, natural killer cells 7Á3 ± 2Á5%, and macrophages 2%. Freshly isolated T cells were in resting (G 1 ) state, with 25% expressing the activationassociated cell surface antigens major histocompatibility complex II and CD25. When stimulated in vitro, a significant proportion entered S and G 2 phase of the cell cycle, up-regulated CD25, and secreted tumour necrosis factor-a, interferon-c, interleukin (IL)-5 and IL-6. Despite patient differences, the composition of the AAA inflammatory infiltrate was remarkably consistent, and when re-stimulated ex-vivo T cells produced a stereotypical cytokine response, consistent with the hypothesis that AAA T cells can promote tissue inflammation by secretion of proinflammatory cytokines, and in addition provide signals for B-cell help.

Introduction Provide evidence to support evolving management strategies for high-risk (B3) breast lesions by assessing risk of carcinoma in subgroups of B3 lesions using systematic review and meta-analysis. MethodsDatabases identified observational studies between 1980 and 2015 that reported on underestimation of malignancy following B3 lesion diagnosis at core needle biopsy.Critical appraisal, quality assessment, data extraction and meta-analysis was undertaken to calculate rate of malignancy of the whole B3 group and individual lesions. Study heterogeneity and association between variables and underestimation rate was investigated using random effects logistic modelling.Results Meta-analysis, using data from 129 studies, assessed 11 423 lesions of which 2160 were upgraded to malignancy after surgical excision biopsy (17% malignancy rate, 95% CI 15 to 19%). Malignancy rates varied from 6% in radial scars with no atypia (95% CI 2 to 13%, I2 72.8%), to 32% in papillomas with atypia (95% CI 23 to 41%, I2 57.4%). Differences in upgrade rates between atypical and non-atypical lesions were statistically significant (p<0.05). Study heterogeneity could not be explained by differences in core biopsy size or year of publication.Conclusions This comprehensive, inclusive assessment of all published literature, provides an accurate estimate of malignancy risk in subgroups of B3 lesions, to guide tailored management strategies. Some lesions have a high risk of malignancy, while others have a much lower risk, and could be safely managed with surveillance strategies rather than surgery.

These data support the hypothesis that increased NK cytotoxicity could be a contributing factor in the generation or potentiation of inflammation in patients with an AAA.

No abstract