Periprostatic Fat Thickness on MRI: Correlation With Gleason Score in Prostate Cancer

Transcription

1 Genitourinary Imaging Original Research Woo et al. Periprostatic Fat Thickness on MRI Genitourinary Imaging Original Research Sungmin Woo 1 Jeong Yeon Cho 1,2 Sang Youn Kim 1 Seung Hyup Kim 1,2 Woo S, Cho JY, Kim SY, Kim SH Keywords: Gleason score, MRI, periprostatic fat, prostate cancer, radical prostatectomy DOI: /AJR Received February 10, 2014; accepted after revision May 1, Department of Radiology, Seoul National University College of Medicine, 101 Daehak-ro, Jongno-gu, Seoul , Korea. Address correspondence to J. Y. Cho (radjycho@snu.ac.kr). 2 Institute of Radiation Medicine and Kidney Research Institute, Seoul National University Medical Research Center, Seoul, Korea. WEB This is a web eclusive article. AJR 2015; 204:W43 W X/14/2041 W43 American Roentgen Ray Society Periprostatic Fat Thickness on MRI: Correlation With in Prostate Cancer OBJECTIVE. The purpose of this study was to retrospectively evaluate the relationship between periprostatic fat thickness on MRI and Gleason score of prostate cancer using radical prostatectomy as the reference standard. MATERIALS AND METHODS. This study included 190 patients (mean age [± SD], 66.9 ± 7.0 years) who underwent MRI before radical prostatectomy. Two radiologists measured the subcutaneous and periprostatic fat thickness on midsagittal T2-weighted MR images as the shortest perpendicular distance from the pubic symphysis to the skin and prostate, respectively. Subcutaneous and periprostatic fat along with age, height, weight, body mass inde, and prostate-specific antigen (PSA) were correlated with Gleason score by using Pearson (r) or Spearman (ρ) correlation coefficients and compared between low- (Gleason score = 6) and high- ( 7) grade prostate cancer by using univariate and multivariate logistic regression analyses. RESULTS. The mean subcutaneous and periprostatic fat thicknesses were 24.0 ± 8.4 mm and 5.0 ± 2.0 mm, respectively. The Gleason score was significantly correlated with age (ρ = 0.181, p = 0.012), PSA (ρ = 0.345, p < 0.001), and periprostatic fat thickness (ρ = 0.228, p = 0.002). Multivariate analysis revealed that age, height, PSA level, and periprostatic fat thickness (odds ratio, 1.331; 95% CI, ) were independently predictive of high-grade (p 0.013) disease. CONCLUSION. Periprostatic fat thickness on MRI showed a mild to modest but significant correlation with Gleason score of prostate cancer with radical prostatectomy as the reference standard and was an independent predictive factor for high-grade prostate cancer. T here is increasing evidence implicating a relationship between obesity and prostate cancer [1, 2]. Although there are several confounding factors, such as the effect of obesity on cancer detection, treatment decisions, treatment efficacy, and treatment complications, it is becoming increasingly clear that obesity is associated with a greater risk of high-grade prostate cancer [2 4]. Because both obesity and prostate cancer are highly prevalent, understanding the biologic features of their relationship is of crucial importance. Until now, most of the literature eploiting the role of obesity on prostate cancer has used the body mass inde (BMI) as a surrogate marker for general obesity and has provided controversial results [5 7]. Nevertheless, visceral fat as opposed to subcutaneous fat, which is the most metabolically active fat is gaining more interest. Although the eact mechanism is yet to be elucidated, visceral fat is thought to play a role in prostate carcinogenesis by producing hormones and cytokines, such as tumor necrosis factor-α, interleukin-6, leptin, and adiponectin [8]. A few investigators have recently focused on the periprostatic fat using transrectal sonography or CT and their association with prostate cancer [9 11]. Bhindi et al. [9] found that the amount of retropubic periprostatic fat on transrectal sonography was a predictor of prostate cancer and particularly of high-grade prostate cancer. On the other hand, Roermund et al. [10] showed that the periprostatic fat measured on CT did not correlate with prostate cancer aggressiveness. That the reference standard for the aggressiveness of prostate cancer was based on the Gleason score from biopsy results and the investigators used different approaches in quantifying the periprostatic fat may have attributed to the discrepant results. Therefore, the purpose of our study was to retro- W43

