- Email: [email protected]
Malignant melanoma and pregnancy: Second thoughts
Journal of Plastic, Reconstructive & Aesthetic Surgery (2010) 63, 1163e1168
Malignant melanoma and pregnancy: Second thoughts*,** Ehud Miller a, Yoav Barnea a, Eyal Gur a, David Leshem a, Eliad Karin c, Jerry Weiss a, Schlomo Schneebaum b,* a
Tel-Aviv Sourasky Medical Center, affiliated to the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel Department of Surgery ‘‘A’’, Tel-Aviv Sourasky Medical Center, affiliated to the Sackler Faculty of Medicine, Tel-Aviv University, 6 Weizmann Street, Tel Aviv 64239, Israel c Radio-guided Surgery Unit, Department of Surgery, Tel-Aviv Sourasky Medical Center, affiliated to the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel b
Received 30 December 2008; accepted 29 May 2009
KEYWORDS Melanoma; Pregnancy; Hormone; Sentinel node
Summary Malignant melanoma (MM) was considered a hormone-sensitive tumour, and pregnancy was thought to increase its risk and cause faster progression and earlier metastasis. Several controlled studies demonstrated similar survival rates between pregnant and non-pregnant patients and concluded that early reports of advanced MM of pregnancy were probably due to late diagnosis. We retrieved information from our database between 1997 and 2006 on all patients diagnosed as having MM during and up to 6 months after pregnancy (n Z 11) and compared them to age-matched, non-pregnant, MM patients (n Z 65, controls) treated by us during that period. The mean Breslow thickness was 4.28 mm for the pregnant patients and 1.69 mm for the controls (p Z 0.15). The sentinel nodes were metastatic in five pregnant patients compared to four controls (p < 0.0001). Two patients in the pregnancy group and one control died of MM (p Z 0.0532). Our results indicate a negative effect of pregnancy on the course of MM. ª 2009 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published by Elsevier Ltd. All rights reserved.
* Financial Disclosure and Products: Nothing whatsoever to declare. ** Part of this work was presented at the Plastic Surgery at the Red Sea Conference, Eilat, Israel, March 2005. * Corresponding author. Tel.: þ972 3 697 3727; fax: þ972 3 697 4820. E-mail address: [email protected] (S. Schneebaum).
Malignant melanoma (MM) during pregnancy is rare, appearing in 0.14 to 2.8 in 1000 births and accounting for 8% of all malignancies during pregnancy.1,2 Pregnancy was considered for many years to have an adverse effect on the course of MM; several reports in the 1950s suggested that pregnancy increased the risk of developing melanoma and caused the disease to advance more rapidly and metastasise earlier, especially in patients with a history of melanoma.3,4 The rationale behind these findings was that
1748-6815/$ - see front matter ª 2009 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.bjps.2009.05.050
1164 MM is a hormone-dependent tumour that occurs more frequently in women than in men,5 especially during childbearing years,6 and is extremely rare prior to puberty.7,8 Moreover, melasma, or increased pigmentation, is common during pregnancy and among women using oral contraceptives.9 The belief was that similar changes take place in pigmented nevi, causing more malignant transformations. Furthermore, early laboratory findings showed that melanocytes have the ability to bind oestrogen (E2) and progesterone, supportive of these claims.10 Several controlled trials were carried out for the World Health Organization (WHO) and shed new light on the issue. About 400 pregnant MM patients were compared to 1800 matched, non-pregnant, MM patients.11e16 Two studies demonstrated thicker melanoma tumours in the pregnant group compared to controls, but related this finding to later diagnosis of the tumour in pregnant women.13,16 No trial showed a difference in survival between the groups.11e16 More recent investigations on large groups of patients (about 450 women who were pregnant or who were shortly after pregnancy compared to 3000 non-pregnant women) also demonstrated no difference in survival between the groups.17e20 In vitro studies added inconclusive data, with two trials finding neither E2 nor progesterone receptors on several melanoma cell lines,8,21 while other trials found hormone down-regulating cell activity.22,25 Another investigation demonstrated an influence of hormones on nuclear morphology.23 Following two recent cases of MM in pregnant women seen fairly close in time, we began to question the lack of any association between the condition and the pathology. We searched our own institutional database and retrieved the files of 11 MM patients during or shortly after pregnancy, and compared them to an age-matched group of non-pregnant MM women.
Patients and methods The study was approved by our institutional Helsinki Committee. We retrospectively reviewed the institutional database between 1997 and 2006 and retrieved information on all the patients who had been diagnosed of having MM during pregnancy or shortly afterwards. Information on age, marital status, medical history, the anatomic location and the tumour staging of the melanoma was available for each patient as was pregnancy status at diagnosis, history of fertility hormonal therapy, and oncological follow-up. Data on age-matched, non-pregnant, MM women treated at our institute during the same period of time were also available and retrieved as well. Treatment had included wide excision of the primary tumour area and sentinel lymph node biopsy (SNLB) when indicated. SLNB for MM was introduced into our institution in 1997. Inclusion criteria for SLNB were MM with a Breslow thickness of 1 mm or deeper and no clinical evidence of enlarged regional lymph nodes or metastatic disease. Mapping of the sentinel lymph node (SLN) was performed by lymphoscintigraphy using radio-guided surgery and intraoperative patent blue dye injection. Pregnant MM patients had only intra-operative patent blue dye injection, with no
E. Miller et al. radioactive material injection. All women in whom metastasis was found in the SLN underwent formal lymph node dissection.
Statistical methods Comparison between pregnant women with MM and nonpregnant MM controls was by the chi-square test for categorical variables (nodes, location, survival) and the nonparametric ManneWhitney test for age and the Breslow staging. The SAS for Windows version 9.1 was used for all statistical analysis. Significance was set at P < 0.05.
