Search

Proton high-resolution magic angle spinning (1H HR-MAS) NMR spectroscopy and quantitative histopathology were performed on the same 54 MRI/3D-MRSI-targeted postsurgical prostate tissue samples. Presurgical MRI/3D-MRSI targeted healthy and malignant prostate tissues with an accuracy of 81%. Even in the presence of substantial tissue heterogeneity, distinct 1H HR-MAS spectral patterns were observed for different benign tissue types and prostate cancer. Specifically, healthy glandular tissue was discriminated from prostate cancer based on significantly higher levels of citrate (P = 0.04) and polyamines (P = 0.01), and lower (P = 0.02) levels of the choline-containing compounds choline, phosphocholine (PC), and glycerophosphocholine (GPC). Predominantly stromal tissue lacked both citrate and polyamines, but demonstrated significantly (P = 0.01) lower levels of choline compounds than cancer. In addition, taurine, myo-inositol, and scyllo-inositol were all higher in prostate cancer vs. healthy glandular and stromal tissues. Among cancer samples, larger increases in choline, and decreases in citrate and polyamines (P = 0.05) were observed with more aggressive cancers, and a MIB-1 labeling index correlated (r = 0.62, P = 0.01) with elevated choline. The elucidation of spectral patterns associated with mixtures of different prostate tissue types and cancer grades, and the inclusion of new metabolic markers for prostate cancer may significantly improve the clinical interpretation of in vivo prostate MRSI data. Magn Reson Med 50:944–954, 2003. © 2003 Wiley-Liss, Inc.

In previous in situ point-resolved spectroscopy (PRESS) three-dimensional (3D) 1H magnetic resonance (MR) spectroscopic imaging studies, it has been demonstrated that the ratio of prostatic metabolites can noninvasively discriminate prostate cancer from surrounding normal tissue. However, in these studies, conventional chemical shift selective suppression (CHESS) and short-time inversion recovery (STIR) techniques often resulted in inadequate water and lipid suppression. To improve suppression and spatial coverage, the newly developed T1 insensitive dual band selective inversion with gradient dephasing (BASING) Bandstop Filter and dual phase-compensating spectral/spatial spin-echo pulses have been implemented in a clinical setting. In phantom studies, no change in metabolic profiles was observed with application of either BASING or spectral/spatial pulses. In a study of 17 prostate cancer patients, the use of either BASING or spectral/spatial pulses allowed for suppression of water (BASING 99.80 +/- 0.14% and spectral/spatial 99.73 +/- 0.47%) and lipid (BASING 98.56 +/- 1.03% and spectral/spatial 98.44 +/- 1.90%) without a significant difference in the prostatic metabolite ratios. Spectral/spatial suppression has the added advantage of reducing the chemical shift dependence of the PRESS volume, but optimal performance requires high-speed gradients with negligible eddy current effects. BASING suppression is less reliant on accurate pulse and gradient timings and can be implemented easily with no loss in performance on clinical MR scanners with conventional gradients.

Combined MRI and 3D spectroscopic imaging (MRI/3D-MRSI) was used to study the metabolic effects of hormone-deprivation therapy in 65 prostate cancer patients, who underwent either short, intermediate, or long-term therapy, compared to 30 untreated control patients. There was a significant time-dependent loss of the prostatic metabolites choline, creatine, citrate, and polyamines during hormone-deprivation therapy, resulting in the complete loss of all observable metabolites (total metabolic atrophy) in 25% of patients on long-term therapy. The amount and time-course of metabolite loss during therapy significantly differed for healthy and malignant tissues. Citrate levels decreased faster than choline and creatine levels during therapy, resulting in an increase in the mean (choline + creatine)/citrate ratio with duration of therapy. Due to a loss of all MRSI detectable citrate, this ratio could not be used to identify cancer in 69% of patients on long-term therapy. In the absence of citrate, however, residual prostate cancer could still be detected by elevated choline levels (choline/creatine ratio ≥1.5), or the presence of only choline in the proton spectrum. The loss of citrate and the presence of total metabolic atrophy correlated roughly with decreasing serum prostatic specific antigen levels with increasing therapy. In summary, MRI/3D-MRSI provided both a measure of residual cancer and a time-course of metabolic response following hormone-deprivation therapy. Magn Reson Med 46:49–57, 2001. © 2001 Wiley-Liss, Inc.