Clinical Efficacy of Proton Magnetic Resonance Spectroscopy ( 1 H-MRS) in the Diagnosis of Localized Prostate Cancer

Transcription

1 Clinical Efficacy of Proton Magnetic Resonance Spectroscopy ( 1 H-MRS) in the Diagnosis of Localized Prostate Cancer KEISUKE SAITO 1, TATSURO KAMINAGA 2, SATORU MUTO 1, HISAMITSU IDE 1, KOUJIRO NISHIO 1, YUTAKA KAMIYAMA 1, HIROSHI OKADA 3, YUICHI TERADO 4, SHIGERU FURUI 2 and SHIGEO HORIE 1 Departments of 1 Urology and 2 Radiology, Teikyo University School of Medicine, Tokyo; 3 Department of Urology, Dokkyo Medical University, Saitama; 4 Department of Pathology, Kyorin University School of Medicine, Tokyo, Japan Abstract. Background: The objective of this study was to evaluate the efficacy of proton magnetic resonance spectroscopy ( 1 H-MRS) in the detection of prostate cancer. Materials and Methods: The experimental group consisted of 20 patients with localized prostate cancer who underwent radical prostatectomy. The sensitivity, positive predictive value and accuracy of the detection of prostate cancer in the 12 sections of the peripheral zone were calculated for prostate biopsy alone, 1 H-MRS alone and the combination of these methods, respectively. Results: The sensitivity, the positive predictive value, and the accuracy of the preoperative diagnosis of prostate cancer were 43.7%, 68.9% and 54.6%, respectively, with prostate biopsy alone; 60.6%, 87.8% and 71.4% with 1 H-MRS alone; and 88.7%, 88.7% and 87.4% with the combination of biopsy and 1 H-MRS, respectively. Combined use of biopsy and 1 H-MRS significantly improved the sensitivity, positive predictive value and accuracy of the diagnosis. Conclusion: 1 H-MRS together with biopsy might improve the diagnostic accuracy in prostate cancer. The incidence of prostate cancer was traditionally lower in Japan than in Western countries, although the mortality rate of prostate cancer is increasing in Japan, paralleling the increased performance of the PSA (prostate-specific antigen) test (1). Multi regional needle biopsy of the prostate has been the gold standard for achieving a definitive diagnosis of prostate cancer, but it has the drawback of often yielding false-negative results. The falsenegative rate could be reduced in number by increasing the Correspondence to: Shigeo Horie, MD, Department of Urology, Teikyo University School of Medicine, , Kaga, Itabashi-ku, Tokyo , Japan. Tel: , Fax: , shorie@med.teikyo-u.ac.jp Key Words: Prostate cancer, proton magnetic resonance spectroscopy, prostate biopsy, accuracy of diagnosis. number of biopsy sites within the gland such as saturation biopsies (1), but that leads to a proportional increase in complications for the patient (2). In addition, prostate cancer is often multi focal within the gland and comparison with the histopathological findings after total prostatectomy shows that in many cases the biopsied sites represent only a portion of the entire malignancy (3). Accordingly, a method that is less invasive yet has better diagnostic power for prostate cancer is sorely needed. However, although progress has been made with such techniques as transrectal ultrasound (TRUS) (4), computed tomography (CT), magnetic resonance imaging (MRI) (5), and positron emission computed tomography (PET) (3), the results of these modalities remain unsatisfactory. The conventional body coil MRI has a moderate power for the diagnosis of prostate cancer. However, it showed low accuracy in identifying the distribution of malignant tissue within the prostate (4). The diagnostic power of endorectal coil MRI was investigated for localized prostate cancer that was confirmed by total prostatectomy, and it was reported that the accuracy was 61-90%, the sensitivity %, the specificity 36-83%, the positive predictive value 53-93% and the negative predictive value was % (6-10), however these values vary widely. Normal prostate epithelial cells contain a high concentration of zinc. Zinc inhibits aconitase activity, thereby inhibiting the intracellular citric acid cycle. As a result, citric acid is not consumed, accumulates in the cells and is secreted into the glandular cavity (11). In contrast, in prostate cancer cells, citric acid metabolism via the citric acid cycle is activated, and the citrate in the cells decreases. In addition, in cancer cells, there is an increase in choline containing substances, which cause vigorous cell membrane synthesis and breakdown (12). MR spectroscopy is considered one of the most promising areas of prostate imaging research (13). Voxels, color-encoded to reflect abnormal ratios of metabolites, can be viewed directly as tumor maps (14). In the case of the normal prostate, 1 H /2008 $