2 Woo et al. spectively evaluate the relationship between periprostatic fat thickness on MRI and the Gleason score of prostate cancer using radical prostatectomy as the reference standard. 201 Patients satisfying inclusion criteria: - Preoperative prostate MRI - Radical prostatectomy at our institution within 6 months of MRI Fig. 1 Flowchart shows patient selection process for study. Materials and Methods Patient Selection Institutional Review Board approval was obtained for this retrospective study; requirement for informed consent was waived. A computerized search of our electronic medical records was performed to identify patients who met the following inclusion criteria: prostate MRI performed between January 2013 and December 2013 and radical prostatectomy performed within 6 months at our institution. Patients were ecluded if they had any history of treatment of prostate cancer (Fig. 1). As a result, 190 patients were included in our study population. The mean age (± SD) of the study group was 66.9 ± 7.0 years (range, years). The median time between MRI and radical prostatectomy was 29 days (range, days). MRI Technique MR images were acquired using various MRI scanners. Because our hospital is a tertiary referral center and the results of the image analysis methods in our study were not epected to be affected by the type of scanner (in terms of magnetic field strength, TR/TE parameters, and so on), outside MR images (n = 34) that had been imported into our electronic medical system were included. All MRI studies performed at our institution (n = 156) were performed with 3-T MR systems (Ingenia, Philips Healthcare [n = 32] and Magnetom Trio [n = 19] and Magnetom Verio [n = 105], both Siemens Healthcare). For suppression of bowel peristalsis, 20 mg of butyl scopolamine (Buscopan; Boehringer- Ingelheim) was injected intramuscularly before MRI. Only the sagittal T2-weighted images were evaluated as part of this study. Generally, the T2-weighted imaging parameters were as follows: TR range/te range, / ; slice thickness, 3 mm; interslice gap, 0.3 mm; matri; FOV, 22 cm; number of signals acquired, 2; and parallel imaging acceleration factor, 2. Aial and coronal T2- weighted imaging, aial T1-weighted imaging, aial fat-suppressed single-shot echo-planar diffusionweighted imaging with b values of 0 and 1000 s/ mm 2, and dynamic contrast-enhanced imaging were also routinely performed as part of the standard multiparametric prostate MRI protocol at our institution, although not assessed as part of this study. Image Analysis Two radiologists with 3 and 21 years of eperience in prostate MRI reviewed the images in consensus at a PACS system (Maroview, Marosis). Both 11 Patients ecluded: - Prior hormonal or radiation therapy 190 Patients included in final study population knew that all patients had undergone radical prostatectomy but were blinded to the clinical, surgical, and histologic findings. Using the sagittal T2-weighted image, the radiologists measured the subcutaneous and periprostatic fat thicknesses. The subcutaneous and periprostatic fat were determined by measuring the shortest perpendicular distance from the pubic symphysis to the skin and from the pubic symphysis to the prostate, respectively, on the midsagittal plane as shown in Figure 2. The shortest perpendicular distance was used with the intention of avoiding overestimation of these measurements [9]. Clinical and Pathologic Data Collection One author reviewed the patients electronic medical records before radical prostatectomy to evaluate for clinical data including height, weight, BMI, and serum prostate-specific antigen (PSA) level. In addition, information from the pathologic reports, including the Gleason score, was obtained from the pathology database. The pathologic reports were written according to a standardized processing and reporting