- Email: [email protected]
Evaluation of the role of magnetization transfer imaging in prostate: a preliminary study
Available online at www.sciencedirect.com
Magnetic Resonance Imaging 26 (2008) 644 – 649
Evaluation of the role of magnetization transfer imaging in prostate: a preliminary study Virendra Kumar a , Naranamangalam R. Jagannathan a,⁎, Rajeev Kumar b , Sanjay Thulkar c , Siddhartha D. Gupta d , Ashok K. Hemal b , Narmada P. Gupta b a Department of NMR, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India Department of Urology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India c Department of Radio-diagnosis, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India d Department of Pathology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India Received 16 May 2007; revised 8 December 2007; accepted 6 January 2008 b
Abstract Results of the preliminary study on the evaluation of the role of magnetization transfer imaging (MTI) of prostate in men who had raised prostate-specific antigen (PSA) (N4 ng/ml) or abnormal digital rectal examination (DRE) are reported. MT ratio (MTR) was calculated for 20 patients from the hyper- (normal) and hypo-intense regions (area suspicious of malignancy as seen on T2-weighted MRI) of the peripheral zone (PZ) and the central gland (CG) at 1.5 T. In addition, MTR was calculated for three healthy controls. Mean MTR was also calculated for the whole of the PZ (including hyper- and hypo-intense area) in all patients. Out of 20 patients, biopsy revealed malignancy in 12 patients. Mean MTR value (8.29±3.49) for the whole of the PZ of patients who were positive for malignancy on biopsy was statically higher than that observed for patients who were negative for malignancy (6.18±3.15). The mean MTR for the whole of the PZ of controls was 6.18±1.63 and is similar to that of patients who were negative for malignancy. Furthermore, for patients who showed hyper- (normal portion) and hypointense (region suspicious of malignancy) regions of the PZ, the MTR was statistically significantly different. These preliminary results reveal the potential role of MT imaging in the evaluation of prostate cancer. © 2008 Elsevier Inc. All rights reserved. Keywords: MRI; Magnetization transfer imaging (MTI); Magnetization transfer ratio (MTR); Prostate cancer
1. Introduction Prostate cancer is a major health problem in elderly men. The diagnosis of prostate cancer is based mostly on transrectal ultrasound guided biopsy. MRI and MR spectroscopic imaging (MRSI) are used for noninvasive, anatomic and metabolic evaluation of prostate [1–6]. MRI with endorectal coil provides improved sensitivity compared to that obtained with a pelvic coil. Dynamic contrast MRI has also been used to achieve a higher accuracy in prostate cancer localization and staging [7]. MRSI provides metabolic information by giving the relative concentration of citrate, creatine and choline. Normal prostate tissue contains ⁎ Corresponding author. Tel.: +91 11 2659 3253, 2658 8533; fax: +91 11 2658 8663, 2658 8641. E-mail addresses: [email protected], [email protected] (N.R. Jagannathan). 0730-725X/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.mri.2008.01.030
high level of citrate, and in the presence of cancer, this level decreases or is undetectable. Studies have shown that compared to MRI alone, MRS has higher specificity but lower sensitivity [8]. Combination of these modalities can lead to high sensitivity and specificity for tumor localization [9]. In conventional MRI, the contrast between different tissues is dependent mainly on the relaxation properties of water present in tissues. However, to increase the information content, other properties of tissues like diffusion are utilized for image contrast. The role of diffusion-weighted MRI has been evaluated for the diagnosis of prostate cancer. Studies have shown the measurement of ADC as an important parameter for diagnosis of prostate cancer [10–13]. Yet another contrast mechanism is with the use of magnetization transfer (MT) that utilizes the interaction of water protons with macromolecular protons. This method can be used to generate a contrast that is different from T1 or T2 contrast [14]. Studies using MT imaging (MTI) has been
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649
shown to have potential in improving image contrast and tissue characterization [15–17]. The population of protons in tissues is from ‘free’ or highly mobile water and that associated with macromolecules or immobile water. MRI signal arises from mobile protons that have sufficiently long T2 relaxation times (N10 ms). The T2 of less mobile protons associated with macromolecules is too short (b1 ms) to be detected in MRI. However, in tissues, macromolecules interact with water and influence its relaxation properties by a variety of mechanisms. The dipole–dipole interaction between mobile protons and macromolecular protons perturbs the equilibrium between the two pools of protons, thereby affecting the MR signal. Selective saturation of the macromolecular resonance results in MT to the water protons resulting in the decrease of the magnitude of water signal observed. The amount of reduction depends upon the rate of exchange between the two spin populations. Thus, images with MTI show loss of signal compared to normal image without MT. The decrease will be larger in regions where exchange of magnetization is more efficient, determined by the relative proportion of protons of the two pools, their intrinsic relaxation times and the exchange rate. Application of an off-resonance RF pulse, called MT pulse, is the common method of MT contrast generation. To allow detection of magnetization resulting from immobile protons of bound water and macromolecules, the MT pulse is applied before slice selection and with a frequency shift from resonance, e.g., 1.5 kHz. The MT effect is quantified by determining the magnetization transfer ratio (MTR). It requires acquiring data sets with the MT pulse turned off and on, while all the other parameters being identical. To date most applications of MTI have been in MR angiography [18–21], brain and various neurological disorders [22–30]. MT ratio has been reported to increase with tumor malignancy [22,23]. For example, low-grade gliomas showed smaller MT effect than high-grade gliomas [22]. Similarly, MT contrast was higher in high-grade than in low-grade astrocytomas due to increased content of highmolecular weight nuclear material in high-grade tumor [23]. MTI has been shown to be a useful additional tool in the assessment of pituitary disorders [31]. Recently, the feasibility of whole-body MTI has been demonstrated [32]. There is only one report (in the Japanese language) on the application of MTI in prostate cancer [33]. The objective of this preliminary study was to demonstrate the role of MTI of prostate in men who had raised prostate-specific antigen (PSA) (N4 ng/ml) or abnormal digital rectal examination (DRE) and to use MTR for differentiation of normal and malignant prostatic tissues.