2 Figure 1. The characteristics of MRS in normal and prostate cancer: a. Normal prostate: The high concentration of intracellular zinc inhibits the citric acid cycle, Leading to accumulation of citric acid in the normal prostate cell. b. Prostate cancer: Because of the change in the metabolic pathway, citric acid does not accumulate in the prostate cancer cell. The choline concentration is elevated in malignant cells. MRS shows a high peak for citrate and a low peak for choline containing substances. Conversely, for prostate cancer cells, 1 H-MRS shows a low peak for citrate and a high peak for choline containing substances (Figure 1). On the basis of these characteristics, it can be surmised that the ratio of the peaks for citrate and choline containing substances can serve as a good index of prostate cancer. Accordingly, 1 H-MRS is a diagnostic modality that is based on in vivo metabolic information. Therefore, MR spectroscopy has been shown to improve detection of prostate cancer (15, 16). For the present study, we performed 1 H-MRS and needle biopsy of the prostate gland and investigated whether the combination of these two techniques has the potential to improve the accuracy of diagnosis of localized prostate cancer compared with each of these modalities alone. Patients and Methods Patients. This study was carried out on 20 patients (median age: 67.0 years old, range; years) who were diagnosed with T1c- T2, N0M0 localized prostate cancer by biopsy during the period from April 2003 to April 2006 and underwent radical prostatectomy. Specimens of multi regional biopsy and radical prostatectomy. The multi-regional prostate biopsies were performed under TRUS guidance, including 12 sections of the peripheral zone of the gland, as shown in Figure 2. The total prostatectomy specimens were handled in accordance with the Japanese guideline for prostate specimens (17) and were sliced into 5-mm-thick horizontal sections perpendicular to the dorsal surface of the urethral mucosa. The sections were fixed in formalin and stained with hematoxylin-eosin. The pathological diagnoses were performed by a single pathologist (Terado Y). Figure 2. Systematic sites of performance of prostate needle biopsy. As shown in the figure, the peripheral zone of the prostate was divided into 12 sections and biopsy was performed at a total of 12 sites under TRUS guidance. From among the slices of the total prostatectomy specimen, three slices were decided, named a basal, mid-gland and apex. Then the left and right peripheral zones (black area) of each of the slices were divided into an outer side and an inner side, making 12 sections. 1 H-MRS. 1 H-MRS was performed using a 1.5 Tesla MRI system (Signa LX Horizon Echospeed, General Electric Medical Systems, Inc., USA). The sequence and imaging conditions that were employed were as follows: Sequence = PRESS based 3DPROSE; TR/TE =2000/130 msec; spectral width =1250 Hz; data acquisition points =256; activation area =8.0 or 16.0 cm 2 ; encode = 8 8 or 16 16; number of excitation =2. A body coil was used as the transmission coil, while the receiver coil was an endorectal coil (BPX-10, Nihon Medrad K.K., Tokyo, Japan) placed on the anterior surface of the rectum in close contact with the posterior aspect of the prostate (18, 19). Spectral data correcting volume (VOI) was set on the axial localization image manually to cover the whole outer gland. The thickness of the VOI was 2.0 cm. Six planes of fat saturation pulse were set at a region adjacent to a VOI. Full width half maximum of water peak after shimming was less than 14 