Skin & Aging is proud to bring you this latest installment in its CME series. This series consists of regular CME activities that qualify you for two category 1 physician credit hours. As a reader of Skin & Aging, this course is brought to you free of charge — you aren’t required to pay a processing fee. Melanoma is a serious and frightening disease that is continuously increasing in incidence in the United States, particularly among Caucasian patients. In fact, it’s estimated that melanoma will be responsible for 7,600 deaths in the United States this year. This article reviews the diagnosis, staging and management of the disease. At the end of this article, you’ll find a five-question exam. Mark your responses in the designated area and fax the page to HMP Communications at (610) 560-0501. I hope this CME contributes to your clinical skills. Cordially, Steven R. Feldman, M.D. Ph.D. CME Editor Dr. Feldman is Professor of Dermatology, Pathology and Public Health Sciences at Wake Forest University Medical Center in Winston-Salem, NC. He’s also Director of the Center for Dermatology Research (funded by a grant from Galderma). I n the United States, the incidence of melanoma has more than tripled among Caucasians since 1980, bringing the estimated new cases for this year to 91,900.1 Melanoma, which has an intrinsic susceptibility to metastasize and is generally resistant to medical treatment, will be responsible for an estimated 7,600 deaths in the United States this year.1 Because of the increasing burden of this disease and its lethality, improved methods of early diagnosis and treatment of melanoma will be of increasing importance. However, the clinical diagnosis, staging, surgical management (in particular sentinel lymph node biopsy), and treatment of melanoma remain evolving, controversial areas of practice and research. Early diagnosis of melanoma should be able to significantly decrease the mortality and morbidity of this disease. Generally, early melanomas should be thinner in depth and have improved survival compared to their thicker counterparts. Increased surveillance and use of ABCD(E) criteria, photography, dermoscopy and computer technology have been advocated for improving early diagnosis of melanoma. Surveillance When patients were under surveillance because of a high risk of melanoma or for other skin problems, melanomas were found at an earlier, thinner state, with a better prognosis than those melanomas found at first patient encounter.2 Also, patients under surveillance due to a history of melanoma had their subsequent melanomas diagnosed at a thinner Breslow depth than their first melanoma.3 Masri et al. also found that surveillance and education can contribute to early diagnosis of melanoma in high-risk populations.4 Several organizations now advocate regular skin examinations for the general population or for high- risk patients.5,6 High-risk patients include those with multiple nevi, atypical nevi, history of many sunburns, excessive ultraviolet exposure, fair skin type, family history of melanoma, and/or a personal history of melanoma.7 The American Cancer Society recommends a monthly skin self-exam and a cancer-related check-up every 3 years for patients between the ages of 20 and 40 years and annually for those 40 and older.1 However, there's still controversy over what might constitute the most effective methods of melanoma detection, especially for the high-risk patient. The clinical diagnosis of melanoma is fraught with difficulty due to the overlap in appearance that often can occur between melanomas and other benign pigmented lesions. Also problematic is that high-risk patients often have a multitude of benign nevi, leading to a ?eedle in the haystack?phenomenon for finding a melanoma. When addressing these concerns, you must take into account the sensitivity and specificity of diagnosis when considering the value of new methodology. The sensitivity of diagnosis, which is the percentage of all melanomas diagnosed correctly, is extremely important to avoid ?issing?melanomas. On the other hand, the specificity of diagnosis, which is the proportion of non-melanoma pigmented lesions recognized as such, is important to minimize unnecessary excisions of benign lesions. The ABCDE Acronym In 1985, the acronym ABCD (asymmetry, border irregularity, color variegation and diameter generally greater than 6 mm), describing clinical features more commonly found in melanoma than benign nevi, was introduced to provide a framework for features to examine in any pigmented lesion.8 In 1998, a prospective clinical study of melanomas examined the accuracy of the ABCDE acronym with the addition of the criterion E, defined as enlargement based on the patient's description of the natural history of the lesion.9 The ABCDE criteria were found to have sensitivity of 89.3% and specificity of 65% when two or more of the criteria were required for diagnosis. Of note, the ABCDE acronym with E for evolution rather than enlargement has also been suggested10 and other diagnostic checklists such as the Glaskow seven-point checklist and the three-point checklist or 3 Cs (color, contour, change) also incorporate change as a criterion.11,12 Baseline Total Body Photography Since including enlargement in the criteria appears to improve accuracy of diagnosis, the question could be posed that use of baseline total body photography might provide a useful objective method of assessing change. While no outcome study has been done to analyze diagnostic accuracy with and without photography, melanomas have been found by change on baseline photos.4,13-17 In addition, melanomas arising unassociated with nevi have been found using total body photography.14-16 Kelly at al. found that two-thirds of melanomas in their series were de novo lesions, and remark that the predominance of de novo melanoma supports photographic monitoring rather than efforts at prophylactic excision.15 Grichnik et al. using total body photographs scanned into a digital format said they've done nine biopsies per one melanoma found,17 and Kelly et al. using full body photos noted they've done 10 biopsies per one melanoma detected.15 These numbers are less than might have been expected for the high-risk patients they were following if baseline photographs had not been available. However, due to current lack of automation for review of pictures and inconsistent insurance reimbursement, this procedure presently tends to be reserved for high-risk patients. Of note, recent gains in the resolution of digital cameras and ease of processing, storage and manipulation of digital images makes total body baseline photography more practical. Dermoscopy In addition to magnification, dermoscopy, due to decreased reflectance and refraction from the lesion, allows some visualization into the dermis. Dermoscopy, while it requires training, appears to improve the accuracy of melanoma diagnosis in pigmented skin lesions.18-21 Dermoscopic images can be analyzed by pattern analysis or by using one of five methods of clinical analysis.22-26 A dermoscopic image can be archived and saved for documentation, review and later comparison using a film camera with a modified lens, a digital camera with special attachment, or a video camera with a frame grabber. The criterion of evolution (either based on the patient? history or more objectively using archived images) can be incorporated into schemes of analysis to increase specificity and sensitivity.27-29 Kittler et al. found that 10% of moles substantially changing using dermoscopy were melanoma.29 Der-moscopy can be combined with clinical criteria, in which a lesion is excised if it meets either clinical ABCDE or dermoscopic criteria.27,30 Computer-Assisted Diagnosis An evolving area of research is the automatization of analysis of dermoscopic images. Automatized analysis of an image involves: • Segmentation. Separation of the lesion from the background. • Feature extraction. Identification and/or measurement of diagnostically important structures or characteristics. • Classification. Analysis of the data acquired during segmentation and feature extraction to determine if the lesion meets a threshold value to be clinically diagnosed as a melanoma. In a recent meta-analysis, the diagnostic accuracy achieved with computer diagnosis was not statistically different from that of human diagnosis.31 The diagnostic performance of the computer diagnosis was found to be better for studies using dermoscopic images than those using clinical images.31 However, keep in mind that these studies were done under experimental conditions and the real-world value is unknown.31 Several systems are currently under study.32-37 Systems are being developed with infrared reflectance, to allow for deeper penetration of light into the dermis.34,38 Confocal laser scanning microscopy is a new technology with high resolution that shows great promise for assisting in the diagnosis of pigmented lesions.39 However, while some of these systems are FDA approved for imaging, none are approved for diagnosis. Overall, surveillance of patients, including those coming to the office for other concerns, likely offers a means of diagnosis of melanomas at an early stage. As part of surveillance, ABCD clinical criterion are a common method of analyzing pigmented lesion and the additional E criterion for enlargement has been shown to increase accuracy of diagnosis. You may choose to utilize full body images, particularly for high-risk patients. Dermoscopy has been shown to improve the accuracy of diagnosis and may also be used in addition to clinical criteria or to monitor for change. In the future, automatization of dermoscopic image analysis, use of infrared reflectance, and confocal laser scanning microscopy of individual lesions will likely have increasing roles in the diagnosis of melanoma. Melanoma