Comparison of Prostate Volume Measured by Endorectal Coil MRI to Prostate Specimen Volume and Mass After Radical Prostatectomy

Comparison of Prostate Volume Measured by Endorectal Coil MRI to Prostate Specimen Volume and Mass After Radical Prostatectomy Yousef Mazaheri, PhD, Debra A. Goldman, MS, Pier Luigi Di Paolo, MD, Oguz Akin, MD, Hedvig Hricak, MD, PhD Rationale and Objectives: To compare prostate volume measurements from 3-Tesla endorectal coil magnetic resonance imaging (ERC MRI) obtained with the prolate ellipsoid volume formula (EVF) and volumetry to pathology-based volume measurements. Methods: The institutional review board waived informed consent for this retrospective, health insurance portability and accountability act (HIPAA) compliant study, which included 195 patients who underwent 3-T ERC MRI between January 2008 and October 2011 and had pathologic prostate measurements available. Two readers in consensus measured the prostate length, height, and width on each MRI. They estimated prostate volumes using the prolate EVF (length  height  width  [p/6]) and also by performing three-dimensional volumetry. Pathologic specimen mass and dimensions were used to calculate prostate volume. Agreement was measured with Lin’s concordance correlation coefficient (CCC). Volume differences were assessed using the Wilcoxon signed-rank test. Correct prostate-specific antigen (PSA) density classification rates were compared between EVF-based and volumetry-based PSA density levels using the exact McNemar test, with pathology-based PSA density as the reference standard. Results: Concordance was high between EVF and volumetry measurements (CCC, 0.950 [95% confidence interval, 0.935–0.962]) and between both kinds of MRI measurements and pathology (both CCC > 0.80). Based on a cut-off of #0.15 ng/mL/cm3, use of EVF-based volume produced correct classification of 46 of 48 PSA density levels >15 ng/mL/cm3 and 113 of 147 PSA density levels #15 ng/mL/cm3; use of volumetry-based volume produced correct classification of 47 of 48 PSA density levels >15 ng/mL/cm3 and 121 of 147 PSA density levels #15 ng/mL/cm3. Rates of underclassification (P > .95) and overclassification (P = .10) did not differ significantly between EVF and volumetry. Conclusions: EVF appears to be suitable for measuring prostate volume from ERC-MRI. Key Words: Endorectal coil MRI; prostate volume; prostate specimen; radical prostatectomy. ªAUR, 2015

T

he accurate measurement of in vivo prostate volume is important for a wide variety of clinical situations. For patients with benign prostatic hypertrophy (BPH), prostate volume measurements are used to monitor the condition and assess the efficacy of treatment (1). For patients with prostate cancer, prostate volume measurements have prognostic significance and may influence diagnosis and management. A number of recent studies have shown that prostate volume can be useful in models for predicting the presence of indolent or clinically ‘‘insignificant’’ tumors (2–4). Furthermore, accurate estimation of prostate volume is Acad Radiol 2015; 22:556–562 From the Department of Medical Physics, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10605 (Y.M.); Department of Radiology, Memorial Sloan-Kettering Cancer Center, New York, NY (Y.M., P.L.D.P., O.A., H.H.); and Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, NY (D.A.G.). Received October 31, 2014; accepted January 10, 2015. Address correspondence to: Y.M. e-mail: [email protected] ªAUR, 2015 http://dx.doi.org/10.1016/j.acra.2015.01.003

556

necessary to calculate the prostate-specific antigen (PSA) density (the PSA level divided by the prostate volume), which may be used to help distinguish BPH from prostate cancer and to predict adverse treatment outcomes (5). Ha et al. (6) recently reported that among patients with low-risk prostate cancer (defined as a biopsy-derived Gleason score #6 in a single positive core, clinical stage #T1c, PSA #10 ng/mL, and unremarkable magnetic resonance imaging [MRI] results) who underwent radical prostatectomy, PSA density was a predictor of advanced disease (the reported cutoff was a PSA density of 0.085 ng/mL/cm3). Although a range of PSA density cutoffs have been reported in the literature, Epstein et al. (7) found that a PSA density of 0.1–0.15 ng/mL/cm3, along with low- to intermediate-grade cancer <3 mm found in only one needle biopsy core specimen, constituted the best model to predict preoperatively insignificant tumor. Prostate volume measurements are also used to determine patient selection for brachytherapy and the number of radioactive seeds used for the procedure (8). For patients with larger prostates, brachytherapy is associated with a higher incidence

