Measurements of human breast cancer using magnetic resonance spectroscopy: a review of clinical measurements and a report of localized31P measurements of response to treatment

A review of the literature has shown that in human breast tumours, large signals from phosphomonoesters (PME) and phosphodiesters (PDE) are evident. In serial measurements in 19 patients with breast cancer, a decrease in PME was significantly associated with a stable or responding disease (p = 0.017), and an increase in PME was associated with disease progression. Extract studies have shown PME to comprise of phosphoethanolamine (PEth) and phosphocholine (PCho), with the PEth to PCho ratio ranging from 1.3 to 12. The PCho content of high grade tumours was found to be higher than low grade tumours. In some animal models, changes in PCho have been shown to correlate with indices of cellular proliferation, and spheroid studies have shown a decrease in PCho content in spheroids with smaller growth fractions. A serial study of 25 patients with advanced primary breast tumours undergoing hormone, chemotherapy or radiotherapy treatments, showed that in this heterogenous group there were significant changes in metabolites that were seen during the first 3 weeks (range 2–4 weeks) of treatment, that correlated with volume change over this period, employed here as a measure of response. Changes in PME (p = 0.003), total phosphate (TP) (p = 0.008) and total nucleoside tri-phosphate (TNTP) (p = 0.02) over 3 (±1) weeks were significantly associated with response, as were the levels of PME (p < 0.001), PDE (p = 0.01), TP (p = 0.001) and TNTP (p = 0.007) at week 3 (±1). PME at week 3 (±1) was also significantly associated with the best volume response to treatment (p = 0.03). A reproducibility analysis of results from the observation of normal breast metabolism in four volunteers showed a mean coefficient of variation of 25%, after correcting for changes resulting from the menstrual cycle. Reproducibility studies in four patients with breast cancer showed a mean coefficient of variation of 33%, with the reproducibility being better in patients measured on different days (difference in TP was −6%) compared with those measured on the same day (difference in TP was −29%). © 1998 John Wiley & Sons, Ltd.