CA 15-3 is predictive of response and disease recurrence following treatment in locally advanced breast cancer

A significant number of women present with locally advanced breast cancer (LABC) despite efforts to promote public awareness. The use of Primary chemotherapy (PC) was initially introduced in patients with LABC, and was first reported in 1978 as part of a multidisciplinary approach [1]. Many studies have subsequently demonstrated the benefits of PC in this setting [2, 3]. Much interest has focused on the role of molecular and biological markers in predicting response to PC [4‐7], however the precise role of these markers remains controversial. CA 15-3 is a serum-based product of the MUC1 gene and is the most widely used serum marker for breast cancer [8]. Its preoperative levels predict adverse outcome in node negative and node positive breast cancer [9]. Kurebayashi et al have demonstrated that CA 15-3 may have a role in monitoring response to chemotherapy in LABC [10]. This was reinforced by Duffy et al, who suggested that the most important role for CA 15-3 is in monitoring therapy in patients with advanced breast cancer [9, 11]. However the precise role of CA15.3 in predicting response to PC and time to disease recurrence in patients with LABC following both PC and surgery is ill defined. Here we provide evidence to support a role for CA15.3 in both conditions.

73 patients who presented with LABC between January 1999 and May 2005 were included. All patients received primary chemotherapy in the oncology day clinic at St. Vincent's University Hospital, Dublin, Ireland. All patients were free from metastatic disease at the time of diagnosis as determined by bone scanning, liver ultrasonography and chest x-ray. The histological diagnosis was established either by core biopsy or excisional biopsy. Data was obtained by reviewing the medical records of these patients. The hospital local ethics committee approval was obtained for the study.

Seventy three patients received primary chemotherapy for LABC. The mean age was 52.3 years with a range from 26 to 96 years. The mean tumour size was 4.9 cm with a range from 2–10 cm. The median follow up was 790 days. Table 1 illustrates the pathological parameters of tumours in the studied group. The chemotherapeutic regimens used did not influence the outcome.

CA 15-3 is a serum based product for the MUC1 gene. While CA 15-3 levels have previously been demonstrated to predict outcome in LABC [15]. Its precise role defined by clinical response, pathological response and disease recurrence following PC treatment requires definition. In a recent review Duffy et al concluded that the most important application of CA 15-3 is in monitoring therapy in patients with advanced breast cancer [11]. Expert panels such as NACB and EGTM recommend the routine use of CA 15-3 in monitoring therapy in patients with LABC [16, 17]. In contrast the routine use of CA 15-3 in LABC is not recommended by the ASCO panel [18]. Here we provide evidence to support a role for CA 15-3 in LABC.

Much interest has focused on predictors of response to treatment in LABC. Chang et al and Mauriac et al [19, 20] showed a favourable outcome in ER negative patients following treatment for LABC. We found no relation between ER and clinical response however a borderline relation was found between ER positive patients and poor pathological response p = 0.052 (table 3), a larger sample of patients may be needed to show a significant relation. However these markers all require tumour tissue. CA 15-3 has the advantage of being a serum based marker and therefore its measurement is less invasive. Kurebayashi et al demonstrated that in pre-treatment CA15-3 positive patients changes in CA 15-3 levels after systemic therapy correlated with response to treatment [10]. Here we show that a high CA 15-3 level before PC along with advanced tumour stage predict a poor clinical response to chemotherapy. A poor pathological response to chemotherapy is predicted by elevated pre chemotherapy CA 15-3 levels, large tumour size and the presence of lympho-vascular invasion. Kurebayashi et al also demonstrated that a greater than 20% reduction in CA 15-3 levels following treatment was a favourable predictive factor for time to disease progression during systemic therapy. In this study we found that post chemotherapy high CA 15-3 levels were associated with reduced time to disease progression. We used a Cox proportional hazards model to identify predictors of recurrence following both chemotherapy and surgery in LABC. Detection of conventional markers such as tumour size, lympho-vascular invasion and hormone receptor status in addition to more recent parameters such as HER2, KI-67 and p53 have all been associated with outcome in LABC [19, 20, 4]. We found that lympho-vascular invasion, HER2 status and elevated CA15-3 levels post chemotherapy were independent predictors of a poor outcome. De La Lande et al reported a relation between CA 15-3 lead time and prognosis [21]. In this study we report a predictive value of high CA 15-3 levels in LABC following PC in relation to TTP and recurrence after surgery. These findings may have important implications as they suggest that surgery may be of minimal benefit in patients who have persistently elevated CA15-3 levels following chemotherapy in LABC. It is not surprising that HER2 status is an independent predictor of poor outcome and this subgroup of patients would undoubtedly benefit from Herceptin.

Elevated CA 15-3 levels predict a poor response to primary chemotherapy in locally advanced breast cancer. In addition persistent elevation of CA 15-3, post chemotherapy, in conjunction with the presence of lympho-vascular invasion and HER2 positivity predict a reduced disease free survival following treatment in locally advanced breast cancer.