Staging After diagnosing melanoma, it? important to identify the stage that the disease is in so that the best course of treatment can be taken. Several pathological and clinical factors have been identified to be of prognostic significance in patients with cutaneous melanoma.40-42 • Tumor thickness and level of invasion. These have long been identified as significant prognostic factors, and have been incorporated into previous staging systems for this disease.43-46 Breslow was the first to describe the significance of tumor thickness, and since that time, this has been identified as the most significant prognostic factor for all relevant outcomes, including risk for relapse either locally or at distant sites, as well as survival.45 Level of invasion appears to be to be most significant as a prognostic factor for thin lesions (less than 1.0 mm).47 For the remaining tumor thickness subgroups, the level of invasion does not appear to add prognostic information to the overall tumor thickness.47-48 •Tumor ulceration. In contrast, the presence of tumor ulceration, as defined by the absence of an intact epidermis, has been shown to add additional prognostic information, regardless of tumor thickness.47,49 Tumor ulceration has been associated with aggressive biological behavior and increased potential for the development of both local and distant metastases.49 Incidence of ulceration increases with tumor thickness. However, when ulceration is identified, the overall prognosis for the individual patient is clearly worse than would have been predicted by the tumor thickness alone.47,49 •Age, gender and tumor location. The most common clinical prognostic factors include age, gender and tumor location.41 Both advanced age (older than 60 years of age) and male gender have been associated with poorer prognosis than either younger or female patients, respectively.41,50 High-risk tumor locations generally include the trunk, proximal upper extremities, neck and, particularly, the scalp. Distal upper extremities and lower extremities have been associated with lower risk for relapse and improved survival.41 Recently, the American Joint Committee on Cancer (AJCC) revised the staging system for melanoma.51 (See ?pdated AJCC Staging System for Cutaneous Melanoma,?above.) The new staging system has been validated through an analysis of a combined melanoma database, which included clinical, demographic and pathological data collected prospectively from 13 centers.47 In the multivariate analysis of patients with localized melanoma, tumor thickness (risk ratio 1.558), ulceration (risk ratio 1.90), age (risk ratio 1.1), site (trunk and head and neck versus extremities, risk ratio 1.34), level (risk ratio 1.21), and gender (female, risk ratio .84) were significant (p<0.00001, except gender, p=0.001) prognostic factors for survival. When the analysis was restricted to patients whose nodal status was defined pathologically (i.e. with a sentinel lymph node biopsy), the level of invasion was no longer significant. In a comparison of the level of invasion with ulceration, level of invasion was more predictive of survival only in thin (less than 1.0 mm) melanomas. In all remaining thickness groups (greater than 1.0 mm), the presence of ulceration was of greater prognostic significance. Prognostic Factors Once melanoma has spread to the local lymph node basin, the number of lymph nodes involved with a tumor has generally been identified as the most significant prognostic factor for survival.52 Characteristics of the primary tumor have also retained prognostic significance in the setting of lymph node metastases, including tumor thickness and ulceration. In the analysis of the AJCC Melanoma Database, the number of lymph nodes (risk ratio 1.26), tumor burden (microscopic versus macroscopic, risk ratio 1.79) and ulceration of the primary tumor (risk ratio 1.58) were identified as significant prognostic factors for survival (p< 0.0001).47 Presence of melanoma satellites around the primary tumor or the presence of in-transit metastases between the primary mel-anoma and the regional lymph node basin are also associated with a poor prognosis that is comparable to patients with metastases to local lymph nodes. Finally, once melanoma has spread to distant sites, both the location (visceral versus soft tissue) and number of different sites of metastases have been identified as significant prognostic factors for survival. In addition, the disease-free interval and stage before the development of distant metastases, as well as gender and abnormal serum levels of serum lactate dehydrogenase (LDH) and albumin, have also been identified as significant prognostic factors in multivariate analyses.53-55 In the AJCC Melanoma Database, non-visceral sites of metastatic disease (skin, subcutaneous, distant lymph nodes) were associated with significantly improved survival over lung (p=0.003) or other visceral sites of disease (p<0.0001).47 Sentinel Lymph Node Biopsy Management of the regional lymph nodes in cutaneous melanoma patients without clinical lymphadenopathy has been controversial. Some investigators have proposed removal of all the regional lymph nodes, elective lymph node dissection (ELND), for patients without evidence of palpable regional lymphadenopathy. Proponents of ELND cite the difference in survival for patients with histopathological node-positive disease undergoing ELND versus patients with palpable nodal disease undergoing therapeutic lymph node dissection,56 as well as nonrandomized trials suggesting improved survival after elective ELND.57,58 Opponents of this approach object to the significant morbidity of the operation and cite three randomized trials that show no survival benefit after elective ELND59-61 as reasons to pursue nodal observation for patients with clinically negative nodes, and they perform therapeutic lymph node dissection only for patients with clinically positive nodes. Fortunately, the controversy surrounding ELND was laid to rest with the development of a less morbid, more accurate, method to stage the regional nodal basin, intraoperative lymphatic mapping and sentinel lymphadenectomy (LM/SL).62 The Theory LM/SL The concept of lymphatic mapping began in the late 1970s as investigators sought ways to determine the lymphatic basin at risk for patients with truncal melanomas.63-65 Investigators injected various agents, colloidal gold,63 technetium-99m sulfur colloid,64 or technetium-99m antimony sulfur colloid,65 to identify the basin at risk. This nodal basin would then be surgically excised with an elective lymph node dissection. Although this approach was more precise in identifying the basin at risk, the patient still incurred the morbidity associated with removal of the entire nodal basin. Morton and colleagues66 reasoned that if there could be more precise lymphatic mapping performed, then this would perhaps obviate the need to remove all of the lymph nodes in the draining nodal basin. The concept of the sentinel node (SN) is predicated on the fact that the efferent lymphatic channel draining a primary tumor will lead directly to the first, or sentinel lymph node in the regional lymphatic basin. This lymphatic channel can carry malignant cells from a primary tumor to the SN. The tumor cells can then lodge in the subcapsular sinus of the lymph node and proliferate into a nodal metastasis. So the SN is the lymph node most likely to harbor metastatic disease if a regional nodal metastasis is present. Cabanas66,67 used the term ?entinel node?in penile cancer to indicate a node in a fixed anatomic location adjacent to the interior epigastric vein.66 The concept of an anatomically fixed SN doesn? apply to melanoma, breast cancer or other solid neoplasms where the precise anatomic location determines which lymphatic drainage pathways are utilized and, therefore, the site of the SN. Until techniques to study the physiology of lymphatic drainage were developed, it wasn? possible to perform lymphatic mapping and precisely identify the first lymph node receiving efferent lymphatic drainage. A feline model demonstrated the feasibility of intraoperative lymphatic mapping and sentinel lymphadenectomy using blue vital dyes63 and in 1985 the first clinical trial of lymphatic mapping and sentinel lymphadenectomy began. Using isosulfan blue dye, Morton et al. demonstrated that the lymphatic drainage of a primary melanoma could be predicted intraoperatively and used as a less invasive technique for nodal staging.62 All patients underwent LM/SL with isosulfan blue dye. If an SN was identified, it was sent to pathology as a separate specimen. The patient then underwent a complete regional nodal dissection. In this feasibility trial, 237 lymphatic basins were studied in 223 patients with clinical node-negative melanoma. (see Distribution of Metastases in SN and NSN.") A SN was identified in 82% (194/237) of nodal basins, and 225 (38/194) of SN contained metastases. Only 1% (2/194) of the basins had metastases in nonsentinel nodes (NSN) when the SN was tumor-free. When the SN was identified, it predicted the tumor status of the nodal basin in 99% of cases. Other groups have independently confirmed the biologic significance of intraoperative lymphatic mapping and sentinel lymphadenectomy. Alex and Krag were the first to describe lymphatic mapping with a radiopharmaceutical.68,69 However, as investigators became more sophisticated with the techniques, it was clear that a combination of both isosulfan blue dye and radioactive technetium sulfur colloid optimized the ability to identify the sentinel node. Albertini et al.70,71 demonstrated superiority of using dual agent mapping as opposed to either just the blue dye or radiopharmaceutical alone. It? generally believed that the mapping techniques are complimentary because lymphatic mapping with the