Center for OPTical IMagery Analysis and Learning (OPTIMAL), State Key Laboratory of Transient Optics and Photonics, Xi'an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, Shaanxi, China
Peter L. Choyke, Molecular Imaging Program, NCI, NIH, 10 Center Dr, MSC 1182 Bldg 10, Room B69, Bethesda, MD 20892-1088, USA. e-mail: firstname.lastname@example.org
Study Type – Diagnosis (case series)
Level of Evidence 4
What's known on the subject? and What does the study add?
Benign prostatic hyperplasia is the most common symptomatic disorder of the prostate and its severity varies greatly in the population. Various methods have been used to estimate prostate volumes in the past including the digital rectal examination and ultrasound measurements.
High-resolution T2 weighted MRI can provide accurate measurements of zonal volumes and total volumes, which can be used to better understand the etiology of lower urinary tract symptoms of men.
• To use ability of magnetic resonance imaging (MRI) to investigate age-related changes in zonal prostate volumes.
PATIENTS AND METHODS
• This Institutional Review Board approved, Health Insurance Portability and Accountability Act-compliant study consisted of 503 patients who underwent 3 T prostate MRI before any treatment for prostate cancer.
• Whole prostate (WP) and central gland (CG) volumes were manually contoured on T2-weighted MRI using a semi-automated segmentation tool. WP, CG, peripheral zone (PZ) volumes were measured for each patient.
• WP, CG, PZ volumes were correlated with age, serum prostate-specific antigen (PSA) level, International Prostate Symptom Score (IPSS), Sexual Health Inventory for Men (SHIM) scores.
• Linear regression analysis showed positive correlations between WP, CG volumes and patient age (P < 0.001); there was no correlation between age and PZ volume (P= 0.173).
• There was a positive correlation between WP, CG volumes and serum PSA level (P < 0.001), as well as between PZ volume and serum PSA level (P= 0.002).
• At logistic regression analysis, IPSS positively correlated with WP, CG volumes (P < 0.001).
• SHIM positively correlated with WP (P= 0.015) and CG (P= 0.023) volumes.
• As expected, the IPSS of patients with prostate volumes (WP, CG) in first decile for age were significantly lower than those in tenth decile.
• Prostate MRI is able to document age-related changes in prostate zonal volumes.
• Changes in WP and CG volumes correlated inversely with changes in lower urinary tract symptoms.
• These findings suggest a role for MRI in measuring accurate prostate zonal volumes; have interesting implications for study of age-related changes in the prostate.
Unlike many other organs that atrophy with age, as men age, the volume of their prostate usually increases mostly due to BPH [1–6]. However, the rate at which BPH develops is highly variable. Methods of measuring prostate volume include estimates from the DRE and TRUS measurements that assume that the prostate is a three-dimensional (3D) ellipsoid. Both methods are acknowledged to be inaccurate and neither method can differentiate central gland (CG) volume, the site of most BPH, from peripheral zone (PZ) volume. Prostate MRI, using high resolution T2-weighted (T2W) MRI readily discriminates between the CG and the PZ and therefore, it is straightforward to determine both different zonal and overall prostate volumes. However, few studies have attempted to correlate MRI-related zonal prostate volumes with age, voiding symptoms or sexual health. A large cross-sectional study of men at different ages could result in a better understanding of the natural history of BPH, which could be used to design preventive and treatment strategies. In the present study, we examine changes in prostate zonal volumes with age and its association with clinical symptoms by using volumes obtained by prostate MRI in a large cohort of men.
PATIENTS AND METHODS
This Institutional Review Board-approved, Health Insurance Portability and Accountability Act (HIPAA)-compliant retrospective study was conducted in one institution. It included a cross-sectional cohort of 503 patients who were referred for prostate MRI due to rising PSA levels without any prior cancer treatment between November 2005 and October 2010. The mean (median, range) age of the participants was 60.5 (60, 38–83) years, and the mean (median, range) serum PSA level was 8.2 (5.5, 0.2–103) ng/mL.