2. Methods The patient study population consisted of 26 men with raised PSA (N4 ng/ml) or abnormal DRE; age range,
645
50–83 years; mean age, 64.2 years; PSA range, 0.48–1507 ng/ml; mean PSA, 93.18 ng/ml; the other details are as presented in Table 1. Patients were recruited from the urology clinic and institute ethical committee approved the study. The exclusion criteria used for this study include patients with metallic implant or claustrophobic. Three healthy young men (age range, 28–30 years; mean age, 29.3 years) were recruited to serve as controls for the study. This age group was chosen to avoid contamination of normal control data from either BPH or asymptomatic prostate cancer [34]. MR investigations were carried out at 1.5 T using a whole-body scanner (Sonata/Avanto, Siemens, Erlangen, Germany). The pulse sequence used for MTC consisted of an off-resonance saturation pulse immediately before the first 90° RF pulse. As discussed earlier, two data sets of transverse images with and without MT pulse were acquired keeping all the other parameters identical. Details of the various parameters used are follows: pulse length, 7680 μs; bandwidth, 250 Hz; frequency offset, 1500 Hz; and flip angle, 500°. Images with and without MT pulse used for MTR calculation were acquired using a TR of 1000 ms and a TE of 13 or 14 ms with a slice thickness of 4 or 5 mm, without inter-slice gap. These values of TR and TE were chosen to minimize T1 and T2 effects [16]. T2-weighted images were acquired using TR 4000–5000 ms and TE 98 ms with slice thickness of 4 or 5 mm without inter-slice gap. The slice position of MT images and T2-weighted transverse images was identical. The amount of MT is quantified by the calculation of MTR using the equation MTR =1−Ms/Mo]×100, where Mo is the signal intensity of a given ROI without MTC and Ms is the signal intensity of the same ROI with MT. Signal Table 1 Details of patients used for MTR investigation Patient no.
Age (years)
PSA (ng/ml)
Biopsy data (Gleason grade or score)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
65 52 83 62 69 67 75 67 55 67 55 68 72 54 60 70 64 60 75 64
4.2 5.5 155 49.6 10.3 5.4 4.2 1.73 30.9 0.48 5.08 12.2 20.96 11.01 110 38.56 26.77 1507 13.7 NA
Adenocarcinoma Adenocarcinoma 5 6 Negative Negative Negative Negative 5b Negative Adenocarcinoma Negative 2 Negative 9(4+5) 7(3+4) 3a+3c 3a 8(4+4) Negative
646
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649
Fig. 1. Representative example of prostate MR images of a patient whose PSA level is 38.56 ng/ml; Gleason score, 7 (3+4); and age, 70 years. (A) T2-weighted image; (B) T2-weighted image showing the ROIs for the calculation MTR; (C) image without MT pulse; and (D) image with MT pulse. The figure also shows the placement of circular ROIs for calculation of MT ratios from CG, PZ and the malignant region. The hypo-intense (dark — shown by arrows) signal area of the PZ shown in (A) is indicative of malignancy which was confirmed on biopsy. MTR between the boundary of the PZ and the CG was not included in the calculation to avoid contamination of data between tissue types. Similarly, a small portion between the hypo- and hyper-intense areas of the PZ was not included either in the MTR calculation. Accordingly in the figure, the ROIs are not shown in these regions.
intensities were measured by drawing circular ROIs of uniform size of 0.04 cm2 (12 pixels). For placement of ROIs the T2-weighted images were used (vide infra). MTR was calculated from the peripheral zone (PZ) and the central gland (CG) of the prostate by drawing nonoverlapping consecutive ROIs. Care was taken while drawing ROIs to sample exclusively either from hyper- (normal) or hypointense (suspicious of malignancy seen as low signal intensity on T2-weighted images) regions of the PZ or CG, to minimize contamination of data between tissue types. Furthermore, mean MTR was calculated for the whole of the PZ [including hyper- (normal) and hypo-intense regions (suspicious of malignancy)] in all patients. One-way analysis of variance was applied to compare MTR calculated from normal PZ, CG and malignant portion of the PZ tissue. Average value for the whole of the PZ and CG was calculated for comparison between patient and controls. Mann–Whitney test was applied to compare MTR between patients positive for malignancy and negative for malignancy. 3. Results Out of 26 patients, images acquired with and without MT pulse showed a mismatch due to movement in four patients. Biopsy report was not available for one patient and MR
examination could not be completed in one patient due to technical reasons. Thus a total of 3725 ROIs from 20 patients and three controls (2153 and 1572 ROIs for CG and PZ, respectively) were used for MTR calculation and further analysis. The age, PSA and biopsy data of these patients are summarized in Table 1. Table 2 Comparison of mean MTR calculated for the whole of the PZ and the whole of the CG in control and patient groups Region
Mean MTR±SD
1. Control (n=3) (a) PZ (b) CG ⁎
6.18±1.63 6.97±1.90
2. Patients positive for malignancy on histopathology (n=12) (a) PZ 8.29±3.49 ⁎⁎ (b) CG ⁎ 7.23±3.80 3. Patients negative for malignancy on histopathology (n=8) (a) PZ 6.18±3.15 ⁎⁎⁎ (b) CG 7.01±4.17 ⁎ CG was not distinguished clearly in prostate of one control and one patient and hence not included in the analysis. ⁎⁎ Denotes the Pb.05 with respect to control for PZ. ⁎⁎⁎ Denotes the Pb.05 with respect to the value observed for PZ for patients who were positive for malignancy on biopsy.
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649
647
Table 3 Comparison of MTR values of patients (n=4) who showed hyper- (normal) and hypo-intense (region suspicious of malignancy) signal on T2-weighted image of the PZ and CG Patient no.
Region
No. of ROIs
Mean MTR±S.D.
P value
1.
a. Hyper-intense area of the PZ (normal) b. Hypo-intense region (suspicious of malignancy) c. CG a. Hyper-intense area of the PZ (normal) b. Hypo-intense region (suspicious of malignancy) c. CG a. Hyper-intense area of the PZ (normal) b. Hypo-intense region (suspicious of malignancy) c. CG a. Hyper-intense area of the PZ (normal) b. Hypo-intense region (suspicious of malignancy) c. CG
136 35 33 38 28 75 8 11 297 6 21 22
4.67±2.02 7.69±1.66 6.81±2.32 2.07±1.77 6.91±1.36 5.57±2.12 5.81±0.49 9.56±2.02 6.83±2.33 2.13±2.02 7.04±1.16 4.87±2.11
.0001 (a vs. b) .035 (b vs. c) .0001 (c vs. a) .0001 (a vs. b) .002 (b vs. c) .0001 (c vs. a) .0001 (a vs. b) .0001 (b vs. c) Not significant .0001 (a vs. b) .0001 (b vs. c) .001 (c vs. a)
2.
