| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Johns Hopkins Medical Center (CMB), Baltimore, Maryland; University of Alabama at Birmingham, Birmingham, Alabama (S-jS, MMU, RD); Morristown Memorial Hospital, Morristown, New Jersey (TS); University of Texas MD Anderson Center (MIR), Houston, Texas; Roswell Park Cancer Institute (CPK), Buffalo, New York; University of Calgary (WJT), Calgary, Alberta, Canada; Massachusetts General Hospital (MCM, RLB), Boston, Massachusetts; University of Kansas (WRJ), Kansas City, Kansas; Roger Williams Hospital (HJW), Providence, Rhode Island.
Correspondence: Address correspondence and reprint requests to: Charles M. Balch, MD, 1900 Duke St., Suite 200, Alexandria, VA 22314; Fax: 703-299-1044; E-mail: balchc{at}asco.org
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: There were two cohorts entered into a prospective multi-institutional trial: (1) 468 patients with melanomas on the trunk or proximal extremity who randomly received a 2 cm or 4 cm radial excision margin and (2) 272 patients with melanomas on the head, neck, or distal extremities who received a 2 cm radial excision margin.
Results: A local recurrence (LR) was associated with a high mortality rate, with a 5-year survival rate of only 9% (as a first relapse) or 11% (anytime) compared with an 86% survival for those patients who did not have a LR (P < .0001). The 10-year survival for all patients with a LR was 5%. The 10-year survival rates were not significantly different when comparing 2 cm vs. 4 cm margins of excision (70% vs. 77%) or comparing the management of the regional lymph nodes (observation vs. elective node dissection). The incidences of LR were the same for patients having a 2 cm vs. 4 cm excision margin regardless of whether the comparisons were made as first relapse (0.4% vs. 0.9%) or at anytime (2.1% vs. 2.6%). When analyzed by anatomic site, the LR rates were 1.1% for melanomas arising on the proximal extremity, 3.1% for the trunk, 5.3% for the distal extremities, and 9.4% for the head and neck. The most profound influence on LR rates was the presence or absence of ulceration; it was 6.6% vs. 1.1% in the randomized group involving the trunk and proximal extremity and was 16.2% vs. 2.1% in the non-randomized group involving the distal extremity and head and neck (P < .001). A multivariate (Cox) regression analysis showed that ulceration was an adverse and independent factor (P = .0001) as was head and neck melanoma site (P = .01), while the remaining factors were not significant (all with P > .12).
Conclusion: For this group of melanoma patients, a local recurrence is associated with a high mortality rate, a 2-cm margin of excision is safe and ulceration of the primary melanoma is the most significant prognostic factor heralding an increased risk for a local recurrence.
Key Words: Melanoma • Surgical excision margins • Neoplasm staging • Cox regression • Risk factors.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The World Health Organization Melanoma Program conducted the first randomized trial that addressed this issue.9,10 Their surgical protocol (Trial #10) compared 1-cm vs. 3-cm radial margin of excision in a limited eligibility group of patients with melanoma thickness of 2.0 mm or less. This patient group was selected because the risk of a local recurrence was low, and therefore it would be safe to test a more skin-sparing margin of excision. Long-term results of this trial showed that a 1-cm margin of excision is safe, and that there was no difference in survival rates compared with a 3-cm radial margin. However, there was a clustering of local recurrence rates up to 4.5% confined to patients who had a tumor thickness of 1.1 to 2.0 mm and who had a 1-cm margin of excision (Cascinelli, N; personal communication).
The Intergroup Melanoma Surgical Trial has been conducted since 1983 in a cohort of 740 patients with intermediate thickness melanomas (i.e., 1.0–4.0 mm) to prospectively determine: (1) whether a 2-cm radial margin of excision is equivalent (or safe) compared with a 4-cm radial margin with respect to survival and local recurrence rates; and (2) whether currently available prognostic factors could identify subgroups of patients with intermediate thickness melanomas who had a high risk for local recurrence and for whom a wider margin of excision would reduce the risk of local recurrence. Patient eligibility was confined to those patients with melanomas located on the trunk or proximal extremity to ensure that the more radical 4-cm margin could be carried out if they were randomly assigned to receive such a treatment. All patients were also randomly selected to receive an elective lymph node dissection (ELND) or observation of their clinically uninvolved nodes as their initial management. The results of the ELND portion of the randomized surgical trial have been published.11,12
This is the largest randomized surgical trial ever conducted to address the definitive surgical management of a primary melanoma and to confirm the validity of prognostic factors to define risk groups. For the purposes of identification, these patients will be referred to in this article as "Group A" or the randomized group. An interim report that showed no overall survival difference in 2-cm vs. 4-cm margins were compared and published after 6 years and 7.6 years of follow-up with the limitation that a number of patients were still developing potentially fatal distant metastasis.13,14 The data set is now matured to make definitive conclusions because the relapse rates for all stages have plateaued.