ppm in all subjects. The processing of the 1 H-MRS data was performed with the supplied processing software (SAGE Ⅱ, General Electric Medical Systems, Milwaukee, USA) on the MRI console. Each spectral signal was zero filled to 512 points and multiplied by a 2 Hz of exponential function. Fourier transformation, phase adjustment, frequency adjustment, baseline correction and calculation of the CCr ratio were automatically performed. The median interval from the biopsy to performance of 1 H-MRS was 80 days, and the median interval between 1 H-MRS and total prostatectomy was 7.5 days. Results Cut-off value for CCr [(choline + creatine)/citrate] ratio in prostate cancer. The slices of the total prostatectomy specimens were compared with the 1 H-MRS images at the 1900

3 Saito et al: MRS in the Diagnosis of Localized Prostate Cancer Figure 3. Determination of the regions of interest (ROI) in prostate cancer sites and normal tissue sites. a. Total prostatectomy specimen: The diagonal red lines indicate the presence of cancer on the basis of the histopathological findings. b. 1 H-MRS image: 1 H-MRS image at the level corresponding to the slice of the total prostatectomy specimen (a). For the slice of the total prostatectomy specimen (a), the mean of each CCr ratio [= (choline + creatine)/citrate] was calculated from the ROI (blue encircled area) corresponding to the site with pathologically confirmed malignancy and the ROI (yellow rectangular area) corresponding to the normal prostate site. Figure 5. The total prostatectomy specimen and the corresponding 1 H- MRS image (Case 2). a. Total prostatectomy specimen: The area with red lines shows the region containing cancerous tissue. In this patient, a large tumor was confirmed in the right lobe, while a daughter lesion was confirmed in the left lobe. b. 1 H-MRS image: The orange areas indicate regions in which the CCr ratio was H-MRS thus also confirmed the presence of a large tumor in the right lobe and a daughter lesion in the left lobe. Figure 6. Sensitivity, positive predictive values and accuracy with biopsy alone, 1 H-MRS alone and biopsy and 1 H-MRS combined. For 10 patients (i.e. excluding the 10 patients used to decide the cut-off value), the peripheral zone of the prostate was divided into 12 sections (total n=120 sections). For each of the sections, the sensitivity and positive predictive values are shown for biopsy alone, 1 H-MRS alone and biopsy with 1 H-MRS combined. *p< White bar: prostate biopsy; shaded bar: MRS; black bar: prostate biopsy + MRS. Figure 4. CCr ratio at malignant prostate sites and normal prostate sites. The CCr values were plotted for each malignant prostate site and normal prostate site. The cut-off value for the diagnosis of prostate cancer by 1 H- MRS on the basis of the CCr ratio was set at 1.07, which is equivalent to the mean + 2 SD of the CCr ratio for cancerous prostate tissue. corresponding level. The regions of interest (ROI) for the prostate cancer lesions and the normal prostate sites were decided and the CCr ratio was calculated (Figure 3a, b). Data were obtained from 51 loci of 10 patients, and the CCr ratios between the prostate cancer lesions and the normal peripheral zone were compared (Figure 4). The mean CCr ratio for the prostate cancer lesions was 1.40±0.20, compared with 0.75±0.16 for the normal prostate sites. On the basis of these data, it was decided that the cut off value for diagnosis of prostate cancer by 1 H-MRS would be 1.07, which is equivalent to the mean + 2 standard deviation of the CCr ratio for a normal (noncancerous) peripheral zone (Figure 4). Sensitivity, positive predictive value and accuracy in cases of diagnosis by biopsy, 1 H-MRS and their combination. The cutoff value decided above was employed to retrospectively determine the location of lesions in the prostates of the remaining 10 patients by 1 H-MRS. Next, the entire peripheral 1901

