New information prompts old question: is sentinel lymph node sampling equivalent to axillary lymph node dissection?

Abstract

In 2010, the results of American College of Surgeons Oncology Group Z0011 trial were published in Annals of Surgery. A second publication appeared in Journal of the American Medical Association the following year. Z0011 concluded, in a select population and using a noninferiority design, that axillary lymph node dissection (ALND) did not provide any clinical advantage over sentinel lymph node sampling (SLNS) alone. Thus, it was felt that SLNS can replace an ALND (SLNS ALND) in a select population. Although some called for caution in the acceptance of this conclusion due to concerns about the details of the study, surgeons quickly began omitting routine ALND in patients with positive SLNS. In fact, much has been written acknowledging the astonishing alacrity at which the results of this trial influenced practice. Now, with the systematic review of the Z0011 radiation records by Jagsi et al, published with this editorial, one must ask if the results of Z0011 are still valid and wonder if the rapid acceptance of Z0011, despite critical absence of treatment information, was because the conclusions reached were indeed accurate or because the results of this trial simply agreed with preconceptions of the practitioners. Briefly, in Z0011 women undergoing breast-conserving therapy (BCT) with fewer than three positive sentinel lymph nodes (SLN) were randomly assigned to ALND or not. All patients were to receive whole-breast irradiation only (2 fields). Specifically, nodal-directed or third-field radiation was prohibited. Although powered for 1,900 patients, this noninferiority study accrued only 891 patients. The study concluded that there was no significant difference between the ALND and SLNS cohorts with respect to locoregional, disease-free survival (DFS), and overall survival. Z0011 was criticized for not meeting its accrual goal, having too many patients lost to follow-up, and lacking radiation therapy details. Some suspected that radiation oncologists adjusted their breast tangents (high tangents) to cover more nodal tissue in patients known to have an undissected axilla, which could have resulted in an unfair advantage to the SLNS arm. Jagsi et al undertook the tedious process of reviewing the radiation records to determine how often patients were treated with high tangents. A surprise finding awaited the investigators. Radiation case report forms were available for 605 of the 856 evaluable patients on the study. From the 605 case report forms, 89 of 540 patients, who received whole-breast radiation, also received a protocolprohibited treatment to the supraclavicular field. To further evaluate the radiation delivered, Jagsi et al requested detailed records for all evaluable patients. Although they only received 29% of the requested charts, this quality control evaluation revealed that 18.9% of patients received three fields of radiation. This protocol violation was present in both cohorts at not significantly different rates. Interestingly, when categorized by the number of nodes positive, the SLNS cohort was more likely to receive the extra field of radiation in each category. Jagsi et al conclude that further study may be necessary to answer the question posed by Z0011 and suggest that standardized collection of locoregional treatment common data elements, as developed by the NCI Breast Oncology-Local Disease (BOLD) Task Force, would allow better monitoring of treatment delivered in clinical trials. The revelation that three fields were used in a subset of patients provides evidence that should compel the re-examination of Z0011. Although, the third-field protocol violation rate did not differ significantly between cohorts, there is no solace to be found in this fact. It is the undissected axilla that is far more likely to benefit from additional radiation than the dissected axilla. This previously unrecognized third-field radiation finding may explain the apparent equivalence between the treatment arms. In response to criticisms, Z0011 results are said to be similar to those of National Surgical Adjuvant Breast and Bowel Project B-04. It has even been referred to as a “contemporary reaffirmation” of B-04. In B-04, clinically node-negative patients were randomly assigned to radical mastectomy, total mastectomy (TM) with axillary radiation, or TM alone. Based on the rate of positive nodes in the dissected axilla, it was estimated that 40% of the TM-only patients had positive axillary nodes. However, only approximately 20% developed clinically evident axillary recurrences. B-04 implies that not all axillary disease will manifest itself—a finding the Z0011 investigators suggest supports the validity of their study. What is often overlooked when referencing B-04 is the fact that it too suffered from significant protocol violations. One third of the patients in B-04 who were randomly assigned to no ALND actually did have axillary nodes removed. The unauthorized removal of these nodes could have influenced the axillary recurrence rate in JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 32 NUMBER 32 NOVEMBER 1