Summary Phospholipid metabolism of human breast cancer was studied by 31p magnetic resonance spectroscopy (MRS). In vivo localised 31p MR spectra were obtained from the tumour alone using phase modulated rotating frame imaging. For 31 tumours, median (range) phosphomonoester (PME) to ATP ratio was 1.48 (0.57-3.78) and phosphodiester (PDE) to ATP ratio was 1.65 (0.44-3.89). DNA index and S phase fraction (SPF) were measured by flow cytometry of paraffin embedded tissue. Twelve (39%) tumours were diploid and 19 aneuploid. Median (range) SPF for 29 assessable tumours was 5.3% (0.6-28%), with significantly greater median SPF for aneuploid tumours (9.3%) than diploid (3.8%, P = 0.007). There was a significant association between PME/ATP and SPF (P= 0.03) due to a significant correlation for aneuploid tumours (P= 0.01). High resolution 31P MRS of extracts from 18 tumours (including seven studied in vivo) demonstrated that the PME peak consists predominantly of phosphoethanolamine (PE) with a smaller contribution from phosphocholine (PC) (median (range) PE/PC: 3.02 (1.13-5.09)). Changes in PME/ATP were observed for two tumours where tamoxifen stablised disease and may be consistent with the cytostatic effects of this drug.Breast cancer is the commonest cancer in women with an annual mortality rate in the United Kingdom of over 14,000 (OPCS Monitor, 1992), and its incidence appears to be rising in the western hemisphere (Glass & Hoover, 1990). Reduction in mortality has been reported for mammographic screening (Shapiro et al., 1982;Tabar et al., 1985) and adjuvant chemotherapy (Early Breast Cancer Trialists' Collaborative Group, 1992a,b (Glaholm et al., 1989). In another report of three patients with locally advanced breast tumours, PME/ATP decreased in response to therapy (Ng et al., 1989). PC and PE are intermediates in membrane phospholipid biosynthesis and can also be generated by growth factor signalling (Fisher et al., 1991). The concentration of PME may reflect the rate of cell membrane synthesis and therefore cellular proliferation . This is consistent with the observation of high concentration of the phosphomonoester, PE in the developing infant brain (Hope et al., 1984;Younkin et al., 1984) and liver (Moorcroft et al., 1991), and is further supported by the increase in PME to ATP ratio during liver regeneration in the rat following 70% resection (Murphy et al., 1992). However, alteration in PME associated with therapy could also be explained by a change in cellular number, particularly as in most cases there was an associated change in tumour size. The biochemical basis for therapy induced changes in PME in human breast cancer detected by 31p MRS is not clear.The aim of this study was to assess the relationship between the level of PME measured by 31P MRS in vivo and proliferation in untreated, primary human breast cancer. Patients were studied the day before surgery using 31P MRS and where possible, PME composition of extracts of tumour were studied by high resolution 31P MRS to aid interpretation of the in...