blue dye is visual whereas the radiopharmaceutical is an auditory process. Identification rates using both agents approach 99%.70-73 Histopathologic Validation of LM/SL Early critics of LM/SL felt the technique was a sophisticated histopathologic technique (step sectioning, immunohistochemical stains) applied to the SNs allowing an increased accuracy rate of detecting nodal metastases.74-76 To validate the SN hypothesis histopathologically, sentinel and nonsentinel nodes must be examined with a similar and sensitive histopathologic technique. In the first published description of LM/SL in melanoma,62 each SN was evaluated by frozen-section examination using routine hematoxylin-eosin (H&E) staining and by permanent section examination using H&E and immunohistochemical (IHC) stains with antibodies to S100 protein and NKI/C3. Nonsentinel nodes were evaluated by permanent section examination using H&E and IHC with antibodies to S-100 protein, HMB-45, and NKI/C3. LM/SL performed in 194 lymphatic basins yielding 259 SNs (1.3/basin); completion lymphadenectomy yielded 3079 NSN from the same basins. Only 2 (0.06%) NSN had metastatic tumor when the SN in the same basin was tumor-free. Thus, the false-negative rate of the procedure was less than 1%. Sophisticated histopathologic techniques, including multiple sections and IHC of both the sentinel and nonsentinel nodes, support the concept that the SN is the lymph node most likely to harbor metastatic disease in melanoma patients who have nodal metastasis. Forty-seven of 259 (18%) SN had metastases as compared with only two of 3,079 (0.1%) NSN. This confirms that the predictive ability of the SN is not due to chance alone: Lymphatic metastasis in melanoma is an orderly process which can be identified through lymphatic mapping. Current Techniques for LM/SL • Patient Selection for LM/SL. Inclusion criteria for LM/SL for patients with invasive melanoma at the University of North Carolina at Chapel Hill are as follows: 1.) Invasive melanoma measuring at least Breslow thickness of at least 0.76 mm 2.) Breslow thickness less than 0.76 mm with evidence of ulceration regression or Clark Level IV or V 3.) Melanoma less than 0.76 mm without regression or ulceration or Clark Level IV or V are given a balanced discussion regarding the merits of lymphatic mapping and sentinel lymphadenectomy 4.) Patient entry into clinical trials. For patients with thin melanomas (less than 0.76 mm), we?e found the risk for nodal metastases to be 6.3%.87 By multivariate analysis there? no predictor (i.e anatomic location, ulceration etc.) of which patients with thin melanomas will harbor metastatic disease. We present these patients with a balanced discussion of the risks and benefits for LM/SL in their particular case. • Nuclear Medicine and Surgical Methods. LM/SL is scheduled after diagnostic biopsy but before wide excision and before any type of tissue rearrangement for repair that might disrupt lymphatic flow. LM/SL is performed using a combined technetium-99m sulfur colloid and isosulfan blue dye technique. Techniques for LM/SL combining blue dye (Lymphazurin) and technetium-99m sulfur colloid (Nicomed) have been described previously.70,71,76-78 Patients are brought to the Nuclear Medicine Department about 3 hours before surgery for a 450 µCi (±10%) injection of filtered technetium-99m sulfur colloid around the primary melanoma site or biopsy scar. Four injections are made circumferentially around the primary site. The radiocolloid injections must be given in the dermis instead of the subcutaneous tissue in order to map the dermal lymphatics, which is the network by which melanoma cells metastasize. In order to facilitate the lymphatic mapping process, patients undergo pre-operative lymphoscintigraphy in order to identify all basins at risk for disease. Dynamic scans of all lymphatic basins at risk for metastatic disease are performed 5 to 10 minutes after the injection. In our office, we use a large field-of-view gamma camera set at 20% window and fitted with a low-energy, high-resolution parallel hole collimator. It takes 20 minutes to 2 hours to obtain images. In the operating room, 0.5 ml to 1.5 ml of isosulfan blue dye is injected intradermally around the patient? primary melanoma or previous biopsy site. The area is then compressed for 3 minutes to 10 minutes to augment the lymphatic pump. Prior to skin incision, the area within the nodal basin with the greatest number of counts per second is located with a commercially available handheld gamma detection probe. An incision is made in this area. Careful dissection of the underlying tissue is performed until a blue-stained lymphatic channel is located. Once identified, the blue lymphatic channel is tracked proximally and distally until a blue node is located. This node is excised and labeled as the SN. If the blue-stained lymphatics are difficult to find or follow, the handheld probe is used to direct the dissection. Once the SN is excised, the ex vivo SN activity and the residual background activity in the nodal basin are then documented. If the ratio of the ex vivo SN counts per second to the background counts per second remains greater than 10:1 after removal of the SN, the dissection should be continued to identify and remove all additional SN. If hot spots have been identified in multiple nodal basins on pre-operative lymphoscintigraphy, we repeat the above procedure for each involved nodal basin. • Histopathologic Examination of Sentinel Nodes. The SN is first bivalved along the shortest axis. Each half is then rotated so that the cut surface faces up. The halves are placed adjacent to each other, and the node is then cut into 1-mm slices parallel to the cut surface. We send alternating slices to surgical pathology and the University of North Carolina Lineberger Comprehensive Cancer Center Tissue Procurement and Analysis Core Facility.79 The tissue submitted to surgical pathology is fixed in formalin, processed overnight and embedded in paraffin. Initial evaluation of each SN entails examination of a single H&E-stained slide from each 1-mm SN slice. If a metastasis is discovered during the initial H&E examination, no further sections are obtained from the SN. For patients with tumor-free SN on this initial H&E examination, additional levels are obtained and stained with S100 or MART-1 immunohistochemical stains. Since the detection of micrometastatic disease is difficult using just H&E stain, we rely on the combination of S100 and MART-1 to detect micrometastatic melanoma in the SN.
ADVERTISEMENT
Part 1: CME # 115 — June 2003 Advances and Challenges in Melanoma
Skin & Aging is proud to bring you this latest installment in its CME series. This series consists of regular CME activities that qualify you for two category 1 physician credit hours. As a reader of Skin & Aging, this course is brought to you free of charge — you aren’t required to pay a processing fee. Melanoma is a serious and frightening disease that is continuously increasing in incidence in the United States, particularly among Caucasian patients. In fact, it’s estimated that melanoma will be responsible for 7,600 deaths in the United States this year. This article reviews the diagnosis, staging and management of the disease. At the end of this article, you’ll find a five-question exam. Mark your responses in the designated area and fax the page to HMP Communications at (610) 560-0501. I hope this CME contributes to your clinical skills. Cordially, Steven R. Feldman, M.D. Ph.D. CME Editor Dr. Feldman is Professor of Dermatology, Pathology and Public Health Sciences at Wake Forest University Medical Center in Winston-Salem, NC. He’s also Director of the Center for Dermatology Research (funded by a grant from Galderma). I n the United States, the incidence of melanoma has more than tripled among Caucasians since 1980, bringing the estimated new cases for this year to 91,900.1 Melanoma, which has an intrinsic susceptibility to metastasize and is generally resistant to medical treatment, will be responsible for an estimated 7,600 deaths in the United States this year.1 Because of the increasing burden of this disease and its lethality, improved methods of early diagnosis and treatment of melanoma will be of increasing importance. However, the clinical diagnosis, staging, surgical management (in particular sentinel lymph node biopsy), and treatment of melanoma remain evolving, controversial areas of practice and research. Early diagnosis of melanoma should be able to significantly decrease the mortality and morbidity of this disease. Generally, early melanomas should be thinner in depth and have improved survival compared to their thicker counterparts. Increased surveillance and use of ABCD(E) criteria, photography, dermoscopy and computer technology have been advocated for improving early diagnosis of melanoma. Surveillance When patients were under surveillance because of a high risk of melanoma or for other skin problems, melanomas were found at an earlier, thinner state, with a better prognosis than those melanomas found at first patient encounter.2 Also, patients under surveillance due to a history of melanoma had their subsequent melanomas diagnosed at a thinner Breslow depth than their first melanoma.3 Masri et al. also found that surveillance and education can contribute to early diagnosis of melanoma in high-risk populations.4 Several organizations now advocate regular skin examinations for the general population or for high- risk patients.5,6 High-risk patients include those with multiple nevi, atypical nevi, history of many sunburns, excessive ultraviolet exposure, fair skin type, family history of melanoma, and/or a personal history of melanoma.7 The American Cancer Society recommends a monthly skin self-exam and a cancer-related check-up every 3 years for patients between the