The British Association of Plastic Surgeons (2005) 58, 348 352 GLUT-1: an extra diagnostic tool to differentiate between haemangiomas and vascular malformations q J. Leon-Villapalos a, K. Wolfe b, L. Kangesu a, * a Department of Plastic and Reconstructive Surgery, St Andrews Centre for Plastic Surgery and Burns, Court Road, Broomfield, Chelmsford CM1 7ET, UK b Department of Histopathology, Southend General Hospital, Prittlewell Chase, Westcliff-on-Sea, Essex SSO ORY, UK Received 22 October 2003; accepted 27 May 2004 KEYWORDS GLUT-1; Haemangiomas; Vascular malformations; Vascular anomalies Summary The differential diagnosis between juvenile haemangiomas, vascular malformations, pyogenic granulomas and normally proliferative endothelium (granulation tissue) on the basis of histology alone is sometimes difficult. This is important because haemangiomas, are self-limiting and vascular malformations are not. We report our experience of using the immunohistochemical marker GLUT-1 to distinguish haemangiomas from vascular malformations following the initial report by North and Colleagues (1998). We studied a total of 50 specimens from patients with vascular anomalies, and found that GLUT-1 reactivity was positive in 18 out of 19 juvenile haemangiomas, negative in two out of two noninvoluting congenital haemangiomas (NICH) and negative in 29 out of 29 vascular malformations, that included capillary malformations, lymphatic malformations, venous malformations and arteriovenous malformations (95% sensitivity, 100% specificity). Pyogenic granulomas ðn ¼ 4Þ and granulation tissue samples ðn ¼ 4Þ were used as negative controls. Placenta tissue was used as positive control. GLUT-1 accurately distinguishes haemangiomas from vascular malformations, and as a result from this work, we use this technique in routine histopathological differentiation of vascular anomalies. Q 2004 The British Association of Plastic Surgeons. Published by Elsevier Ltd. All rights reserved. The seminal study from Mulliken and Glowacki 1 categorised vascular anomalies as either juvenile haemangiomas or vascular malformations based on q This work presented at the 14th International Workshop on Vascular Anomalies of the International Society for the Study of Vascular Anomalies (ISSVA) Nijmegan, Netherlands, 2002. *Corresponding author. Tel.: þ44-1245-516229; fax: þ44-1245-516132. E-mail address: loshan.kangesu@meht.nhs.uk their cellular characteristics, and attempted to provide a classification of vascular lesions that could act as a guide for diagnosis, management, and further research. Currently, the term benign vascular tumours is preferred to include juvenile haemangioma but also other lesions which were considered previously to be atypical, but are now recognized as separate entities (see below). 3 Accurate diagnosis of haemangiomas and vascular S0007-1226/$ - see front matter Q 2004 The British Association of Plastic Surgeons. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.bjps.2004.05.029
GLUT-1: an extra diagnostic tool to differentiate between haemangiomas and vascular malformations 349 malformations is sometimes a challenge for the clinician and the pathologist. This has significant implications for the patient s management, as the natural history of the two lesions, is quite different. 3 Also, there can be confusion of vascular anomalies with vascular tumours (benign or malignant). Vascular malformations (capillary, lymphatic, venous and arteriovenous) are permanent malformations of the vascular system present at birth, whereas juvenile haemangiomas are common tumours of infancy (not present at birth) that have three characteristic stages: a rapidly proliferating stage during the first year of life; a prolonged involuting stage lasting till the age of four generally; and a final involuted stage. The works of Enjolras 5 and Mulliken, 7 leaders in the field of vascular anomalies recently suggested that congenital haemangiomas (i.e. those present at birth) should be a separate group. These lesions seem to be of two varieties, either rapidly involuting congenital haemangiomas (RICH) 2 or noninvoluting congenital haemangiomas (NICH) 4,5 The evaluation of these type of patients may require multiple fields of expertise and adequate tools for the correct diagnosis 6 8 The dilemma is that histologically, late staged haemangiomas can be difficult to differentiate from venous or arteriovenous malformations. This absence of diagnostic histopathological features often leads to the management of these lesions being planned on the basis of clinical findings alone. Any histological marker for haemangiomas would be of obvious benefit to both the pathologist and clinician. Previous authors tried to identify cellular markers distinguishing the