This review assessed the relationship between mastectomy resection margins and relapse. The author concluded that close resection margin was associated with a significantly increased risk of local relapse. The review had some methodological problems that included a lack of quality assessment and the possibility of several review biases, hence caution is warranted when interpreting the review findings.

To explore the relationship between margin status and subsequent relapse in women with breast cancer.

MEDLINE, EMBASE and ProQuest were searched to August 2009 for published articles. Search terms were reported. ASCO meetings (1998 to 2008), San Antonio Breast Cancer Symposia (2005 to 2008) and St Gallen Consensus Conferences (2000 to 2007) were searched electronically. Reference lists of retrieved articles were searched.

Randomised controlled trials (RCTs) and non-randomised studies in women with stage T1-4N1-3M0 breast cancer treated with radical or modified radical mastectomy with or without simultaneous reconstruction were eligible for inclusion. Interventions could be with or without axillary sampling, clearance or sentinal node biopsy. Patients could have received neoadjuvant or adjuvant chemotherapy, endocrine therapy or radiotherapy and could be of any age and menopausal status. Inflammatory breast cancer was eligible for inclusion. Relevant outcomes included local relapse, regional relapse and locoregional relapse.

Included studies involved patients with breast cancer (some with inflammatory breast cancer) who had undergone radical or modified mastectomy or skin-sparing mastectomy. Stage of breast cancer ranged from stage I/II to IIIb/IV. In most studies most patients had positive nodes and received chemotherapy (where reported). The proportion of patients who received radiotherapy was variable; in most studies either all or no patients received chest wall radiotherapy. Reported outcomes included muscle or fascia invasion, involved or close margin rates and relapse rates.

One author was involved in this review, thus it appeared that study selection was undertaken by one reviewer only.

The author did not state that he assessed validity.

Data that related to the incidence of involved or close margins and occurrence of local or locoregional relapse were used to calculate relative risks (RRs) and 95% confidence intervals (CIs).

One author was involved in this review, thus it appeared that data extraction was undertaken by one reviewer only.

Pooled relative risks, together with 95% CIs, were calculated using a random-effects meta-analysis. Statistical heterogeneity was assessed using Χ² and I² statistics. Studies were grouped according to clinical group: studies of non-inflammatory breast cancer treated by radical or modified radical mastectomy with or without radiotherapy; studies of inflammatory breast cancer; and studies of skin sparing mastectomy. The relationship of risk of relapse with distance from the resection margin was analysed.

Thirty non-randomised studies were included in the review (n=23,683): 25 studies of radical/modified mastectomy (n=23,051); three studies of inflammatory breast cancer (n=243); and three studies of skin-sparing mastectomy (n=389). Some studies had overlapping populations or involved different patient groups. The number of patients included in studies ranged from 23 to 4,181. Median length of follow-up ranged from 21 to 216 months. The years studied varied from 1970 to 2006, where stated.

All clinical groups: There was a significantly greater risk of relapse in women with close margin compared to women without close margin (RR 2.57, 95% CI 1.94 to 3.41, I²=19%; 11 studies, n=3,646).

Non-inflammatory breast cancer treated with radical/modified mastectomy: With no radiotherapy, there was a significantly greater risk of relapse in women with close margin compared to women without close margin (RR 2.60, 95% CI 1.79 to 3.77, I²=19%; five studies, n=2,825). With some use of radiotherapy (one study, n=100) or all patients receiving radiotherapy (one study, n=268), there was no statistically significant difference in risk of relapse in women with close margin compared to women without close margin. With no radiotherapy there was a significantly greater risk of relapse in women with muscle or fascia invasion compared to women without invasion (RR 1.72, 95% CI 1.03 to 2.86, I²=76%; four studies, n=5,071).

Inflammatory breast cancer: There was a significantly greater risk of relapse in women with close margin compared to women without close margin (RR 3.08, 95% CI 1.77 to 5.36, I²=36%; three studies, n=243).

Skin-sparing mastectomy: There was no statistically significant difference in relapse rate in women with close margin compared to women without close margin (one study, n=210).

Close resection margin was associated with a significantly increased risk of local relapse and chest wall radiotherapy should be considered, especially when other risk factors are present.

Inclusion criteria for the review were broadly defined. Several relevant data sources were searched. There was potential for language and publication biases as the author did not state whether articles in languages other than English language and unpublished studies were searched for; publication bias was not assessed. One author undertook the review and presumably performed study selection and data extraction alone, which increased risks of error and bias. It appeared that no quality assessment was undertaken, which made the quality of the included studies difficult to verify. Studies were combined using random-effects meta-analysis and statistical heterogeneity was assessed, which appeared to be appropriate. Heterogeneity was not investigated when present.

The author’s conclusions were based on non-randomised studies of unknown quality, hence caution is warranted when interpreting the review findings.

Practice: The author stated that for women with close resection margin and one or more additional risk factors, chest wall radiotherapy should be recommended. For women with close resection margin and no additional risk factors, chest wall radiotherapy should be considered. In the absence of further data, a 5mm cut-off seemed reasonable.

Research: The author stated that further research was needed to explore the relationship between distance from the resection margin and risk of local recurrence. An approach similar to that used in the reanalysis of Mentzer’s data was suggested. The relative infrequency of close margins and the impact of other risk factors suggested that a study would need thousands of women.

None stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.