3.
4.
Combined Analysis of these 4 patients Region
Mean MTR±S.D.
Hyper-intense area of the PZ (normal)
3.43±2.36
Hypo-intense region (suspicious of malignancy) CG
7.53±1.70 6.57±2.35
Fig. 1 shows the representative example of T2-weighted, without and with MT pulse, images of prostate of a patient whose PSA level is 38.56 ng/ml. The hyper-intense area (with respect to muscle tissue/CG) on T2-weighted image corresponds to normal PZ while the low signal intensity area (hypo-intense) corresponds to a region suspicious of malignancy (see Fig. 1A) [1–5]. Later, the histopathology evaluation of biopsy specimens of these patients confirmed malignancy in this area. Out of 20 patients, biopsy revealed malignancy in 12 patients. The mean MTR was calculated for the whole of the PZ and the whole of the CG in controls, patients positive for malignancy (on biopsy, n=12) and patients negative for malignancy (n=8), and the values are presented in Table 2. A mean MTR value of 8.29±3.49 was obtained for the whole of the PZ of patients who were positive for malignancy on biopsy which was statically higher than that observed for patients who were negative for malignancy (6.18±3.15). The mean MTR for the whole of the PZ of controls was 6.18±1.63 and is similar to that of patients who were negative for malignancy. Four out of 12 patients who were positive for malignancy on biopsy showed certain regions suspicious of malignancy of the PZ on T2-weighted image while the rest of the portion of the PZ appeared hyper-intense (normal). The biopsy revealed normal and malignant PZ tissue in the respective cores in these 4 patients and their MTR values are summarized in Table 3. The mean MTR calculated from the hypo-intense region suspicious of malignancy of the PZ was significantly higher than that observed for the hyperintense normal portion of the PZ and the CG (see Table 3). In addition, the mean MTR obtained for the hyper-intense portion of the PZ was significantly lower than that observed for the CG. No significant difference between mean MTR of
P value Hyper-intense area of the PZ (normal) vs. hypo-intense region (suspicious of malignancy) Hyper-intense area of the PZ (normal) vs. CG Hypo-intense region (suspicious of malignancy) vs. CG
.0001 .0001
the PZ and the CG was observed in one patient. Furthermore, the combined analysis of these four patients showed statistically significant difference in the mean MTR of the hyper- and hypo-intense regions of the PZ and the CG.
4. Discussion In this preliminary study, we investigated the diagnostic potential MTI of prostate in patients who had raised PSA or abnormal DRE. MTI is a promising method of tissue characterization and reflects the microstructural changes that accompany the pathological process. Thus, MTR provides a quantitative index of the structural integrity of tissues. Different tissues display different amounts of MT and the degree of MT within a tissue is primarily governed by the local concentration of macromolecules. Cell membrane proteins and phospholipids mainly contribute to the MT signal. Change in MTR may reflect abnormal cell membrane structure or decreased cell number or cell size. Our data indicate that mean MTR within a patient showed a significant difference between the normal (hyper-intense area) and suspicious region of malignancy (hypo-intense area) of the PZ (see Table 3). Inter-patient comparison revealed a similar behavior (Table 2). The CG was not distinguishable in one control; therefore, comparison of CG with controls could not be carried out. The finding of significantly higher MTR for the hypo-intense region compared to the hyper-intense region of the PZ is further supported by the statistically higher mean MTR obtained for the whole of the PZ of patients who were positive for malignancy on histopathology compared to those negative for malignancy. The MTR values calculated and reported in our study are from one tissue type. Contamination from
648
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649
different tissues is not expected in our study, since the small portion between the PZ and the CG and between the hypointense and the hyper-intense region of the PZ was not included for the MTR calculation, as shown in Fig. 1. Furthermore, the mean MTR for the whole of the PZ of patients who were positive for malignancy on biopsy was significantly different from that observed for controls and patients who were negative for malignancy. As expected, no significant difference was observed in the mean MTR calculated for the whole of the PZ of controls and patients who were negative for malignancy. The only study that is available in the literature by Arima et al. [33] also reported similar findings. Progress in malignancy of prostate adenocarcinoma is associated with histopathological features of cellular atypia, pleomorphism, de-differentiation and increased cellularity. With increasing malignancy, an increase in the amount of high molecular weight material in cellular nuclei is reported [22]. The glandular structure of prostate is disrupted during malignancy and luminar spaces are replaced with proliferating malignant cells, leading to higher MTR in the malignant regions of the prostate. Various MRI methods have been used for prostate cancer detection and localization, as well as to study their biological behavior [1–9]. Routine MRI demonstrates zonal anatomy with excellent soft tissue contrast resolution and allows assessment of the local extent of disease. MRS improves the prostate cancer detection and localization [1–9]. Lower citrate to choline and creatine ratio is observed in cancerous tissue than in normal tissue. In addition, because the cellular structure of malignant tumors differs from that of normal tissue, it is documented that diffusion MRI is also a useful modality in the diagnosis of prostate cancer with the apparent diffusion coefficient of tumor being less than that observed for the normal PZ tissue [10–13]. In this direction, our preliminary data using MTI suggest that this technique may have the potential to provide additional information in the diagnosis of prostate cancer in addition to the above existing MR methodologies. 5. Conclusion This preliminary study demonstrated the possible role of MT imaging in the differentiation of normal (hyperintense) and suspicious region of malignancy (hypointense) of the PZ of prostate cancer patients. The mean MTR calculated for the whole of the PZ of patients who were positive for malignancy on biopsy is higher than that observed for the normal PZ of controls and patients who were negative for malignancy. Furthermore, for patients who showed hyper- (normal) and hypo-intense regions (suspicious of malignancy) of the PZ, the MTR was statistically significantly different. These interesting results should be further investigated in a large cohort of patients and controls.