Academic Radiology, Vol 22, No 5, May 2015

of side effects and is typically not recommended owing to technical difficulties primarily related to the encroachment of the pubic arch over the prostate gland (9). Prostate volume is typically measured with transrectal ultrasonography (TRUS) during prostate biopsies or before interstitial brachytherapy seed implantation. Other methods to measure prostate volume include digital rectal examination, computed tomography (CT), and MRI. MRI is considered more accurate (10) than CT (11,12) or ultrasound for measuring prostate volume (13,14). Although three-dimensional (3D) volumetry is believed to be the most accurate means of calculating prostate volume from imaging, it is generally considered too time consuming for routine clinical use. As a result, formula-derived methods have been proposed for prostate volume measurement with TRUS (14–17) and MRI (13,14,18,19). One of the most commonly used formulas is the conventional prolate ellipsoid volume formula (EVF), which incorporates measurements of the (maximum) length, (maximum) height, and (maximum) width as follows: length  height  width  (p/6) (16,18,20). However, a concern is that by deforming the prostate, the use of an endorectal coil (ERC) to acquire MRI might interfere with the accuracy of the EVF for measuring prostate volume. The purpose of our study was to compare prostate volume measurements from 3-T ERC MRI obtained with the prolate EVF and volumetry to pathology-based prostate volume measurements.

MATERIALS AND METHODS Our institutional review board waived the requirement for informed consent for this retrospective study, which was compliant with the Health Insurance Portability and Accountability Act. We searched the radiology and institutional databases to identify patients who underwent 3-T ERC MRI between January 2008 and October 2011 and subsequently had prostatectomy (n = 958). Nine hundred forty-three of these patients were found to have complete mass and dimension data available from pathology. Using SAS 9.2 (SAS Institute, Cary, NC), 200 patients were randomly chosen. Of these, 195 had complete clinical information and were included in our study. MRI Data Acquisition

MRI examinations were performed on a 3-T whole-body MRI unit (GE Medical Systems, Milwaukee, WI). A commercially available balloon-covered expandable ERC (Medrad, Pittsburgh, PA), which consists of an outer sheath covering an inner balloon, was inflated with 60–80 mL of air and inserted in the rectum. A body coil was used for excitation, and a pelvic phased-array coil was used in conjunction with the ERC for signal reception. As per the standard clinical prostate MRI examination at our institution, the images obtained included transverse T1-weighted images (repetition time (TR)/echo time (TE) = 400–1100/6–14 ms; 5-mm

COMPARISON OF VOLUME BY ERC MRI

slice thickness; 0-mm interslice gap; field of view [FOV] = 24–38 cm; matrix, 320  224) and transverse, coronal, and sagittal T2-weighted fast spin-echo images (TR/effective TE, 3200–4800/90–122 ms; echo train length, 12–16; 3-mm slice thickness; no interslice gap; FOV, 16 cm; matrix, 512  512) of the prostate and seminal vesicles. Pathology Preparation

Prostatectomy specimens were processed according to standard institutional protocol (21). Specimens underwent fixation and were subsequently fine-needle–injected with formalin, cut at regular intervals perpendicular to the posterior capsule, embedded in paraffin on 3- to 4-micron slides, and stained with hematoxylin and eosin. The slices were photographed using a digital camera. Analysis

MR Measurements. Measurements were made in consensus by a senior radiologist and a medical student, who jointly reviewed the transverse and sagittal T2-weighted images displayed on a picture and archiving communication system (PACS) workstation (Centricity RA 1000; GE Healthcare). They performed volumetry by first tracing the prostate circumference on each slice that contained prostate; the areas drawn on the image sections were then recorded from the PACS workstation, multiplied by the slice thickness, and summed to obtain an estimate of the prostate volume. The radiologist and medical student also recorded the prostate dimensions—specifically the maximum length (millimeters) in the right-to-left (RL) dimension from the axial images, maximum height (millimeters) in the superior–inferior (SI) dimension from the axial images, and the depth (millimeters) in the anterior–posterior (AP) dimension from the sagittal images. The prostate volume was calculated using the EVF (16,18,20). Pathology Measurements. The gross mass of the pathologic specimen (prostate gland with attached seminal vesicles) was obtained from the pathology report. Shortly after radical prostatectomy, while the specimen is still fresh, it is weighed. The reported average weight of seminal vesicles is 3.8 g (22). To compensate for the attached seminal vesicles, we subtracted this number from the specimen mass. Prior reports have suggested that the prostate mass (grams) can be used to estimate the volume (obtained otherwise through measuring water displacement). The reported conversion factor for prostatic tissue, or specific gravity, is 1.05 g/mL (23,24), and the estimated volume using the specific gravity formula (SGF) is mass/specific gravity. Using the presurgery PSA, PSA density was calculated based on pathologic (SGF-based) volume, EVF-based volume, and MRI volumetry. Statistical Analysis