ages of 20 and 40 years and annually for those 40 and older.1 However, there's still controversy over what might constitute the most effective methods of melanoma detection, especially for the high-risk patient. The clinical diagnosis of melanoma is fraught with difficulty due to the overlap in appearance that often can occur between melanomas and other benign pigmented lesions. Also problematic is that high-risk patients often have a multitude of benign nevi, leading to a ?eedle in the haystack?phenomenon for finding a melanoma. When addressing these concerns, you must take into account the sensitivity and specificity of diagnosis when considering the value of new methodology. The sensitivity of diagnosis, which is the percentage of all melanomas diagnosed correctly, is extremely important to avoid ?issing?melanomas. On the other hand, the specificity of diagnosis, which is the proportion of non-melanoma pigmented lesions recognized as such, is important to minimize unnecessary excisions of benign lesions. The ABCDE Acronym In 1985, the acronym ABCD (asymmetry, border irregularity, color variegation and diameter generally greater than 6 mm), describing clinical features more commonly found in melanoma than benign nevi, was introduced to provide a framework for features to examine in any pigmented lesion.8 In 1998, a prospective clinical study of melanomas examined the accuracy of the ABCDE acronym with the addition of the criterion E, defined as enlargement based on the patient's description of the natural history of the lesion.9 The ABCDE criteria were found to have sensitivity of 89.3% and specificity of 65% when two or more of the criteria were required for diagnosis. Of note, the ABCDE acronym with E for evolution rather than enlargement has also been suggested10 and other diagnostic checklists such as the Glaskow seven-point checklist and the three-point checklist or 3 Cs (color, contour, change) also incorporate change as a criterion.11,12 Baseline Total Body Photography Since including enlargement in the criteria appears to improve accuracy of diagnosis, the question could be posed that use of baseline total body photography might provide a useful objective method of assessing change. While no outcome study has been done to analyze diagnostic accuracy with and without photography, melanomas have been found by change on baseline photos.4,13-17 In addition, melanomas arising unassociated with nevi have been found using total body photography.14-16 Kelly at al. found that two-thirds of melanomas in their series were de novo lesions, and remark that the predominance of de novo melanoma supports photographic monitoring rather than efforts at prophylactic excision.15 Grichnik et al. using total body photographs scanned into a digital format said they've done nine biopsies per one melanoma found,17 and Kelly et al. using full body photos noted they've done 10 biopsies per one melanoma detected.15 These numbers are less than might have been expected for the high-risk patients they were following if baseline photographs had not been available. However, due to current lack of automation for review of pictures and inconsistent insurance reimbursement, this procedure presently tends to be reserved for high-risk patients. Of note, recent gains in the resolution of digital cameras and ease of processing, storage and manipulation of digital images makes total body baseline photography more practical. Dermoscopy In addition to magnification, dermoscopy, due to decreased reflectance and refraction from the lesion, allows some visualization into the dermis. Dermoscopy, while it requires training, appears to improve the accuracy of melanoma diagnosis in pigmented skin lesions.18-21 Dermoscopic images can be analyzed by pattern analysis or by using one of five methods of clinical analysis.22-26 A dermoscopic image can be archived and saved for documentation, review and later comparison using a film camera with a modified lens, a digital camera with special attachment, or a video camera with a frame grabber. The criterion of evolution (either based on the patient? history or more objectively using archived images) can be incorporated into schemes of analysis to increase specificity and sensitivity.27-29 Kittler et al. found that 10% of moles substantially changing using dermoscopy were melanoma.29 Der-moscopy can be combined with clinical criteria, in which a lesion is excised if it meets either clinical ABCDE or dermoscopic criteria.27,30 Computer-Assisted Diagnosis An evolving area of research is the automatization of analysis of dermoscopic images. Automatized analysis of an image involves: • Segmentation. Separation of the lesion from the background. • Feature extraction. Identification and/or measurement of diagnostically important structures or characteristics. • Classification. Analysis of the data acquired during segmentation and feature extraction to determine if the lesion meets a threshold value to be clinically diagnosed as a melanoma. In a recent meta-analysis, the diagnostic accuracy achieved with computer diagnosis was not statistically different from that of human diagnosis.31 The diagnostic performance of the computer diagnosis was found to be better for studies using dermoscopic images than those using clinical images.31 However, keep in mind that these studies were done under experimental conditions and the real-world value is unknown.31 Several systems are currently under study.32-37 Systems are being developed with infrared reflectance, to allow for deeper penetration of light into the dermis.34,38 Confocal laser scanning microscopy is a new technology with high resolution that shows great promise for assisting in the diagnosis of pigmented lesions.39 However, while some of these systems are FDA approved for imaging, none are approved for diagnosis. Overall, surveillance of patients, including those coming to the office for other concerns, likely offers a means of diagnosis of melanomas at an early stage. As part of surveillance, ABCD clinical criterion are a common method of analyzing pigmented lesion and the additional E criterion for enlargement has been shown to increase accuracy of diagnosis. You may choose to utilize full body images, particularly for high-risk patients. Dermoscopy has been shown to improve the accuracy of diagnosis and may also be used in addition to clinical criteria or to monitor for change. In the future, automatization of dermoscopic image analysis, use of infrared reflectance, and confocal laser scanning microscopy of individual lesions will likely have increasing roles in the diagnosis of melanoma. Melanoma Staging After diagnosing melanoma, it? important to identify the stage that the disease is in so that the best course of treatment can be taken. Several pathological and clinical factors have been identified to be of prognostic significance in patients with cutaneous melanoma.40-42 • Tumor thickness and level of invasion. These have long been identified as significant prognostic factors, and have been incorporated into previous staging systems for this disease.43-46 Breslow was the first to describe the significance of tumor thickness, and since that time, this has been identified as the most significant prognostic factor for all relevant outcomes, including risk for relapse either locally or at distant sites, as well as survival.45 Level of invasion appears to be to be most significant as a prognostic factor for thin lesions (less than 1.0 mm).47 For the remaining tumor thickness subgroups, the level of invasion does not appear to add prognostic information to the overall tumor thickness.47-48 •Tumor ulceration. In contrast, the presence of tumor ulceration, as defined by the absence of an intact epidermis, has been shown to add additional prognostic information, regardless of tumor thickness.47,49 Tumor ulceration has been associated with aggressive biological behavior and increased potential for the development of both local and distant metastases.49 Incidence of ulceration increases with tumor thickness. However, when ulceration is identified, the overall prognosis for the individual patient is clearly worse than would have been predicted by the tumor thickness alone.47,49 •Age, gender and tumor location. The most common clinical prognostic factors include age, gender and tumor location.41 Both advanced age (older than 60 years of age) and male gender have been associated with poorer prognosis than either younger or female patients, respectively.41,50 High-risk tumor locations generally include the trunk, proximal upper extremities, neck and, particularly, the scalp. Distal upper extremities and lower extremities have been associated with lower risk for relapse and improved survival.41 Recently, the American Joint Committee on Cancer (AJCC) revised the staging system for melanoma.51 (See ?pdated AJCC Staging System for Cutaneous Melanoma,?above.) The new staging system has been validated through an analysis of a combined melanoma database, which included clinical, demographic and pathological data collected prospectively from 13 centers.47 In the multivariate analysis of patients with localized melanoma, tumor thickness (risk ratio 1.558), ulceration (risk ratio 1.90), age (risk ratio 1.1), site (trunk and head and neck versus extremities, risk ratio 1.34), level (risk ratio 1.21), and gender (female, risk ratio .84) were significant (p<0.00001, except gender, p=0.001) prognostic factors for survival. When the analysis was restricted to patients whose nodal status was defined pathologically (i.e. with a sentinel lymph node biopsy), the level of invasion was no longer significant. In a comparison of the level of invasion with ulceration, level of invasion was more predictive of survival only in thin (less than 1.0 mm) melanomas. In all remaining thickness groups (greater than 1.0 