Results Eleven women who had been diagnosed as having MM during pregnancy or up to 6 months after pregnancy were available for analysis. We compared them to 65 age-matched, nonpregnant, MM patients who served as controls. All the patients in this study were Caucasian, except for one African patient who was in the pregnant group. The mean standard deviation (SD) age at the time of diagnosis was 34 3.7 years for the pregnancy group and 34 7.7 years for the control group. Ten patients in the pregnancy group (91%) had primary melanoma and one (9%) had recurrent melanoma during pregnancy. MM was diagnosed during pregnancy in six patients (54.5%), with an average of 26 gestational weeks (range: 11e36 weeks, median 28 weeks). The diagnosis of MM was made after delivery in the remaining five patients (45.5%), with an average of 3.8 months post-delivery (range: 2e6 months). Three women (27.2%) had undergone fertility hormonal treatments prior to conception (Table 1). The mean Breslow thickness was 4.28 4.48 mm in the pregnancy group, and two patients had ulceration of the tumour on pathological examination. The mean Breslow thickness was 1.69 1.13 mm in the control group (p Z 0.15). Eight women in the pregnancy group were treated with combined wide local excision and SLNB and the other three with only wide excision during or shortly after pregnancy. Two of these patients had a Breslow of 0.5 mm, and a third patient had recurrent melanoma in the cheek area. All 65 patients in the control group underwent SLNB. The SLNs were metastatic in five of the 11 patients in the pregnancy group (45.5%) and in only four of the 65 (6.1%) patients in the control group (p < 0.0001) (Table 2). These nine patients underwent radical dissection of the involved lymph node basin. Two patients died of the disease in the pregnancy group (18.1%), compared to one patient (1.5%) in the control group (p Z 0.0532) (Table 2). The anatomic location of the melanoma differed between the groups. Most of the tumours were located in the upper limb in the pregnancy group (six of 11 patients, 55%), while the most common site in the control group was the lower limb (22 of 65 patients, 34%).
Discussion Early reports in the 1950s on MM during pregnancy suggested that there was a greater risk of developing the disease and a more aggressive manifestation in these women compared
Malignant melanoma and pregnancy Table 1
1165
Summary of our reported patients with pregnancy-related melanoma
Age at diagnosis, y
Pregnancy status at diagnosis
Site of melanoma
Staging of melanoma
SLNB
Lymph node dissection
Follow-up
1
37
36 weeks
Clitoris
þ
þ
Disseminated Disease
2
41
11 weeks
Back
þ
No Recurrence
3 4 5
26 44 29
Back Upper Limb Upper Limb
þ þ þ
þ
Disseminated Disease Disseminated Disease No Recurrence
6 7
34 33
þ þ
þ
No Recurrence No Recurrence
8
30
Upper limb
No Recurrence
9 10 11
34 29 35
26 weeks 26 weeks 4 months postpartum 28 weeks 6 months postpartum 3 months postpartum 28 weeks 36 weeks 5 weeks
Bres. 11.8 mm Ulcerated Bres. 2 mm Clark-4 Bres. 2.1 mm Bres. 6.6 mm Bres. 1.5 mm Clark-3 Bres. 6.8 mm Bres. 1 mm Clark-2e3 Bres 0.5 mm
Cheek Arm Hand
Bres. 2.2 mm Bres. 12 mm Bres. 0.5 mm
þ
þ þ
Recurrence Disseminated Disease No Recurrence
Back Back
SLNB, sentinel lymph node biopsy; Bres, Breslow thickness.
to non-pregnant women. Pack and Scharnagel3 reported 10 patients with melanoma during pregnancy, of whom five died within 30 months of diagnosis, suggesting unusual rapidity and earlier metastatic spread associated with pregnancy. George and Fortner4 noted that the disease was more advanced among 115 patients with MM during pregnancy compared to a non-pregnant, age-matched, control group, although there was no difference in the 5-year survival rate between the two groups. These early findings seemed logical, and so the consensus was that MM was a hormonesensitive tumour that occurs more in women and mostly during their fertile years, 5e8 and that increased pigmentation is common during pregnancy and under oral contraceptive treatment.9 This view was also supported by an early in vitro experiment that showed that melanoma cells expressed hormone receptors.10 The more than half-century-old association between MM and pregnancy was then challenged by a number of controlled studies carried out for the WHO, in which around 400 pregnant MM patients were compared to around 1800 non-pregnant MM patients. In 1985, Reintgen et al. published the first case-control study of 58 women with MM during pregnancy compared with 585 non-pregnant MM patients at the same disease stage. Their results showed thicker tumours in the pregnant group (1.9 mm vs. 1.5 mm for the non-pregnant group), but no difference in survival between the two groups.11 Slinguff et al. enlarged the Reintgen et al. group to 100 patients, and found no
significant change in survival and overall recurrence rate between the study group and 86 age-matched controls.12 In both trials, the disease-free interval (DFI) was shorter in the study group compared to the controls. A multivariant analysis showed a significant influence of tumour thickness, Clark’s level and diagnosis during pregnancy on the shorter DFI. The controversy between thicker tumours and shorter DFI, with no difference in survival, could be attributed to a too-short follow-up or insufficient statistical power.11,12 Mackie et al.13 also reported thicker tumours in a group of 92 patients with MM during pregnancy, and a greater frequency of tumours in truncal regions. Again, there was no change in DFI or survival between the pregnancy and non-pregnancy groups.13 According to all these authors, the thicker tumours of pregnant MM patients could be attributed to later diagnosis, since patients consider changes in nevi to be part of normal pregnancy changes.11e13 In 1997, Pennoyer et al.19 examined changes of nevi during pregnancy: 129 nevi of 22 pregnant women showed no difference in size between the first and third trimesters, suggesting a lack of hormonal effect on nevi during pregnancy. Wong et al.14 reviewed 66 cases of stage 1 melanoma and found no difference in 5-year survival compared to 619 controls matched for tumour thickness, anatomic distribution, age and histological subtypes. McManamny et al.15 noted a similar lack of difference in survival, DFI or tumour thickness between 23 stage 1 pregnant MM patients and 243 non-pregnant controls. Travers et al.,16 similar to
Table 2 Comparison of tumour thickness, metastatic lymph nodes and mortality between pregnancy-related melanoma group and the non-pregnancy-related melanoma (control) group
Tumor thickness (Breslow) Metastatic lymph nodes Mortality
Pregnancy (n Z 11)
Control (n Z 65)
P value
4.28 mm 5 patients (45.5%) 2 patients (18.1%)
1.69 mm 4 patients (6.1%) 1 patient (1.5%)
0.15 <0.0001 0.0532
1166 Table 3
E. Miller et al. Summary of the published literature on melanoma and pregnancy
First author
Melanoma and pregnancy (n)
Melanoma and no pregnancy (control) (n)
Reintgen11 Slingluff12 MacKie13 Wong14 McManamy15
58 100 92 66 23
585 86 296 619 243
Travers16 Darayanani17 O’Meara18
45 46 412
420 368 2451
Breslow study group
Breslow control group
Survival
1.9 0.39 mm 2.17 mm 2.38 mm
1.51 0.78 mm 1.52 mm (P Z 0.052) 1.71 mm (P Z 0.002)
No diff.
