Fax: (716) 845-2320
The benefits of adjuvant radiation therapy after therapeutic lymphadenectomy for clinically advanced, high-risk, lymph node-metastatic melanoma
Article first published online: 21 AUG 2009
Copyright © 2009 American Cancer Society
Volume 115, Issue 24, pages 5836–5844, 15 December 2009
How to Cite
Agrawal, S., Kane, J. M., Guadagnolo, B. A., Kraybill, W. G. and Ballo, M. T. (2009), The benefits of adjuvant radiation therapy after therapeutic lymphadenectomy for clinically advanced, high-risk, lymph node-metastatic melanoma. Cancer, 115: 5836–5844. doi: 10.1002/cncr.24627
- Issue published online: 2 DEC 2009
- Article first published online: 21 AUG 2009
- Manuscript Accepted: 5 MAY 2009
- Manuscript Revised: 27 APR 2009
- Manuscript Received: 5 JAN 2009
- lymphatic metastasis;
The objective of this study was to evaluate the impact of adjuvant radiation therapy (RT) on regional recurrence and survival after therapeutic lymphadenectomy (TL) for clinically advanced, lymph node-metastatic melanoma.
Six hundred fifteen patients who had clinically advanced, regional lymph node-metastatic disease underwent TL. All patients were appropriate potential candidates for adjuvant RT (enlarged or multiple positive lymph nodes, extracapsular extension) because of a high risk for regional recurrence regardless of whether or not they received RT. Patient-related, tumor-related, and treatment-related variables that were associated with recurrence, survival, and treatment-related morbidity with and without RT were analyzed.
The median follow-up was 5 years. The actuarial 5-year regional lymph node basin control rate was 81%. On multivariate analysis, the number of positive lymph nodes, the number of lymph nodes removed, and the use of adjuvant RT were associated with improved regional control. Treatment-related morbidity, particularly lymphedema, was increased with the use of adjuvant RT and an inguinal site of lymph node metastases. At last follow-up, 268 patients were alive with actuarial 5-year distant metastasis-free survival (DMFS) and disease-specific survival (DSS) rates of 40% and 48%, respectively. On multivariate analysis, DMFS and DSS both were influenced by the number of positive lymph nodes and the number of lymph nodes removed. In addition, DSS was influenced by primary tumor thickness and the receipt of adjuvant RT.
Adjuvant RT was associated with improved regional lymph node basin control compared with TL alone in patients with high-risk, clinically advanced, lymph node-metastatic melanoma. Although it is a regional therapy, adjuvant RT also may have an impact on DSS. Cancer 2009. © 2009 American Cancer Society.
Approximately 15% to 20% of patients with melanoma will develop lymph node-metastatic disease. Therapeutic lymphadenectomy (TL) is potentially curative for patients who have clinical lymph node metastases; however, regional recurrences develop in 30% to 50% of these patients, and the 10-year OS (OS) rate is relatively poor (range, 25%-40%).1-12 Factors that are associated with regional recurrence include large lymph nodes, multiple positive lymph nodes, and the presence of extracapsular extension (ECE).8 The University of Texas M. D. Anderson Cancer Center has demonstrated that the use of adjuvant, regional radiation therapy (RT) in such patients is associated with an 89% 5-year regional lymph node basin control rate and has refined selection criteria for patients who are appropriate candidates for adjuvant RT.13-16 Although adjuvant RT can be effective in achieving regional control after TL, patients remain at risk for developing distant metastatic disease with 5-year distant metastasis-free survival (DMFS) and disease-specific survival (DSS) rates of 44% and 49%, respectively.16 Chronic lymphedema is the major morbidity associated with TL, and adjuvant RT increases this long-term toxicity, particularly in the inguinal region.16-20
To date, no randomized, prospective study of TL with and without adjuvant RT has been reported for high-risk, clinically advanced, lymph node-metastatic melanoma. Therefore, the objective of the current study was to analyze patient-related, tumor-related, and treatment-related variables associated with recurrence, OS, and morbidity in a large cohort of patients with melanoma who underwent TL with or without receiving adjuvant regional RT and to further identify subsets of patients who may benefit from this multidisciplinary approach.