protocol by a genitourinary pathologist with 17 years of eperience in uropathology after analyzing step-section histologic slides obtained at 4-mm-intervals from radical prostatectomy specimens [12]. Fig year-old man with prostate cancer. Sagittal T2-weighted image shows measurement of subcutaneous and periprostatic fat thickness. Subcutaneous and periprostatic fat were determined by measuring shortest perpendicular distance from pubic symphysis to skin and from pubic symphysis to prostate, respectively, on midsagittal plane. Statistical Analysis The Spearman correlation coefficient (ρ) was used to determine the association of the Gleason score and the potential predictors including BMI, PSA, subcutaneous fat thickness, and periprostatic fat thickness. For evaluation of the relationship between BMI and subcutaneous or periprostatic fat thickness, the Pearson correlation coefficient (r) was used. In addition, we compared the low- (Gleason score = 6) and high- (Gleason score 7) grade prostate cancers to ascertain the differences in these potential predictive variables [13]. Then, multivariate logistic regression analysis including all variables was performed using the stepwise forward procedure to determine the variables that were independent predictors of high-grade prostate cancer. The Hosmer-Lemeshow goodness-of-fit test was used to assess the quality of the fitted model to the observed data, and a p value less than 0.05 indicated that the model corresponded well to the actual data [14]. Statistical analyses were performed with software (MedCalc, version , MedCalc and PASW, version 18, SPSS). A p value less than 0.05 was considered to indicate a significant difference. Results Patient and Tumor Characteristics The mean height and weight of the patients were 167 ± 5.8 cm (range, cm) and, 67.8 ± 8.3 kg (range, kg), respectively. The calculated BMIs of the patients were 24.4 kg/m 2 ± 2.6 (range, kg/m 2 ). The patients were normal (BMI < 25 kg/m 2 ), overweight (BMI kg/m 2 ), and obese (BMI 30 kg/m 2 ) in 117 (61.6%), 69 (36.3%), and four (2.1%) patients, respectively, according to the World Health Organization categories. Whereas BMI was significantly correlated with subcutaneous fat (r = 0.668, p < 0.001), it did not show a significant relationship with periprostatic fat (p = 0.110). The mean preoperative PSA level was 11.5 ± 11.6 ng/ml (range, ng/ml). The mean subcutaneous and periprostatic fat on MRI were 24.0 ± 8.4 mm (range, W44

3 Periprostatic Fat Thickness on MRI mm) and 5.0 ± 2.0 mm (range, mm), respectively. The Gleason score of prostate cancer on the basis of radical prostatectomy were (n = 46), (n = 71), (n = 61), and 8 (n = 12; in 3, in 7, in 1, and in one patient, respectively). Correlation With The Gleason score was significantly correlated with age (ρ = 0.181, p = 0.012), PSA level (ρ = 0.345, p < 0.001), and periprostatic fat thickness (ρ = 0.228, p = 0.002) (Fig. 3). Height, weight, BMI, or subcutaneous fat thickness did not show significant correlation with Gleason score (p = 0.131, 0.400, 0.784, and 0.788, respectively). Differentiation of 7 From 6 The univariate analyses of clinical, pathologic, and fat measurements stratified to low- Age (y) Periprostatic Fat Thickness (mm) * * and high-grade prostate cancer are summarized in Table 1. Patients with aggressive prostate cancer had significantly greater age, serum PSA levels, and periprostatic fat thickness (p = 0.06, 0.02, and 0.03, respectively). BMI and subcutaneous fat thickness were not significantly different between patients with low- and high-grade prostate cancer. TABLE 1: Comparison of Clinical, Pathologic, and Fat Measurements According to Characteristic 6 (n = 47) 7 (n = 155) Age (y) 64.2 ± 6.9 (48 75) 68.0 ± 6.8 (47 97) Height (cm) ± 5.8 ( ) ± 5.6 ( ) Weight (kg) 66.2 ± 6.8 (53 84) 68.3 ± 8.7 (47 97) BMI (kg/m 2 ) 24.4 ± 2.1 ( ) 24.4 ± 2.7 ( ) PSA (ng/ml) 6.8 ± 4.6 ( ) 13.1 ± 12.7 ( ) < Subcutaneous fat (mm) 22.6 ± 6.5 ( ) 24.5 ± 8.9 ( ) Periprostatic fat (mm) 4.2 ± 2.0 (0 8.6) 5.2 ± 1.9 (0 13.3) Note Ecept for p, data are mean ± SD, with range in parentheses. BMI = body mass inde, PSA = prostatespecific antigen. A PSA (ng/ml) * p B 0 C Fig. 3 Bo-and-whisker plots of age, prostate-specific antigen (PSA), and periprostatic fat thickness in prostate cancers with different Gleason scores. A C, Gleason score was significantly correlated with (A) age (ρ = 0.181, p = 0.012), (B) PSA level (ρ = 0.345, p < 0.001), and (C) periprostatic fat thickness (ρ = 0.228, p = 0.002). W45