Acknowledgment VK thanks the Indian Council of Medical Research (ICMR), New Delhi, for providing Junior Research Fellowship. References [1] Katz S, Rosen M. MR imaging and MR spectroscopy in prostate cancer management. Radiol Clin North Am 2006;44:723–34. [2] Choi YJ, Kim JK, Kim N, Kim KW, Choi EK, Cho KS. Functional MR imaging of prostate cancer. Radiographics 2007;27:63–75. [3] Akin O, Hricak H. Imaging of prostate cancer. Radiol Clin North Am 2007;45:207–22. [4] Hricak H, Choyke PL, Eberhardt SC, Leibel SA, Scardino PT. Imaging prostate cancer: a multidisciplinary perspective. Radiology 2007;243: 28–53. [5] Carroll PR, Coakley FV, Kurhanewicz J. Magnetic resonance imaging and spectroscopy of prostate cancer. Rev Urol 2006;8:S4–S10. [6] Rajesh A, Coakley FV, Kurhanewicz J. 3D MR spectroscopic imaging in the evaluation of prostate cancer. Clin Radiol 2007;62:921–9. [7] Noworolski SM, Henry RG, Vigneron DB, Kurhanewicz J. Dynamic contrast-enhanced MRI in normal and abnormal prostate tissues as defined by biopsy, MRI, and 3D MRSI. Magn Reson Med 2005;53: 249–55. [8] Scheidler J, Hricak H, Vigneron DB, Yu KK, Sokolov DL, Huang LR, et al. Prostate cancer: localization with three-dimensional proton MR spectroscopic imaging—clinicopathologic study. Radiology 1999;213: 473–80. [9] Wefer AE, Hricak H, Vigneron DB, Coakley FV, Lu Y, Wefer J, et al. Sextant localization of prostate cancer: comparison of sextant biopsy, magnetic resonance imaging and magnetic resonance spectroscopic imaging with step section histology. J Urol 2000;164:400–4. [10] Sato C, Naganawa S, Nakamura T, Kumada H, Miura S, Takizawa O, et al. Differentiation of noncancerous tissue and cancer lesions by apparent diffusion coefficient values in transition and peripheral zones of the prostate. J Magn Reson Imaging 2005;21:258–62. [11] Issa B. In vivo measurement of the apparent diffusion coefficient in normal and malignant prostatic tissues using echo-planar imaging. Magn Reson Imaging 2002;16:196–200. [12] Hosseinzadeh K, Schwarz SD. Endorectal diffusion-weighted imaging in prostate cancer to differentiate malignant and benign peripheral zone tissue. J Magn Reson Imaging 2004;20:654–61. [13] Kumar V, Jagannathan NR, Kumar R, Thulkar S, Gupta SD, Dwivedi SN, et al. Apparent diffusion coefficient of the prostate in men prior to biopsy: determination of a cut-off value to predict malignancy of the peripheral zone. NMR Biomed 2007;20:505–11. [14] Wolff SD, Balaban RS. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med 1989;10: 135–44. [15] Wolff SD, Balaban RS. Magnetization transfer imaging: practical aspects and clinical applications. Radiology 1994;192:593–9. [16] Finelli DA. Magnetization transfer in neuroimaging. Magn Reson Imaging Clin N Am 1998;6:31–52. [17] Henkelman RM, Stanisz GJ, Graham SJ. Magnetization transfer in MRI: a review. NMR Biomed 2001;14:57–64. [18] Edelman RR, Ahn SS, Chien D, Li W, Goldmann A, Mantello M, et al. Improved time-of-flight MR angiography of the brain with magnetization transfer contrast. Radiology 1992;184:395–9. [19] Pike GB, Hu BS, Glover GH, Enzmann DR. Magnetization transfer time-of-flight magnetic resonance angiography. Magn Reson Med 1992;25:372–9. [20] Atkinson D, Brant-Zawadzki M, Gillan G, Purdy D, Laub G. Improved MR angiography: magnetization transfer suppression with variable flip angle excitation and increased resolution. Radiology 1994;190:890–4.
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649 [21] Dousset V, Franconi JM, Degreze P, Balderrama J, Lexa F, Caille JM. Use of magnetization transfer contrast to improve cerebral 3D MR angiography. Neuroradiology 1994;36:188–92. [22] Kurki T, Lundbom N, Kalimo H, Valtonen S. MR classification of brain gliomas: value of magnetization transfer and conventional imaging. Magn Reson Imaging 1995;13:501–11. [23] Lundbom N. Determination of magnetization transfer contrast in tissue: an MR imaging study of brain tumors. AJR Am J Roentgenol 1992;159:1279–85. [24] Dousset V. Magnetization transfer imaging in vivo study of normal brain tissues and characterization of multiple sclerosis and experimental allergic encephalomyelitis lesions. J Neuroradiol 1993;20:297. [25] Filippi M. The role of magnetization transfer and diffusion-weighted MRI in the understanding of multiple sclerosis evolution. Neurol Sci 2000;21:S877–81. [26] Filippi M. Magnetization transfer MRI in multiple sclerosis and other central nervous system disorders. Eur J Neurol 2003;10:3–10. [27] Filippi M, Rocca MA. Magnetization transfer magnetic resonance imaging in the assessment of neurological diseases. J Neuroimaging 2004;14:303–13.
649
[28] Filippi M, Agosta F. Magnetization transfer MRI in multiple sclerosis. J Neuroimaging 2007;17(Suppl 1):22S–6S. [29] Pui MH, Wang Y. Diffusion and magnetization transfer MRI of brain infarct, infection, and tumor in children. Clin Imaging 2005;29:162–71. [30] Ridha BH, Tozer DJ, Symms MR, Stockton KC, Lewis EB, Siddique MM, et al. Quantitative magnetization transfer imaging in Alzheimer disease. Radiology 2007;244:832–7. [31] Argyropoulou MI, Kiortsis DN. Magnetization transfer imaging of the pituitary gland. Hormones 2003;2:98–102. [32] Boss A, Martirosian P, Kuper K, Fierlbeck G, Claussen CD, Schick F. Whole-body magnetization transfer contrast imaging. J Magn Reson Imaging 2006;24:1183–7. [33] Arima K, Hayashi N, Yanagawa M, Kawamura J, Kobayashi S, Takeda K, et al. The progress in diagnostic imaging for staging of bladder and prostate cancer: endorectal magnetic resonance imaging and magnetization transfer contrast. Hinyokika Kiyo 1999;45:553–7. [34] Kumar V, Jagannathan NR, Kumar R, Das SC, Jindal L, Thulkar S, et al. Correlation between metabolite ratios and ADC values of prostate in men with increased PSA level. Magn Reson Imaging 2006; 24:541–8.
