|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology
Sentinel-Node Technique Will Change the Design of Clinical Trials in Malignant MelanomaUniversity of Gottingen, Gottingen, Germany To the Editor:Kirkwood et al1 recently published the results of the intergroup trial E1694, designed to test the efficacy of the GMK ganglioside vaccine versus high-dose interferon alfa-2b (HDI) in malignant melanoma. Analyzing a heterogeneous population of high-risk patients (141 T4N0M0 patients with no lymph node surgery, 56 T4N0M0 patients with negative sentinel nodes, 315 patients with microscopic lymph node metastasis, and 306 patients with clinically detectable node metastasis), they found a minor overall survival benefit for HDI of 5% by 2 years (median duration of follow-up, 16 months). The recurrence-free survival rate was significantly improved by HDI treatment (13% by 2 years). However, only the subset of T4N0M0 patients benefited significantly from HDI treatment. As shown by studies dealing with sentinel-node dissection,2,3 this group consists of two subpopulations with quite different prognosis: patients with occult micrometastases to their regional lymph nodes and patients with tumor-free regional nodes. The 3-year survival of sentinel-negative T4N0M0 patients was recently observed to be as high as 89.8%.2 Thus, the inclusion of these patients in trial E1694 is questionable. Moreover, epidermal ulceration has been found to be a significant prognostic factor in T4N0M0 patients with thick primary tumors2,3 as well as in node-positive patients.4 Unfortunately, in trial E1684 this feature was not considered in the multivariate analysis. Thus, an alternative interpretation of trial E1684 could be that ulceration or occult regional lymph node involvement were not equally distributed among two therapy groups in the subsets of T4N0M0 patients because of low case numbers in this heterogeneous stratum. Therefore, it would be important to assess the frequency of recurrence to the regional lymph nodes in the T4N0M0 patients who did not receive a sentinel biopsy. If the T4N0M0 patients with HDI fared better because of a lower proportion of recurrences to their regional lymph nodes, a lower proportion of micrometastases to their sentinel nodes at the time of randomization is the most likely explanation. In T4N0M0 patients, the most frequent pathway of first recurrence is locoregional. Because skin and lymph node metastases can be easily treated by surgery with low morbidity, the effect of HDI treatment on the development of visceral metastases and on overall survival seems to be more important. Therefore, a longer duration of follow-up is mandatory to assess the real clinical benefit. In trial E1694, the survival benefit of HDI in node-negative patients was so strong that it outweighed the missing benefit in node-positive patients. Consequently, the authors considered HDI treatment significant for the whole study population. Regarding node-positive patients, some restrictions have to be made as both patients with micro- and macrometastasis were put together in the same strata. However, the clinical stage of the disease (micro- v macrometastases) has been shown to be an independent predictor of survival, besides the number of positive nodes.5 As illustrated by our own findings,6 the different time points of diagnosis in patients with micro- and macrometastases result in significantly different survival rates because of the so called lead time bias. In patients with delayed lymph node dissection, we observed an apparent survival benefit of 27.1% by 2 years if the survival rate as calculated from primary tumor excision (when micrometastases are expected to be present) was compared with the survival rate as calculated from the appearance of palpable nodes. Taken together, it would be interesting to know whether the overall survival benefit of HDI for the whole study population holds true in a multifactorial analysis taking into consideration clinical stage, ulceration, the number of positive nodes, Breslow thickness, age, and sex. In conclusion, mixing sentinel node–negative patients and patients with unknown lymph node status as well as patients with micro- and macrometastasis in the same strata makes it difficult to draw conclusions. In future studies, the regional lymph nodes should be generally staged by sentinel-node biopsy. Moreover, the prognosis of sentinel-negative patients needs further exploration to answer the question if systemic therapies, associated with significant toxicity as HDI, are justified for these patients. REFERENCES
1.