Materials and Methods
After we received Institutional Review Board approval from each institution, a review of the melanoma databases at the University of Texas M. D. Anderson Cancer Center Department of Radiation Oncology and the Roswell Park Cancer Institute Department of Surgical Oncology identified 862 patients who underwent potentially curative TL for regional lymph node metastases from cutaneous melanoma. Further patient selection was based on 3 additional criteria: 1) the years of treatment (to provide at least 5 years of potential follow-up), 2) the disease presentation, and 3) defined “high-risk” criteria for regional lymph node recurrence. The years of treatment were from 1983 to 2003, and lymph node-metastatic disease had to be clinically advanced (either clinically or radiographically apparent) at presentation. Pathologic characteristics for high-risk, clinically advanced lymph node metastases varied by anatomic lymph node basin and have been described previously.8, 16 For cervical metastases, the criteria included the largest lymph node ≥2 cm, or ≥2 involved lymph nodes, or the presence of ECE. For axillary metastases, the criteria were the largest lymph node ≥3 cm, or ≥4 involved lymph nodes, or the presence of ECE. For inguinal metastases, the criteria were a combination of any 2 criteria, including the largest lymph node ≥3 cm, or ≥4 involved lymph nodes, or the presence of ECE. For epitrochlear metastases, the criteria were the largest lymph node >3 cm, or ≥4 involved lymph nodes, or the presence of ECE. These inclusion criteria identified the patients who were at high risk for regional recurrence after TL alone, allowing for an analysis of the potential benefit of adjuvant RT while minimizing the selection bias inherent in any single-institution, retrospective review. These criteria resulted in a final cohort of 615 patients.
The adjuvant RT techniques employed in this study have been described previously.13-15 Patients with cervical metastases were treated in an open-neck position with appositional electron fields covering the primary site, the parotid, and the ipsilateral cervical basin. Patients with axillary metastases were treated supine with their arm akimbo and received radiation to anterior and posterior photon fields to cover axillary levels I through III. Patients with inguinal metastases were treated supine with their leg abducted and externally rotated and received radiation to both electron and photon fields to cover the surgical scar and inguinal lymph nodes. The deep pelvic lymph nodes were treated if they were involved. The prescribed dose was almost entirely a hypofractionated regimen of 30 gray (Gy) delivered twice weekly at 6 Gy per fraction over 2.5 weeks. Only 5 patients were treated using more standard doses of radiation at 2 Gy per fraction (range, 50-54 Gy).
Treatment-related morbidity, including the severity of lymphedema, was classified retrospectively as grade 1, asymptomatic (noted on routine follow-up examination); grade 2, symptomatic, requiring medical therapy (compressive sleeve for lymphedema, physical therapy for neuropathy, or long-term use of pain medication); or grade 3, requiring surgical intervention.
OS was defined as the time from TL to either death or last follow-up, and DSS was defined as the time from TL to either death from melanoma or last follow-up. Recurrence of disease was defined as any clinical or radiographic evidence of relapse. Regional recurrence was defined as evidence of dermal, subcutaneous, soft tissue, or lymph node tumor within or around the lymph node basin where TL was performed. The date of distant failure was used to calculate DMFS. Actuarial data for regional control, DMFS, DSS, OS, and treatment-related morbidity curves were calculated by using Kaplan-Meier method, and tests of significance were based on the log-rank statistic. Multivariate analysis was done with the proportional hazards model using the Cox log-linear relative hazard function.
Patient, Tumor, and Treatment Characteristics
Of the 615 patients who underwent TL for clinical lymph node-metastatic melanoma, 472 were men, 143 were women, and the median age was 54 years (range, 14-89 years). The primary tumor site was head and neck in 205 patients (33%), trunk in 153 patients (25%), lower extremity in 53 patients (9%), upper extremity in 60 patients (10%), and unknown in 144 patients (23%). All patients who had a known primary tumor underwent wide excision with negative pathologic margins. The median primary tumor thickness was 2.3 mm (range, 0.2-50 mm). Clark level of invasion was known in 418 patients and was level II in 34 patients, level III in 111 patients, level IV in 203 patients, and level V in 49 patients. Disease stage at initial presentation (including before the development of lymph node metastases) was stage I in 65 patients, stage II in 195 patients, stage III (synchronous) in 321 patients, and not determined in 36 patients. Ulceration of the primary tumor was documented in 82 patients, and satellitosis was identified in 13 patients.
