase I data*
Table III. Summary of phase II data No. (%)
Total slides Exact match with original diagnosis Slides without pigmented lesion (edges of specimen)y Slides with a pigmented lesion Exact match with original diagnosis Diagnosis upgraded to more atypical lesion Diagnosis downgraded to more benign lesion Three diagnostic categories: Benign, atypical, malignant Correlate with original diagnosis Diagnosis upgraded to more atypical lesion Diagnosis downgraded to more benign lesion
140 99/140 (71) 19 121 99/121 (82) 17/121 (14)
No. (%)
Total cases Exact match with original diagnosis Diagnosis upgraded to more atypical lesion Diagnosis downgraded to more benign lesion Three diagnostic categories: Benign, atypical, malignant Correlate with original diagnosis Diagnosis upgraded to more atypical lesion
51 45 (88) 5 (10) 1 (2)
50 (98) 1 (2)
5/121 (4)
111/121 (92) 7/121 (6) 3/121 (2)
*When 3 diagnostic categories are considered, melanocytic lesions are diagnosed as benign (category 1, 2, and 5 nevi), atypical (category 3 nevi), and malignant (category 4 in situ or invasive malignant melanoma). y These are slides that did not demonstrate a melanocytic lesion and generally represented normal skin at the edges of the specimen.
melanoma could have been missed. In sum, these results suggest that retaining a small portion of a melanocytic nevus for research purposes does not compromise diagnosis. In phase II, all of the slides for each case were reviewed together, similar to the way the lesion was diagnosed originally. The purpose of phase II was to determine the level of concordance with the way the lesion was diagnosed originally. If the concordance was poor, the phase I results would accordingly be less reliable. The results indicate good concordance5,6 with the original diagnosis. A clinical focus of eccentric hyperpigmentation was seen in 12 of the 49 specimens, all of which were associated with benign pathologic findings, most commonly keratinocyte pigmentation and melanophage deposition in the superficial dermis. The results are similar to those published by Bolognia, Lin, and Shapiro,7 although in their series of 59 melanocytic lesions with an eccentric focus of hyperpigmentation, 3 contained foci of melanoma arising within a melanocytic nevus. The relatively small number of biopsy specimens limits this study and multi-institutional expansion of the study would permit greater assurance of the safety of utilizing a portion of nevi for research. In addition, it is possible that the clinician chose clinically homogeneous nevi for biopsy rather than those with clinically atypical features, which could
skew the data in favor of histologic homogeneity; however, the clinicians submitting melanocytic lesions did not attempt to grade clinical homogeneity. Some of the diagnostic discrepancy between the original diagnosis and phase I and II diagnoses could be related to interobserver and intraobserver variability. This was likely minimized when fewer diagnostic categories were considered (ie, benign, atypical, and malignant). Because the study was not designed to be an interobserver concordance study and the known lack of interobserver concordance among expert pathologists in diagnosis of melanocytic lesions,8-12 only one dermatopathologist evaluated slides in phase I and phase II, similar to our experience that research collaborations involving sampling of melanocytic lesions usually enlist the assistance of one dermatopathologist. In summary, we found that, for the majority of specimens in this study, the diagnostic information present in one section of a melanocytic nevus could be extrapolated to the remainder of the specimen without adverse consequences to the patient from a diagnostic or therapeutic perspective. Broader issues are raised by the findings in this study. The first is that collaboration with colleagues in pathology and dermatopathology is critical for studies in which nevi are sampled for research purposes. A pathologist should be available to assist the clinician in the clinical evaluation of the patient and the lesion(s) to be sampled, then gross the specimen to adequately sample any areas that the pathologist deems important to evaluate. The second is whether histologic evaluation of one part of the lesion can be correlated with molecular findings in other parts of the specimen. According to our results, the histologic findings are usually homogeneous, suggesting that the molecular findings would also be similar, but this was not specifically studied here. The third issue is that in a small number of biopsies, the histologic findings were not homogeneous throughout the lesion and sampling part of a nevus in this setting could potentially affect patient care. There is a small,
820 Florell et al
J AM ACAD DERMATOL NOVEMBER 2008
Fig 3. Compound nevus with congenital features demonstrating homogeneous pathologic findings in all 4 sections. The lesion was found on the left side of the upper part of the neck of a 42-year-old woman. The clinical impression was ‘‘atypical nevus.’’ Phase I diagnoses were uniform and correlated with the phase II and original diagnosis.
