Section 2 -

Angiotropic Melanoma: An Update Raymond L. Barnhill Departments of Dermatology and Pathology University of Miami Miller School of Medicine Miami, FL, USA

The propensity for melanoma to migrate along anatomical structures such as nerves (neurotropism) and skin appendages has been recognized as a common phenomenon for many years [1]. On the other hand this same capacity of melanoma to migrate along the external surfaces of blood vessels and lymphatics has received almost no attention in the literature. The origins of this potential mechanism of tumor dissemination were perhaps first noted in the 18th century French medical literature. In his original use of the term "metastasis" Recamier specifically referred to the spread of tumor cells along the external surfaces of vascular channels rather than within them [inadventently he was misquoted by future authors] [2, 3]. Subsequently in his historic paper laying the foundations for surgical margins for melanoma in 1907 W. Sampson Handley referred to Borst who had noted " the tendency of melanotic sarcoma to spread along the perivascular tissues immediately outside the blood-vessels" [4]. According to Borst this attraction of melanoma cells to blood vessels resulted from a "chemiotaxis", i.e., that blood was a necessary food for the production of melanin. Handley explained the phenomenon as anatomical since lymphatic vessels are in close proximity to arteries and veins, and he believed that melanoma initially disseminated by intra-lymphatic spread [4]. However, Handley's observations were based on the study of a single lymph node metastasis and the regional spread of tumor from that lymph node, rather than from a primary melanoma. In recent years Lugassy and Barnhill have proposed for the first time that an important mechanism of melanoma metastasis may be via this of migration of tumor cells along the external surfaces of vessels or "Extravascular migratory metastasis" [5, 6]. This hypothesized mechanism of tumor spread has been based on ultrastuctural and immunopathological studies; in the latter melanoma cells are closely apposed to the external surfaces of the endothelial cells of blood vessels. Ultrastructurally the melanoma cells are linked to the endothelium by an amorphous matrix confimed to contain laminin (not organized in a basement membrane) by immunohistochemistry. The latter morphological structure has been termed the "Angio-tumoral Complex" [7, 8, 9]. According to this proposed mechanism, tumors cells begin the process of local spread by competing with pericytes for the periendothelial position (of pericytes) or pericytic-like location for migration along the external surfaces of vessels. Lugassy and Barnhill have proposed that the histomorphological counterpart of the Angio-tumoral Complex is angiotropic melanoma [10]. Angiotropic melanoma has been referred to and reported anecdotally in the literature, more as a curiosity than as an important biological phenomenon [12, 13, 14]. However, the authors have recently drawn attention to the importance of angiotropism as a biological phenomenon and prognostic factor in localized melanoma and as the likely correlate of extravascular migratory metastasis [15, 16, 17]. In a recently published study of metastasing melanomas carefully matched with non-metastasizing melanomas for Breslow thickness, age, gender, and site, the presence of angiotropism strongly correlated with the development of metastases whereas vascular invasion was not observed in any cases [18]. Primary Angiotropic Melanoma We have recently reported a series of 36 primary melanomas with angiotropism [19]. Clinical features The clinical findings from the 36 patients are summarized in Table 1. There were 18 men and 18 women. The ages at the time of diagnosis ranged from 13 to 89 years with a mean and median of 47.2 and 48.5 years, respectively. Twelve tumors involved the trunk, 7 the head and neck, 7 the upper and 6 the lower extremities (non glabrous skin), 3 glabrous skin (the feet), and one the uterine cervix. In virtually all cases a nevus, atypical nevus, or melanoma was suspected clinically. No other clinical features were noteworthy. Morphological Findings The critical morphological feature for inclusion in the study was unequivocal angiotropism in the primary melanoma. Angiotropism was defined as follows: 1) clearly recognizable (unequivocal) melanoma cells closely opposed to (cuffing) the external surfaces of the endothelium of microvascular and/or lymphatic channels, either in linear array or in aggregates, 2) the latter occurring in at least two or more foci at the advancing front of the melanoma or in the nearby tissue (usually within 1 to 2 mm of the melanoma), and 3) no evidence of intravascular or intralymphatic tumor aggregates. If the findings were equivocal, the melanomas were not included in the study. The histopathological findings are summarized in Table 1. All conventional types of cutaneous melanoma were observed with pagetoid, lentiginous, and nested intraepidermal components, or "nodular" morphologies. The 35 cutaneous tumors ranged in Breslow thickness from 0.46 to 8.25 mm with a mean of 1.64 mm. One patient had a mucosal melanoma involving the uterine cervix with lentiginous intraepithelial melanoma component and extensive invasive component which measured 35 mm in thickness (not included with cutaneous melanomas). All cutaneous melanomas (31/35) were level IV with the exception of two that were level II and two that were level V. Six (17%) of the tumors also showed neurotropism. In all cases angiotropism was easily observed at the advancing front of the tumor or in nearby tissue. Small numbers of melanoma cells were closely associated with the external surfaces of the endothelium of microvascular channels and in some cases with lymphatic channels, often in a circumferential pattern. There was no evidence of vascular damage, intravasation, or intraluminal aggregates of tumor cells. Immunostaining with S 100 protein was carried out in eight melanomas and highlighted the angiotropism of melanoma cells. However, such immunostaining provided no distinct advantage over light microscopy. Although determining clinical outcomes was not an objective of this study, the patients had an average follow-up of 17.2 months (range 1 to 26 months) and disease-free follow-up of 15.1 months. Thirteen patients underwent sentinel lymph node biopsy and two of the 13 were positive for melanoma. Two patients presented simultaneously with a local micrometastasis and pulmonary metastases, respectively. Two additional patients developed regional lymph node metastases one year after initial diagnosis of melanoma, and one subsequently died with widespread metastases to the lungs, liver, and brain. Although angiotropism in malignant melanoma has been observed at least dating back to the publication of Handley in 1907, there have been only five cases subsequently reported in the literature to our knowledge outside of our studies on this phenomenon, including six angiotropic melanomas described in 2000 [10, 11, 12, 13, 14]. The latter 11 cases comprised 10 primary melanomas and one metastasis. Among their 45 cases of desmoplastic melanoma, Jain and Allen noted that some tumors demonstrated invasion of vascular walls but made no further observations about this finding [20]. The observations of the various authors in the latter reports are all consist with our current definition and understanding of angiotropism in melanoma [10]. This study delineates for the first time a large series of angiotropic melanomas, i.e., primary melanomas with tumor cells tropic for the pericytic position along the external surfaces of microvascular channels. Our findings also illustrate that angiotropism can be easily recognized by conventional microscopy in routine tissue sections, and suggest that this feature is probably more frequent than previously thought. Since histopathologists are currently not recording angiotropism in melanoma, they are not accustomed to actively searching for this finding; this may explain the sparse number of cases reported to date. In our studies we have been careful to distinguish angiotropism at the advancing front of the tumor or in nearby tissue, from the entrapment or engulfment of vessels within the main body of the melanoma. Although the latter feature may in fact constitute angiotropism, we have no means at present to verify if and when such entrapment is biologically significant or not. From these results and in our experience, angiotropism is observed much more frequently than vascular/lymphatic invasion (which was not found in any of our 36 cases). In a series of 650 consecutive invasive melanomas the prevalence of vascular/lymphatic invasion was only 1.4% [21]. Vascular and/or lymphatic invasion, i.e., the presence of tumor aggregates within the latter channels, has been thought to be a direct manifestation of metastasis in progress and accordingly a prognostic factor. In fact, the observation of vascular/lymphatic invasion in tissue sections is exceedingly rare as mentioned above and often an artifact which can be attributed to tortuous vascular channels folding back into the vascular lumina, tissue shrinkage resulting in the appearance of a (false) vascular space, or both. At least in some instances the tumor