different phases of juvenile haemangiomas, 9 but it was the study by North and colleagues 10 that suggested that the erythrocytetype glucose transporter protein GLUT-1 could be useful in the differentiation between juvenile haemangiomas as it is highly expressed in these lesions, but not in congenital haemangiomas, vascular malformations, pyogenic granulomas or granulation tissue. GLUT-1 is not expressed in the microvascular endothelia of normal skin but it is able to stain not only red blood cells and nerve endoneurium and perineurium, but also blood-barrier tissues such as placenta. This is because of its persistence only in certain tissues after fetal development and the similarity of juvenile haemangiomas to these tissues. 11 We have been using GLUT-1 stains on vascular anomalies since 1998, and in this study, we report our experience. The aim of this study is to examine GLUT-1 as a specific marker for juvenile haemangiomas. Patients and methods Patients data Fifty blocks of formalin-fixed, paraffin waxembedded biopsy specimens from patients having surgery on vascular anomalies, and identified from the records of the histopathology department at Southend Hospital from 1997 to 2003 were evaluated for expression of GLUT-1 protein. This was our test group. All patients gave consent for the study. As positive controls we used placenta tissue and as negative controls, we used samples of pyogenic granulomas and granulation tissue, as the histology of these is known to be very similar to that of juvenile haemangiomas. There were 50 samples altogether. They consisted of 19 cases of juvenile haemangiomas: three proliferating, five involuting and 11 involuted; and two cases of NICH. There were 29 cases of vascular malformations: eight capillary, 11 lymphatic, seven venous and three arteriovenous; and eight cases of negative controls: four pyogenic granulomas and four granulation tissue samples (Tables 1 3). Staining procedure Four micrometer sections were cut from formalin fixed paraffin wax-embedded tissue. Sections were de-waxed in xylene and rehydrated through graded alcohols and water. Antigen retrieval was performed by pressure cooking the slides in antigen unmasking solution for 2 min. Automated immunohistochemical staining was performed using a Dako autostain Techmate 500 Plus (Dako Corp, Carpinteria, CA). The primary antibody used was rabbit antihuman GLUT-1 (Dako) at a concentration of 1/400. This included using the standard streptavidin biotin complex technique, using dyaminobenzidine as the chromogen. Negative controls were processed in the same manner as the test samples and the positive controls, except that phosphate buffered saline was used instead of the primary antibody. GLUT-1 immunoreactivity was scored in a blinded manner by one of the authors (K.W.) who did not know the clinical diagnoses, using as criteria Table 1 Hemangiomas 21 Cases n Age range Juvenile proliferating 3 8 12 months Juvenile Involuting 5 1 4 years Juvenile Involuted 11 4 11 years Noninvoluting congenital 2 8 years
350 J. Leon-Villapalos et al. Table 2 Vascular malformations 29 Cases n Age range Capillary 8 4 12 years Lymphatic 11 3 33 years Venous 7 1 3 years Arteriovenous 3 11 47 the intensity of immunoreaction in comparison to that of positive external (placenta tissue) and internal (perineurium of peripheral nerves and erythrocytes) controls. The scoring system was similar to the one used in the original study performed by North and colleagues: 10 1 ¼ No immunoreaction 2 ¼ Weak immunoreaction (less intense than positive controls) 3 ¼ Strong immunoreaction (as intense as positive controls) Figure 1 Results of GLUT-1 positive (red) and negative (yellow) immunoreactivity in haemangioma samples. Results GLUT-1 expression in haemangiomas From a total of 19 juvenile haemangioma samples, 18 were positive for expression of GLUT-1, including three in the proliferating, five in the involuting, and 10 in the involuted phase (Figs. 1 3). All positive cases showed strong immunoreactivity. Both cases of NICH were negative for GLUT-1 expression. GLUT-1 expression in nonhaemangiomas All 29 cases in this category were negative for GLUT-1 expression, including eight capillary malformations, 11 lymphatic malformations, seven venous malformations and three arteriovenous malformations (Figs. 4 and 5). Figure 2 Strong GLUT-1 immunoreaction (dark brown) in haemangioma vascular channels. Positive controls Placenta tissue was used as positive internal control. Strong GLUT-1 staining around placental chorionic villi can be seen in Fig. 6. Table 3 Negative controls Eight cases n Age range Pyogenic granuloma 4 1 3 years Granulation tissue 4 1 2 years Figure 3 High resolution picture of strong GLUT-1 immunoreactivity (dark brown) in a haemangioma.