No diff.
2.28 mm 1.7 mm Localised disease, Preg. 0.77 mm 1 y post. 0.9 mm
1.22 mm 2.0 mm (NS) 0.81 mm (NS)
No diff. No diff. No diff. No diff. 74% trial, 87% control (NS) No diff. No diff. No diff.
NS, not significant; Diff., difference; Preg., pregnancy group; y post., year postpartum.
other studies, found no difference in survival between the two groups, with a slight trend towards better survival in the pregnant group. Those authors suggested that their findings could be related to hormonal changes, immunological inhibitory effect of pregnancy or late diagnosis. A recent retrospective trial by Daryanani et al.17 compared 46 pregnant MM patients to 368 non-pregnant MM patients and reported no difference in tumour thickness or in survival between the groups. The largest and latest trial on this issue, by O’Meara et al.,18 compared 412 pregnant and up to 1 year post-pregnancy MM patients to 2451 non-pregnant MM controls. This large trial also failed to show any difference in tumour thickness or survival between the groups (Table 3). In spite of such convincing and consistent evidence that emerged from all these trials, our growing impression was that pregnancy-related MMs were more aggressive than the non-pregnancy-related ones and so we decided to check our files to see if there was any factual basis for this line of thought. Similar to some of the former trials, we also found the pregnancy group to have more advanced disease than the control group. The mean Breslow thickness for the pregnancy group was higher than the thickness for the control group (4.28 mm vs. 1.69 mm, respectively), although the difference did not reach a level of significance, probably due to the small trial group. There was a significantly higher percentage of positive nodes in the pregnancy group, and a greater mortality rate (of borderline significance) among these women compared to the controls. These findings are in contrast to the WHOsupported trials and the latest published trials. Although our results could be attributed to a delay in diagnosis and to our small study group, the aggressiveness of the pregnancy-related tumours, as demonstrated by patient number 9 (Table 1), must be taken into account: she had recurrent MM only 2 months after undergoing wide excision and SLNB. Patient number 10 in Table 1 is a 29-year-old African woman diagnosed as having a metastasised 12-mm MM of her left arm. MM is rare in black populations and is usually diagnosed in later stages than in non-black populations.31,32 A similar advanced-stage melanoma of a pregnant African-American woman was reported by Hu et al.33
Since clinical trials have not reached conclusive results, further investigation at the cellular level is all the more important. Several in vitro studies that have been published in recent years seem, however, only to add to the confusion. Duncan et al.8 used advanced monoclonal antibody staining techniques and reported finding no E2 or progesterone receptors in 14 pregnancy-related MM. Other in vitro studies showed a variation of hormonal factors. Morvillo et al.21 recently reported a lack of E2 or progesterone receptors, but they did demonstrate the presence of androgen receptors on two human melanoma cell lines. They showed that testosterone, E2 and progesterone significantly stimulated cell proliferation in those cells. Kanda and Watanabe,22 however, found that melanoma cell lines with hormone receptors were down-regulated by 17 beta-oestradiol, progesterone and testosterone, and that other cell lines lacking these receptors were not influenced by the hormones. Yet another in vitro study showed testosterone (DHT) and E2 to change nuclear morphology in melanoma cells,23 while Miller et al.24 noted that there were limited E2 receptors in MM (two out of 69 melanoma specimens). Finally, luteinising hormonereleasing hormone (LHRH) receptors were demonstrated in a BLM melanoma cell line that showed inhibited proliferation and reduced metastatic activity in response to an LHRH agonist.25 Another question that had been addressed in several trials is the risk of recurrent MM in subsequent pregnancies. Mackie et al.13 compared pregnant MM patients to women with MM who had never been pregnant, were between pregnancies or had completed all planned pregnancies. They found no difference in survival between the latter three groups. Reintgen et al.11 compared 43 women who became pregnant within 5 years of being diagnosed as having stage 1 melanoma to a cohort of 2938 non-pregnant melanoma patients and found no effect of subsequent pregnancy on recurrence or survival. Since most recurrences of stage 1 melanoma occur within the first 2 years after excision,13,15 several authors recommend delaying the next pregnancy for 2e3 years after treatment.5,6,13,15,20 Survival seems to be mostly related to the stage of melanoma at diagnosis and whether there is ulceration.11,12 Based on these considerations, recommendations for
Malignant melanoma and pregnancy treatment are similar to those of non-pregnant patients.6 Excisional biopsy under local anaesthesia poses little risk to either mother or foetus,16 and sentinel node biopsy using dye and lymphoscintigraphy is generally not contraindicated during pregnancy.6,35 Lloyd et al.36 of the St. George Healthcare NHS Trust Melanoma Unit, London, recommend the avoidance of lymphoscintigraphy before 30 weeks of gestation, and usually to postpone SLNB until delivery. We chose to avoid lymphoscintigraphy during pregnancy and to perform the SLNB with only intra-operative dye injection. Several other aspects of the melanomaehormone issue have also been addressed by a number of authors. One is hormone replacement therapy (HRT), which does not seem to be a risk factor for melanoma,26,27,34 and MacKie et al.28 reported that the use of HRT after treatment for stage 1e2 melanoma does not have adverse effects on patient outcome. Another is the use of oral contraceptives, which also seems to have no effect on the incidence of MM.29,30,34 The effect of high-dose hormones, such as in the case of protocols for in vitro fertilisation or intrauterine insemination, on the incidence of melanoma has not yet been addressed in the literature. The long-held belief that MM is a hormone-sensitive tumour and one that is affected by pregnancy was later rejected for not having any scientific basis. Our current results indicate that there is a significant relation between pregnancy, more advanced disease stage and greater mortality. We have shown that MM does seem to be influenced by pregnancy in certain cases, in contradiction to the findings of controlled trials that concluded that there is no greater risk of MM during pregnancy and that thicker tumours in pregnant women are most likely related to later diagnosis stemming from the woman’s expectation of changes to occur in her pigmentation during pregnancy, the lack of awareness of an existing lesion or delay of medical attention due to fear. Only two of the controlled trials in the literature were prospective,11,12 and the data that emerged from in vitro studies were inconclusive. While the lack of effect of HRT and contraceptive pills on the course of MM has been established, the effect of increasingly popular hormonal fertility treatments on MM is unknown and warrants investigation.