A second group of patients were prospectively entered into the surgical protocol whose melanoma was located on the distal extremity or head and neck area, as long as they could have a 2-cm radial margin of excision. Such patients were assigned randomly to receive either an ELND or nodal observation as described above. These patients are referred to as "Group B" or non-randomized patients for a standardized surgical margin because it was technically not feasible to perform a 4-cm radial surgical margin; all such patients had a 2-cm surgical margin of excision.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
All surgeons were board certified and accredited members of an established cancer cooperative group. Rigorous quality control was used throughout the study, which included standardized descriptions of the lymphadenectomy procedures, educational sessions at group meetings, centralized surgical quality control review, and use of quality control data forms as previously described.13 An experienced pathologist from the participating cooperative group reviewed the slides and at least two thirds of the slides were also reviewed by a central panel of melanoma pathologists who independently reviewed a representative histologic section of each primary melanoma. Regular audits were performed on all research records by one or more members of the Intergroup Melanoma Surgical Committee. The principle investigator reviewed the circumstances and medical documentation of all deaths and was blinded as to the surgical treatment received.
A local recurrence (LR) was defined as any melanoma recurrence within 2 cm of the surgical scar after a definitive surgical excision. The LR rates were also recorded separately when they were the first manifestation of relapse (1st LR) or that occurred at anytime in the course of progressive metastatic disease (anytime LR group). The rates of LR are expressed as the cumulative rate with an average follow-up of 10 years. Because all the local recurrences occurred within 8 years, it is unlikely that these rates would change with a longer follow-up period beyond the mean 10-year follow-up available now.
Statistical Analysis
Standard statistical techniques were used. Proportions were compared using 
2 analysis. Disease-specific survival from melanoma and disease recurrence curves were constructed by using the Kaplan-Meir product limit method. These curves were analyzed for comparisons by the log-rank procedure. The P value indicated after comparison of the 10-year survival rates was calculated based on comparisons of the survival curves. Multivariate analysis based on Cox’s regression model was used to associate covariates to time-dependent endpoints such as survival. Results are reported based on "randomized intent" which included those patients who refused the randomized treatment assigned.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Patient accrual began in 1983 and stopped in 1989. More than 94% of the patients entered into the study were eligible and were able to be evaluated. There is now a 92% long-term follow-up of at least 5 years or until death. This is a mature database with a median follow-up of 10 years and a range up to 16 years. The clinical and pathological characteristics of the patients are exactly balanced among prognostic factors as previously published.11,13,14
Overall Survival and Incidence Rates of Local Recurrences (LR)
Overall, there were 13 patients who had a LR as a first relapse and a total of 28 patients who had a LR at anytime in the course of their metastatic melanoma. A local recurrence was associated with a highly significant difference in survival compared with patients who never experienced a LR (Fig. 1), regardless of whether survival rates compare 1st LR events alone or those that occurred at anytime (Fig. 2). Thus, those patients who had a LR all died of metastatic melanoma within 10 years.
|
|
For the entire group of 740 patients, the anatomic site of the primary melanoma significantly correlated with the LR rate (P < .001). Thus, the LR rate (anytime) for 177 patients with a melanoma arising on the proximal extremity was only 1.1%, while it rose to 3.1% for the 291 patients with trunk melanoma, to 5.3% among the 208 patients with distal extremity melanomas, and 9.4% among the 64 patients with melanomas of the head or neck.