The median number of lymph nodes removed by TL was 22 (range, 1-114 lymph nodes). The median number of positive lymph nodes was 2 (range, 1-98 positive lymph nodes), and the median size of the largest positive lymph node was 3 cm (range, 0.2-26 cm). The clinical and pathologic characteristics of the patients with lymph node-metastatic disease are categorized further in Table 1.
|Characteristic||No. of Patients (%)|
|Anatomic lymph node basin|
|No. of involved lymph nodes|
|Size of largest lymph node, cm|
One hundred six of 615 patients (17%) who underwent TL in this study underwent surgery alone, while 509 patients (83%) also received adjuvant regional RT. Table 2 categorizes TL alone versus TL with adjuvant RT based on a high risk for lymph node basin relapse (anatomic lymph node basin and high-risk lymph node criteria). Two hundred fifteen patients also received adjuvant systemic therapy, including immunotherapy in 119 patients, cytotoxic chemotherapy in 43 patients, and combined biochemotherapy in 56 patients.
|Regional LN Basin||High-Risk Criteria*||No. of Patients|
|TL Alone||TL and Adjuvant RT|
|Cervical||LN ≥2 cm, or ≥2 LN(+), or ECE||24||268|
|Axilla||LN ≥3 cm, OR ≥4 LN(+), or ECE||59||194|
|Inguinal||LN ≥3 cm and ECE||14||36|
|LN ≥3 cm and ≥4 LN(+)||7||1|
|≥4 LN(+) and ECE||2||7|
|Epitrochlear||LN ≥3 cm, or ≥4 LN(+), or ECE||0||3|
Recurrence and Survival
At a median follow-up of 60 months (range, 5.3-284 months), 95 patients (15%) developed a regional recurrence, and 360 patients (59%) developed distant metastatic disease. By treatment group, only 52 of 509 patients (10.2%) who underwent TL and received adjuvant RT developed a regional recurrence versus 43 of 106 patients (40.6%) in the TL-alone group. For distant recurrence, the rate was 55.4% (282 of 509 patients) in the TL-plus-adjuvant-RT group versus 73.6% (78 of 106 patients) in the TL-alone group. The site of first recurrence was regional in 49 patients and distant in 312 patients. There were 14 synchronous recurrences. The actuarial 5-year regional control rate for all patients was 81%. Of the 95 regional recurrences, 38 were from cervical lymph node-metastatic disease, 44 were from axillary disease, and 13 were from inguinal disease. Table 3 shows the 5-year regional control rate according to which lymph node basin was treated and the presence or absence of adjuvant RT.
|Regional LN Basin||TL Alone, %||TL and Adjuvant RT, %||P|
The median time to regional recurrence was 9.2 months (range, 1-78 months). The 5-year OS for the 95 patients who developed regional recurrence was 20% compared with 53% for patients without regional recurrence (P<.001). On multivariate analysis, the number of lymph nodes removed at TL, the number of positive lymph nodes, and the administration of adjuvant RT were associated with improved regional control (Table 4). Figure 1 shows the regional control curves for TL alone versus TL plus adjuvant RT. Higher regional control rates were observed with adjuvant RT regardless of the number of lymph nodes removed at TL. In patients who had <15 lymph nodes removed, the regional control rate with and without RT was 82% and 52%, respectively (P<.001). In patients who had ≥15 lymph nodes removed, the regional control rate with and without RT was 90% and 53%, respectively (P<.001).
|Factor||RC, %||P||DMFS, %||P||DSS, %||P|
|No. of LN(+)|
|Size of largest LN, cm|
|No. of LNs removed|
|Anatomic lymph node basin|
The median time to distant recurrence was 9.7 months (range, 1-187 months), and the actuarial 5-year DMFS rate was 40%. In the 326 patients who had a distant site of first recurrence, the most common site was the lung in 120 patients (40%), followed by the brain in 52 patients (16%), distant skin and subcutaneous in 43 patients (13%), the liver in 39 patients (12%), nonregional lymph nodes in 24 patients (7%), and bone in 22 patients (7%). At the time of last follow-up, 268 patients remained alive with a corresponding 5-year actuarial OS rate of 44%. The actuarial 5-year DSS rate for all patients was 48%. Table 4 shows that several tumor-related and treatment-related variables were associated with differences in DMFS and DSS. On multivariate analysis, the number of lymph nodes removed at TL and the number of positive lymph nodes were associated with DMFS. In addition to those variables, primary tumor thickness and the use of adjuvant RT also were associated with the DSS. Figure 2 shows the DSS curves for patients who underwent TL alone versus patients who underwent TL plus adjuvant RT.