Descriptive statistics were used to summarize clinical and demographic data; medians, ranges, and interquartile 557

MAZAHERI ET AL

ranges (IQRs) were used to summarize continuous variables, and frequencies and percents were used to summarize categorical variables. As previously mentioned, the prostate volume was assessed two ways on MRI (with EVF and volumetry) and one way on pathology (with SGF). We assessed the relationship between measurements on MRI using Lin’s concordance correlation coefficient (CCC) with corresponding 95% confidence intervals (95% CIs). The CCC was calculated using variance component measurements developed by Carrasco et al. (25). The CCC was chosen because it measures both precision, through the line of best fit, and accuracy, through the variation from the 45 line (26). Furthermore, the precision of measurements was graphically assessed using a scatter plot containing the line of best fit and a 45 line through the origin. The line of best fit was calculated using the least squares regression. In addition, the Wilcoxon signed-rank test was used to examine whether the difference between the measurements on MRI were significantly different from zero. The CCC was also used to measure the agreement between MRI volume measurements and the pathology volume. Bland–Altman plots, along with mean difference and 95% limits of agreement (LOA), were created to provide a graphical representation of the relationship between volumes measured on MRI and on pathology. We dichotomized the three PSA density measurements (we used a cutoff of 0.15 ng/mL/cm3, which is generally recommended for assessing the effect of volume shifts on PSA density.) Cross tabulations were created to demonstrate the frequency of cases that were correctly and incorrectly classified, using the pathology-based PSA density calculated with the SGF volume formula as the standard of reference. Underclassified cases were considered those for which MRI-based PSA density was <0.15 ng/mL/cm3 when pathology PSA density was >0.15 ng/mL/cm3 and overclassified cases were those for which MRI-based PSA density was >0.15 ng/mL/cm3 when pathology PSA density was <0.15 ng/mL/cm3. Additionally, we compared the >0.15 ng/mL/cm3 and #0.15 ng/mL/cm3 correct classification rates between EVF- and volumetry-based PSA density levels using the exact McNemar test. P < .05 was considered statistically significant. All analyses were performed using SAS 9.2 (SAS Institute, Cary, NC) and R 3.0.1, including the package ‘‘cccrm’’ (The R Foundation, http://www.r-project.org/).

RESULTS Patient Characteristics

One hundred ninety-five patients with a median age of 62.4 years (range, 43.1–75.9 years) and a prebiopsy PSA level of 5.3 ng/mL (range, 0.2–60.9 ng/mL) were included in this study.

558

Academic Radiology, Vol 22, No 5, May 2015

TABLE 1. Volume Measurements Volumes (cm3)

Median (IQR)

MRI volumetry MRI EVF Pathology SGF

41.6 (30.42–60.21) 41.5 (29.08–55.04) 51.6 (41.14–66.86)

EVF, ellipsoid volume formula; IQR, interquartile range; MRI, magnetic resonance imaging; SGF, specific gravity formula.

TABLE 2. Differences between Volume Measurements Volume Differences (cm3) MRI EVF MRI volumetry Path SGF MRI EVF Path SGF MRI volumetry

Median (IQR)

P Value

2.17 ( 1.64 to 6.24) 10.2 ( 4.10 to 15.56) 7.74 ( 3.25 to 13.28)

<.0001 <.0001 <.0001

EVF, ellipsoid volume formula; IQR, interquartile range; MRI, magnetic resonance imaging; SGF, specific gravity formula.