The correlation of thickness and ulceration remained when the analysis was conducted within the prospectively defined cohorts of Group A and Group B patients. Among Group A patients (i.e.: with trunk or proximal melanoma melanomas), the local recurrence rates (anytime) increased from 1.0% for melanomas of 1.0–2.0 mm thick, to 4.6% for 2.1–3.0 mm melanoma, and to 4.1% for those 3.1–4.0 in thickness (Table 1)(P < .05). Among Group B patients (i.e.: with distal extremity or head and neck melanomas), the LR rates were higher, ranging from 3.8% for 1.0 to 2.0 mm melanomas, 5.6% for 2.0 to 3.0 mm melanomas and 23.1% for 3.0 to 4.0 mm melanomas (P = .05 by single factor analysis, but P = .14 by multifactorial analysis).
|
Survival and Incidence Rates Among the Randomized Patients (Group A)
There were no differences in survival rates overall or incidence rates of LR comparing patients who had a 2-cm vs. 4-cm surgical margin (Fig. 3; Table 2). For the 238 patients who had a 2-cm margin, the incidence of LR was 0.4% as a first relapse and 2.1% at anytime. For those 230 patients that had a 4-cm margin, the LR rates was 0.9% as a first relapse and 2.6% at anytime. Ten year survival rates for the two groups had no significant differences, being 70% 10 year disease-specific survival for those who had a 2-cm margin vs. 77% for those who had a 4-cm margin (P = .074, Figure 3). When the survival rates were further divided based upon whether the patients had ELND or not, there was still no survival difference (P = .16) (Table 2). The median time to local recurrence was 1.7 years and the median survival after LR was only 1 year.
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Two randomized trials in Europe compared 2-cm vs. 5-cm radial margins for melanomas up to 2.0 mm in thickness and showed no survival advantage for the more radical surgical margin.16,17 The Swedish Melanoma Study Group patients had local recurrence rates of 0.8% for those 373 patients who had a 2-cm margin and a 1.0% rate for those 396 patients who had a 5-cm margin after a follow-up period of 5.8 years. These rates are comparable with those of the present Intergroup study, although their follow-up period is shorter.16
This study also demonstrated that a LR is associated with a strikingly high mortality rate. The 10 year survival rate was only 5% for those patients who had a LR, either as a first relapse or at anytime in the course of metastatic disease. We had postulated at the outset of the trial that a LR may have been a manifestation of retained primary melanoma cells or neighboring intralymphatic satellites. However, the patterns of metastases analysis showed that a LR was more likely a manifestation of Stage IV melanoma. First, the next site of relapse in 62% of patients was at distant sites, not regional sites as would be expected if a LR followed the patterns of recurrences of a primary melanoma or satellite metastasis which would next relapse at a regional site in the majority of patients. Second, a LR was associated with a higher rate of metastases in the skin and subcutaneous tissues than with any other metastatic site. It is probable, therefore, that a LR is equivalent to a skin or subcutaneous metastasis that coincidentally occurs in or around the surgical scar of a primary melanoma excision. This does not exclude the possibly that some LR reflect the biological manifestation of regrowth of retained primary melanoma cells or intralymphatic metastases that are very aggressive with regard to their metastatic behavior.
This study extends previous reports and validates the powerful predictive value of melanoma ulceration that is associated with a strikingly high incidence of local recurrences and metastatic melanoma.3,4,6,18–24 An ulcerated melanoma is, in every respect, a locally advanced and aggressive cancer. The presence of ulceration thus portends a more invasive metastatic property of a melanoma and a substantially higher risk for regional and distant metastases. This study demonstrates the striking influence of ulceration on local recurrence rates where the incidence of LR increased 6-fold to 8-fold compared with non-ulcerated melanomas of intermediate thickness. Increasing tumor thickness correlated significant with the risk of a LR, with the risk increasing by 4-fold to 6-fold as the melanoma thickness increased from 1.0 up to 4.0 mm The risk of a local recurrence was substantially increased when comparing the Group A and Group B patients of equivalent thicknesses, demonstrating the importance of anatomic site of the primary melanoma as a predictive factor for a LR.