but definite, risk of missing significant pathology when sampling melanocytic lesions for research purposes because such tissues cannot be used (in most cases) for additional sectioning and histologic examination. This risk can be minimized by carefully selecting lesions for biopsy and by rapid histopathologic examination of the tissue by the dermatopathologist at the time of biopsy. On the basis of the results of this study, our melanoma program has developed guidelines for selection, sampling, and processing of melanocytic nevi for research purposes (after obtaining Institutional Review Board [IRB] approval): 1. Select homogeneous lesions without significant asymmetry in border, color, or foci of eccentric pigmentation 2. Biopsy of ‘‘typical’’ atypical nevus—biopsy of a clinically atypical nevus is permissible when other similar-appearing nevi are present (ie, signature nevus) 3. When a clinically atypical nevus undergoes biopsy and there is return of pigment or the presence of a persistent pink macule or papule within the scar, a repeat biopsy is performed. 4. Clinically atypical lesions suspected of being malignant melanoma are submitted in their entirely for histopathologic analysis due to variation in histologic features unless express IRB approval has been obtained for collection for these lesions. Specimen #28 (Fig 2) was clinically
suspected of being malignant melanoma and thus would not have been a candidate for research sampling. In our program, following these guidelines, portions of clinically atypical nevi are held for potential research use until cleared by the dermatopathologist (with express patient consent). There are various methods for preserving tissue while awaiting histologic diagnosis.4 Usually, the remainder of the specimen is retained in tissue culture medium and/or RNA preservative. Then next morning, the routine hematoxylin-eosin profiles are reviewed by a dermatopathologist and a diagnostic report is issued. If there are any suspect histologic features, the retained portions of the nevus are submitted entirely for hematoxylin-eosin processing and not used for research. If the lesion demonstrates benign histologic features, the researcher may then use the remaining specimen for research purposes. We gratefully acknowledge the willing participation of all individuals in this study.
REFERENCES 1. Feldman MY. Reactions of nucleic acids and nucleoproteins with formaldehyde. Prog Nucleic Acid Res Mol Biol 1973;13: 1-49. 2. Moller K, Rinke J, Ross A, Buddle G, Brimacombe R. The use of formaldehyde in RNA-protein cross-linking studies with ribosomal subunits from Escherichia coli. Eur J Biochem 1977;76: 175-87.
Florell et al 821
J AM ACAD DERMATOL VOLUME 59, NUMBER 5
3. Werner M, Chott A, Fabiano A, Battifora H. Effect of formalin tissue fixation and processing on immunohistochemistry. Am J Surg Pathol 2000;24:1016-9. 4. Florell SR, Coffin CM, Holden JA, Zimmermann JW, Gerwels JW, Summers BK, et al. Preservation of RNA for functional genomic studies: a multidisciplinary tumor bank protocol. Mod Pathol 2001;14:116-28. 5. Kendall BS, Ronnett BM, Isacson C, Cho KR, Hedrick L, Diener-West M, Kurman RJ. Reproducibility of the diagnosis of endometrial hyperplasia, atypical hyperplasia, and welldifferentiated carcinoma. Am J Surg Pathol 1998;22:1012-9. 6. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74. 7. Bolognia JL, Lin A, Shapiro PE. The significance of eccentric foci of hyperpigmentation (’small dark dots’) within melanocytic nevi. Analysis of 59 cases. Arch Dermatol 1994;130:1013-7. 8. Corona R, Mele A, Amini M, De Rosa G, Coppola G, Piccardi P, et al. Interobserver variability on the histopathologic
9.
10.
11.
12.
diagnosis of cutaneous melanoma and other pigmented skin lesions. J Clin Oncol 1996;14:1218-23. Duncan LM, Berwick M, Bruijn JA, Byers HR, Mihm MC, Barnhill RL. Histopathologic recognition and grading of dysplastic melanocytic nevi: an interobserver agreement study. J Invest Dermatol 1993;100:318S-21S. Farmer ER, Gonin R, Hanna MP. Discordance in the histopathologic diagnosis of melanoma and melanocytic nevi between expert pathologists. Hum Pathol 1996;27:528-31. Florell SR, Boucher KM, Leachman SA, Azmi F, Harris RM, Malone JC, et al. Histopathologic recognition of involved margins of lentigo maligna excised by staged excision: an interobserver comparison study. Arch Dermatol 2003;139: 595-604. Wechsler J, Bastuji-Garin S, Spatz A, Bailly C, Cribier B, AndracMeyer L, et al. Reliability of the histopathologic diagnosis of malignant melanoma in childhood. Arch Dermatol 2002;138: 625-8.