MRI and Pathology Volume Analyses

The median volumes derived from MRI with volumetry and the EVF and from pathology with the SGF were 41.6 cm3 (IQR, 30.43–60.21 cm3), 41.52 cm3 (IQR, 29.08–55.04 cm3), and 51.62 cm3 (IQR, 41.14– 66.86 cm3), respectively (Table 1). The median difference in total prostate volumes estimated on ERC-MRI with the EVF and volumetry was 2.17 cm3 (IQR, 1.57 to 6.24 cm3), indicating that half the measurements fell within approximately 8 cm3. The difference between the two MRI measurement mean ranks was significantly different from zero (P < .0001), as were the differences between the mean ranks of MRI measurements and the mean rank of measurements made from pathology with the SGF (Table 2). The CCC for MRI measurements was 0.950 (95% CI, 0.935–0.962), indicating very strong reproducibility. As demonstrated on the plot, the line of best fit and 45 line are nearly indistinguishable from one another (Fig 1). The majority of values cluster along the line of best fit; however, we see a bit more dispersion as the volumes increase. The CCCs of pathology-based (SGF based) measurements with EVF- and volumetry-based measurements were 0.849 (95% CI, 0.810–0.881; Fig 2) and 0.872 (95% CI, 0.836–0.900; Fig 3), respectively. These CCCs were lower than the CCCs for the MRI measurements but similar to each other. Using the Bland–Altman plots, no discernible pattern was present between mean volume measurements and differences between volume measurements. As illustrated in the plots (Fig 4), the mean (bias) and 95% LOA volume differences between volumetry and EVF-based measurements were 2.21 mm3 (95% LOA, 14.23 to 18.66 mm3), indicating that volumetry measurements were higher on average and the LOA spanned 33 mm3. For SGF and EVF, the mean difference was9.88 mm3 (95% LOA, 11.63 to 31.40 mm3), indicating that SGF

Academic Radiology, Vol 22, No 5, May 2015

Figure 1. Scatterplot and line of best fit for prostate volume measured using ellipsoid volume formula (EVF; dependent variable) versus prostate volume measured using volumetry (explanatory variable). Also shown is the relationship based on the linear regression (y = 0.95x + 4.77) and the line of 45 angle passing through the origin.

COMPARISON OF VOLUME BY ERC MRI

Figure 3. Scatterplot and line of best fit for prostate magnetic resonance volume measured using ellipsoid volume formula (dependent variable) versus prostate specimen volume measured using specimen mass (explanatory variable). Also shown is the relationship based on the linear regression (y = 0.82x + 16.21), and the line of 45 angle passing through the origin. SGF, specific gravity formula.

MRI and Pathology PSA Density Analyses

When PSA density was calculated based on pathologic volume, 48 of 195 patients (25%) had a PSA density >15 ng/mL/cm3, and 147 patients (75%) had a density #15 ng/mL/cm3. The use of EVF-based volume measurements resulted in correct classification of 46 of 48 (95.8%) PSA density levels >15 ng/mL/cm3 and 113 of 147 (76.9%) PSA density levels #15 ng/mL/cm3; 2 patients were underclassified (2 of 48; 4.2%) and 34 patients were overclassified (34 of 147; 23.1%). The use of volumetry measurements resulted in correct classification of 47 of 48 (97.9%) PSA density levels >15 ng/mL/cm3 and 121 of 147 (82.3%) PSA density levels #15 ng/mL/cm3; only 1 patient was underclassified (2.1%), and 26 patients (17.7%) were overclassified. The differences in the rates of underclassification (P > .95) and overclassification (P = .10) with EVF and volumetry were not significant.

Figure 2. Scatterplot and line of best fit for specimen prostate volume measured using ellipsoid volume formula (EVF; dependent variable) versus specimen prostate volume measured using specimen mass (explanatory variable). Also shown is the relationship based on the linear regression (y = 0.82x + 18.28), and the line of 45 angle passing through the origin. SGF, specific gravity formula.

measurements were higher on average and the LOA spanned 43 mm3 of each other. For SGF and volumetry, mean difference was 7.67 mm3 (95% LOA, 13.87 to 29.21 mm3), indicating that SGF measurements were higher on average and the LOA spanned 43 mm3.