aggregates are external to the vascular/lymphatic channel rather than being intra-luminal, i.e., they are angiotropic. We consequently question the practical value of observing vascular invasion as a prognostic factor in cutaneous melanoma. We have recently investigated the prognostic significance of angiotropism as a qualitative variable, i.e., one that is simply observed as present or not, versus one that is quantified such as numbers of microvessels. A series of patients with primary cutaneous melanoma and documented metastasis were matched with a similar group of patients with non metastasizing primary melanoma for the latter study [18]. Angiotropism as defined above was observed histopathologically more much more frequently than vascular/lymphatic invasion in patients with metastasizing melanoma. Furthermore the results strongly suggested that angiotropism is an important prognostic factor correlating with metastasis. Recent experimental studies strongly suggested a correlation of angiotropism/pericytic location of melanoma cells with extravascular migratory metastasis (EVMM) [15, 16, 17]. With the description of the "Angio-Tumoral Complex" in melanoma in 1997 as mentioned in the Introduction, Lugassy et al were the first investigators to formally recognize and articulate the importance of angiotropism as a marker (and extravascular mechanism) of melanoma (and other solid tumor) metastasis [7]. EVMM proposed for melanoma has strong analogies with the migration of neoplastic glial invasion of the nervous system. Indeed it is known that invading glioma cells follow distinct anatomic structures such as the basement membranes of blood vessels [15]. We have observed a similar pericytic location of glioma and melanoma cells in human specimens [15]. We have also demonstrated that melanoma and glioma cells used in a murine brain tumor model exhibit the same pattern of invasion along the chorois plexus and along vessels [21]. In addition, using time-lapse video microscopy in cocultures of melanoma cells with capillary-like structures in vitro, we have observed the angiotropism of melanoma cells migrating toward and then along the external surface of the vascular tubules, occupying a pericyte-like location in a pattern analogous to the angio-tumoral complex [16]. In recent studies, we have observed the same phenomenon using ex vivo rat aortic rings, and in vivo melanoma cells growing on the chick chorioallantoic membrane [17]. These experiments have demonstrated the migration of angiotropic melanoma cells in a pericyte-like location along the vascular channels, supporting the concept of EVMM. Angiotropic Metastatic Melanoma Recently we have examined 26 melanoma metastases (including in transit, satellite, and epidermotropic metastases) for angiotropism [23]. Among 23 of the total 26 melanoma metastases studied, angiotropism of melanoma cells was observed in some portion of the metastasis. Thirteen out of 16 in transit metastases, all seven epidermotropic metastases, and two of three satellite metastases showed angiotropism. In general the in transit metastases were small and some were epidermotropic while others involved solely the reticular dermis and possibly the subcutaneous fat. Melanoma cells cuffed the external surfaces of microvessels within the metastasis, at the peritumoral interface of the metastasis, or in immediate proximity to the main portion of the metastasis, in a pattern analogous to the angio-tumoral complex. The melanoma cells were present in one or more layers and occasionally in small aggregates juxtaposed to the external vascular wall. There was no evidence of intravasation. Simultaneous (double) immunostaining of five specimens with S 100 protein and CD31 highlighted melanoma cell angiotropism, i.e., melanomas cells expressing S 100 were observed along the external surfaces of microvessels labeling with CD31. In conclusion, the spread of melanoma cells along vessels in human primary, in transit and other melanoma metastases, suggest that EVMM may be a mechanism by which some melanoma cells spread to nearby and even distant organs. Understanding the molecular basis of extravascular migratory metastasis may be of critical importance for the identification and evaluation of new therapeutic approaches to melanoma. References Barnhill RL: Pathology of melanocytic nevi and malignant melanoma. Boston: Butterworth. Heinemann, 1995. Recamier JCA : Recherches sur le traitement du cancer. 1829, Gabon, Paris. Lugassy C, Escande JP. The hematogenous theory of metastasis: Recamier did not propose it. Virchows Arch. 1997,431: 371 Handley WS. The pathology of melanotic growths in relation to their operative treatment. Lancet 11907;927: 996-998. Lugassy C, Barnhill RL, Christensen L. Melanoma and extravascular migratory metastasis. J Cutan Pathol. 