4 Woo et al. Multivariate logistic regression analysis (Table 2) revealed that age, height, PSA level, and periprostatic fat thickness were independently predictive of high-grade prostate cancer (p = 0.001, 0.006, 0.001, and 0.013, respectively). An increase in one millimeter of periprostatic fat thickness corresponded to a 31% (odds ratio, 1.331; 95% CI, ) increase in odds for prostate cancer with Gleason score 7. The Hosmer-Lemeshow goodness-of-fit test revealed that the model corresponded well with the observed data (p = 0.279). Discussion In our study, we have shown that the thickness of retropubic periprostatic fat measured on midsagittal MR images was significantly correlated with Gleason score and was able to differentiate Gleason score 7 or higher prostate cancer from Gleason score 6 prostate cancer. Moreover, the relationship between periprostatic fat and Gleason score was shown to be independent on multivariate regression analysis. This is in agreement with previous reports that used CT and transrectal sonography and showed that the amount of periprostatic fat was associated with aggressive prostate cancer [9, 11]. Because Gleason score is the best single predictor of outcome in patients with prostate cancer among all clinical and pathologic markers, our study, along with the literature, suggests that measurement of periprostatic fat, if properly validated, may have important clinical implications in preoperative assessment of prostate cancer aggressiveness and help direct the appropriate management and prognosis in patients with prostate cancer [15]. However, the measurement of periprostatic fat using our method should be interpreted with consideration of the preliminary and eploratory nature of our study, given the relatively low degree of correlation (r = 0.228) with Gleason score and relatively wide overlap between low- and high-grade prostate cancer. TABLE 2: Results of Multivariate Logistic Regression Analysis Characteristic Odds Ratio p Age (y) ( ) Height (cm) ( ) PSA (ng/ml) ( ) Periprostatic fat (mm) ( ) Note Data in parentheses are 95% CIs. PSA = prostate-specific antigen. Unlike periprostatic fat, BMI and subcutaneous fat thickness were not correlated with Gleason score. In addition, BMI showed a significant relationship with subcutaneous fat but not periprostatic fat thickness. These findings support the recent notion that visceral fat, including periprostatic fat, is the most metabolically active fat and may represent a distinct parameter from other measures rather than a surrogate marker for general obesity. In fact, there is emerging evidence on the role of periprostatic fat and adipokines, including cytokines, hormones, and growth factors produced by the periprostatic fat, on prostate cancer progression on the basis of radical prostatectomy specimens and in vitro culture models of prostate cancer and periprostatic fat [16 19]. We recognize that our study population did not include many patients who were obese. Most (61.6%) were of healthy weight (BMI < 25 kg/m 2 ). In comparison, % of the study population was of healthy weight in previous reports [9 11]. This could partially eplain the finding that BMI was shown to be a risk factor for high-risk disease in previous studies but not in ours [9, 11]. Thus, further studies that include a sufficient number of patients with BMI in the overweight and obesity categories are warranted to ascertain the significance of obesity with regard to prostate cancer aggressiveness. Of the several variables other than periprostatic fat thickness that were evaluated in our study, age, height, and PSA were also independent predictive factors associated with prostate cancer with Gleason score of 7 or higher. This was an epected finding because these factors are known to be associated with the aggressiveness of prostate cancer [20 