Study limitations We recognise that a series of only 11 MM patients diagnosed during pregnancy or shortly after delivery precludes our arriving at firm conclusions about a direct link between MM and pregnancy. We can, however, hope to raise the level of awareness of MM in pregnancy for the sake of early detection and prompt treatment: new suspected lesions or changing moles should be evaluated with an excisional biopsy under local anaesthesia and SLNB should be performed when MM is discovered. We endorse the recommendation that a subsequent pregnancy should be delayed for at least 2 years after treatment.
Acknowledgement Esther Eshkol is thanked for editorial assistance.
1167
Conflict of interest None.
Funding None.
References 1. Derek JW, Strassner HT. Melanoma in pregnancy. Clin Obstet Gynecol 1990;33:782e91. 2. Colbourn DS, Nathanson L, Belilos E. Pregnancy and malignant melanoma. Semin Oncol 1989;16:377e87. 3. Pack GT, Scharnagel IM. The prognosis for malignant melanoma in the pregnant woman. Cancer 1951;4:324e34. 4. George PA, Fortner IG. Melanoma with pregnancy: a report of 115 cases. Cancer 1960;13:854e9. 5. MacKie RM. Pregnancy and exogenous hormones in patients with cutaneous malignant melanoma. Curr Opin Oncol 1999; 11:129e31. 6. Borden EC. Melanoma and pregnancy. Semin Oncol 2000;27: 654e6. 7. Kjems E, Krag C. Melanoma and pregnancy e a review. Acta Oncol 1993;32:371e8. 8. Duncan LM, Travers RL, Koerner FC, et al. Estrogen and progesterone receptor analysis in pregnancy-associated melanoma: absence of immunohistochemically detectable hormone receptors. Hum Pathol 1994;25:36e41. 9. Katz VL, Farmer RM, Dotters D. Focus on primary care: from nevus to neoplasm: myths of melanoma in pregnancy. Obstet Gynecol Surv 2002;57:112e9. 10. Stoica A, Hoffman M, Marta L. Estradiol and progesterone receptors in human cutaneous melanoma. Neoplasma 1991;38: 137e45. 11. Reintgen DS, McCarty KS, Vollmer R. Malignant melanoma and pregnancy. Cancer 1985;55:1340e4. 12. Slingluff CL, Reintgen DS, Vollmer RT. Malignant melanoma arising during pregnancy e a study of 100 patients. Ann Surg 1990;211:552e9. 13. MacKie RM, Bufalino R, Morabito A. Lack of effect of pregnancy on outcome of melanoma. Lancet 1991;337:653e5. 14. Wong JH, Sterns EE, Kopald KH. Prognostic significance of pregnancy in stage 1 melanoma. Arch. Surg 1989;124:1227e31. 15. McManamny DS, Moss ALH, Pocock PV. Melanoma and pregnancy: a long term follow-up. Br J Obstet Gynaecol 1989;96: 1419e23. 16. Travers RL, Sober AJ, Berwick M. Increased thickness of pregnancy-associated melanoma. Br J Dermatol 1995;132:876e83. 17. Daryanani D, Plukker JT, De Hullu JA, et al. Pregnancy and early stage melanoma. Cancer 2003;97:2248e53. 18. O’Meara AT, Cress R, Xing G, et al. Malignant melanoma in pregnancy e a population-based evaluation. Cancer 2005;103: 1217e26. 19. Pennoyer JW, Grin CM, Driscoll MS. Changes in size of melanocytic nevi during pregnancy. J Am Acad Dermatol 1997;36: 378e82. 20. Grin CM, Driscoll MS, Grant-Kels JM. Pregnancy and the prognosis of malignant melanoma. Semin Oncol 1996;23:734e6. 21. Morvillo V, Luthy IA, Bravo AI. Androgen receptors in human cell lines 2B-MEL-LES and 2B-MEL IAN and in human melanoma metastases. Melanoma Res 2002;12:529e38. 22. Kanda N, Watanabe S. 17beta-estradiol, progesterone, and dihydrotestosterone suppress growth of human melanoma by inhibiting IL-8 production. J Invest Dermatol 2001;117: 274e83.
1168 23. Lorea P, Decaestecker C, Goldchmidt D. Correlation between gender and cytomorphonuclear characteristics in human melanomas and in vitro evidence of sex steroid-induced modifications in the morphonuclear characteristics of three human melanoma cell lines. Melanoma Res 1997;7:382e92. 24. Miller JG, Gee J, Price A. Investigation of oestrogen receptors, sex steroids and soluble adhesion molecules in the progression of malignant melanoma. Melanoma Res 1997;7: 197e208. 25. Moretti RM, Marelli MM, Van Groeninghen JC. Locally expressed LHRH receptors mediate the oncostatic and antimetastatic activity of LHRH agonists on melanoma cells. J Clin Endocrinol Metab 2002;87:3791e7. 26. Osterlind A, Tucker MA, Stone BJ. The Danish case control study on cutaneous malignant melanoma. Hormonal and reproductive factors in women. Int J Cancer 1988;42:821e4. 27. Persson I, Yuen J, Bergkvist L, et al. Cancer incidence and mortality in women receiving oestrogen or oestrogen-progestin replacement therapy e long term follow up in Swedish cohort. Int J Cancer 1996;67:327e32. 28. Mackie RM, Bray CA. Hormone replacement therapy after surgery for stage 1 or 2 cutaneous melanoma. Br J Cancer 2004; 90:770e2.
E. Miller et al. 29. Smith MA, Fine JA, Barnhill RL, et al. Hormonal and reproductive influences and risk of melanoma in women. Int J Epidemiol 1998;27:751e7. 30. Grin CM, Driscoll MS, Grant-Kels JM. The relationship of pregnancy, hormones and melanoma. Semin Cutan Med Surg 1998; 17:167e71. 31. Gloster HM, Neal K. Skin cancer in skin of color. J Am Acad Dermatol 2006;55:741e60. 32. Byrd KM, Wilson DC, Hoyler SS, et al. Advanced presentation of melanoma in African Americans. J Am Acad Dermatol 2004;51: 1031e3. 33. Hu W, Nelson JE, Mohney CA, et al. Malignant melanoma in a pregnant African American woman with congenital blue nevus. Dermatol Surg 2004;30:1530e2. 34. Lea CS, Holly EA, Hartge P, et al. Reproductive risk factors for cutaneous melanoma: a case-control study. Am J Epidemiol 2007;165:505e13. 35. Mondi MM, Cuenca RE, Ollila DW, et al. Sentinel lymph node biopsy during pregnancy: initial clinical experience. Annal Surg Oncol 2007;14:218e21. 36. Lloyd MS, Topping A, Allan R, et al. Contraindications to sentinel lymph node biopsy in cutaneous malignant melanoma. Br J Plast Surg 2004;57:725e7.