In this study, as with the preponderance of other publications, anatomic site was an independent predictor of outcome, with patients having their melanomas arising on the head and neck having a worse prognosis compared with other anatomical sites.3,14,25-29
Given this knowledge of prognostic factors predicting the risk of a LR, can the surgeon prevent a local recurrence with a wider surgical margin in those patients with a "high risk" for this event? Intuitively, one would postulate that a wider surgical margin for an ulcerated melanoma might be associated with a lower risk of a LR. This would especially be true if a wider excision of skin and subcutaneous tissue were removing lateral extensions of primary melanoma cells or microsatellites to a greater extent than a narrower surgical margin. This was not the case, however, further substantiating the conclusion that a LR is a first manifestation of a Stage IV melanoma. Thus, the analysis comparing local recurrence rates in patients having a 4-cm excision was no lower than those having a 2-cm excision when compared for equivalent thickness groups or ulceration status (Table 5).
This trial also dispels the previous notion that a skin graft would enable the clinician to better detect a local recurrence earlier through a thin coverage of the surgical defect. First, the incidence is too low to justify an approach that is more expensive and disfiguring. Second, since all patients with a LR eventually died, regardless of width of margin or closure by STSG, it is highly unlikely that any patients would be salvaged by a slightly earlier detection of a LR. And finally, the multifactorial analysis showed that closure by STSG or a primary closure with a full-thickness flap had no impact on survival rates, even after accounting for all the other factors that influence the risk of a LR (Table 3).
This surgical trial does not specifically address the surgical management of several important groups of melanoma patients. Those patients with lentigo maligna melanoma and desmoplastic melanomas were excluded as were patients with thick melanomas (>4.0 mm). Subsequent surgical trials will need to address the optimal surgical management of these patients as well as examining whether even narrower margins might be safe in some subgroups.24 For the present, however, those patients comprising the vast majority of melanoma patients can have an appropriate excision margin of their melanoma with a risk of local recurrence of <1.0% and no compromise in their survival rates if they have a 1-cm radial margin for a melanoma up to 1 mm in thickness and a 2-cm margin for up to 4.0 mm melanomas. In both circumstances, the surgical defect can (and should) be closed primarily with a full thickness flap.
APPENDIX
Surgeons from the following cancer cooperative groups participated in this study: Eastern Cooperative Oncology Group, Southwest Oncology Group, National Surgical Adjuvant Breast Project, Cancer and Leukemia Group B, NCIC, MAOP, Pediatric Oncology Group, and the Danish Melanoma Group.
The following institutions participated in this melanoma surgical trial. The first 15 institutions listed below entered 77% of the total patient accrual. In addition to the authors, those surgeons who entered five or more patients are listed in parentheses after each institution.
Roswell Park Memorial Institute
University of Texas-M.D. Anderson Cancer Center (A. Boddie, D. Hohn, K. Jessup, D.D. Ota, R. Pollock, M. Romsdahl, M. Ross)
University of Alabama at Birmingham (W. Maddox)
McGill University (A. Loutfi, H. Shibata)
University of Florida (K. Bland)
Case Western Reserve University (E. Mansour)
City of Hope Medical Center (J. Terz)
Ohio State University (W. Farrar)
Tom Baker Cancer Center, University of Calgary
Danish Melanoma Group (C. Krag)
Emory University (D. Murray)
Washington Hospital Center (M. Cohen)
Scott and White Medical Center (C. Verheyden)
Medical Center of Delaware
The University of Virginia
Spartanburg CCOP (J. McCulloch)
Indiana University Medical Center (J. Bennett)
University of Kansas Medical Center
University of Pretoria South Africa
Tufts University
University of California at San Diego
Washington University Medical Center
Tulsa University (J. Lockhart)
University of Arkansas (D. Chu)
Good Samaritan Hospital (R. Welling)
Baptist Medical Center, Oklahoma (K. Boatman)
Letterman Army Center
|