23]. Therefore, they were incorporated in the multivariate logistic regression model along with periprostatic fat thickness, which did not alter the effect of periprostatic fat on Gleason score. Regarding the modality and methodology of measuring the thickness of periprostatic fat, we measured the thickness of retropubic fat in the midsagittal plane using MRI. We believe that using MRI in place of transrectal sonography and CT used in previous studies has certain strengths [9 11]. Transrectal ultrasound (TRUS) is highly operator dependent. Not only will there be a possibility of variation in selecting the image plane for measurement, but pressure applied against the prostate during the TRUS eamination could have an effect on the periprostatic fat thickness. CT, although an ecellent modality to distinguish and quantify subcutaneous and visceral fat, is not often used in patients with prostate cancer unless they plan to undergo radiation therapy. Quantification of fat using CT is highly accurate, although calculation of the total periprostatic fat volume and periprostatic fat density (%) requires dedicated software. Furthermore, CT has an inherent risk from ionizing radiation. On the other hand, MRI is routinely used for diagnosis, localization, and risk stratification in prostate cancer patients and is free of ionizing radiation. Moreover, our method is relatively simple and easy and does not require epertise in measurement or interpretation. In our study, we used the pathologic reports with radical prostatectomy as the reference standard. Other studies correlating periprostatic fat thickness with the aggressiveness of prostate cancer used biopsy as the reference standard [9, 11]. There is a risk of missing tissues with a high Gleason score in the prostate during biopsy. As a result, upgrading of Gleason score between prostate biopsy and radical prostatectomy is a common event, occurring in about 30% of patients who undergo radical prostatectomy [24]. Therefore, using radical prostatectomy for correlation of periprostatic fat with Gleason score may be a more adequate approach compared with using biopsy. Our study has several limitations. First, it was a retrospective analysis. Second, because Gleason score was based on radical prostatectomy as the reference standard, we included only patients who had undergone prostatectomy. This may have caused selection bias, and the results of our study may not be etrapolated to a broader patient population. For eample, in our study population, the number of patients who had prostate cancer with very high Gleason scores ( 8) was small (n = 12). However, we were able to show a statistically significant trend between periprostatic fat thickness and Gleason score and to show significant differences in periprostatic fat thickness between lowand high-grade prostate cancers. Fourth, measuring the thickness of periprostatic fat was only performed in one plane for practicality purposes. To take into account the total amount of periprostatic fat, an alternative approach, such as volumetric quantification, may be more accurate. However, this approach would be laborious, and the primary goal of our study was not to establish W46

5 Periprostatic Fat Thickness on MRI a standardized method but rather to identify an association. Lastly, causality cannot be assigned in this association between periprostatic fat and Gleason score because of our study design. A future step would be to evaluate the biologic activity of periprostatic fat in terms of adipokines from radical prostatectomy specimens and to correlate them with the thickness of periprostatic fat and the Gleason score of prostate cancer. By linking these data together, better insight may be gained in the eact role of periprostatic fat in prostate cancer. In conclusion, periprostatic fat thickness on MRI showed a mild to moderate but significant correlation with Gleason score of prostate cancer based on radical prostatectomy as the reference standard and was an independent predictive factor for high-grade (Gleason score 7) prostate cancer. These findings suggest that the measurement of periprostatic fat thickness on MRI, if validated