Magnetic Resonance Imaging 26 (2008) 644 – 649
Evaluation of the role of magnetization transfer imaging in prostate: a preliminary study Virendra Kumar a , Naranamangalam R. Jagannathan a,⁎, Rajeev Kumar b , Sanjay Thulkar c , Siddhartha D. Gupta d , Ashok K. Hemal b , Narmada P. Gupta b a Department of NMR, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India Department of Urology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India c Department of Radio-diagnosis, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India d Department of Pathology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India Received 16 May 2007; revised 8 December 2007; accepted 6 January 2008 b
Abstract Results of the preliminary study on the evaluation of the role of magnetization transfer imaging (MTI) of prostate in men who had raised prostate-specific antigen (PSA) (N4 ng/ml) or abnormal digital rectal examination (DRE) are reported. MT ratio (MTR) was calculated for 20 patients from the hyper- (normal) and hypo-intense regions (area suspicious of malignancy as seen on T2-weighted MRI) of the peripheral zone (PZ) and the central gland (CG) at 1.5 T. In addition, MTR was calculated for three healthy controls. Mean MTR was also calculated for the whole of the PZ (including hyper- and hypo-intense area) in all patients. Out of 20 patients, biopsy revealed malignancy in 12 patients. Mean MTR value (8.29±3.49) for the whole of the PZ of patients who were positive for malignancy on biopsy was statically higher than that observed for patients who were negative for malignancy (6.18±3.15). The mean MTR for the whole of the PZ of controls was 6.18±1.63 and is similar to that of patients who were negative for malignancy. Furthermore, for patients who showed hyper- (normal portion) and hypointense (region suspicious of malignancy) regions of the PZ, the MTR was statistically significantly different. These preliminary results reveal the potential role of MT imaging in the evaluation of prostate cancer. © 2008 Elsevier Inc. All rights reserved. Keywords: MRI; Magnetization transfer imaging (MTI); Magnetization transfer ratio (MTR); Prostate cancer
1. Introduction Prostate cancer is a major health problem in elderly men. The diagnosis of prostate cancer is based mostly on transrectal ultrasound guided biopsy. MRI and MR spectroscopic imaging (MRSI) are used for noninvasive, anatomic and metabolic evaluation of prostate [1–6]. MRI with endorectal coil provides improved sensitivity compared to that obtained with a pelvic coil. Dynamic contrast MRI has also been used to achieve a higher accuracy in prostate cancer localization and staging [7]. MRSI provides metabolic information by giving the relative concentration of citrate, creatine and choline. Normal prostate tissue contains ⁎ Corresponding author. Tel.: +91 11 2659 3253, 2658 8533; fax: +91 11 2658 8663, 2658 8641. E-mail addresses: [email protected], [email protected] (N.R. Jagannathan). 0730-725X/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.mri.2008.01.030
high level of citrate, and in the presence of cancer, this level decreases or is undetectable. Studies have shown that compared to MRI alone, MRS has higher specificity but lower sensitivity [8]. Combination of these modalities can lead to high sensitivity and specificity for tumor localization [9]. In conventional MRI, the contrast between different tissues is dependent mainly on the relaxation properties of water present in tissues. However, to increase the information content, other properties of tissues like diffusion are utilized for image contrast. The role of diffusion-weighted MRI has been evaluated for the diagnosis of prostate cancer. Studies have shown the measurement of ADC as an important parameter for diagnosis of prostate cancer [10–13]. Yet another contrast mechanism is with the use of magnetization transfer (MT) that utilizes the interaction of water protons with macromolecular protons. This method can be used to generate a contrast that is different from T1 or T2 contrast [14]. Studies using MT imaging (MTI) has been
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649
shown to have potential in improving image contrast and tissue characterization [15–17]. The population of protons in tissues is from ‘free’ or highly mobile water and that associated with macromolecules or immobile water. MRI signal arises from mobile protons that have sufficiently long T2 relaxation times (N10 ms). The T2 of less mobile protons associated with macromolecules is too short (b1 ms) to be detected in MRI. However, in tissues, macromolecules interact with water and influence its relaxation properties by a variety of mechanisms. The dipole–dipole interaction between mobile protons and macromolecular protons perturbs the equilibrium between the two pools of protons, thereby affecting the MR signal. Selective saturation of the macromolecular resonance results in MT to the water protons resulting in the decrease of the magnitude of water signal observed. The amount of reduction depends upon the rate of exchange between the two spin populations. Thus, images with MTI show loss of signal compared to normal image without MT. The decrease will be larger in regions where exchange of magnetization is more efficient, determined by the relative proportion of protons of the two pools, their intrinsic relaxation times and the exchange rate. Application of an off-resonance RF pulse, called MT pulse, is the common method of MT contrast generation. To allow detection of magnetization resulting from immobile protons of bound water and macromolecules, the MT pulse is applied before slice selection and with a frequency shift from resonance, e.g., 1.5 kHz. The MT effect is quantified by determining the magnetization transfer ratio (MTR). It requires acquiring data sets with the MT pulse turned off and on, while all the other parameters being identical. To date most applications of MTI have been in MR angiography [18–21], brain and various neurological disorders [22–30]. MT ratio has been reported to increase with tumor malignancy [22,23]. For example, low-grade gliomas showed smaller MT effect than high-grade gliomas [22]. Similarly, MT contrast was higher in high-grade than in low-grade astrocytomas due to increased content of highmolecular weight nuclear material in high-grade tumor [23]. MTI has been shown to be a useful additional tool in the assessment of pituitary disorders [31]. Recently, the feasibility of whole-body MTI has been demonstrated [32]. There is only one report (in the Japanese language) on the application of MTI in prostate cancer [33]. The objective of this preliminary study was to demonstrate the role of MTI of prostate in men who had raised prostate-specific antigen (PSA) (N4 ng/ml) or abnormal digital rectal examination (DRE) and to use MTR for differentiation of normal and malignant prostatic tissues.
