Liver metastases are found in 6–25% of patients with metastatic breast cancer (Tampellini et al, 1997). Most authors quote a median survival of 1–14 months (Jaffe et al, 1968; Zinser et al, 1987; Patanaphan et al, 1988; O’Reilly et al, 1990; Hoe et al, 1991) which compares unfavourably with metastases at other sites; isolated soft tissue metastases have a median survival of 50 months (Tampellini et al, 1997), bone 33–48 months (Sherry et al, 1986) and pleural disease 6–15 months (Johnson et al, 1998). Only cerebral metastases confer a worse prognosis with a median survival of only 4 months (Johnson et al, 1998). Survival may be prolonged by chemotherapy or endocrine therapy and a small proportion of patients may survive for 5 years (3%) or even 10 years (1%) with these therapies (Sledge and Miller, 1999). Current recommendations are that patients with asymptomatic, oestrogen receptor (ER) positive liver metastases may be treated with endocrine therapy (Buzdar, 2001). Those with symptomatic metastases or ER negative tumours are treated with combination chemotherapy such as FEC (5-fluorouracil, epirubicin and cyclophosphamide) or CMF (cyclophosphamide, methotrexate, 5-fluorouracil) (Fossati et al, 1998; Sledge and Miller, 1999). Taxane chemotherapy may be used in anthracycline-resistant disease (Hortobagyi et al, 2000). Novel therapies with monoclonal antibodies such as trastuzumab (for c-erb-B2/Her-2 positive breast cancers) are also being used often in combination with taxane chemotherapy (Fournier et al, 2000). Chemotherapy will achieve a response in 50–80% of patients and survival may be prolonged (Hortobagyi, 2000). However, the side effects of these agents may be severe and occasionally result in fatal complications.
There have been few studies of surgical treatment of liver metastases from breast cancer, either for attempted cure or debulking. Several studies suggest a small subgroup of patients may gain a survival advantage from surgery and there are reports of rare long-term survivors (Pocard et al, 2000; Selzner et al, 2000).
The aim of this study was to identify factors that enable prediction of those patients unlikely to survive for prolonged periods, for whom chemotherapy may be of no benefit and impair their remaining quality of life.
Liver metastases may present asymptomatically during a metastatic screen, or may present with upper abdominal fullness, a mass, ascites, jaundice or weight loss (O’Reilly et al, 1990). Ultrasound or CT scan usually confirms the diagnosis. Liver function tests are deranged in 92% of patients at presentation with gamma glutamyl transferase (GGT) and alkaline phosphatase being the most commonly elevated enzymes. Overt jaundice is less common (13%) (O’Reilly et al, 1990).
Factors adversely affecting prognosis include: jaundice (Hoe et al, 1991), deranged liver function tests (Zinser et al, 1987; O’Reilly et al, 1990), ascites, palpable hepatomegaly (O’Reilly et al, 1990), poor performance status and disease confined to the liver (Zinser et al, 1987; Hoe et al, 1991). The interval between primary presentation and metastatic disease is an important predictor of survival in bone metastases (Sherry et al, 1986) but may not be important in liver metastases (O’Reilly et al, 1990). The tumour marker CA15-3 is often higher in patients who do poorly, but is not reported to be an independent predictor of survival (Tampellini et al, 1997). Carcinoembryonic antigen (CEA) has not been previously studied in this respect although it is recognised as useful in monitoring disease progression (Cheung et al, 2000).
The influence of disease pattern, both outside the liver and within the liver has received little attention. Two studies suggest that extrahepatic disease may impair survival (Zinser et al, 1987; Hoe et al, 1991), but there are no data on the prognostic significance of disease distribution within the liver.
This study has examined the cases of all patients presenting in the last 5 years to a single breast unit with metastatic breast cancer involving the liver at metastatic diagnosis. Survival from the time of metastatic diagnosis was compared with primary disease information, patient characteristics and pattern of metastatic disease in an attempt to establish factors predicting outcome. It is hoped that these prognostic factors may be of benefit in tailoring treatment to avoid toxicity to patients with only a short life expectancy for whom palliative support would be most appropriate.