DISCUSSION The accurate measurement of in vivo prostate volume is important in the evaluation of BPH and influences numerous aspects of prostate cancer care, including the assessment of prognosis. A study of a cohort of 1071 patients found that, when combined with other clinical and pathologic variables, the measurement of total prostate volume with MRI, as well as the PSA density calculated with that measurement, was useful for predicting indolent cancer (3). In a cohort of 1251 patients whose disease was low risk according to the D’Amico scoring system (4), Davies et al. (2) found that smaller prostate size was

559

MAZAHERI ET AL

Academic Radiology, Vol 22, No 5, May 2015

Figure 4. Bland–Altman plots show comparisons between (a) prostate magnetic resonance (MR) volumes measured using ellipsoid volume formula (EVF) and prostate MR volume using volumetry, (b) specimen prostate volume measured using EVF and specimen prostate volume measured using specimen mass (specific gravity formula [SGF]), (c) prostate MR volume measured using EVF and prostate specimen volume measured using SGF.

an independent predictor of upgrading. Larger prostate volumes are associated with greater loss of blood and perioperative complications at radical prostatectomy (27), whereas smaller volumes are associated with more advanced disease and worse prognosis (28). Although 3D volumetry is thought to be the most accurate method for measuring the prostate on MRI, it is generally considered too time consuming for routine clinical use. Formulas are intended to help radiologists measure prostate volumes efficiently, but the reproducibility of the resulting volume measurements may be limited by factors such as inaccuracies of the formulas themselves and variations in image interpretation and in the measurements of parameters incorporated into the formulas. The ideal prostate volume formula would be straightforward, requiring few parameters to be measured, and would yield accurate prostate volumes under a wide range of conditions. In this study, we compared prostate volumes measured from MRI using volumetry and EVF with prostate specimen volumes measured with the SGF (based on the mass). 560

We found high levels of concordance between these measurements, and importantly, PSA density risk classifications did not differ significantly between EVF- and volumetry-based PSA density levels, although rates of overclassification were high with both. Given that larger measured volumes lead to lower PSA density levels, it is possible that the high rates of overclassification were due to SGF-based pathologic volume estimates being larger than MRI-based volume measurements. In the EVF versus SGF plot (Fig 2), the values cluster well along the line of best fit. It appears that in volumes below approximately 100 cm3, the SGF provides a higher estimate than EVF. However, we see a shift in this through the line of best fit crossing the 45 line. According to the line of best fit equation, the lines cross at approximately 102 cm3. Few patients had a volume >100 cm3, so it is possible that this shift was driven by a few outliers. One patient had an EVF-derived volume of 181 cm3 and an SGF-derived volume of 151 cm3, and another had an EVF-derived volume of 189 cm3 with an SGF-derived volume of 172 cm3. Unsurprisingly, the volumetry versus SGF plot showed a pattern

Academic Radiology, Vol 22, No 5, May 2015

extremely similar to that of the EVF versus SGF plot; volumetry- and SGF-derived measurements generally hover around the line of best fit, and the line of best fit crosses the 45 line at approximately 93 cm3. For the two patients described previously as outliers, the volumetry-based measurements were 178 cm3 and 188 cm3, respectively. Several studies have looked into the accuracy of 3-T MRIbased measurements of prostate volume (19). Sosna et al. (19) compared in vivo prostate volumes measured from 3-T MRI acquired with an external torso phased-array imaging coil to ex vivo surgical specimen volumes. Six different data sets (which consisted of different measurements of the prostate dimensions) were considered; the correlation coefficient with the specimen volume varied from 0.325 to 0.751. The best estimate (correlation coefficient, 0.751) was obtained with the formula using the AP and SI dimensions from the sagittal images and the RL dimension from the axial images (correlation coefficient of 0.751). Our study was motivated by the concern that by deforming the prostate, the use of an ERC (rather than a phased-array coil (PAC), as in the study by Sosna et al.) to acquire MRI might render the EVF inaccurate for measuring prostate volume. Our EVF uses prostate dimensions measured along the planes typically used in the clinical setting for calculating volume (namely, RL and AP dimensions measured in the axial plane and SI dimensions measured in the sagittal plane). Sosna et al. argue that measuring the AP dimension from sagittal rather than axial images could result in a more precise estimate of prostate volume because the shape of the prostate is more oval or ellipsoid in the sagittal plane, whereas in the axial (and coronal) planes, it is more rounded. Although there is some evidence of this in their report on PAC-MRI, in our study, there was limited evidence that measurement in the sagittal plane (as opposed to the axial plane) results in improved prostate volume estimation with ERC-MRI. A number of investigators have focused on developing novel automatic or semiautomatic methods to automatically segment the prostate and measure its volume from MR images (29–31). Klein et al. (30) used atlas images to segment 3D images of the prostate. Pasquier et al. (29) applied a 3D, deformable-model approach and a seeded region-growing algorithm for automatic delineation of the prostate. Recently, Toth et al. (31) reported a multifeature active shape model (ASM) to measure prostate volume, which automatically determines the location of the prostate boundary on in vivo T2-weighted MRI. Separately, the same group investigated the segmentation of the prostate boundary on clinical images using ASM (31). These methods are promising, and when validated prospectively, they can provide useful time-saving tools in the management of prostate cancer. Our study had a number of limitations. First, we did not investigate interobserver variability of MRI measurements for prostate volume calculation, nor did we incorporate the level of observer experience in our analysis. Rather, all the measurements were performed in consensus by a radiologist and a medical student. There could be variations be-