2. Methods The patient study population consisted of 26 men with raised PSA (N4 ng/ml) or abnormal DRE; age range,
645
50–83 years; mean age, 64.2 years; PSA range, 0.48–1507 ng/ml; mean PSA, 93.18 ng/ml; the other details are as presented in Table 1. Patients were recruited from the urology clinic and institute ethical committee approved the study. The exclusion criteria used for this study include patients with metallic implant or claustrophobic. Three healthy young men (age range, 28–30 years; mean age, 29.3 years) were recruited to serve as controls for the study. This age group was chosen to avoid contamination of normal control data from either BPH or asymptomatic prostate cancer [34]. MR investigations were carried out at 1.5 T using a whole-body scanner (Sonata/Avanto, Siemens, Erlangen, Germany). The pulse sequence used for MTC consisted of an off-resonance saturation pulse immediately before the first 90° RF pulse. As discussed earlier, two data sets of transverse images with and without MT pulse were acquired keeping all the other parameters identical. Details of the various parameters used are follows: pulse length, 7680 μs; bandwidth, 250 Hz; frequency offset, 1500 Hz; and flip angle, 500°. Images with and without MT pulse used for MTR calculation were acquired using a TR of 1000 ms and a TE of 13 or 14 ms with a slice thickness of 4 or 5 mm, without inter-slice gap. These values of TR and TE were chosen to minimize T1 and T2 effects [16]. T2-weighted images were acquired using TR 4000–5000 ms and TE 98 ms with slice thickness of 4 or 5 mm without inter-slice gap. The slice position of MT images and T2-weighted transverse images was identical. The amount of MT is quantified by the calculation of MTR using the equation MTR =1−Ms/Mo]×100, where Mo is the signal intensity of a given ROI without MTC and Ms is the signal intensity of the same ROI with MT. Signal Table 1 Details of patients used for MTR investigation Patient no.
Age (years)
PSA (ng/ml)
Biopsy data (Gleason grade or score)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
65 52 83 62 69 67 75 67 55 67 55 68 72 54 60 70 64 60 75 64
4.2 5.5 155 49.6 10.3 5.4 4.2 1.73 30.9 0.48 5.08 12.2 20.96 11.01 110 38.56 26.77 1507 13.7 NA
Adenocarcinoma Adenocarcinoma 5 6 Negative Negative Negative Negative 5b Negative Adenocarcinoma Negative 2 Negative 9(4+5) 7(3+4) 3a+3c 3a 8(4+4) Negative
646
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649
Fig. 1. Representative example of prostate MR images of a patient whose PSA level is 38.56 ng/ml; Gleason score, 7 (3+4); and age, 70 years. (A) T2-weighted image; (B) T2-weighted image showing the ROIs for the calculation MTR; (C) image without MT pulse; and (D) image with MT pulse. The figure also shows the placement of circular ROIs for calculation of MT ratios from CG, PZ and the malignant region. The hypo-intense (dark — shown by arrows) signal area of the PZ shown in (A) is indicative of malignancy which was confirmed on biopsy. MTR between the boundary of the PZ and the CG was not included in the calculation to avoid contamination of data between tissue types. Similarly, a small portion between the hypo- and hyper-intense areas of the PZ was not included either in the MTR calculation. Accordingly in the figure, the ROIs are not shown in these regions.
intensities were measured by drawing circular ROIs of uniform size of 0.04 cm2 (12 pixels). For placement of ROIs the T2-weighted images were used (vide infra). MTR was calculated from the peripheral zone (PZ) and the central gland (CG) of the prostate by drawing nonoverlapping consecutive ROIs. Care was taken while drawing ROIs to sample exclusively either from hyper- (normal) or hypointense (suspicious of malignancy seen as low signal intensity on T2-weighted images) regions of the PZ or CG, to minimize contamination of data between tissue types. Furthermore, mean MTR was calculated for the whole of the PZ [including hyper- (normal) and hypo-intense regions (suspicious of malignancy)] in all patients. One-way analysis of variance was applied to compare MTR calculated from normal PZ, CG and malignant portion of the PZ tissue. Average value for the whole of the PZ and CG was calculated for comparison between patient and controls. Mann–Whitney test was applied to compare MTR between patients positive for malignancy and negative for malignancy. 3. Results Out of 26 patients, images acquired with and without MT pulse showed a mismatch due to movement in four patients. Biopsy report was not available for one patient and MR
examination could not be completed in one patient due to technical reasons. Thus a total of 3725 ROIs from 20 patients and three controls (2153 and 1572 ROIs for CG and PZ, respectively) were used for MTR calculation and further analysis. The age, PSA and biopsy data of these patients are summarized in Table 1. Table 2 Comparison of mean MTR calculated for the whole of the PZ and the whole of the CG in control and patient groups Region
Mean MTR±SD
1. Control (n=3) (a) PZ (b) CG ⁎
6.18±1.63 6.97±1.90
2. Patients positive for malignancy on histopathology (n=12) (a) PZ 8.29±3.49 ⁎⁎ (b) CG ⁎ 7.23±3.80 3. Patients negative for malignancy on histopathology (n=8) (a) PZ 6.18±3.15 ⁎⁎⁎ (b) CG 7.01±4.17 ⁎ CG was not distinguished clearly in prostate of one control and one patient and hence not included in the analysis. ⁎⁎ Denotes the Pb.05 with respect to control for PZ. ⁎⁎⁎ Denotes the Pb.05 with respect to the value observed for PZ for patients who were positive for malignancy on biopsy.
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649
647
Table 3 Comparison of MTR values of patients (n=4) who showed hyper- (normal) and hypo-intense (region suspicious of malignancy) signal on T2-weighted image of the PZ and CG Patient no.
Region
No. of ROIs
Mean MTR±S.D.
P value
1.
a. Hyper-intense area of the PZ (normal) b. Hypo-intense region (suspicious of malignancy) c. CG a. Hyper-intense area of the PZ (normal) b. Hypo-intense region (suspicious of malignancy) c. CG a. Hyper-intense area of the PZ (normal) b. Hypo-intense region (suspicious of malignancy) c. CG a. Hyper-intense area of the PZ (normal) b. Hypo-intense region (suspicious of malignancy) c. CG
136 35 33 38 28 75 8 11 297 6 21 22
4.67±2.02 7.69±1.66 6.81±2.32 2.07±1.77 6.91±1.36 5.57±2.12 5.81±0.49 9.56±2.02 6.83±2.33 2.13±2.02 7.04±1.16 4.87±2.11
.0001 (a vs. b) .035 (b vs. c) .0001 (c vs. a) .0001 (a vs. b) .002 (b vs. c) .0001 (c vs. a) .0001 (a vs. b) .0001 (b vs. c) Not significant .0001 (a vs. b) .0001 (b vs. c) .001 (c vs. a)
2.