4 zone was divided into 12 sections according to the regions of biopsies as shown in Figure 2. Each of these regions was examined for the presence/absence of malignancy on the basis of biopsy and 1 H-MRS alone as well as the combination of biopsy and 1 H-MRS. In addition, the generated results were compared with the histopathological findings for the total prostatectomy specimens and the sensitivity, positive predictive value and accuracy were compared among the three diagnostic approaches of biopsy and 1 H-MRS alone and in combination. In previous reports, the efficacy of MRS in the diagnosis of Transition zone carcinoma of the prostate gland was controversial, hence this investigation was limited to the peripheral zone (20). Comparison of the distribution of peripheral zone sites showing a CCr ratio 1.07 on the 1 H-MRS images with the corresponding slices of the total prostatectomy specimen showed that there was good agreement between the sites diagnosed as malignant by 1 H-MRS and the sites of cancer loci upon histopathological inspection of the total prostatectomy specimens (Figure 5). The power of detection of cancer by 1 H-MRS was investigated by comparing the sensitivity, positive predictive value and accuracy among the 12 sections of the prostate peripheral zone (Figure 6). Examination of the total prostatectomy specimens revealed the presence of malignancy in 71 sections in 10 patients: biopsy detected malignancy in 31 of these sections while 1 H-MRS detected malignancy in 43 of them. The sensitivity of detection was thus 43.7% with biopsy compared to 60.6% with 1 H-MRS (p<0.001). Combining the results of biopsy and 1 H-MRS increased the sensitivity of detection up to 88.7% (p<0.001 vs. biopsy, and 1 H-MRS alone). Examination of the total prostatectomy specimens revealed the presence of malignancy in 31 of the 44 sections in which biopsy had detected malignancy and in 43 of the 49 sections containing malignancy detected by 1 H-MRS. Therefore, the positive predictive value was 68.9% with biopsy and 87.8% with 1 H- MRS, which indicated that 1 H-MRS had a higher positive predictive value than did biopsies. The combination of the findings obtained by biopsy and 1 H-MRS indicated malignancy in a total of 70 prostate sections, and 63 of those sections also showed malignancy in the examination of the total prostatectomy specimens. Therefore, the positive predictive value was 88.7% with the combination of biopsy and 1 H-MRS (Figure 6). The accuracy was 54.6% with biopsy, 71.4% with 1 H-MRS and 87.4% with the combination of these two diagnostic modalities. Discussion At present, prostate biopsy is the only definitive means for the diagnosis of prostate cancer, with significant morbidity and mortality. In addition, false-negative results require repeated biopsies. Literature reports that around 30% of cancers are detected in the second biopsies (21). Current imaging modalities are not satisfactory enough to support biopsies. Our present study was thus designed and carried out to elucidate whether the combined use of 1 H-MRS and biopsy improve the accuracy of identifying the localized distribution. We first established a cut-off value for the CCr ratio in prostate cancer. Previous reports have determined the cut-off value as a value exceeding the mean + 2 SD of the CCr ratio for noncancerous prostate tissue (14, 22). By applying the same principle, a cut-off value for the CCr ratio in this study included no false-positives sites and only one false-negative site (Figure 4). In our study, the sensitivity, positive predictive value and accuracy were all significantly higher with 1 H-MRS alone than with biopsy alone. Moreover, with the combination of 1 H-MRS and biopsy each of the sensitivity, positive predictive value and accuracy was further increased. The combined application of 1 H-MRS and biopsy was thus proposed to be a good method for improving the accuracy of diagnosis of localized prostate cancer. More practically, this approach allows the clinician to consider performing targeted biopsy of the prostate only at the regions that are found to be 1 H-MRS-positive (23), or re biopsy of sections found to be 1 H-MRS-positive but biopsy-negative. Scheidler et al. (14) divided total prostatectomy specimens into sextants and found that with MRS the ranges of values in each section were 63~86% for the