COMPARISON OF VOLUME BY ERC MRI

tween measurements made by radiologists of different experience levels that should be examined. Second, our pathology measurements were limited to specimen dimensions and specimen mass. Measurement of water displacement could provide a more precise estimate of prostate size (32). To convert prostate mass to prostate volume, we used literature values for the mass of seminal vesicles, along with dissection of periprostatic fat, connective tissue, and residual bladder neck tissue, which are only approximate and can influence the accuracy of the findings. We also relied on the literature value for the specific gravity, which could not be independently verified. As noted graphically and numerically, SGF-based volume measurements were systematically higher than MRI volume measurements obtained with both EVF and volumetry methods. It is possible that our use of the textbook specific gravity and ancillary mass measurements created this bias. Third, volume calculation using axial images with either formula was prone to some error owing to the use of 3- to 4-mm acquisition increments (33). The use of isotropic 3D images would allow for more accurate measurements and improved characterization of prostate shape with and without the ERC. In conclusion, prostate volume measured from ERC MRI with the conventional prolate EVF showed a high degree of concordance with those measured with volumetry, although both differed significantly from prostate specimen volume. This suggests that use of the prolate EVF is appropriate for calculating prostate volume from 3-T ERC MRI.

ACKNOWLEDGMENTS The authors are grateful to Ada Muellner, M.S., for editing this article.

REFERENCES 1. Nickel JC. Benign prostatic hyperplasia: does prostate size matter? Reviews in urology 2003; 5(Suppl 4):S12–S17. 2. Davies JD, Aghazadeh MA, Phillips S, et al. Prostate size as a predictor of Gleason score upgrading in patients with low risk prostate cancer. The Journal of urology 2011; 186(6):2221–2227. 3. Bianco FJ, Jr, Mallah KN, Korets R, et al. Prostate volume measured preoperatively predicts for organ-confined disease in men with clinically localized prostate cancer. Urology 2007; 69(2):343–346. 4. D’Amico AV, Whittington R, Malkowicz SB, et al. A prostate gland volume of more than 75 cm3 predicts for a favorable outcome after radical prostatectomy for localized prostate cancer. Urology 1998; 52(4):631–636. 5. Radwan MH, Yan Y, Luly JR, et al. Prostate-specific antigen density predicts adverse pathology and increased risk of biochemical failure. Urology 2007; 69(6):1121–1127. 6. Ha YS, Yu J, Salmasi AH, et al. Prostate-specific antigen density toward a better cutoff to identify better candidates for active surveillance. Urology 2014; 84(2):365–371. 7. Epstein JI, Chan DW, Sokoll LJ, et al. Nonpalpable stage T1c prostate cancer: prediction of insignificant disease using free/total prostate specific antigen levels and needle biopsy findings. The Journal of urology 1998; 160(6 Pt 2):2407–2411. 8. Al-Qaisieh B, Brearley E, St Clair S, et al. A study of a pretreatment method to predict the number of I-125 seeds required for prostate brachytherapy. International journal of radiation oncology, biology, physics 2006; 65(1): 304–307.