mm), the presence of ulceration was of greater prognostic significance. Prognostic Factors Once melanoma has spread to the local lymph node basin, the number of lymph nodes involved with a tumor has generally been identified as the most significant prognostic factor for survival.52 Characteristics of the primary tumor have also retained prognostic significance in the setting of lymph node metastases, including tumor thickness and ulceration. In the analysis of the AJCC Melanoma Database, the number of lymph nodes (risk ratio 1.26), tumor burden (microscopic versus macroscopic, risk ratio 1.79) and ulceration of the primary tumor (risk ratio 1.58) were identified as significant prognostic factors for survival (p< 0.0001).47 Presence of melanoma satellites around the primary tumor or the presence of in-transit metastases between the primary mel-anoma and the regional lymph node basin are also associated with a poor prognosis that is comparable to patients with metastases to local lymph nodes. Finally, once melanoma has spread to distant sites, both the location (visceral versus soft tissue) and number of different sites of metastases have been identified as significant prognostic factors for survival. In addition, the disease-free interval and stage before the development of distant metastases, as well as gender and abnormal serum levels of serum lactate dehydrogenase (LDH) and albumin, have also been identified as significant prognostic factors in multivariate analyses.53-55 In the AJCC Melanoma Database, non-visceral sites of metastatic disease (skin, subcutaneous, distant lymph nodes) were associated with significantly improved survival over lung (p=0.003) or other visceral sites of disease (p<0.0001).47 Sentinel Lymph Node Biopsy Management of the regional lymph nodes in cutaneous melanoma patients without clinical lymphadenopathy has been controversial. Some investigators have proposed removal of all the regional lymph nodes, elective lymph node dissection (ELND), for patients without evidence of palpable regional lymphadenopathy. Proponents of ELND cite the difference in survival for patients with histopathological node-positive disease undergoing ELND versus patients with palpable nodal disease undergoing therapeutic lymph node dissection,56 as well as nonrandomized trials suggesting improved survival after elective ELND.57,58 Opponents of this approach object to the significant morbidity of the operation and cite three randomized trials that show no survival benefit after elective ELND59-61 as reasons to pursue nodal observation for patients with clinically negative nodes, and they perform therapeutic lymph node dissection only for patients with clinically positive nodes. Fortunately, the controversy surrounding ELND was laid to rest with the development of a less morbid, more accurate, method to stage the regional nodal basin, intraoperative lymphatic mapping and sentinel lymphadenectomy (LM/SL).62 The Theory LM/SL The concept of lymphatic mapping began in the late 1970s as investigators sought ways to determine the lymphatic basin at risk for patients with truncal melanomas.63-65 Investigators injected various agents, colloidal gold,63 technetium-99m sulfur colloid,64 or technetium-99m antimony sulfur colloid,65 to identify the basin at risk. This nodal basin would then be surgically excised with an elective lymph node dissection. Although this approach was more precise in identifying the basin at risk, the patient still incurred the morbidity associated with removal of the entire nodal basin. Morton and colleagues66 reasoned that if there could be more precise lymphatic mapping performed, then this would perhaps obviate the need to remove all of the lymph nodes in the draining nodal basin. The concept of the sentinel node (SN) is predicated on the fact that the efferent lymphatic channel draining a primary tumor will lead directly to the first, or sentinel lymph node in the regional lymphatic basin. This lymphatic channel can carry malignant cells from a primary tumor to the SN. The tumor cells can then lodge in the subcapsular sinus of the lymph node and proliferate into a nodal metastasis. So the SN is the lymph node most likely to harbor metastatic disease if a regional nodal metastasis is present. Cabanas66,67 used the term ?entinel node?in penile cancer to indicate a node in a fixed anatomic location adjacent to the interior epigastric vein.66 The concept of an anatomically fixed SN doesn? apply to melanoma, breast cancer or other solid neoplasms where the precise anatomic location determines which lymphatic drainage pathways are utilized and, therefore, the site of the SN. Until techniques to study the physiology of lymphatic drainage were developed, it wasn? possible to perform lymphatic mapping and precisely identify the first lymph node receiving efferent lymphatic drainage. A feline model demonstrated the feasibility of intraoperative lymphatic mapping and sentinel lymphadenectomy using blue vital dyes63 and in 1985 the first clinical trial of lymphatic mapping and sentinel lymphadenectomy began. Using isosulfan blue dye, Morton et al. demonstrated that the lymphatic drainage of a primary melanoma could be predicted intraoperatively and used as a less invasive technique for nodal staging.62 All patients underwent LM/SL with isosulfan blue dye. If an SN was identified, it was sent to pathology as a separate specimen. The patient then underwent a complete regional nodal dissection. In this feasibility trial, 237 lymphatic basins were studied in 223 patients with clinical node-negative melanoma. (see Distribution of Metastases in SN and NSN.") A SN was identified in 82% (194/237) of nodal basins, and 225 (38/194) of SN contained metastases. Only 1% (2/194) of the basins had metastases in nonsentinel nodes (NSN) when the SN was tumor-free. When the SN was identified, it predicted the tumor status of the nodal basin in 99% of cases. Other groups have independently confirmed the biologic significance of intraoperative lymphatic mapping and sentinel lymphadenectomy. Alex and Krag were the first to describe lymphatic mapping with a radiopharmaceutical.68,69 However, as investigators became more sophisticated with the techniques, it was clear that a combination of both isosulfan blue dye and radioactive technetium sulfur colloid optimized the ability to identify the sentinel node. Albertini et al.70,71 demonstrated superiority of using dual agent mapping as opposed to either just the blue dye or radiopharmaceutical alone. It? generally believed that the mapping techniques are complimentary because lymphatic mapping with the blue dye is visual whereas the radiopharmaceutical is an auditory process. Identification rates using both agents approach 99%.70-73 Histopathologic Validation of LM/SL Early critics of LM/SL felt the technique was a sophisticated histopathologic technique (step sectioning, immunohistochemical stains) applied to the SNs allowing an increased accuracy rate of detecting nodal metastases.74-76 To validate the SN hypothesis histopathologically, sentinel and nonsentinel nodes must be examined with a similar and sensitive histopathologic technique. In the first published description of LM/SL in melanoma,62 each SN was evaluated by frozen-section examination using routine hematoxylin-eosin (H&E) staining and by permanent section examination using H&E and immunohistochemical (IHC) stains with antibodies to S100 protein and NKI/C3. Nonsentinel nodes were evaluated by permanent section examination using H&E and IHC with antibodies to S-100 protein, HMB-45, and NKI/C3. LM/SL performed in 194 lymphatic basins yielding 259 SNs (1.3/basin); completion lymphadenectomy yielded 3079 NSN from the same basins. Only 2 (0.06%) NSN had metastatic tumor when the SN in the same basin was tumor-free. Thus, the false-negative rate of the procedure was less than 1%. Sophisticated histopathologic techniques, including multiple sections and IHC of both the sentinel and nonsentinel nodes, support the concept that the SN is the lymph node most likely to harbor metastatic disease in melanoma patients who have nodal metastasis. Forty-seven of 259 (18%) SN had metastases as compared with only two of 3,079 (0.1%) NSN. This confirms that the predictive ability of the SN is not due to chance alone: Lymphatic metastasis in melanoma is an orderly process which can be identified through lymphatic mapping. Current Techniques for LM/SL • Patient Selection for LM/SL. Inclusion criteria for LM/SL for patients with invasive melanoma at the University of North Carolina at Chapel Hill are as follows: 1.) Invasive melanoma measuring at least Breslow thickness of at least 0.76 mm 2.) Breslow thickness less than 0.76 mm with evidence of ulceration regression or Clark Level IV or V 3.) Melanoma less than 0.76 mm without regression or ulceration or Clark Level IV or V are given a balanced discussion regarding the merits of lymphatic mapping and sentinel lymphadenectomy 4.) Patient entry into clinical trials. For patients with thin melanomas (less than 0.76 mm), we?e found the risk for nodal metastases to be 6.3%.87 By multivariate analysis there? no predictor (i.e anatomic location, ulceration etc.) of which patients with thin melanomas will harbor metastatic disease. We present these patients with a balanced discussion of the risks and benefits for LM/SL in their particular case. • Nuclear Medicine and Surgical Methods. LM/SL is scheduled after diagnostic biopsy but before wide excision and before any type of tissue rearrangement for repair that might disrupt lymphatic flow. LM/SL is performed using a combined technetium-99m sulfur colloid and isosulfan blue dye technique. Techniques for LM/SL combining blue dye (Lymphazurin) and technetium-99m sulfur colloid (Nicomed) have been described previously.70,71,76-78 Patients are brought to the Nuclear