3.
4.
Combined Analysis of these 4 patients Region
Mean MTR±S.D.
Hyper-intense area of the PZ (normal)
3.43±2.36
Hypo-intense region (suspicious of malignancy) CG
7.53±1.70 6.57±2.35
Fig. 1 shows the representative example of T2-weighted, without and with MT pulse, images of prostate of a patient whose PSA level is 38.56 ng/ml. The hyper-intense area (with respect to muscle tissue/CG) on T2-weighted image corresponds to normal PZ while the low signal intensity area (hypo-intense) corresponds to a region suspicious of malignancy (see Fig. 1A) [1–5]. Later, the histopathology evaluation of biopsy specimens of these patients confirmed malignancy in this area. Out of 20 patients, biopsy revealed malignancy in 12 patients. The mean MTR was calculated for the whole of the PZ and the whole of the CG in controls, patients positive for malignancy (on biopsy, n=12) and patients negative for malignancy (n=8), and the values are presented in Table 2. A mean MTR value of 8.29±3.49 was obtained for the whole of the PZ of patients who were positive for malignancy on biopsy which was statically higher than that observed for patients who were negative for malignancy (6.18±3.15). The mean MTR for the whole of the PZ of controls was 6.18±1.63 and is similar to that of patients who were negative for malignancy. Four out of 12 patients who were positive for malignancy on biopsy showed certain regions suspicious of malignancy of the PZ on T2-weighted image while the rest of the portion of the PZ appeared hyper-intense (normal). The biopsy revealed normal and malignant PZ tissue in the respective cores in these 4 patients and their MTR values are summarized in Table 3. The mean MTR calculated from the hypo-intense region suspicious of malignancy of the PZ was significantly higher than that observed for the hyperintense normal portion of the PZ and the CG (see Table 3). In addition, the mean MTR obtained for the hyper-intense portion of the PZ was significantly lower than that observed for the CG. No significant difference between mean MTR of
P value Hyper-intense area of the PZ (normal) vs. hypo-intense region (suspicious of malignancy) Hyper-intense area of the PZ (normal) vs. CG Hypo-intense region (suspicious of malignancy) vs. CG
.0001 .0001
the PZ and the CG was observed in one patient. Furthermore, the combined analysis of these four patients showed statistically significant difference in the mean MTR of the hyper- and hypo-intense regions of the PZ and the CG.
4. Discussion In this preliminary study, we investigated the diagnostic potential MTI of prostate in patients who had raised PSA or abnormal DRE. MTI is a promising method of tissue characterization and reflects the microstructural changes that accompany the pathological process. Thus, MTR provides a quantitative index of the structural integrity of tissues. Different tissues display different amounts of MT and the degree of MT within a tissue is primarily governed by the local concentration of macromolecules. Cell membrane proteins and phospholipids mainly contribute to the MT signal. Change in MTR may reflect abnormal cell membrane structure or decreased cell number or cell size. Our data indicate that mean MTR within a patient showed a significant difference between the normal (hyper-intense area) and suspicious region of malignancy (hypo-intense area) of the PZ (see Table 3). Inter-patient comparison revealed a similar behavior (Table 2). The CG was not distinguishable in one control; therefore, comparison of CG with controls could not be carried out. The finding of significantly higher MTR for the hypo-intense region compared to the hyper-intense region of the PZ is further supported by the statistically higher mean MTR obtained for the whole of the PZ of patients who were positive for malignancy on histopathology compared to those negative for malignancy. The MTR values calculated and reported in our study are from one tissue type. Contamination from
648
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649
different tissues is not expected in our study, since the small portion between the PZ and the CG and between the hypointense and the hyper-intense region of the PZ was not included for the MTR calculation, as shown in Fig. 1. Furthermore, the mean MTR for the whole of the PZ of patients who were positive for malignancy on biopsy was significantly different from that observed for controls and patients who were negative for malignancy. As expected, no significant difference was observed in the mean MTR calculated for the whole of the PZ of controls and patients who were negative for malignancy. The only study that is available in the literature by Arima et al. [33] also reported similar findings. Progress in malignancy of prostate adenocarcinoma is associated with histopathological features of cellular atypia, pleomorphism, de-differentiation and increased cellularity. With increasing malignancy, an increase in the amount of high molecular weight material in cellular nuclei is reported [22]. The glandular structure of prostate is disrupted during malignancy and luminar spaces are replaced with proliferating malignant cells, leading to higher MTR in the malignant regions of the prostate. Various MRI methods have been used for prostate cancer detection and localization, as well as to study their biological behavior [1–9]. Routine MRI demonstrates zonal anatomy with excellent soft tissue contrast resolution and allows assessment of the local extent of disease. MRS improves the prostate cancer detection and localization [1–9]. Lower citrate to choline and creatine ratio is observed in cancerous tissue than in normal tissue. In addition, because the cellular structure of malignant tumors differs from that of normal tissue, it is documented that diffusion MRI is also a useful modality in the diagnosis of prostate cancer with the apparent diffusion coefficient of tumor being less than that observed for the normal PZ tissue [10–13]. In this direction, our preliminary data using MTI suggest that this technique may have the potential to provide additional information in the diagnosis of prostate cancer in addition to the above existing MR methodologies. 5. Conclusion This preliminary study demonstrated the possible role of MT imaging in the differentiation of normal (hyperintense) and suspicious region of malignancy (hypointense) of the PZ of prostate cancer patients. The mean MTR calculated for the whole of the PZ of patients who were positive for malignancy on biopsy is higher than that observed for the normal PZ of controls and patients who were negative for malignancy. Furthermore, for patients who showed hyper- (normal) and hypo-intense regions (suspicious of malignancy) of the PZ, the MTR was statistically significantly different. These interesting results should be further investigated in a large cohort of patients and controls.