Ninety-eight of 615 patients (16%) developed clinically significant, treatment-related morbidity that required either medical management (grade 2) or surgical management (grade 3). At 3 years, 5 years, and 10 years, the overall treatment-related morbidity rates were 18%, 19%, and 22%, respectively. There were significant differences in treatment-related morbidity based on which anatomic lymph node basin was involved. For epitrochlear, cervical, axillary, and inguinal lymph node metastases, the 5-year morbidity rates were 0%, 11%, 21%, and 49%, respectively (P<.0001). The addition of adjuvant RT was associated with a 5-year morbidity rate of 20% versus 13% for TL alone (P = .004). On multivariate analysis, both the receipt of adjuvant RT and an inguinal site of lymph node metastases were associated with increased treatment-related morbidity.
The most common treatment-related morbidity was symptomatic grade 2 lymphedema (53 patients). The 3-year, 5-year, and 10-year rates of symptomatic lymphedema were 17%, 19%, and 23%, respectively. The risk of lymphedema varied according to which anatomic lymph node was basin treated. The 5-year rates for epitrochlear, cervical, axillary, and inguinal disease were 0%, 1%, 13%, and 44%, respectively (P<.0001), and the rate remained significant on multivariate analysis (P<.0001). Grade 2 lymphedema was noted in 45 of 509 patients (9%) who received adjuvant RT versus 8 of 106 patients (7.5%) who underwent TL alone. On multivariate analysis, the 5-year rate of lymphedema differed significantly between adjuvant RT versus TL alone (P = .001).
In the current retrospective analysis, we compared the patterns of failure and outcomes between 509 patients who underwent TL with adjuvant RT and 106 patients who underwent TL alone for clinically advanced, high-risk, lymph node-metastatic melanoma. Patients were treated at 2 tertiary cancer centers that had widely divergent approaches to the treatment of regional lymph node metastases from melanoma. At Roswell Park Cancer Institute, patients historically underwent lymphadenectomy alone; whereas, at The University of Texas M. D. Anderson Cancer Center, patients were traditionally treated with adjuvant RT in addition to surgery. At both institutions, this treatment schema was applied regardless of pathologic characteristics. By matching the years of treatment and selecting only those patients with lymph node disease who were at high risk for recurrence, we were able to focus on the potential effects of adjuvant RT while minimizing, as much as possible, any associated selection bias. In support of our hypothesis, adjuvant RT was associated with improved regional lymph node basin control compared with TL alone for high-risk, clinically advanced, lymph node-metastatic melanoma. It is noteworthy that, although it is a regional therapy, we also observed that adjuvant RT also may have a positive impact on DSS.
Regional lymph node basin recurrence after TL for lymph node-metastatic melanoma is a significant problem, and the rates are as high as 20% to 50%, especially among patients who have high-risk factors, such as large lymph nodes, multiple positive lymph nodes, or the presence of ECE.1, 7-12 Karakousis et al observed that the size of the largest metastatic lymph node was a significant predictor of outcome and reported a 5-year survival rate of 73% for patients who had palpable lymph nodes that measured <2 cm in greatest dimension versus 46% for patients who had palpable lymph nodes that measured between 2 cm and 4 cm.21 In contrast, others have reported that the size of the largest positive lymph node on physical examination or final pathology was not associated significantly with survival.22 Several studies have demonstrated that ECE is an independent predictor of OS in lymph node-metastatic melanoma.2, 4, 19, 23-25 This is in contrast to Rao et al, who reported that ECE was associated with decreased recurrence-free survival but not with decreased OS.26 In the current study, only the number of positive lymph nodes was associated with the risk of regional recurrence.
Several investigators have demonstrated that the number of lymph nodes removed at TL, the total number of positive lymph nodes, and the lymph node ratio (positive lymph nodes/total number of lymph nodes removed) can be predictors of both regional recurrence and OS for patients with lymph node-metastatic melanoma.2-4, 7, 19, 20, 23-25, 27, 28 In 197 patients who underwent TL for axillary lymph node-metastatic melanoma, Bevilacqua et al observed that both the total number of lymph nodes and the percentage of positive lymph nodes were associated significantly with survival on univariate analysis.2 However, only the percentage of positive lymph nodes was predictive of OS on multivariate analysis. Chan et al demonstrated that both tumor burden and the quartile of lymph nodes removed at TL were independent factors predictive of OS, suggesting that the extent of lymph node dissection is important, particularly among patients with higher tumor burdens.27 In the current series, the total number of positive lymph nodes and the number of lymph nodes removed at TL were associated significantly with the risk of regional recurrence and DSS. The predictive value of the total number of lymph nodes removed at TL (independent of the total number of positive lymph nodes) for regional recurrence and survival in this and other studies suggests not only that an oncologically complete regional lymphadenectomy is important for prognostic purposes but also that it may have an effect on long-term survival.