sensitivity, 49~75% for the specificity, 77~83% for the positive predictive value, 51~65% for the negative predictive value and 67~74% for the accuracy. Wefer et al. (24) similarly divided total prostatectomy specimens into sextants and compared the findings with those of sextant prostate biopsy, MRI and MRS. They reported that the sensitivity of MRS was the highest (MRI 67%, MRS 76% and biopsy 50% ), whereas its specificity was the lowest (MRI 69%, MRS 68% and biopsy 82% ). However, both of those research groups carried out their investigations on the premise of sextant biopsy based on division of the peripheral zone of the prostate into six sections, whereas today the recommendation is to use a biopsy core number of at least eight (25). Our present study is the first reported use of at least 10 cores and comparison of the diagnostic capacities of biopsy and MRS, and it is also the first documentation that addition of MRS to biopsy is useful for diagnosing localized prostate cancer. 1 H-MRS is a diagnostic modality that is based on in vivo metabolic information, and for this reason it provides information that differs from the information obtained with conventional MRI. Several lines of evidence show that 1 H- MRS would be useful for detecting tumors even after administration of hormonal therapy (22) or radiotherapy 1902

5 Saito et al: MRS in the Diagnosis of Localized Prostate Cancer (26) without the signal intensity being affected. The limitation of 1 H-MRS in this study was the use of the endorectal coil. It has been reported that endorectal coil increases diagnostic ability, (1, 5, 15, 26) but it places considerable stress on the patient. Moreover, it has been reported that it is often difficult to differentiate cancer from benign hyperplasia in the transition zone in MR imaging, even when the endorectal coil is used (27), which negates the application of 1 H-MRS in the transitional and central zones. The capability of the spatial resolution of the endorectal coil is lowered in the transition zone by the distance from the coil. Further modalities to circumvent endorectal coils warrant further studies. In conclusion, 1 H-MRS resulted in better accuracy in the diagnosis of prostate cancer than biopsy alone for the peripheral zone of the gland. 1 H-MRS together with biopsy might improve the diagnostic accuracy and the location of primary carcinomas. References 1 Nakata S, Takahashi H, Ohtake N, Takei T and Yamanaka H: Trends and characteristics in prostate cancer mortality in Japan. Int J Urol 7: , Presti JC Jr, O Dowd GJ, Miller MC, Mattu R and Veltri RW: Extended peripheral zone biopsy schemes increase cancer detection rates and minimize variance in prostate-specific antigen and age-related cancer rates: results of a community multi-practice study. J Urol 169: , de Jong IJ, Pruim J, Elsinga PH, Vaalburg W and Mensink HJ: 11C-Choline positron emission tomography for the evaluation after treatment of localized prostate cancer. Eur Urol 44: 32-38, Rifkin MD, Zerhouni EA, Gatsonis CA, Quint LE, Paushter DM, Epstein JI, Hamper U, Walsh PC and McNeil BJ: Comparison of magnetic resonance imaging and ultrasonography in staging early prostate cancer results of a multi-institutional cooperative trial. N Engl J Med 323: , Clarke DH, Banks SJ, Wiederhorn AR, Klousia JW, Lissy JM, Miller M, Able AM, Artiles C, Hindle WV, Blair DN, Houk RR and Sheridan MJ: The role of endorectal coil MRI in patient selection and treatment planning for prostate seed implants. Int J Radiat Oncol Biol Phys 52: , Ikonen S, Kivisaari L, Tervahartiala P, Vehmas T, Taari K and Rannikko S: Prostatic MR imaging: Accuracy in differentiating cancer from other prostatic disorders. Acta Radiol 42: , Comet-Batlle J, Vilanova-Busquets JC, Saladie-Roig JM, Gelabert- Mas A and Barcelo-Vidal C: The value of endorectal MRI in the early diagnosis of prostate cancer. Eur Urol 44: , Tanaka N, Samma S, Joko M, Akiyama T, Takewa M, Kitano S and Okajima E: Diagnostic usefulness of endorectal magnetic resonance imaging with dynamic contrast-enhancement in patients with localized prostate cancer: mapping studies with biopsy