561

MAZAHERI ET AL

9. Langley SE, Laing R. Prostate brachytherapy has come of age: a review of the technique and results. BJU international 2002; 89(3):241–249. 10. Tewari A, Indudhara R, Shinohara K, et al. Comparison of transrectal ultrasound prostatic volume estimation with magnetic resonance imaging volume estimation and surgical specimen weight in patients with benign prostatic hyperplasia. Journal of clinical ultrasound: JCU 1996; 24(4): 169–174. 11. Milosevic M, Voruganti S, Blend R, et al. Magnetic resonance imaging (MRI) for localization of the prostatic apex: comparison to computed tomography (CT) and urethrography. Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology 1998; 47(3):277–284. 12. Rasch C, Barillot I, Remeijer P, et al. Definition of the prostate in CT and MRI: a multi-observer study. International journal of radiation oncology, biology, physics 1999; 43(1):57–66. 13. Lee JS, Chung BH. Transrectal ultrasound versus magnetic resonance imaging in the estimation of prostate volume as compared with radical prostatectomy specimens. Urologia internationalis 2007; 78(4):323–327. 14. Jeong CW, Park HK, Hong SK, et al. Comparison of prostate volume measured by transrectal ultrasonography and MRI with the actual prostate volume measured after radical prostatectomy. Urologia internationalis 2008; 81(2):179–185. 15. Eri LM, Thomassen H, Brennhovd B, et al. Accuracy and repeatability of prostate volume measurements by transrectal ultrasound. Prostate cancer and prostatic diseases 2002; 5(4):273–278. 16. Kimura A, Kurooka Y, Kitamura T, et al. Biplane planimetry as a new method for prostatic volume calculation in transrectal ultrasonography. International journal of urology: official journal of the Japanese Urological Association 1997; 4(2):152–156. 17. MacMahon PJ, Kennedy AM, Murphy DT, et al. Modified prostate volume algorithm improves transrectal US volume estimation in men presenting for prostate brachytherapy. Radiology 2009; 250(1):273–280. 18. Hricak H, Jeffrey RB, Dooms GC, et al. Evaluation of prostate size: a comparison of ultrasound and magnetic resonance imaging. Urologic radiology 1987; 9(1):1–8. 19. Sosna J, Rofsky NM, Gaston SM, et al. Determinations of prostate volume at 3-Tesla using an external phased array coil: comparison to pathologic specimens. Academic radiology 2003; 10(8):846–853. 20. Bazinet M, Karakiewicz PI, Aprikian AG, et al. Reassessment of nonplanimetric transrectal ultrasound prostate volume estimates. Urology 1996; 47(6):857–862.

562

Academic Radiology, Vol 22, No 5, May 2015

21. Fine SW, Amin MB, Berney DM, et al. A contemporary update on pathology reporting for prostate cancer: biopsy and radical prostatectomy specimens. European urology 2012; 62(1):20–39. 22. Rodriguez E, Jr, Skarecky D, Narula N, et al. Prostate volume estimation using the ellipsoid formula consistently underestimates actual gland size. The Journal of urology 2008; 179(2):501–503. 23. Ohlsen H, Ekman P, Ringertz H. Assessment of prostatic size with computed tomography. Methodologic aspects. Acta radiologica: diagnosis 1982; 23(3A):219–223. 24. Varma M, Morgan JM. The weight of the prostate gland is an excellent surrogate for gland volume. Histopathology 2010; 57(1):55–58. 25. Carrasco JL, Jover L. Estimating the generalized concordance correlation coefficient through variance components. Biometrics 2003; 59(4): 849–858. 26. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989; 45(1):255–268. 27. Meraj S, Nagler HM, Homel P, et al. Radical prostatectomy: size of the prostate gland and its relationship with acute perioperative complications. The Canadian journal of urology 2003; 10(1):1743–1748. 28. Freedland SJ, Isaacs WB, Platz EA, et al. Prostate size and risk of highgrade, advanced prostate cancer and biochemical progression after radical prostatectomy: a search database study. Journal of clinical oncology: official journal of the American Society of Clinical Oncology 2005; 23(30):7546–7554. 29. Pasquier D, Lacornerie T, Vermandel M, et al. Automatic segmentation of pelvic structures from magnetic resonance images for prostate cancer radiotherapy. International journal of radiation oncology, biology, physics 2007; 68(2):592–600. 30. Klein S, van der Heide UA, Lips IM, et al. Automatic segmentation of the prostate in 3D MR images by atlas matching using localized mutual information. Medical physics 2008; 35(4):1407–1417. 31. Toth R, Bloch BN, Genega EM, et al. Accurate prostate volume estimation using multifeature active shape models on T2-weighted MRI. Academic radiology 2011; 18(6):745–754. 32. Park SB, Kim JK, Choi SH, et al. Prostate volume measurement by TRUS using heights obtained by transaxial and midsagittal scanning: comparison with specimen volume following radical prostatectomy. Korean journal of radiology: official journal of the Korean Radiological Society 2000; 1(2): 110–113. 33. Miyashita H, Watanabe H, Ohe H, et al. Transrectal ultrasonotomography of the canine prostate. The Prostate 1984; 5(4):453–457.