2000 Oct;27(9):481. Barnhill RL. The biology of melanoma micrometastases. Recent Results Cancer Res. 2001;158:3-13. Lugassy C, BP Eyden, L Christensen and JP Escande. Angio-tumoral complex in human malignant melanoma characterised by free laminin: ultrastructural and immuno-histochemical observations. J Submicrosc Cytol Pathol 1997;29: 19-28. Lugassy C, Dickersin GR, Christensen L, Karaoli T, LeCharpentier M, Escande JP, Barnhill RL. Ultrastructural and immunohistochemical studies of the periendothelial matrix in human melanoma: evidence for an amorphous matrix containing laminin. J Cutan Pathol. 1999 Feb;26(2):78-83. Lugassy C, Shahsafaei A, Bonitz P, Busam KJ, Barnhill RL. Tumor microvessels in melanoma express the beta-2 chain of laminin. Implications for melanoma metastasis. J Cutan Pathol. 1999 May;26(5):222-6. Barnhill RL, Sagebiel RW, Lugassy C. Angiotropic malignant melanoma. Report of six additional cases. J Cutan Pathol 2000 ;27 :548. Moreno A, Espanol I, Ramogosa V. Angiotropic malignant melanoma. Report of two cases. J Cutan Pathol 1992 ;19 :325-329. Kerr S, Going JJ. Angiocentric invasion by lentigo maligna melanoma. J Clin Pathol 1994;47:183-184. Shea CR, Kline MA, Lugo J, McNutt NS. Angiotropic metastatic malignant melanoma. Am J Dermatopathol 1995;17:58-62. Saluja A, Money N, Zivoney DI, Solomon AR. Angiotropic malignant melanoma : A rare pattern of local metastases. J Am Acad Dermatol 2001;44:829-832. Lugassy C, Haroun RI, Brem H, Tyler BM, Jones RV, Fernandez PM, Patierno SR, Kleinman HK, Barnhill RL. Pericytic-like angiotropism of glioma and melanoma cells. Am J Dermatopathol. 2002 Dec;24(6):473-8. Lugassy C, Kleinman HK, Fernandez PM, Patierno SR, Webber MM, GhanemG, Spatz A, Barnhill RL. Human melanoma cell migration along capillary-like structures in vitro: A new dynamic model for studying extravascular migratory metastasis. J Invest Dermatol. 2002;119(3):703-4. Lugassy C, Kleinman HK, Engbring JA, Welch DR, Harms JF, Rufner R, Ghanem G, Patierno SR, Barnhill RL. Pericyte-like location of GFP-tagged melanoma cells: ex vivo and in vivo studies of extravascular migratory metastasis. Am J Pathol. 2004 164(4):1191-8. Barnhill RL, Dy K, Lugassy C. Angiotropism in Cutaneous Melanoma: A Prognostic Factor Strongly Predicting Risk for Metastasis. J Invest Dermatol 2002 119: 705-706. Barnhill RL, Lugassy C. Angiotropic malignant melanoma and extravascular migratory metastasis: description of 36 cases with emphasis on a new mechanism of tumour spread. Pathology. 2004 36:485. 90. Jain S, Allen PW. Desmoplastic malignant melanoma and its variants: A study of 45 cases. Am J Surg Pathol 1989;13:358-373. Barnhill RL, Fine J, Roush GC, Berwick M. Predicting five-year outcome from cutaneous melanoma in a population-based study. Cancer 78:427-432,1996. Wesseling P, Ruiter DJ, Burger PC. Angiogenesis in brain tumors; pathobiological and clinical aspects. J Neurooncol. 1997;32:253-65. Lugassy C, Vernon SE, Busam K, Engbring JA, Welch DR, Poulos EG, Kleinman HK, Barnhill RL. Angiotropism of human melanoma: Studies involving in transit and other cutaneous metastases and the chicken chorioallantoic membrane Implications for extravascular melanoma invasion and metastasis. Am J Dermatopathol 2006;28:187-193. Table 1. Clinical and Histopathological Characteristics of 36 Patients with Angiotropic Melanoma Case Age Sex Site Thickness Level SLN Metastases Status Follow-up (months) 1 38 F R. UPPER ARM 0.61 IV 18 2 69 M R. EAR 1.60 IV SLN- 18 3 17 F L. CHEEK 2.20 IV 16 4 31 F R. POST ARM 1.0 IV 15 5 47 F R. ANT THIGH 1.61 IV SLN- 18 6 51 M KNEE 0.71 IV 16 7 42 M R. SHOULDER 0.75 IV 20 8 49 M BUTTOCK 1.58 IV SLN- 16 9 14 M L. CHEST 3.13 IV SLN- 18 10 34 F L. FOOT 2.31 IV 18 11 16 M R.CHEST 1.25 IV 23 12 43 F LOWER LEG 1.47 IV SLN- 24 13 55 M R. ARM 0.46 II 26 14 81 F R. ARM 1.0 IV SLN- 15 15 37 F LOW BACK 1.31 IV 16 16 66 M ARM 0.92 IV 15 17 21 F L.THIGH 2.62 IV SLN- 24 18 18 F L. SCAPULA 0.54 IV 15 19 52 M L. LEG 0.84 IV 24 20 49 F CENTRAL BACK 3.21 IV SLN- Widespread Dead 12/02 24 21 89 M R. SIDE FACE 1.44 IV 16 22 54 M NECK 1.41 IV SLN- 16 23 60 M L. BACK 1.00 IV SLN- 18 24 69 M L. HEEL 0.88 IV 25 25 52 F R. POST SHOULDER 0.68 IV 21 26 71 M L. FOREARM 0.88 IV 21 27 39 M BACK 1.76 IV SLN+ 24 28 33 F R. LOW.BACK 0.84 IV 16 29 13 M BACK 2.33 IV SLN+ 24 30 71 M L. TEMP SCALP 2.11 IV Lungs 16 31 56 F R. ARM 0.78 II In-transit 2 32 27 F R. LEG 0.84 IV 3 33 89 F UTERUS 35 NA Micromet 1 34 48 F L. HEEL 8.25 V Regional LN 14 35 75 F NOSE 3.95 V 12 36 23 M L. TEMP SCALP 1.16 IV SLN- 11