Footnotes |
|---|
Received for publication July 27, 2000. Accepted for publication October 26, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. Ariyan, M. S. Brady, M. Gonen, K. Busam, and D. Coit Positive Nonsentinel Node Status Predicts Mortality in Patients with Cutaneous Melanoma Ann. Surg. Oncol., January 1, 2009; 16(1): 186 - 190. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Thirlwell and P. Nathan Melanoma--Part 2: management BMJ, December 1, 2008; 337(dec01_1): a2488 - a2488. [Full Text]
|
|
|
|
 |
|
 F. Grange, F. Vitry, F. Granel-Brocard, D. Lipsker, F. Aubin, G. Hedelin, S. Dalac, F. Truchetet, C. Michel, M.-L. Batard, et al. Variations in Management of Stage I to Stage III Cutaneous Melanoma: A Population-Based Study of Clinical Practices in France Arch Dermatol, May 1, 2008; 144(5): 629 - 636. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Erickson, J. M. Velasco, and T. J. Hieken Compliance with Melanoma Treatment Guidelines in a Community Teaching Hospital: Time Trends and Other Variables Ann. Surg. Oncol., April 1, 2008; 15(4): 1211 - 1217. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. B. Lens, P. Nathan, and V. Bataille Excision Margins for Primary Cutaneous Melanoma: Updated Pooled Analysis of Randomized Controlled Trials Arch Surg, September 1, 2007; 142(9): 885 - 891. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M. Balch, M. I. Ross, N. Cascinelli, and S.-j. Soong Excision Margins for Primary Cutaneous Melanoma Invited Critique Arch Surg, September 1, 2007; 142(9): 891 - 893. [Full Text] [PDF]
|
|
|
|
|
|
K. M. Dalal, A. Patel, M. S. Brady, D. P. Jaques, and D. G. Coit Patterns of First-Recurrence and Post-recurrence Survival in Patients with Primary Cutaneous Melanoma After Sentinel Lymph Node Biopsy Ann. Surg. Oncol., June 1, 2007; 14(6): 1934 - 1942. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. R. Yurkovetsky, J. M. Kirkwood, H. D. Edington, A. M. Marrangoni, L. Velikokhatnaya, M. T. Winans, E. Gorelik, and A. E. Lokshin Multiplex Analysis of Serum Cytokines in Melanoma Patients Treated with Interferon-{alpha}2b Clin. Cancer Res., April 15, 2007; 13(8): 2422 - 2428. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. N. Markovic, L. A. Erickson, R. D. Rao, R. H. Weenig, B. A. Pockaj, A. Bardia, C. M. Vachon, S. E. Schild, R. R. McWilliams, J. L. Hand, et al. Malignant Melanoma in the 21st Century, Part 2: Staging, Prognosis, and Treatment Mayo Clin. Proc., April 1, 2007; 82(4): 490 - 513. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Kretschmer, I. Beckmann, K.-M. Thoms, C. Mitteldorf, H. P. Bertsch, and C. Neumann Factors Predicting the Risk of In-Transit Recurrence After Sentinel Lymphonodectomy in Patients With Cutaneous Malignant Melanoma Ann. Surg. Oncol., August 1, 2006; 13(8): 1105 - 1112. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. N. Cormier, Y. Xing, M. Ding, J. E. Lee, P. F. Mansfield, J. E. Gershenwald, M. I. Ross, and X. L. Du Population-Based Assessment of Surgical Treatment Trends for Patients With Melanoma in the Era of Sentinel Lymph Node Biopsy J. Clin. Oncol., September 1, 2005; 23(25): 6054 - 6062. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Tsao, M. B. Atkins, and A. J. Sober Management of Cutaneous Melanoma N. Engl. J. Med., September 2, 2004; 351(10): 998 - 1012. [Full Text] [PDF]
|
|
|
|
|
|
J. M. Thomas, J. Newton-Bishop, R. A'Hern, G. Coombes, M. Timmons, J. Evans, M. Cook, J. Theaker, M. Fallowfield, T. O'Neill, et al. Excision Margins in High-Risk Malignant Melanoma N. Engl. J. Med., February 19, 2004; 350(8): 757 - 766. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. E. Krown and P. B. Chapman Defining Adequate Surgery for Primary Melanoma N. Engl. J. Med., February 19, 2004; 350(8): 823 - 825. [Full Text] [PDF]
|
|
|
|
|
|
B. A. Pockaj, D. E. Jaroszewski, D. J. DiCaudo, J. G. Hentz, E. W. Buchel, R. J. Gray, S. N. Markovic, and U. Bite Changing Surgical Therapy for Melanoma of the External Ear Ann. Surg. Oncol., July 1, 2003; 10(6): 689 - 696. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. B. Lens, M. Dawes, T. Goodacre, and J. A. N. Bishop Excision Margins in the Treatment of Primary Cutaneous Melanoma: A Systematic Review of Randomized Controlled Trials Comparing Narrow vs Wide Excision Arch Surg, October 1, 2002; 137(10): 1101 - 1105. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