In the era of sentinel lymph node biopsy, many patients with clinically occult lymph node-metastatic melanoma are at low risk for developing a regional recurrence after TL. Therefore, not every patient with lymph node-metastatic disease should undergo adjuvant RT. In a review of 253 patients who underwent TL, Lee et al noted a 36% regional recurrence rate at 10 years and reported that patients who had certain features, such as cervical lymph node basin, large lymph nodes, multiple positive lymph nodes, and ECE, were at even higher risk for recurrence.8 Their conclusion was that these high-risk patients may be appropriate candidates for adjuvant RT. Several nonrandomized studies, including those from The University of Texas M. D. Anderson Cancer Center, have demonstrated that adjuvant RT for lymph node-metastatic melanoma can reduce the risk of regional recurrence to between 5% and 20%.13-17, 29-34 In a prospective, phase 2 study by the Trans Tasman Radiation Oncology Group (TROG Study 96.06) of adjuvant RT after surgical resection for patients with lymph node-metastatic melanoma, the regional control rate was 91%, and the 5-year OS rate was 36%.17 Despite the significant reductions in regional recurrence, those studies have not suggested a survival benefit with adjuvant RT, because >50% of the patients in those studies developed distant metastatic disease by 5 years.16, 17 In the current series, adjuvant RT not only was associated with a significant decrease in the risk for regional recurrence, but it also was statistically significant on multivariate analysis in terms of an improved DSS. A similar benefit in OS with improved regional control also has been reported in prospective randomized trials for other nonmelanoma cancers, such as breast cancer, head and neck cancer, and nonsmall cell lung cancer.35-37
Adjuvant RT was well tolerated by most of the patients in our study. The most common complication was symptomatic lymphedema requiring some form of medical management. The risk for lymphedema clearly increased with the use of adjuvant RT but also was highly dependent on treatment of the anatomic lymph node basin (13% for the axilla vs 44% for the groin at 5 years). TROG Study 96.06 also demonstrated mild-to-moderate lymphedema as the most significant late morbidity; it occurred in 9% of patients who had axillary metastases and in 19% of patients who had inguinal disease.17 These differences in the rates almost certainly are because of the various lengths of follow-up and the various definitions of clinically significant lymphedema.
In conclusion, our retrospective analysis of a large group of well matched patients with high-risk, clinically advanced, lymph node-metastatic melanoma demonstrated that the addition of adjuvant RT to TL was associated with a significantly lower risk of regional recurrence. This benefit also was observed in patients who underwent extensive lymphadenectomy with the removal of large numbers of lymph nodes, suggesting that adjuvant RT may have a role in the treatment of all patients who have high-risk, lymph node-metastatic disease (and not simply as salvage therapy for patients who undergo inadequate surgery). In addition to the improvement in regional control, we also observed that adjuvant RT may improve survival from melanoma. One cannot simply dismiss this provocative finding as retrospective “selection bias,” because both institutions involved in this study approached the treatment of patients with clinically advanced, lymph node-metastatic melanoma in a different but uniform way during the defined period. In addition, if there was any inherent treatment bias, then the tendency would have been to give adjuvant RT to the highest risk patients (large lymph nodes, multiple positive lymph nodes, ECE, etc) who also would be at the greatest risk of dying from distant metastatic disease. From a tumor biology standpoint, although lymph node-metastatic melanoma often is thought of conceptually as a marker of a more aggressive primary tumor that carries a concomitant higher risk of early distant metastases (rather than as the cause of the distant metastatic disease), the findings in other types of cancer (breast, head and neck, and nonsmall cell lung cancers) that therapies that increase locoregional control also may improve OS suggest that the latter tumor biology may play a role in certain subsets of patients. Based on our current findings and those in other cancers, it is time to strongly consider a multicenter, prospective, randomized study of TL with and without adjuvant RT for patients with high-risk, lymph node-metastatic melanoma in an attempt to better define the potential benefits in regional control and possibly even survival.
Conflict of Interest Disclosures
The authors made no disclosures.