Neoadjuvant chemotherapy (NACT) is increasingly used in the setting of early-stage breast cancer, with up to 30 % of patients now receiving this treatment [1,2]. Although the long-term survival outcomes for patients treated with NACT are comparable to those treated with adjuvant chemotherapy, the benefits associated with NACT are numerous [3]. NACT provides an opportunity to assess tumor chemo-responsiveness and gain valuable prognostic information, with pCR acting as a surrogate marker for improved long-term survival [4]. Furthermore, results from the KATHERINE and CREATE-X trials have demonstrated improved disease-free survival and overall survival respectively in those who have not achieved pCR [5,6]. From an operative perspective NACT can de-escalate surgery and make tumors amenable to breast conservation and avoid the morbidity of axillary lymph node dissection [7]. The evolving role of targeted axillary dissection means that an increasing number of patients are benefiting from more minimalistic approaches to surgery following NACT [8].

Given the trend in surgical de-escalation it is increasingly important to have an accurate imaging in order to accurately plan post-NACT surgical management. For precise surgical planning an accurate means of assessing response to NACT is required to ensure that patients receive the appropriate surgery both from an oncological and quality of life perspective. Several radiological modalities, such as ultrasound (US), mammography (MG) and MRI have been proposed to predict pathological response following NACT. Hylton et al. suggest that MRI can be a valuable and effective tool in monitoring response to NACT. After analysis of data of 216 women who underwent NACT for breast cancer, they concluded that MRI findings are a stronger predictor of pathologic response to NACT than clinical assessment alone [9]. Similarly, Rosen et al.'s analysis showed better correlation between MRI after NACT and histology compared to clinical examination with Pearson's correlation coefficient 0.75 versus 0.65 (SE = 0.15, CI = 0.32–0.9) respectively [10]. Negrão et al. performed a retrospective analysis of 308 patients and reported that MRI displayed a good accuracy (acc) of 79 % in predicting clinical response with the highest sensitivity and PPVs in TNBC and HER2+ subtypes [11]. These findings were corroborated by Londero who deduced that MRI performed better than MG in predicting a clinical response to NACT. MRI compared to mammography correctly identified residual disease in 100 % versus 86 % respectively [12].

Whilst MRI might be advantageous over US or MG, its sensitivity might be reduced by tumor response causing reactive inflammation, surrounding sclerosis and necrosis, perilesional oedema and the presence of accompanying ductal carcinoma in situ (DCIS) [13,15]. It has also been suggested that MRI may underestimate tumor response to NACT due to the radiological appearance of fibrosis, scaring, oedema and tumor fragmentation instead of concentric shrinkage [14]. In contrast, overestimation of response can occur due to the antiangiogenic effects of the chemotherapeutic agent on contrast kinetics producing reassuring radiological images [14,15]. These technical challenges may reduce the absolute sensitivity of MRI and prohibit the de-escalation of surgery based on radiological complete response (rCR) post-NACT [16].

PCR is used as a potential surrogate for long-term outcomes, including improved disease-free and overall survival. There is, however, variability in this definition amongst institutions with respect to residual DCIS (ypT0 versus ypT0/is). In our centre pCR is defined as no residual invasive or in situ disease following NACT.

The aim of this study was to perform a retrospective analysis to determine whether MRI accurately predict pathological response to NACT in a cohort of Irish women.