Acknowledgment VK thanks the Indian Council of Medical Research (ICMR), New Delhi, for providing Junior Research Fellowship. References [1] Katz S, Rosen M. MR imaging and MR spectroscopy in prostate cancer management. Radiol Clin North Am 2006;44:723–34. [2] Choi YJ, Kim JK, Kim N, Kim KW, Choi EK, Cho KS. Functional MR imaging of prostate cancer. Radiographics 2007;27:63–75. [3] Akin O, Hricak H. Imaging of prostate cancer. Radiol Clin North Am 2007;45:207–22. [4] Hricak H, Choyke PL, Eberhardt SC, Leibel SA, Scardino PT. Imaging prostate cancer: a multidisciplinary perspective. Radiology 2007;243: 28–53. [5] Carroll PR, Coakley FV, Kurhanewicz J. Magnetic resonance imaging and spectroscopy of prostate cancer. Rev Urol 2006;8:S4–S10. [6] Rajesh A, Coakley FV, Kurhanewicz J. 3D MR spectroscopic imaging in the evaluation of prostate cancer. Clin Radiol 2007;62:921–9. [7] Noworolski SM, Henry RG, Vigneron DB, Kurhanewicz J. Dynamic contrast-enhanced MRI in normal and abnormal prostate tissues as defined by biopsy, MRI, and 3D MRSI. Magn Reson Med 2005;53: 249–55. [8] Scheidler J, Hricak H, Vigneron DB, Yu KK, Sokolov DL, Huang LR, et al. Prostate cancer: localization with three-dimensional proton MR spectroscopic imaging—clinicopathologic study. Radiology 1999;213: 473–80. [9] Wefer AE, Hricak H, Vigneron DB, Coakley FV, Lu Y, Wefer J, et al. Sextant localization of prostate cancer: comparison of sextant biopsy, magnetic resonance imaging and magnetic resonance spectroscopic imaging with step section histology. J Urol 2000;164:400–4. [10] Sato C, Naganawa S, Nakamura T, Kumada H, Miura S, Takizawa O, et al. Differentiation of noncancerous tissue and cancer lesions by apparent diffusion coefficient values in transition and peripheral zones of the prostate. J Magn Reson Imaging 2005;21:258–62. [11] Issa B. In vivo measurement of the apparent diffusion coefficient in normal and malignant prostatic tissues using echo-planar imaging. Magn Reson Imaging 2002;16:196–200. [12] Hosseinzadeh K, Schwarz SD. Endorectal diffusion-weighted imaging in prostate cancer to differentiate malignant and benign peripheral zone tissue. J Magn Reson Imaging 2004;20:654–61. [13] Kumar V, Jagannathan NR, Kumar R, Thulkar S, Gupta SD, Dwivedi SN, et al. Apparent diffusion coefficient of the prostate in men prior to biopsy: determination of a cut-off value to predict malignancy of the peripheral zone. NMR Biomed 2007;20:505–11. [14] Wolff SD, Balaban RS. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med 1989;10: 135–44. [15] Wolff SD, Balaban RS. Magnetization transfer imaging: practical aspects and clinical applications. Radiology 1994;192:593–9. [16] Finelli DA. Magnetization transfer in neuroimaging. Magn Reson Imaging Clin N Am 1998;6:31–52. [17] Henkelman RM, Stanisz GJ, Graham SJ. Magnetization transfer in MRI: a review. NMR Biomed 2001;14:57–64. [18] Edelman RR, Ahn SS, Chien D, Li W, Goldmann A, Mantello M, et al. Improved time-of-flight MR angiography of the brain with magnetization transfer contrast. Radiology 1992;184:395–9. [19] Pike GB, Hu BS, Glover GH, Enzmann DR. Magnetization transfer time-of-flight magnetic resonance angiography. Magn Reson Med 1992;25:372–9. [20] Atkinson D, Brant-Zawadzki M, Gillan G, Purdy D, Laub G. Improved MR angiography: magnetization transfer suppression with variable flip angle excitation and increased resolution. Radiology 1994;190:890–4.
V. Kumar et al. / Magnetic Resonance Imaging 26 (2008) 644–649 [21] Dousset V, Franconi JM, Degreze P, Balderrama J, Lexa F, Caille JM. Use of magnetization transfer contrast to improve cerebral 3D MR angiography. Neuroradiology 1994;36:188–92. [22] Kurki T, Lundbom N, Kalimo H, Valtonen S. MR classification of brain gliomas: value of magnetization transfer and conventional imaging. Magn Reson Imaging 1995;13:501–11. [23] Lundbom N. Determination of magnetization transfer contrast in tissue: an MR imaging study of brain tumors. AJR Am J Roentgenol 1992;159:1279–85. [24] Dousset V. Magnetization transfer imaging in vivo study of normal brain tissues and characterization of multiple sclerosis and experimental allergic encephalomyelitis lesions. J Neuroradiol 1993;20:297. [25] Filippi M. The role of magnetization transfer and diffusion-weighted MRI in the understanding of multiple sclerosis evolution. Neurol Sci 2000;21:S877–81. [26] Filippi M. Magnetization transfer MRI in multiple sclerosis and other central nervous system disorders. Eur J Neurol 2003;10:3–10. [27] Filippi M, Rocca MA. Magnetization transfer magnetic resonance imaging in the assessment of neurological diseases. J Neuroimaging 2004;14:303–13.
649
[28] Filippi M, Agosta F. Magnetization transfer MRI in multiple sclerosis. J Neuroimaging 2007;17(Suppl 1):22S–6S. [29] Pui MH, Wang Y. Diffusion and magnetization transfer MRI of brain infarct, infection, and tumor in children. Clin Imaging 2005;29:162–71. [30] Ridha BH, Tozer DJ, Symms MR, Stockton KC, Lewis EB, Siddique MM, et al. Quantitative magnetization transfer imaging in Alzheimer disease. Radiology 2007;244:832–7. [31] Argyropoulou MI, Kiortsis DN. Magnetization transfer imaging of the pituitary gland. Hormones 2003;2:98–102. [32] Boss A, Martirosian P, Kuper K, Fierlbeck G, Claussen CD, Schick F. Whole-body magnetization transfer contrast imaging. J Magn Reson Imaging 2006;24:1183–7. [33] Arima K, Hayashi N, Yanagawa M, Kawamura J, Kobayashi S, Takeda K, et al. The progress in diagnostic imaging for staging of bladder and prostate cancer: endorectal magnetic resonance imaging and magnetization transfer contrast. Hinyokika Kiyo 1999;45:553–7. [34] Kumar V, Jagannathan NR, Kumar R, Das SC, Jindal L, Thulkar S, et al. Correlation between metabolite ratios and ADC values of prostate in men with increased PSA level. Magn Reson Imaging 2006; 24:541–8.