Kirkwood JM, Ibrahim JG, Sosman JA, et al: High-dose interferon alfa-2b significantly prolongs relapse-free and overall survival compared with the GM2-KLH/QS-21 vaccine in patients with resected stage IIB-III melanoma: Results of Intergroup Trial EI694/S9512/C509801. J Clin Oncol 19: 2370-2380, 2001 2. Gershenwald JE, Mansfield PF, Lee JE, et al: Role for lymphatic mapping and sentinel lymph node biopsy in patients with thick (> or = 4 mm) primary melanoma. Ann Surg Oncol 7: 160-165, 2000[Abstract] 3. Cherpelis BS, Haddad F, Messina J, et al: Sentinel lymph node micrometastasis and other histologic factors that predict outcome in patients with thicker melanomas. J Am Acad Dermatol 44: 762-766, 2001[CrossRef][Medline] 4. Balch CM, Soong SJ, Murad TM, et al: A multifactorial analysis of melanoma: III. Prognostic factors in melanoma patients with lymph node metastases (stage II). Ann Surg 193: 377-388, 1981[Medline] 5. Coit DG, Rogatko A, Brennan MF: Prognostic factors in patients with melanoma metastatic to axillary or inguinal lymph nodes: A multivariate analysis. Ann Surg 214: 627-636, 1991[Medline] 6. Kretschmer L, Preußer K-P, Marsch WC, et al: Prognostic factors of overall survival in patients with delayed lymph node dissection for cutaneous malignant melanoma. Melanoma Res 10: 483-489, 2000[CrossRef][Medline]
Sentinel-Node Technique Will Change the Design of Clinical Trials in Malignant Melanoma
Cincinnati, OH To the Editor:We have reviewed the article by Kirkwood et al1 with great interest. We note that there isn’t a balance between the two arms of the study with respect to Breslow’s measurement data. Specifically, there are fewer numbers of patients with very thin (< 0.75 mm) melanomas randomized to the CMK vaccine arm of the study versus a larger number of patients with thin melanomas that were randomized to the high-dose interferon alfa-2b (HDI) arm. Because Breslow’s measurement is a significant prognostic factor, a significant bias might have been given to the HDI group, particularly the node-negative subset. It seems, from the survival charts, that the N0 group, where this maldistribution occurred, was also the group with the most dramatic benefit from interferon therapy. In fact, it seems that the N0 group was the only group that showed statistically significant improvement in relapse-free survival (RFS) and overall survival (OS). In addition, it is stated that an analysis of RFS and OS in all three trials (E1684, E1690, and E1694), indicated that the benefit associated with HDI improved with each successive trial. It follows then that the vaccine group in E1694 should have had an improved OS and RFS when compared with observation groups of the other two studies (E1690 and E1684). Because this was not the case, this suggests to us that there may have been a detrimental effect on OS and RFS produced by the vaccine. These two observations are deserving of further explanation and may invalidate some of the conclusions of the study. REFERENCES 1. Kirkwood JM, Ibrahim JG, Sosman JA, et al: High-dose interferon alfa-2b significantly prolongs relapse-free and overall survival compared with the GM2-KLH/QS vaccine in patients with resected stage IIB-III melanoma: Results of Intergroup Trial E1694/S9512/C509801. J Clin Oncol 19: 2370-2380, 2001
Response
University of Pittsburgh Medical Center, Pittsburgh, PA In Reply:We would like to respond to the above letters in regard to our article.1 First, Drs Kretschmer and Neumann erroneously conclude from our article that "only the subset of T4N0M0 patients benefited significantly from HDI treatment," and that "because ulceration in trial E1694 was not included in the multivariate analysis, imbalances in ulceration might have accounted for the benefit observed with high-dose interferon alfa-2b in Intergroup trial E1694."