Medicine Department about 3 hours before surgery for a 450 µCi (±10%) injection of filtered technetium-99m sulfur colloid around the primary melanoma site or biopsy scar. Four injections are made circumferentially around the primary site. The radiocolloid injections must be given in the dermis instead of the subcutaneous tissue in order to map the dermal lymphatics, which is the network by which melanoma cells metastasize. In order to facilitate the lymphatic mapping process, patients undergo pre-operative lymphoscintigraphy in order to identify all basins at risk for disease. Dynamic scans of all lymphatic basins at risk for metastatic disease are performed 5 to 10 minutes after the injection. In our office, we use a large field-of-view gamma camera set at 20% window and fitted with a low-energy, high-resolution parallel hole collimator. It takes 20 minutes to 2 hours to obtain images. In the operating room, 0.5 ml to 1.5 ml of isosulfan blue dye is injected intradermally around the patient? primary melanoma or previous biopsy site. The area is then compressed for 3 minutes to 10 minutes to augment the lymphatic pump. Prior to skin incision, the area within the nodal basin with the greatest number of counts per second is located with a commercially available handheld gamma detection probe. An incision is made in this area. Careful dissection of the underlying tissue is performed until a blue-stained lymphatic channel is located. Once identified, the blue lymphatic channel is tracked proximally and distally until a blue node is located. This node is excised and labeled as the SN. If the blue-stained lymphatics are difficult to find or follow, the handheld probe is used to direct the dissection. Once the SN is excised, the ex vivo SN activity and the residual background activity in the nodal basin are then documented. If the ratio of the ex vivo SN counts per second to the background counts per second remains greater than 10:1 after removal of the SN, the dissection should be continued to identify and remove all additional SN. If hot spots have been identified in multiple nodal basins on pre-operative lymphoscintigraphy, we repeat the above procedure for each involved nodal basin. • Histopathologic Examination of Sentinel Nodes. The SN is first bivalved along the shortest axis. Each half is then rotated so that the cut surface faces up. The halves are placed adjacent to each other, and the node is then cut into 1-mm slices parallel to the cut surface. We send alternating slices to surgical pathology and the University of North Carolina Lineberger Comprehensive Cancer Center Tissue Procurement and Analysis Core Facility.79 The tissue submitted to surgical pathology is fixed in formalin, processed overnight and embedded in paraffin. Initial evaluation of each SN entails examination of a single H&E-stained slide from each 1-mm SN slice. If a metastasis is discovered during the initial H&E examination, no further sections are obtained from the SN. For patients with tumor-free SN on this initial H&E examination, additional levels are obtained and stained with S100 or MART-1 immunohistochemical stains. Since the detection of micrometastatic disease is difficult using just H&E stain, we rely on the combination of S100 and MART-1 to detect micrometastatic melanoma in the SN.
Skin & Aging is proud to bring you this latest installment in its CME series. This series consists of regular CME activities that qualify you for two category 1 physician credit hours. As a reader of Skin & Aging, this course is brought to you free of charge — you aren’t required to pay a processing fee. Melanoma is a serious and frightening disease that is continuously increasing in incidence in the United States, particularly among Caucasian patients. In fact, it’s estimated that melanoma will be responsible for 7,600 deaths in the United States this year. This article reviews the diagnosis, staging and management of the disease. At the end of this article, you’ll find a five-question exam. Mark your responses in the designated area and fax the page to HMP Communications at (610) 560-0501. I hope this CME contributes to your clinical skills. Cordially, Steven R. Feldman, M.D. Ph.D. CME Editor Dr. Feldman is Professor of Dermatology, Pathology and Public Health Sciences at Wake Forest University Medical Center in Winston-Salem, NC. He’s also Director of the Center for Dermatology Research (funded by a grant from Galderma). I n the United States, the incidence of melanoma has more than tripled among Caucasians since 1980, bringing the estimated new cases for this year to 91,900.1 Melanoma, which has an intrinsic susceptibility to metastasize and is generally resistant to medical treatment, will be responsible for an estimated 7,600 deaths in the United States this year.1 Because of the increasing burden of this disease and its lethality, improved methods of early diagnosis and treatment of melanoma will be of increasing importance. However, the clinical diagnosis, staging, surgical management (in particular sentinel lymph node biopsy), and treatment of melanoma remain evolving, controversial areas of practice and research. Early diagnosis of melanoma should be able to significantly decrease the mortality and morbidity of this disease. Generally, early melanomas should be thinner in depth and have improved survival compared to their thicker counterparts. Increased surveillance and use of ABCD(E) criteria, photography, dermoscopy and computer technology have been advocated for improving early diagnosis of melanoma. Surveillance When patients were under surveillance because of a high risk of melanoma or for other skin problems, melanomas were found at an earlier, thinner state, with a better prognosis than those melanomas found at first patient encounter.2 Also, patients under surveillance due to a history of melanoma had their subsequent melanomas diagnosed at a thinner Breslow depth than their first melanoma.3 Masri et al. also found that surveillance and education can contribute to early diagnosis of melanoma in high-risk populations.4 Several organizations now advocate regular skin examinations for the general population or for high- risk patients.5,6 High-risk patients include those with multiple nevi, atypical nevi, history of many sunburns, excessive ultraviolet exposure, fair skin type, family history of melanoma, and/or a personal history of melanoma.7 The American Cancer Society recommends a monthly skin self-exam and a cancer-related check-up every 3 years for patients between the ages of 20 and 40 years and annually for those 40 and older.1 However, there's still controversy over what might constitute the most effective methods of melanoma detection, especially for the high-risk patient. The clinical diagnosis of melanoma is fraught with difficulty due to the overlap in appearance that often can occur between melanomas and other benign pigmented lesions. Also problematic is that high-risk patients often have a multitude of benign nevi, leading to a ?eedle in the haystack?phenomenon for finding a melanoma. When addressing these concerns, you must take into account the sensitivity and specificity of diagnosis when considering the value of new methodology. The sensitivity of diagnosis, which is the percentage of all melanomas diagnosed correctly, is extremely important to avoid ?issing?melanomas. On the other hand, the specificity of diagnosis, which is the proportion of non-melanoma pigmented lesions recognized as such, is important to minimize unnecessary excisions of benign lesions. The ABCDE Acronym In 1985, the acronym ABCD (asymmetry, border irregularity, color variegation and diameter generally greater than 6 mm), describing clinical features more commonly found in melanoma than benign nevi, was introduced to provide a framework for features to examine in any pigmented lesion.8 In 1998, a prospective clinical study of melanomas examined the accuracy of the ABCDE acronym with the addition of the criterion E, defined as enlargement based on the patient's description of the natural history of the lesion.9 The ABCDE criteria were found to have sensitivity of 89.3% and specificity of 65% when two or more of the criteria were required for diagnosis. Of note, the ABCDE acronym with E for evolution rather than enlargement has also been suggested10 and other diagnostic checklists such as the Glaskow seven-point checklist and the three-point checklist or 3 Cs (color, contour, change) also incorporate change as a criterion.11,12 Baseline Total Body Photography Since including enlargement in the criteria appears to improve accuracy of diagnosis, the question could be posed that use of baseline total body photography might provide a useful objective method of assessing change. While no outcome study has been done to analyze diagnostic accuracy with and without photography, melanomas have been found by change on baseline photos.4,13-17 In addition, melanomas arising unassociated with nevi have been found using total body photography.14-16 Kelly at al. found that two-thirds of melanomas in their series were de novo lesions, and remark that the predominance of de novo melanoma supports photographic monitoring rather than efforts at prophylactic excision.15 Grichnik et al. using total body photographs scanned into a digital format said they've done nine biopsies per one melanoma found,17 and Kelly et al. using full body photos noted they've done 10 biopsies per one melanoma detected.15 These numbers are less than might have been expected for the high-risk patients they were following if baseline photographs had not been available. However, due to current lack of automation for review of pictures and inconsistent insurance reimbursement, this procedure presently tends to be reserved for high-risk patients. Of note, recent gains in the resolution of digital cameras