Malignant melanoma and pregnancy: Second thoughts*,** Ehud Miller a, Yoav Barnea a, Eyal Gur a, David Leshem a, Eliad Karin c, Jerry Weiss a, Schlomo Schneebaum b,* a
Tel-Aviv Sourasky Medical Center, affiliated to the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel Department of Surgery ‘‘A’’, Tel-Aviv Sourasky Medical Center, affiliated to the Sackler Faculty of Medicine, Tel-Aviv University, 6 Weizmann Street, Tel Aviv 64239, Israel c Radio-guided Surgery Unit, Department of Surgery, Tel-Aviv Sourasky Medical Center, affiliated to the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel b
Received 30 December 2008; accepted 29 May 2009
KEYWORDS Melanoma; Pregnancy; Hormone; Sentinel node
Summary Malignant melanoma (MM) was considered a hormone-sensitive tumour, and pregnancy was thought to increase its risk and cause faster progression and earlier metastasis. Several controlled studies demonstrated similar survival rates between pregnant and non-pregnant patients and concluded that early reports of advanced MM of pregnancy were probably due to late diagnosis. We retrieved information from our database between 1997 and 2006 on all patients diagnosed as having MM during and up to 6 months after pregnancy (n Z 11) and compared them to age-matched, non-pregnant, MM patients (n Z 65, controls) treated by us during that period. The mean Breslow thickness was 4.28 mm for the pregnant patients and 1.69 mm for the controls (p Z 0.15). The sentinel nodes were metastatic in five pregnant patients compared to four controls (p < 0.0001). Two patients in the pregnancy group and one control died of MM (p Z 0.0532). Our results indicate a negative effect of pregnancy on the course of MM. ª 2009 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published by Elsevier Ltd. All rights reserved.
* Financial Disclosure and Products: Nothing whatsoever to declare. ** Part of this work was presented at the Plastic Surgery at the Red Sea Conference, Eilat, Israel, March 2005. * Corresponding author. Tel.: þ972 3 697 3727; fax: þ972 3 697 4820. E-mail address: [email protected] (S. Schneebaum).
Malignant melanoma (MM) during pregnancy is rare, appearing in 0.14 to 2.8 in 1000 births and accounting for 8% of all malignancies during pregnancy.1,2 Pregnancy was considered for many years to have an adverse effect on the course of MM; several reports in the 1950s suggested that pregnancy increased the risk of developing melanoma and caused the disease to advance more rapidly and metastasise earlier, especially in patients with a history of melanoma.3,4 The rationale behind these findings was that
1748-6815/$ - see front matter ª 2009 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.bjps.2009.05.050
1164 MM is a hormone-dependent tumour that occurs more frequently in women than in men,5 especially during childbearing years,6 and is extremely rare prior to puberty.7,8 Moreover, melasma, or increased pigmentation, is common during pregnancy and among women using oral contraceptives.9 The belief was that similar changes take place in pigmented nevi, causing more malignant transformations. Furthermore, early laboratory findings showed that melanocytes have the ability to bind oestrogen (E2) and progesterone, supportive of these claims.10 Several controlled trials were carried out for the World Health Organization (WHO) and shed new light on the issue. About 400 pregnant MM patients were compared to 1800 matched, non-pregnant, MM patients.11e16 Two studies demonstrated thicker melanoma tumours in the pregnant group compared to controls, but related this finding to later diagnosis of the tumour in pregnant women.13,16 No trial showed a difference in survival between the groups.11e16 More recent investigations on large groups of patients (about 450 women who were pregnant or who were shortly after pregnancy compared to 3000 non-pregnant women) also demonstrated no difference in survival between the groups.17e20 In vitro studies added inconclusive data, with two trials finding neither E2 nor progesterone receptors on several melanoma cell lines,8,21 while other trials found hormone down-regulating cell activity.22,25 Another investigation demonstrated an influence of hormones on nuclear morphology.23 Following two recent cases of MM in pregnant women seen fairly close in time, we began to question the lack of any association between the condition and the pathology. We searched our own institutional database and retrieved the files of 11 MM patients during or shortly after pregnancy, and compared them to an age-matched group of non-pregnant MM women.
Patients and methods The study was approved by our institutional Helsinki Committee. We retrospectively reviewed the institutional database between 1997 and 2006 and retrieved information on all the patients who had been diagnosed of having MM during pregnancy or shortly afterwards. Information on age, marital status, medical history, the anatomic location and the tumour staging of the melanoma was available for each patient as was pregnancy status at diagnosis, history of fertility hormonal therapy, and oncological follow-up. Data on age-matched, non-pregnant, MM women treated at our institute during the same period of time were also available and retrieved as well. Treatment had included wide excision of the primary tumour area and sentinel lymph node biopsy (SNLB) when indicated. SLNB for MM was introduced into our institution in 1997. Inclusion criteria for SLNB were MM with a Breslow thickness of 1 mm or deeper and no clinical evidence of enlarged regional lymph nodes or metastatic disease. Mapping of the sentinel lymph node (SLN) was performed by lymphoscintigraphy using radio-guided surgery and intraoperative patent blue dye injection. Pregnant MM patients had only intra-operative patent blue dye injection, with no
E. Miller et al. radioactive material injection. All women in whom metastasis was found in the SLN underwent formal lymph node dissection.
Statistical methods Comparison between pregnant women with MM and nonpregnant MM controls was by the chi-square test for categorical variables (nodes, location, survival) and the nonparametric ManneWhitney test for age and the Breslow staging. The SAS for Windows version 9.1 was used for all statistical analysis. Significance was set at P < 0.05.