GLUT-1: an extra diagnostic tool to differentiate between haemangiomas and vascular malformations 351 Figure 4 samples. Negative controls All cases were negative for GLUT-1 expression. These included four cases of pyogenic granulomas and four of granulation tissue (Fig. 7). This was expected, as these tissues act as internal controls for the study due to their histological resemblance with juvenile haemangiomas. Discussion Results of GLUT-1 negative non-haemangioma We confirm the high expression of the immunohistochemical marker GLUT-1 in the endothelia of vascular channels of juvenile haemangiomas. One case of involuted juvenile hemangioma did not express GLUT-1 immunoreactivity; and it appears Figure 6 Positive GLUT-1 control (placental tissue) showing GLUT-1 immunoreactivity in the chorionic villi. that in this rare case, juvenile haemangioma GLUT- 1 expression may have been lost with involution. We also confirm that GLUT-1 is not expressed in any of the vascular malformations that were sampled or in other tissues with proliferative capacity such as pyogenic granulomas and granulation tissue. The results of positive GLUT-1 immunoreactivity in the juvenile haemangiomas we studied are similar to those reported in the original study of North et al. (95% versus 97%); and even though we studied fewer patients (51 versus 209), we feel that: GLUT-1 seems to be a reliable and highly specific immunohistochemical marker for juvenile haemangiomas. GLUT-1 expression seems to be independent of proliferative activity. GLUT-1 is a useful extra diagnostic tool to differentiate between juvenile haemangiomas and vascular malformations. Furthermore, in agreement with Enjolras, 2 our Figure 5 Absence of GLUT-1 staining in capillary endothelia of a vascular malformation. 1 With strong positive staining of red cells, 1 and perineurium of peripheral nerve. 2 Figure 7 Results in the negative control group.
352 J. Leon-Villapalos et al. two cases of noninvoluting congenital haemangiomas were GLUT-1 negative. The reason for this is still unclear. It is important to differentiate between haemangiomas and vascular malformations, as their management is very different. The treatment of haemangiomas involves mainly a watchful waiting approach as their main characteristic is a rapid neonatal growth that leads to spontaneous regression. They are hardly ever life threatening, and pharmacological treatment including steroids, anti-metabolites or surgery in severe cases is only indicated for problematic and endangering haemangiomas. 12 Vascular malformations require different surgical management that ranges from laser for capillary malformations to sclerotherapy and surgery for lymphatic and venous malformations and embolisation, sclerotherapy and surgery 13 for fast flow arteriovenous malformations. We believe that GLUT-1 reactivity is a valuable marker for haemangiomas, and following this study we use it as routine histological marker in specimens of vascular anomalies and vascular tumours. The specificity of GLUT-1 staining maybe useful for difficult cases, but moreover it may give clues as to the aetiology and nature of haemangiomas and bring us close to understanding these fascinating self-destructive tumours. Acknowledgements Our acknowledgments to W. Furley and J. O Brien for their technical help in this study. References 1. Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: a classification based on endothelial characteristics. Plast Reconstr Surg 1982;69(3): 412 22. 2. Enjolras O, Wassef M. A review of benign vascular tumours in the neonate. Presentation at 14th International Workshop on Vascular anomalies Nijmegen, The Netherlands, June; 2002. 3. Gampper TJ, Morgan RF. Vascular anomalies: hemangiomas. Plast Reconstr Surg 2002;110(2):572 85quiz 586; discussion 587 8. 4. North PE, et al. Congenital nonprogressive hemangioma. A distinct clinico-pathological entity unlike infantile hemangiomas. Arch Dermatol 2001;137(12):1607 20. 5. Enjolras O, et al. Non-involuting congenital hemangioma: a rare cutaneous vascular anomaly. Plast Reconstr Surg 2001; 107:1647 54. 6. Burrows PE, Laor T, Paltiel H, Robertson RL. Diagnostic imaging in the evaluation of vascular birthmarks. Dermatol Clin 1998;16(3):455 88. 7. Enjolras O, Mulliken JB. Vascular tumours and vascular malformations (new issues). Adv Dermatol 1997;13: 375 423. 8. Enjolras O. Classification and management of the various superficial vascular anomalies: haemangiomas and vascular malformations. J Dermatol 1997;24:701 10. 9. Takahashi K, et al. Cellular markers that distinguish the phases of hemangioma during infancy and school. J Clin Invest 1994;93(6):2357 64. 10. North PE, Waner M, Mizeracki A, Mihm Jr MC. GLUT-1: a newly discovered immunohistochemical marker for juvenile heangiomas. Hum Pathol 2000;31(1):11 22. 11. North PE, Waner M, Mizeracki A, et al. A unique microvascular phenotype shared by juvenile hemangiomas and human placenta. Arch Dermatol 2001;137(5):559 70. 12. Mueller BU, Mulliken JB. The infant with a vascular tumour. Semin Perinatol 1999;23(4):332 40. 13. Enjolras O, Deffrennes D, Borsik M, et al. Vascular tumors and the rules of their surgical management. Annales de Chirurgie Plastique et Esthetique 1998;43(4):455 89.