specimens. Int J Urol 6: , Ikonen S, Karkkainen P, Kivisaari L, Salo JO, Taari K, Vehmas T, Tervahartiala P and Rannikko S: Magnetic resonance imaging of clinically localized prostatic cancer. J Urol 59: , Hricak H, White S, Vigneron D, Vigneron D, Kurhanewicz J, Kosco A, Levin D, Weiss J, Narayan P and Carroll PR: Carcinoma of the prostate gland: MR imaging with pelvic phased-array coils versus integrated endorectal-pelvic phasedarray coils. Radiology 193: , Costello LC and Franklin RB: Concepts of citrate production and secretion by prostate. Metabolic relationships. Prostate 18: 25-46, Macara IG: Elevated phosphocholine concentration in rastransformed NIH 3T3 cells arises from increased choline kinase activity, not from phosphatidylcholine breakdown. Mol Cell Biol 9: , Thornbury JR, Ornstein DK, Choyke PL, Langlotz CP and Weinreb JC: Prostate cancer: what is the future role for imaging? AJR Am J Roentgenol 176: 17-22, Scheidler J, Hricak H, Vigneron DB, Vigneron DB, Yu KK, Sokolov DL, Huang LR, Zaloudek CJ, Nelson SJ, Carroll PR and Kurhanewicz J: Prostate cancer: localization with threedimensional proton MR spectroscopic imaging-clinicopathologic study. Radiology 213: , Swindle P, McCredie S, Russell P, Himmelreich U, Khadra M, Lean C and Mountford C: Pathologic characterization of human prostate tissue with proton MR spectroscopy. Radiology 228: , Kaji Y, Kurhanewicz J, Hricak H, Sokolov DL, Huang LR, Nelson SJ and Vigneron DB: Localizing prostate cancer in the presence of postbiopsy changes on MR images: role of proton MR spectroscopic imaging. Radiology 206: , Japanese Urological Association and The Japanese Society of Pathology: General Rule for Clinical and Pathological Studies on Prostate Cancer (3rd edition). Pp50-56, Tokyo, Kanehara Huch Boni RA, Meyenberger C, Pok Lundquist J, Trinkler F, Lutolf U and Krestin GP: Value of endorectal coil versus body coil MRI for diagnosis of recurrent pelvic malignancies. Abdom Imaging 21: , Perrotti M, Han KR, Epstein RE, Kennedy EC, Rabbani F, Badani K, Pentuck AJ, Weiss RE and Cummings KB: Prospective evaluation of endorectal magnetic resonance imaging to detect tumor foci in men with prior negative prostastic biopsy: a pilot study. J Urol 162: , Kurhanewicz J, Vigneron DB, Males RG, Swanson MG, Yu KK and Hricak H: The prostate: MR imaging and spectroscopy. Radiol Clin North Am 38: , Satoh A, Matsumoto K and Nakamura S: Is interval from an initial biopsy a significant predictor of prostate cancer at repeat biopsies? Int J Urol 13: , Mueller-Lisse UG, Vigneron DB, Hricak H, Swanson MG, Carroll PR, Bessette A, Scheidler J, Srivastava A, Males RG, Cha I and Kurhanewicz J: Localized prostate cancer: effect of hormone deprivation therapy measured by using combined three-dimensional 1 H-MRspectroscopy and MRimaging: clinicopatho-logic case-controlled study. Radiology 221: , Prando A, Kurhanewicz J, Borges AP, Oliveira EM and Figueiredo E: Prostatic biopsy directed with endorectal MR spectroscopic imaging findings in patients with elevated prostate specific antigen levels and prior negative biopsy findings: Early experience. Radiology 236: ,

6 24 Wefer AE, Hricak H, Vigneron DB, Vigneron DB, Coakley FV, Lu Y, Wefer J, Mueller-Lisse U, Carroll PR and Kurhanewicz J: Sextant localization of prostate cancer: comparison of sextant biopsy, magnetic resonance imaging and magnetic resonance spectroscopic imaging with step section histology. J Urol 164: , Makhlouf AA, Krupski TL, Kunkle D and Theodorescu D: The effect of sampling more cores on the predictive accuracy of pathological grade and tumour distribution in the prostate biopsy. BJU Int 93: , Coakley FV, Hricak H, Wefer AE, Speight JL, Kurhanewicz J and Roach M: Brachytherapy for prostate cancer: endorectal MRimaging of local treatment-related changes. Radiology 219: , Jager GJ, Ruijter ET, van de kaa CA, de la Rosette JJ, Oosterhof GO, Thornbury JR, Ruijs SH and Barentsz JO: Dynamic TurboFLASH subtraction technique for contrast-enhanced MR imaging of the prostate: correlation with histopathologic results. Radiology 203: , Received January 18, 2008 Revised March 27, 2008 Accepted March 31,