properly, may have the potential to contribute to preoperative assessment of Gleason score in prostate cancer patients. References 1. Efstathiou JA, Bae K, Shipley WU, et al. Obesity and mortality in men with locally advanced prostate cancer: analysis of RTOG Cancer 2007; 110: Allott EH, Masko EM, Freedland SJ. Obesity and prostate cancer: weighing the evidence. Eur Urol 2013; 63: Spangler E, Zeigler-Johnson CM, Coomes M, Malkowicz SB, Wein A, Rebbeck TR. Association of obesity with tumor characteristics and treatment failure of prostate cancer in African- American and European American men. J Urol 2007; 178: Discacciati A, Orsini N, Wolk A. Body mass inde and incidence of localized and advanced prostate cancer: a dose-response meta-analysis of prospective studies. Ann Oncol 2012; 23: Amling CL, Riffenburgh RH, Sun L, et al. Pathologic variables and recurrence rates as related to obesity and race in men with prostate cancer undergoing radical prostatectomy. J Clin Oncol 2004; 22: Freedland SJ, Aronson WJ, Kane CJ, et al. Impact of obesity on biochemical control after radical prostatectomy for clinically localized prostate cancer: a report by the Shared Equal Access Regional Cancer Hospital database study group. J Clin Oncol 2004; 22: van Roermund JG, Kok DE, Wildhagen MF, et al. Body mass inde as a prognostic marker for biochemical recurrence in Dutch men treated with radical prostatectomy. BJU Int 2009; 104: van Roermund JG, Witjes JA. The impact of obesity on prostate cancer. World J Urol 2007; 25: Bhindi B, Trottier G, Elharram M, et al. Measurement of peri-prostatic fat thickness using transrectal ultrasonography (TRUS): a new risk factor for prostate cancer. BJU Int 2012; 110: van Roermund JG, Bol GH, Witjes JA, Ruud Bosch JL, Kiemeney LA, van Vulpen M. Periprostatic fat measured on computed tomography as a marker for prostate cancer aggressiveness. World J Urol 2010; 28: van Roermund JG, Hinnen KA, Tolman CJ, et al. Periprostatic fat correlates with tumour aggressiveness in prostate cancer patients. BJU Int 2011; 107: Samaratunga H, Montironi R, True L, et al.; ISUP Prostate Cancer Group. International Society of Urological Pathology (ISUP) Consensus Conference on Handling and Staging of Radical Prostatectomy Specimens Working Group 1: specimen handling. Mod Pathol 2011; 24: Thompson IM Jr, Goodman PJ, Tangen CM, et al. Long-term survival of participants in the Prostate Cancer Prevention Trial. N Engl J Med 2013; 369: Hosmer DW, Lemeshow S. Assessing the fit of the model. In: Hosmer DW, Lemeshow S. Applied logistic regression, 2nd ed. New York, NY: Wiley, 2000: Gleason DF, Mellinger GT. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 1974; 111: Toren P, Venkateswaran V. Periprostatic adipose tissue and prostate cancer progression: new insights into the tumor microenvironment. Clin Genitourin Cancer 2014; 12: Finley DS, Calvert VS, Inokuchi J, et al. Periprostatic adipose tissue as a modulator of prostate cancer aggressiveness. J Urol 2009; 182: Novakovic K, Fitchev P, Brendler C, et al. Active triacylglycerol metabolism and increased infiltrating adipocyte density in high grade prostate cancer. J Urol 2012; 187(suppl):e Ribeiro R, Monteiro C, Cunha V, et al. Human periprostatic adipose tissue promotes prostate cancer aggressiveness in vitro. J Ep Clin Cancer Res 2012; 31: Bechis SK, Carroll PR, Cooperberg MR. Impact of age at diagnosis on prostate cancer treatment and survival. J Clin Oncol 2011; 29: Sun L, Caire AA, Robertson CN, et al. Men older than 70 years have higher risk prostate cancer and poorer survival in the early and late prostate specific antigen eras. J Urol 2009; 182: Pinsky PF, Andriole G, Crawford ED, et al. Prostatespecific antigen velocity and prostate cancer Gleason grade and stage. Cancer 2007; 109: Zuccolo L, Harris R, Gunnell D, et al. Height and prostate cancer risk: a large nested case-control study (ProtecT) and meta-analysis. Cancer Epidemiol Biomarkers Prev 2008; 17: Cohen MS, Hanley RS, Kurteva T, et al. Comparing the Gleason prostate biopsy and Gleason prostatectomy grading system: the Lahey Clinic Medical Center eperience and an international meta-analysis. Eur Urol 2008; 54: W47