The impact of high-dose interferon alfa-2b (HDI) upon relapse interval and survival in Intergroup study E1694 was significant for the study as a whole, both when analyzed using the intent-to-treat population and the eligible population (P = .0015 and P = .009 for relapse-free survival and overall survival, respectively, in eligible patients). The analysis of subsets according to the stage stratification subgroups was only performed as a secondary analysis, after achieving significance in the overall study for survival benefit and relapse-interval prolongation. Subsets were analyzed to explore the possibility of a differential impact of therapy according to stage groups into which patients were stratified at entry to this study. The most significant impact of HDI was observed in patients with T4N0M0 disease, where prolongation of relapse-free survival achieved significance when analyzed for the subset apart from the trial as a whole. The benefit of HDI was also seen in each of the other stratification groups with one, two to three, or Kretschmer and Neumann also posit that ulceration might have been unequally distributed between the GMK and HDI therapy groups, and that, especially in the T4N0M0 patients, this might have related to the outcome of our study. The results of E1694 have in fact been analyzed according to ulceration, as published in Table 1 of our article.1 There is no statistically significant difference in the frequency of ulceration among recipients of GMK (138 of 353 patients, 19.6%) and the frequency of ulceration among recipients of HDI (150 of 351 patients, 21.31%); if anything, the distribution favored the vaccine arm. The published Cox multivariate analyses of E1694 related to factors for which we had generally complete data; for ulceration, there were 20% of cases in which this variable was not available, and for this reason, ulceration was not included in the multivariable analysis as published. The impact of ulceration in the E1694 trial is significant (P = .0001) for both eligible and intent-to-treat analyses. Thus, although ulceration played no part in the trial differences related to interferon, ulceration is an important prognostic factor overall in the study.
Secondly, Drs Coldiron and Dinehart raise additional concerns regarding two aspects of the E1694 trial; "there is no (sic) balance between the two arms of the E1694 trial, with respect to Breslow’s depth," and "the benefit associated with HDI has improved with each successive trial [E1684, E1690, and E1694], and it follows that the vaccine group in E1694 should have had an improved OS and RFS when compared with observation groups of the two [earlier] studies E1684 and E1690. . . because this was not the case, there may have been a detrimental effect on OS and RFS produced by the vaccine."
The arms are quite balanced with respect to the Breslow thickness categories overall (with a two-sided P value comparing the two arms over all four categories of .1430). For the Breslow category T1 (Breslow
Finally, with respect to the comparison of outcomes between trials, it is not possible to conclude that GMK had a detrimental effect on the outcome of patients in E1694. In fact, all information that is available to date from our pooled analyses of all trials that have tested HDI and GMK suggest that there was no adverse impact of GMK. The comparisons of GMK with observation across trials are not valid because they were performed across studies with differing known and unknown prognostic factors. The analysis of outcome for patients who responded with the production of antibody at a titer of 1:80 or greater by day 29 reveals, if anything, a trend to more favorable outcome, albeit significantly less than that for the impact of HDI in the Intergroup E1694 trial. Moreover, there was no evidence of any detrimental effect of vaccination with GM2 ganglioside plus BCG in a prior randomized trial.4 Finally, because HDI significantly improved relapse-free survival in two prior trials compared with observation in a magnitude similar to the E1694 trial, if there were a detrimental effect of GMK on survival, it occurred without resulting in a disproportionate number of relapses, a highly implausible result. REFERENCES 1. Kirkwood JM, Ibrahim JG, Sosman JA, et al: High-dose interferon alfa-2b significantly prolongs relapse-free and overall survival compared with the GM2-KLH/QS-21 vaccine in patients with resected stage IIB-III melanoma: Results of intergroup trial E1694/S9512/C509801. J Clin Oncol 19: 2370-2380, 2001 2. Kirkwood JM, Strawderman MH, Ernstoff MS, et al: Interferon alfa-2b adjuvant therapy of high-risk resected cutaneous melanoma: The Eastern Cooperative Oncology Group Trial EST 1684. J Clin Oncol 14: 7-17, 1996[Abstract]
3.
Kirkwood JM, Ibrahim JG, Sondak VK, et al: High- and low-dose interferon alfa-2b in high-risk melanoma: First analysis of Intergroup Trial E1690/S9111/C9190. J Clin Oncol 18: 2444-2458, 2000
4.
Livingston PO, Wong GYC, Adluri S, et al: Improved survival in stage III melanoma patients with GM2 antibodies: A randomized trial of adjuvant vaccination with GM2 ganglioside. J Clin Oncol 12: 1036-1044, 1994
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2002 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
four nodes involved. Although these subsets did not yield significance when analyzed independently, the trial was not designed with a sample size large enough to address each individual subset with a high degree of statistical power. This subset analysis may provoke speculation that the benefit of HDI is the greatest in patients with the lowest burden of disease. However, it must be noted that the greatest impact of HDI observed in the pivotal E1684 study