and ease of processing, storage and manipulation of digital images makes total body baseline photography more practical. Dermoscopy In addition to magnification, dermoscopy, due to decreased reflectance and refraction from the lesion, allows some visualization into the dermis. Dermoscopy, while it requires training, appears to improve the accuracy of melanoma diagnosis in pigmented skin lesions.18-21 Dermoscopic images can be analyzed by pattern analysis or by using one of five methods of clinical analysis.22-26 A dermoscopic image can be archived and saved for documentation, review and later comparison using a film camera with a modified lens, a digital camera with special attachment, or a video camera with a frame grabber. The criterion of evolution (either based on the patient? history or more objectively using archived images) can be incorporated into schemes of analysis to increase specificity and sensitivity.27-29 Kittler et al. found that 10% of moles substantially changing using dermoscopy were melanoma.29 Der-moscopy can be combined with clinical criteria, in which a lesion is excised if it meets either clinical ABCDE or dermoscopic criteria.27,30 Computer-Assisted Diagnosis An evolving area of research is the automatization of analysis of dermoscopic images. Automatized analysis of an image involves: • Segmentation. Separation of the lesion from the background. • Feature extraction. Identification and/or measurement of diagnostically important structures or characteristics. • Classification. Analysis of the data acquired during segmentation and feature extraction to determine if the lesion meets a threshold value to be clinically diagnosed as a melanoma. In a recent meta-analysis, the diagnostic accuracy achieved with computer diagnosis was not statistically different from that of human diagnosis.31 The diagnostic performance of the computer diagnosis was found to be better for studies using dermoscopic images than those using clinical images.31 However, keep in mind that these studies were done under experimental conditions and the real-world value is unknown.31 Several systems are currently under study.32-37 Systems are being developed with infrared reflectance, to allow for deeper penetration of light into the dermis.34,38 Confocal laser scanning microscopy is a new technology with high resolution that shows great promise for assisting in the diagnosis of pigmented lesions.39 However, while some of these systems are FDA approved for imaging, none are approved for diagnosis. Overall, surveillance of patients, including those coming to the office for other concerns, likely offers a means of diagnosis of melanomas at an early stage. As part of surveillance, ABCD clinical criterion are a common method of analyzing pigmented lesion and the additional E criterion for enlargement has been shown to increase accuracy of diagnosis. You may choose to utilize full body images, particularly for high-risk patients. Dermoscopy has been shown to improve the accuracy of diagnosis and may also be used in addition to clinical criteria or to monitor for change. In the future, automatization of dermoscopic image analysis, use of infrared reflectance, and confocal laser scanning microscopy of individual lesions will likely have increasing roles in the diagnosis of melanoma. Melanoma Staging After diagnosing melanoma, it? important to identify the stage that the disease is in so that the best course of treatment can be taken. Several pathological and clinical factors have been identified to be of prognostic significance in patients with cutaneous melanoma.40-42 • Tumor thickness and level of invasion. These have long been identified as significant prognostic factors, and have been incorporated into previous staging systems for this disease.43-46 Breslow was the first to describe the significance of tumor thickness, and since that time, this has been identified as the most significant prognostic factor for all relevant outcomes, including risk for relapse either locally or at distant sites, as well as survival.45 Level of invasion appears to be to be most significant as a prognostic factor for thin lesions (less than 1.0 mm).47 For the remaining tumor thickness subgroups, the level of invasion does not appear to add prognostic information to the overall tumor thickness.47-48 •Tumor ulceration. In contrast, the presence of tumor ulceration, as defined by the absence of an intact epidermis, has been shown to add additional prognostic information, regardless of tumor thickness.47,49 Tumor ulceration has been associated with aggressive biological behavior and increased potential for the development of both local and distant metastases.49 Incidence of ulceration increases with tumor thickness. However, when ulceration is identified, the overall prognosis for the individual patient is clearly worse than would have been predicted by the tumor thickness alone.47,49 •Age, gender and tumor location. The most common clinical prognostic factors include age, gender and tumor location.41 Both advanced age (older than 60 years of age) and male gender have been associated with poorer prognosis than either younger or female patients, respectively.41,50 High-risk tumor locations generally include the trunk, proximal upper extremities, neck and, particularly, the scalp. Distal upper extremities and lower extremities have been associated with lower risk for relapse and improved survival.41 Recently, the American Joint Committee on Cancer (AJCC) revised the staging system for melanoma.51 (See ?pdated AJCC Staging System for Cutaneous Melanoma,?above.) The new staging system has been validated through an analysis of a combined melanoma database, which included clinical, demographic and pathological data collected prospectively from 13 centers.47 In the multivariate analysis of patients with localized melanoma, tumor thickness (risk ratio 1.558), ulceration (risk ratio 1.90), age (risk ratio 1.1), site (trunk and head and neck versus extremities, risk ratio 1.34), level (risk ratio 1.21), and gender (female, risk ratio .84) were significant (p<0.00001, except gender, p=0.001) prognostic factors for survival. When the analysis was restricted to patients whose nodal status was defined pathologically (i.e. with a sentinel lymph node biopsy), the level of invasion was no longer significant. In a comparison of the level of invasion with ulceration, level of invasion was more predictive of survival only in thin (less than 1.0 mm) melanomas. In all remaining thickness groups (greater than 1.0 mm), the presence of ulceration was of greater prognostic significance. Prognostic Factors Once melanoma has spread to the local lymph node basin, the number of lymph nodes involved with a tumor has generally been identified as the most significant prognostic factor for survival.52 Characteristics of the primary tumor have also retained prognostic significance in the setting of lymph node metastases, including tumor thickness and ulceration. In the analysis of the AJCC Melanoma Database, the number of lymph nodes (risk ratio 1.26), tumor burden (microscopic versus macroscopic, risk ratio 1.79) and ulceration of the primary tumor (risk ratio 1.58) were identified as significant prognostic factors for survival (p< 0.0001).47 Presence of melanoma satellites around the primary tumor or the presence of in-transit metastases between the primary mel-anoma and the regional lymph node basin are also associated with a poor prognosis that is comparable to patients with metastases to local lymph nodes. Finally, once melanoma has spread to distant sites, both the location (visceral versus soft tissue) and number of different sites of metastases have been identified as significant prognostic factors for survival. In addition, the disease-free interval and stage before the development of distant metastases, as well as gender and abnormal serum levels of serum lactate dehydrogenase (LDH) and albumin, have also been identified as significant prognostic factors in multivariate analyses.53-55 In the AJCC Melanoma Database, non-visceral sites of metastatic disease (skin, subcutaneous, distant lymph nodes) were associated with significantly improved survival over lung (p=0.003) or other visceral sites of disease (p<0.0001).47 Sentinel Lymph Node Biopsy Management of the regional lymph nodes in cutaneous melanoma patients without clinical lymphadenopathy has been controversial. Some investigators have proposed removal of all the regional lymph nodes, elective lymph node dissection (ELND), for patients without evidence of palpable regional lymphadenopathy. Proponents of ELND cite the difference in survival for patients with histopathological node-positive disease undergoing ELND versus patients with palpable nodal disease undergoing therapeutic lymph node dissection,56 as well as nonrandomized trials suggesting improved survival after elective ELND.57,58 Opponents of this approach object to the significant morbidity of the operation and cite three randomized trials that show no survival benefit after elective ELND59-61 as reasons to pursue nodal observation for patients with clinically negative nodes, and they perform therapeutic lymph node dissection only for patients with clinically positive nodes. Fortunately, the controversy surrounding ELND was laid to rest with the development of a less morbid, more accurate, method to stage the regional nodal basin, intraoperative lymphatic mapping and sentinel lymphadenectomy (LM/SL).62 The Theory LM/SL The concept of lymphatic mapping began in the late 1970s as investigators sought ways to determine the lymphatic basin at risk for patients with truncal melanomas.63-65 Investigators injected various agents, colloidal gold,63 technetium-99m sulfur colloid,64 or technetium-99m antimony sulfur colloid,65 to identify the basin at risk. This nodal basin would