Sentinel Lymph Node Biopsy in Early Breast Cancer: Has Its Time Come?

The rationale behind the use of sentinel node biopsy in the management of early-stage breast cancer is straightforward—namely, to obtain the information on stage and prognosis needed to guide further treatment in a manner that avoids morbidity associated with full axillary dissection in patients whose sentinel lymph nodes are negative. Because sentinel lymph node biopsy is applicable to many newly diagnosed breast cancer patients, the consequences of its adoption are potentially substantial, and data from clinical trials that compare it with more standard approaches are critical. To be fully informative, such trials must demonstrate that sentinel lymph node biopsy achieves its goal of reducing patient morbidity but does not result in poorer disease control. In this issue of the Journal, Mansel et al. ( 1 ) illustrate that accomplishing these dual goals presents two key challenges: 1) assessing patient morbidity in a reliable and convincing fashion and 2) deciding what to do when the short-term morbidity results show benefit but the longer-term disease outcomes are not known. With respect to the first challenge, Mansel et al. ( 1 ) emphasize the important role that quality-of-life assessments had in supporting their conclusion that the use of sentinel node biopsy reduced patient morbidity. Their report illustrates well the potential advantages and disadvantages of using a patient-reported outcome such as health-related quality of life as a major study endpoint. A key disadvantage to this approach is that it is usually impossible to avoid the problem of missing data. In this regard, the longer-term compliance rates for questionnaire completion in oncology trials are rarely greater than 90% ( 2 ) and are often much lower. It is questionable to assume that failure to complete a questionnaire is a random event that is not related to the subject's status ( 3 ) . Thus, substantial rates of noncompliance can mean that the key advantage of randomization is lost; that is, the groups being compared can no longer be assumed to differ only by chance. The usual approach to avoiding bias in the analysis of randomized trials, the intention-to-treat principle, cannot easily be applied in these circumstances, because patients who do not provide data cannot be “counted” ( 4 ) either one way or the other. In this regard, it is disappointing that Mansel et al. ( 1 ) do not describe clearly how many patients completed questionnaires at various time points in the ALMANAC trial. However, it appears that only 371 (78%) of the 478 sentinel node biopsy patients who were “eligible for the ITT analyses” [table 4 in Mansel et al. ( 1 ) ] contributed data to the assessment of changes of five points or more in the trial outcome index score. It is interesting that this approach—converting health-related quality-of-life data into a form of response—can support a true intention-to-treat analysis ( 5 ) , but that does not appear to have been done here. The ALMANAC trial results reported by Mansel et al. ( 1 ) also demonstrate the major advantage of using health-related quality of life as a study endpoint. As the authors indicate, differences in health-related quality of life between study arms were statistically significant even though differences in arm volumes were not, demonstrating that having patients report on the impact of interventions can provide a more sensitive measure of treatment effects than more objective assessments. In fact, in situations like the one the ALMANAC trial was addressing, it can be argued that health-related quality of life is the real outcome and more objective measures should be viewed as surrogate outcomes. Interpreting the results of a study in which health-related quality of life is a major endpoint, as is the case with the ALMANAC trial, therefore often requires one to balance methodologic concerns about the possible impact of missing data against the inherent value of this approach to assessment. In this particular case, given all the information provided, it seems likely that Mansel et al. are correct in concluding that patients obtain benefit from avoiding axillary dissection. However, results of trials with health-related quality-of-life outcomes cannot always be interpreted unambiguously, and more effort is needed to accomplish the goal of having patient-reported outcomes routinely and reliably incorporated into cancer trials. Regarding the issue of use of sentinel node biopsy prior to knowledge of its impact on relapse free and overall survival, Mansel et al. ( 1 ) could not complete accrual to the ALMANAC trial because the investigators involved in the trial were no longer willing to have their patients be randomly assigned to routine axillary dissection. This situation parallels the widespread acceptance of sentinel node biopsy in the medical community even before publication of the long-term results from randomized clinical studies ( 6 ) . Is this practice justified? The answer appears to be yes on the basis of the following considerations. Is the procedure safe? Adverse events related to the use of dye or radiolabeled colloid appear to be low, with estimated frequencies of up to 1%–2% for allergic reactions and of less than 1% for anaphylaxis, and no substantial radiation risk has been identified to date ( 7 – 9 ) . Also, Mansel et al. ( 1 ) reported that patients in the sentinel node group had shorter mean hospital stays, less axillary drain usage, and fewer infections than patients in the standard axillary treatment group. What is the accuracy of sentinel node biopsy in predicting the presence or absence of disease in the axillary lymph nodes? In the National Surgical Adjuvant Breast and Bowel Project (NSABP)-B32 trial ( 9 ) and the study reported by Veronesi et al. ( 10 ) , two randomized controlled trials in which patients in the control arms were treated with sentinel node biopsy followed immediately by axillary dissection, the accuracy rates were greater than 96%. Both studies had false-negative rates of less than 10% and negative predictive values greater than 95%. Are the results of the randomized studies published to date biologically plausible and clinically relevant? The concept of predictable progression of tumor cells within the lymph nodes of the axilla is supported by the outcomes of the patients who received sentinel node biopsy followed by axillary node dissection. The relatively low false-negative rate argues against the phenomenon of skip metastases. Furthermore, the sentinel node was the only site of disease found in 64.5% of patients in the control arm of the NSABP-B32 trial ( 11 ) , providing more evidence that the sentinel node(s) is the initial site of spread. The consistent finding—that less invasive surgery is associated with better patient outcomes (i.e., morbidity and quality of life) ( 1 , 10 , 12 ) —is relevant to patients, treating physicians, and those responsible for organizing and funding health care. Are the clinical trial results reported to date applicable to current practice? Two issues of importance when considering this question relate to patient selection and surgical and technical abilities. The 2005 American Society for Clinical Oncology guideline (which was based on a systematic literature review conducted by an expert panel) recommends that patients with clinically node-negative tumors that are 5 cm or smaller are appropriate candidates for lymph node mapping and sentinel node sampling ( 6 ) . This recommendation applies to a substantial number of women seen in current practice. An important issue regarding implementation of this recommendation into current practice relates to surgical quality control, i.e., the demonstration of surgical experience and competence (as determined by the number of cases, the rates of sentinel lymph node identification, and the rate of false negatives). Multidisciplinary collaboration with the nuclear medicine and pathology departments within the treating institution is also needed. Will long-term outcome data from the randomized studies be available? A critical and to-date unanswered question relates to disease outcome. As indicated by Mansel et al. ( 1 ) , there are plans to analyze the combined long-term data from completed and closed randomized phase III studies [i.e., ALMANAC, NSABP B-32, and ACOSOG Z0011 ( 13 ) ]. Also, ongoing studies such as EORTC 10981 ( 14 ) and IBCSG-23–01 ( 15 ) will address other important questions, such as the efficacy of completion dissection versus axillary radiotherapy in sentinel lymph node–positive disease and the impact of axillary node dissection in patients with sentinel node micrometastases. The ultimate challenge lies ahead: Will investigators be prepared to alter their practices on the basis of long-term outcome data at a time when sentinel node biopsy has become the standard of care in many centers?

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.