Results Eleven women who had been diagnosed as having MM during pregnancy or up to 6 months after pregnancy were available for analysis. We compared them to 65 age-matched, nonpregnant, MM patients who served as controls. All the patients in this study were Caucasian, except for one African patient who was in the pregnant group. The mean standard deviation (SD) age at the time of diagnosis was 34 3.7 years for the pregnancy group and 34 7.7 years for the control group. Ten patients in the pregnancy group (91%) had primary melanoma and one (9%) had recurrent melanoma during pregnancy. MM was diagnosed during pregnancy in six patients (54.5%), with an average of 26 gestational weeks (range: 11e36 weeks, median 28 weeks). The diagnosis of MM was made after delivery in the remaining five patients (45.5%), with an average of 3.8 months post-delivery (range: 2e6 months). Three women (27.2%) had undergone fertility hormonal treatments prior to conception (Table 1). The mean Breslow thickness was 4.28 4.48 mm in the pregnancy group, and two patients had ulceration of the tumour on pathological examination. The mean Breslow thickness was 1.69 1.13 mm in the control group (p Z 0.15). Eight women in the pregnancy group were treated with combined wide local excision and SLNB and the other three with only wide excision during or shortly after pregnancy. Two of these patients had a Breslow of 0.5 mm, and a third patient had recurrent melanoma in the cheek area. All 65 patients in the control group underwent SLNB. The SLNs were metastatic in five of the 11 patients in the pregnancy group (45.5%) and in only four of the 65 (6.1%) patients in the control group (p < 0.0001) (Table 2). These nine patients underwent radical dissection of the involved lymph node basin. Two patients died of the disease in the pregnancy group (18.1%), compared to one patient (1.5%) in the control group (p Z 0.0532) (Table 2). The anatomic location of the melanoma differed between the groups. Most of the tumours were located in the upper limb in the pregnancy group (six of 11 patients, 55%), while the most common site in the control group was the lower limb (22 of 65 patients, 34%).
Discussion Early reports in the 1950s on MM during pregnancy suggested that there was a greater risk of developing the disease and a more aggressive manifestation in these women compared
Malignant melanoma and pregnancy Table 1
1165
Summary of our reported patients with pregnancy-related melanoma
Age at diagnosis, y
Pregnancy status at diagnosis
Site of melanoma
Staging of melanoma
SLNB
Lymph node dissection
Follow-up
1
37
36 weeks
Clitoris
þ
þ
Disseminated Disease
2
41
11 weeks
Back
þ
No Recurrence
3 4 5
26 44 29
Back Upper Limb Upper Limb
þ þ þ
þ
Disseminated Disease Disseminated Disease No Recurrence
6 7
34 33
þ þ
þ
No Recurrence No Recurrence
8
30
Upper limb
No Recurrence
9 10 11
34 29 35
26 weeks 26 weeks 4 months postpartum 28 weeks 6 months postpartum 3 months postpartum 28 weeks 36 weeks 5 weeks
Bres. 11.8 mm Ulcerated Bres. 2 mm Clark-4 Bres. 2.1 mm Bres. 6.6 mm Bres. 1.5 mm Clark-3 Bres. 6.8 mm Bres. 1 mm Clark-2e3 Bres 0.5 mm
Cheek Arm Hand
Bres. 2.2 mm Bres. 12 mm Bres. 0.5 mm
þ
þ þ
Recurrence Disseminated Disease No Recurrence
Back Back
SLNB, sentinel lymph node biopsy; Bres, Breslow thickness.
to non-pregnant women. Pack and Scharnagel3 reported 10 patients with melanoma during pregnancy, of whom five died within 30 months of diagnosis, suggesting unusual rapidity and earlier metastatic spread associated with pregnancy. George and Fortner4 noted that the disease was more advanced among 115 patients with MM during pregnancy compared to a non-pregnant, age-matched, control group, although there was no difference in the 5-year survival rate between the two groups. These early findings seemed logical, and so the consensus was that MM was a hormonesensitive tumour that occurs more in women and mostly during their fertile years, 5e8 and that increased pigmentation is common during pregnancy and under oral contraceptive treatment.9 This view was also supported by an early in vitro experiment that showed that melanoma cells expressed hormone receptors.10 The more than half-century-old association between MM and pregnancy was then challenged by a number of controlled studies carried out for the WHO, in which around 400 pregnant MM patients were compared to around 1800 non-pregnant MM patients. In 1985, Reintgen et al. published the first case-control study of 58 women with MM during pregnancy compared with 585 non-pregnant MM patients at the same disease stage. Their results showed thicker tumours in the pregnant group (1.9 mm vs. 1.5 mm for the non-pregnant group), but no difference in survival between the two groups.11 Slinguff et al. enlarged the Reintgen et al. group to 100 patients, and found no
significant change in survival and overall recurrence rate between the study group and 86 age-matched controls.12 In both trials, the disease-free interval (DFI) was shorter in the study group compared to the controls. A multivariant analysis showed a significant influence of tumour thickness, Clark’s level and diagnosis during pregnancy on the shorter DFI. The controversy between thicker tumours and shorter DFI, with no difference in survival, could be attributed to a too-short follow-up or insufficient statistical power.11,12 Mackie et al.13 also reported thicker tumours in a group of 92 patients with MM during pregnancy, and a greater frequency of tumours in truncal regions. Again, there was no change in DFI or survival between the pregnancy and non-pregnancy groups.13 According to all these authors, the thicker tumours of pregnant MM patients could be attributed to later diagnosis, since patients consider changes in nevi to be part of normal pregnancy changes.11e13 In 1997, Pennoyer et al.19 examined changes of nevi during pregnancy: 129 nevi of 22 pregnant women showed no difference in size between the first and third trimesters, suggesting a lack of hormonal effect on nevi during pregnancy. Wong et al.14 reviewed 66 cases of stage 1 melanoma and found no difference in 5-year survival compared to 619 controls matched for tumour thickness, anatomic distribution, age and histological subtypes. McManamny et al.15 noted a similar lack of difference in survival, DFI or tumour thickness between 23 stage 1 pregnant MM patients and 243 non-pregnant controls. Travers et al.,16 similar to
Table 2 Comparison of tumour thickness, metastatic lymph nodes and mortality between pregnancy-related melanoma group and the non-pregnancy-related melanoma (control) group
Tumor thickness (Breslow) Metastatic lymph nodes Mortality
Pregnancy (n Z 11)
Control (n Z 65)
P value
4.28 mm 5 patients (45.5%) 2 patients (18.1%)
1.69 mm 4 patients (6.1%) 1 patient (1.5%)
0.15 <0.0001 0.0532
1166 Table 3
E. Miller et al. Summary of the published literature on melanoma and pregnancy
First author
Melanoma and pregnancy (n)
Melanoma and no pregnancy (control) (n)
Reintgen11 Slingluff12 MacKie13 Wong14 McManamy15
58 100 92 66 23
585 86 296 619 243
Travers16 Darayanani17 O’Meara18
45 46 412
420 368 2451
Breslow study group
Breslow control group
Survival
1.9 0.39 mm 2.17 mm 2.38 mm
1.51 0.78 mm 1.52 mm (P Z 0.052) 1.71 mm (P Z 0.002)
No diff.
No diff.