then be surgically excised with an elective lymph node dissection. Although this approach was more precise in identifying the basin at risk, the patient still incurred the morbidity associated with removal of the entire nodal basin. Morton and colleagues66 reasoned that if there could be more precise lymphatic mapping performed, then this would perhaps obviate the need to remove all of the lymph nodes in the draining nodal basin. The concept of the sentinel node (SN) is predicated on the fact that the efferent lymphatic channel draining a primary tumor will lead directly to the first, or sentinel lymph node in the regional lymphatic basin. This lymphatic channel can carry malignant cells from a primary tumor to the SN. The tumor cells can then lodge in the subcapsular sinus of the lymph node and proliferate into a nodal metastasis. So the SN is the lymph node most likely to harbor metastatic disease if a regional nodal metastasis is present. Cabanas66,67 used the term ?entinel node?in penile cancer to indicate a node in a fixed anatomic location adjacent to the interior epigastric vein.66 The concept of an anatomically fixed SN doesn? apply to melanoma, breast cancer or other solid neoplasms where the precise anatomic location determines which lymphatic drainage pathways are utilized and, therefore, the site of the SN. Until techniques to study the physiology of lymphatic drainage were developed, it wasn? possible to perform lymphatic mapping and precisely identify the first lymph node receiving efferent lymphatic drainage. A feline model demonstrated the feasibility of intraoperative lymphatic mapping and sentinel lymphadenectomy using blue vital dyes63 and in 1985 the first clinical trial of lymphatic mapping and sentinel lymphadenectomy began. Using isosulfan blue dye, Morton et al. demonstrated that the lymphatic drainage of a primary melanoma could be predicted intraoperatively and used as a less invasive technique for nodal staging.62 All patients underwent LM/SL with isosulfan blue dye. If an SN was identified, it was sent to pathology as a separate specimen. The patient then underwent a complete regional nodal dissection. In this feasibility trial, 237 lymphatic basins were studied in 223 patients with clinical node-negative melanoma. (see Distribution of Metastases in SN and NSN.") A SN was identified in 82% (194/237) of nodal basins, and 225 (38/194) of SN contained metastases. Only 1% (2/194) of the basins had metastases in nonsentinel nodes (NSN) when the SN was tumor-free. When the SN was identified, it predicted the tumor status of the nodal basin in 99% of cases. Other groups have independently confirmed the biologic significance of intraoperative lymphatic mapping and sentinel lymphadenectomy. Alex and Krag were the first to describe lymphatic mapping with a radiopharmaceutical.68,69 However, as investigators became more sophisticated with the techniques, it was clear that a combination of both isosulfan blue dye and radioactive technetium sulfur colloid optimized the ability to identify the sentinel node. Albertini et al.70,71 demonstrated superiority of using dual agent mapping as opposed to either just the blue dye or radiopharmaceutical alone. It? generally believed that the mapping techniques are complimentary because lymphatic mapping with the blue dye is visual whereas the radiopharmaceutical is an auditory process. Identification rates using both agents approach 99%.70-73 Histopathologic Validation of LM/SL Early critics of LM/SL felt the technique was a sophisticated histopathologic technique (step sectioning, immunohistochemical stains) applied to the SNs allowing an increased accuracy rate of detecting nodal metastases.74-76 To validate the SN hypothesis histopathologically, sentinel and nonsentinel nodes must be examined with a similar and sensitive histopathologic technique. In the first published description of LM/SL in melanoma,62 each SN was evaluated by frozen-section examination using routine hematoxylin-eosin (H&E) staining and by permanent section examination using H&E and immunohistochemical (IHC) stains with antibodies to S100 protein and NKI/C3. Nonsentinel nodes were evaluated by permanent section examination using H&E and IHC with antibodies to S-100 protein, HMB-45, and NKI/C3. LM/SL performed in 194 lymphatic basins yielding 259 SNs (1.3/basin); completion lymphadenectomy yielded 3079 NSN from the same basins. Only 2 (0.06%) NSN had metastatic tumor when the SN in the same basin was tumor-free. Thus, the false-negative rate of the procedure was less than 1%. Sophisticated histopathologic techniques, including multiple sections and IHC of both the sentinel and nonsentinel nodes, support the concept that the SN is the lymph node most likely to harbor metastatic disease in melanoma patients who have nodal metastasis. Forty-seven of 259 (18%) SN had metastases as compared with only two of 3,079 (0.1%) NSN. This confirms that the predictive ability of the SN is not due to chance alone: Lymphatic metastasis in melanoma is an orderly process which can be identified through lymphatic mapping. Current Techniques for LM/SL • Patient Selection for LM/SL. Inclusion criteria for LM/SL for patients with invasive melanoma at the University of North Carolina at Chapel Hill are as follows: 1.) Invasive melanoma measuring at least Breslow thickness of at least 0.76 mm 2.) Breslow thickness less than 0.76 mm with evidence of ulceration regression or Clark Level IV or V 3.) Melanoma less than 0.76 mm without regression or ulceration or Clark Level IV or V are given a balanced discussion regarding the merits of lymphatic mapping and sentinel lymphadenectomy 4.) Patient entry into clinical trials. For patients with thin melanomas (less than 0.76 mm), we?e found the risk for nodal metastases to be 6.3%.87 By multivariate analysis there? no predictor (i.e anatomic location, ulceration etc.) of which patients with thin melanomas will harbor metastatic disease. We present these patients with a balanced discussion of the risks and benefits for LM/SL in their particular case. • Nuclear Medicine and Surgical Methods. LM/SL is scheduled after diagnostic biopsy but before wide excision and before any type of tissue rearrangement for repair that might disrupt lymphatic flow. LM/SL is performed using a combined technetium-99m sulfur colloid and isosulfan blue dye technique. Techniques for LM/SL combining blue dye (Lymphazurin) and technetium-99m sulfur colloid (Nicomed) have been described previously.70,71,76-78 Patients are brought to the Nuclear Medicine Department about 3 hours before surgery for a 450 µCi (±10%) injection of filtered technetium-99m sulfur colloid around the primary melanoma site or biopsy scar. Four injections are made circumferentially around the primary site. The radiocolloid injections must be given in the dermis instead of the subcutaneous tissue in order to map the dermal lymphatics, which is the network by which melanoma cells metastasize. In order to facilitate the lymphatic mapping process, patients undergo pre-operative lymphoscintigraphy in order to identify all basins at risk for disease. Dynamic scans of all lymphatic basins at risk for metastatic disease are performed 5 to 10 minutes after the injection. In our office, we use a large field-of-view gamma camera set at 20% window and fitted with a low-energy, high-resolution parallel hole collimator. It takes 20 minutes to 2 hours to obtain images. In the operating room, 0.5 ml to 1.5 ml of isosulfan blue dye is injected intradermally around the patient? primary melanoma or previous biopsy site. The area is then compressed for 3 minutes to 10 minutes to augment the lymphatic pump. Prior to skin incision, the area within the nodal basin with the greatest number of counts per second is located with a commercially available handheld gamma detection probe. An incision is made in this area. Careful dissection of the underlying tissue is performed until a blue-stained lymphatic channel is located. Once identified, the blue lymphatic channel is tracked proximally and distally until a blue node is located. This node is excised and labeled as the SN. If the blue-stained lymphatics are difficult to find or follow, the handheld probe is used to direct the dissection. Once the SN is excised, the ex vivo SN activity and the residual background activity in the nodal basin are then documented. If the ratio of the ex vivo SN counts per second to the background counts per second remains greater than 10:1 after removal of the SN, the dissection should be continued to identify and remove all additional SN. If hot spots have been identified in multiple nodal basins on pre-operative lymphoscintigraphy, we repeat the above procedure for each involved nodal basin. • Histopathologic Examination of Sentinel Nodes. The SN is first bivalved along the shortest axis. Each half is then rotated so that the cut surface faces up. The halves are placed adjacent to each other, and the node is then cut into 1-mm slices parallel to the cut surface. We send alternating slices to surgical pathology and the University of North Carolina Lineberger Comprehensive Cancer Center Tissue Procurement and Analysis Core Facility.79 The tissue submitted to surgical pathology is fixed in formalin, processed overnight and embedded in paraffin. Initial evaluation of each SN entails examination of a single H&E-stained slide from each 1-mm SN slice. If a metastasis is discovered during the initial H&E examination, no further sections are obtained from the SN. For patients with tumor-free SN on this initial H&E examination, additional levels are obtained and stained with S100 or MART-1 immunohistochemical stains. Since the detection of micrometastatic disease is difficult using just H&E stain, we rely on the combination of S100 and MART-1 to detect micrometastatic melanoma in the SN.