2.28 mm 1.7 mm Localised disease, Preg. 0.77 mm 1 y post. 0.9 mm
1.22 mm 2.0 mm (NS) 0.81 mm (NS)
No diff. No diff. No diff. No diff. 74% trial, 87% control (NS) No diff. No diff. No diff.
NS, not significant; Diff., difference; Preg., pregnancy group; y post., year postpartum.
other studies, found no difference in survival between the two groups, with a slight trend towards better survival in the pregnant group. Those authors suggested that their findings could be related to hormonal changes, immunological inhibitory effect of pregnancy or late diagnosis. A recent retrospective trial by Daryanani et al.17 compared 46 pregnant MM patients to 368 non-pregnant MM patients and reported no difference in tumour thickness or in survival between the groups. The largest and latest trial on this issue, by O’Meara et al.,18 compared 412 pregnant and up to 1 year post-pregnancy MM patients to 2451 non-pregnant MM controls. This large trial also failed to show any difference in tumour thickness or survival between the groups (Table 3). In spite of such convincing and consistent evidence that emerged from all these trials, our growing impression was that pregnancy-related MMs were more aggressive than the non-pregnancy-related ones and so we decided to check our files to see if there was any factual basis for this line of thought. Similar to some of the former trials, we also found the pregnancy group to have more advanced disease than the control group. The mean Breslow thickness for the pregnancy group was higher than the thickness for the control group (4.28 mm vs. 1.69 mm, respectively), although the difference did not reach a level of significance, probably due to the small trial group. There was a significantly higher percentage of positive nodes in the pregnancy group, and a greater mortality rate (of borderline significance) among these women compared to the controls. These findings are in contrast to the WHOsupported trials and the latest published trials. Although our results could be attributed to a delay in diagnosis and to our small study group, the aggressiveness of the pregnancy-related tumours, as demonstrated by patient number 9 (Table 1), must be taken into account: she had recurrent MM only 2 months after undergoing wide excision and SLNB. Patient number 10 in Table 1 is a 29-year-old African woman diagnosed as having a metastasised 12-mm MM of her left arm. MM is rare in black populations and is usually diagnosed in later stages than in non-black populations.31,32 A similar advanced-stage melanoma of a pregnant African-American woman was reported by Hu et al.33
Since clinical trials have not reached conclusive results, further investigation at the cellular level is all the more important. Several in vitro studies that have been published in recent years seem, however, only to add to the confusion. Duncan et al.8 used advanced monoclonal antibody staining techniques and reported finding no E2 or progesterone receptors in 14 pregnancy-related MM. Other in vitro studies showed a variation of hormonal factors. Morvillo et al.21 recently reported a lack of E2 or progesterone receptors, but they did demonstrate the presence of androgen receptors on two human melanoma cell lines. They showed that testosterone, E2 and progesterone significantly stimulated cell proliferation in those cells. Kanda and Watanabe,22 however, found that melanoma cell lines with hormone receptors were down-regulated by 17 beta-oestradiol, progesterone and testosterone, and that other cell lines lacking these receptors were not influenced by the hormones. Yet another in vitro study showed testosterone (DHT) and E2 to change nuclear morphology in melanoma cells,23 while Miller et al.24 noted that there were limited E2 receptors in MM (two out of 69 melanoma specimens). Finally, luteinising hormonereleasing hormone (LHRH) receptors were demonstrated in a BLM melanoma cell line that showed inhibited proliferation and reduced metastatic activity in response to an LHRH agonist.25 Another question that had been addressed in several trials is the risk of recurrent MM in subsequent pregnancies. Mackie et al.13 compared pregnant MM patients to women with MM who had never been pregnant, were between pregnancies or had completed all planned pregnancies. They found no difference in survival between the latter three groups. Reintgen et al.11 compared 43 women who became pregnant within 5 years of being diagnosed as having stage 1 melanoma to a cohort of 2938 non-pregnant melanoma patients and found no effect of subsequent pregnancy on recurrence or survival. Since most recurrences of stage 1 melanoma occur within the first 2 years after excision,13,15 several authors recommend delaying the next pregnancy for 2e3 years after treatment.5,6,13,15,20 Survival seems to be mostly related to the stage of melanoma at diagnosis and whether there is ulceration.11,12 Based on these considerations, recommendations for
Malignant melanoma and pregnancy treatment are similar to those of non-pregnant patients.6 Excisional biopsy under local anaesthesia poses little risk to either mother or foetus,16 and sentinel node biopsy using dye and lymphoscintigraphy is generally not contraindicated during pregnancy.6,35 Lloyd et al.36 of the St. George Healthcare NHS Trust Melanoma Unit, London, recommend the avoidance of lymphoscintigraphy before 30 weeks of gestation, and usually to postpone SLNB until delivery. We chose to avoid lymphoscintigraphy during pregnancy and to perform the SLNB with only intra-operative dye injection. Several other aspects of the melanomaehormone issue have also been addressed by a number of authors. One is hormone replacement therapy (HRT), which does not seem to be a risk factor for melanoma,26,27,34 and MacKie et al.28 reported that the use of HRT after treatment for stage 1e2 melanoma does not have adverse effects on patient outcome. Another is the use of oral contraceptives, which also seems to have no effect on the incidence of MM.29,30,34 The effect of high-dose hormones, such as in the case of protocols for in vitro fertilisation or intrauterine insemination, on the incidence of melanoma has not yet been addressed in the literature. The long-held belief that MM is a hormone-sensitive tumour and one that is affected by pregnancy was later rejected for not having any scientific basis. Our current results indicate that there is a significant relation between pregnancy, more advanced disease stage and greater mortality. We have shown that MM does seem to be influenced by pregnancy in certain cases, in contradiction to the findings of controlled trials that concluded that there is no greater risk of MM during pregnancy and that thicker tumours in pregnant women are most likely related to later diagnosis stemming from the woman’s expectation of changes to occur in her pigmentation during pregnancy, the lack of awareness of an existing lesion or delay of medical attention due to fear. Only two of the controlled trials in the literature were prospective,11,12 and the data that emerged from in vitro studies were inconclusive. While the lack of effect of HRT and contraceptive pills on the course of MM has been established, the effect of increasingly popular hormonal fertility treatments on MM is unknown and warrants investigation.
Study limitations We recognise that a series of only 11 MM patients diagnosed during pregnancy or shortly after delivery precludes our arriving at firm conclusions about a direct link between MM and pregnancy. We can, however, hope to raise the level of awareness of MM in pregnancy for the sake of early detection and prompt treatment: new suspected lesions or changing moles should be evaluated with an excisional biopsy under local anaesthesia and SLNB should be performed when MM is discovered. We endorse the recommendation that a subsequent pregnancy should be delayed for at least 2 years after treatment.
Acknowledgement Esther Eshkol is thanked for editorial assistance.
1167
Conflict of interest None.
Funding None.
References 1. Derek JW, Strassner HT. Melanoma in pregnancy. Clin Obstet Gynecol 1990;33:782e91. 2. Colbourn DS, Nathanson L, Belilos E. Pregnancy and malignant melanoma. Semin Oncol 1989;16:377e87. 3. Pack GT, Scharnagel IM. The prognosis for malignant melanoma in the pregnant woman. Cancer 1951;4:324e34. 4. George PA, Fortner IG. Melanoma with pregnancy: a report of 115 cases. Cancer 1960;13:854e9. 5. MacKie RM. Pregnancy and exogenous hormones in patients with cutaneous malignant melanoma. Curr Opin Oncol 1999; 11:129e31. 6. Borden EC. Melanoma and pregnancy. Semin Oncol 2000;27: 654e6. 7. Kjems E, Krag C. Melanoma and pregnancy e a review. Acta Oncol 1993;32:371e8. 8. Duncan LM, Travers RL, Koerner FC, et al. Estrogen and progesterone receptor analysis in pregnancy-associated melanoma: absence of immunohistochemically detectable hormone receptors. Hum Pathol 1994;25:36e41. 9. Katz VL, Farmer RM, Dotters D. Focus on primary care: from nevus to neoplasm: myths of melanoma in pregnancy. Obstet Gynecol Surv 2002;57:112e9. 10. Stoica A, Hoffman M, Marta L. Estradiol and progesterone receptors in human cutaneous melanoma. Neoplasma 1991;38: 137e45. 11. Reintgen DS, McCarty KS, Vollmer R. Malignant melanoma and pregnancy. Cancer 1985;55:1340e4. 12. Slingluff CL, Reintgen DS, Vollmer RT. Malignant melanoma arising during pregnancy e a study of 100 patients. Ann Surg 1990;211:552e9. 13. MacKie RM, Bufalino R, Morabito A. Lack of effect of pregnancy on outcome of melanoma. Lancet 1991;337:653e5. 14. Wong JH, Sterns EE, Kopald KH. Prognostic significance of pregnancy in stage 1 melanoma. Arch. Surg 1989;124:1227e31. 15. McManamny DS, Moss ALH, Pocock PV. Melanoma and pregnancy: a long term follow-up. Br J Obstet Gynaecol 1989;96: 1419e23. 16. Travers RL, Sober AJ, Berwick M. Increased thickness of pregnancy-associated melanoma. Br J Dermatol 1995;132:876e83. 17. Daryanani D, Plukker JT, De Hullu JA, et al. Pregnancy and early stage melanoma. Cancer 2003;97:2248e53. 18. O’Meara AT, Cress R, Xing G, et al. Malignant melanoma in pregnancy e a population-based evaluation. Cancer 2005;103: 1217e26. 19. Pennoyer JW, Grin CM, Driscoll MS. Changes in size of melanocytic nevi during pregnancy. J Am Acad Dermatol 1997;36: 378e82. 20. Grin CM, Driscoll MS, Grant-Kels JM. Pregnancy and the prognosis of malignant melanoma. Semin Oncol 1996;23:734e6. 21. Morvillo V, Luthy IA, Bravo AI. Androgen receptors in human cell lines 2B-MEL-LES and 2B-MEL IAN and in human melanoma metastases. Melanoma Res 2002;12:529e38. 22. Kanda N, Watanabe S. 17beta-estradiol, progesterone, and dihydrotestosterone suppress growth of human melanoma by inhibiting IL-8 production. J Invest Dermatol 2001;117: 274e83.
1168 23. Lorea P, Decaestecker C, Goldchmidt D. Correlation between gender and cytomorphonuclear characteristics in human melanomas and in vitro evidence of sex steroid-induced modifications in the morphonuclear characteristics of three human melanoma cell lines. Melanoma Res 1997;7:382e92. 24. Miller JG, Gee J, Price A. Investigation of oestrogen receptors, sex steroids and soluble adhesion molecules in the progression of malignant melanoma. Melanoma Res 1997;7: 197e208. 25. Moretti RM, Marelli MM, Van Groeninghen JC. Locally expressed LHRH receptors mediate the oncostatic and antimetastatic activity of LHRH agonists on melanoma cells. J Clin Endocrinol Metab 2002;87:3791e7. 26. Osterlind A, Tucker MA, Stone BJ. The Danish case control study on cutaneous malignant melanoma. Hormonal and reproductive factors in women. Int J Cancer 1988;42:821e4. 27. Persson I, Yuen J, Bergkvist L, et al. Cancer incidence and mortality in women receiving oestrogen or oestrogen-progestin replacement therapy e long term follow up in Swedish cohort. Int J Cancer 1996;67:327e32. 28. Mackie RM, Bray CA. Hormone replacement therapy after surgery for stage 1 or 2 cutaneous melanoma. Br J Cancer 2004; 90:770e2.
E. Miller et al. 29. Smith MA, Fine JA, Barnhill RL, et al. Hormonal and reproductive influences and risk of melanoma in women. Int J Epidemiol 1998;27:751e7. 30. Grin CM, Driscoll MS, Grant-Kels JM. The relationship of pregnancy, hormones and melanoma. Semin Cutan Med Surg 1998; 17:167e71. 31. Gloster HM, Neal K. Skin cancer in skin of color. J Am Acad Dermatol 2006;55:741e60. 32. Byrd KM, Wilson DC, Hoyler SS, et al. Advanced presentation of melanoma in African Americans. J Am Acad Dermatol 2004;51: 1031e3. 33. Hu W, Nelson JE, Mohney CA, et al. Malignant melanoma in a pregnant African American woman with congenital blue nevus. Dermatol Surg 2004;30:1530e2. 34. Lea CS, Holly EA, Hartge P, et al. Reproductive risk factors for cutaneous melanoma: a case-control study. Am J Epidemiol 2007;165:505e13. 35. Mondi MM, Cuenca RE, Ollila DW, et al. Sentinel lymph node biopsy during pregnancy: initial clinical experience. Annal Surg Oncol 2007;14:218e21. 36. Lloyd MS, Topping A, Allan R, et al. Contraindications to sentinel lymph node biopsy in cutaneous malignant melanoma. Br J Plast Surg 2004;57:725e7.