MRI DATA ACQUISITION
All MRI studies were performed using a combination of an endorectal coil (BPX-15 or BPX-30, Medrad, Pittsburgh, PA, USA) tuned to 127.8 MHz and a cardiac coil (6 or 16-channel) (SENSE, Philips Medical Systems, Best, The Netherlands) on a 3 T magnet (Achieva, Philips Medical Systems, Best, the Netherlands) without prior bowel preparation. To use the 1.5 T endorectal coil (BPX-15) at 3 T, it was tuned to 128.78 MHz using a Pi matching network and interfaced to the scanner through a research coil interface box (Philips Healthcare, Cleveland, OH, USA). The MRI protocol included triplanar (sagittal, axial, and coronal) T2W turbo spin-echo, diffusion-weighted MRI, multi-voxel 3D MR Spectroscopy Imaging, axial pre-contrast T1W, axial 3D fast field echo dynamic contrast-enhanced MRI sequences; however for determining prostate volumes only T2W-MRI images were used. The scanning parameters for T2W turbo spin-echo images were as follows: scan resolution 0.461 × 0.598 × 3.0 mm3; field of view, 140 × 140 mm2; repetition time/echo time, 8869/120 ms; flip angle, 90 °; slice thickness, 3 mm without gaps; image reconstruction, 512 × 512.
PROSTATE VOLUME MEASUREMENT
The whole prostate (WP) gland and CG were manually contoured on axial T2W-MRI images using a semi-automated prostate segmentation tool. The semi-automatic segmentation software was developed using C++ programming language based on open source toolkits including Insight Toolkit 3.12 , Visualization Toolkit 5.4.2 , and Fast Light Toolkit 1.1.9 . The algorithm for automatic initialisation and segmentation of the mid-gland slices shares the same framework as the method presented in . The method requires training of an incorporated shape model using a set of prostate contours manually delineated on MRI prostate images obtained using the same imaging protocol as the images to be segmented. The segmented shape is stored as a mesh for output and visualisation. The software allowed calculation of both WP and CG volumes and the volume of the PZ was calculated by subtracting the CG volume from WP volume (Fig. 1).
To validate accuracy of MRI-derived volumes, WP volumes of 47 patients (whom underwent radical prostatectomy) derived with MRI were compared with their specimen weights (Supplementary Data S1).
The IPSS and Sexual Health Inventory for Men (SHIM) scores of the patients were obtained within 1 week of MRI. Scores were correlated with the WP, CG and PZ volumes.
The relationship between the WP, CG, PZ volumes, PZ/CG volume ratio (PZ : CG) and patient age, serum PSA levels were evaluated by linear regression analysis, whereas the correlation between zonal prostate volumes and IPSS, SHIM scores were assessed with logistic regression. Additionally, to evaluate the effects of tumours on prostatic contours, patients were divided into two groups, where group 1 included patients with tumour(s) that did not affect the prostate's outer or zonal contours, while group 2 included patients with tumour(s) affecting either overall or prostate zonal contours. WP, CG, PZ volumes, PZ : CG volume ratio, age, serum PSA levels, IPSS, SHIM scores of these two groups were compared using the t-test
To predict the WP, CG, PZ volumes and PZ : CG volume ratio, a multivariate analysis was conducted using a standard least squares regression model including age, serum PSA levels, IPSS, SHIM scores. Additionally, to predict the IPSS and SHIM scores, a multivariate analysis was performed using age, serum PSA levels, WP, CG volumes and PZ : CG volume ratio.
Finally, a separate analysis was conducted to compare the mean IPSS and SHIM scores of patients in the first and tenth age-specific decile for WP, CG and PZ volumes. The t-test was used for this analysis.
Initial comparison of MRI-derived WP volumes with prostatectomy specimen weights revealed a strong positive correlation (r2=0.863; Supplement Data S1).
In the present study population, the mean (median, range) age was 60.5 (60, 38–83) years. The mean (median, range) serum PSA level was 8.2 (5.5, 0.2–103) ng/mL. The mean (median, range) WP, CG and PZ volumes calculated on MRI were 47.5 (39.6, 13–169.8) mL, 27.6 (19.4, 3.7–159.3) mL, 19.8 (19.8, 6.3–49.7) mL, respectively; however, results varied widely according to age (Fig. 2).
At linear regression analysis, there was a positive correlation between the WP (r2=0.132) and CG (r2=0.161) volume and patient age (P < 0.001). There was a negative correlation between the PZ : CG ratio and patient age (r2=0.148; P < 0.001), indicating that the PZ remains relatively static in volume with age. There was no correlation between age and the PZ volume (P= 0.173); whereas the CG increased with age (Fig. 2). There was a positive correlation between the WP (r2=0.66) and CG (r2=0.52) volume and serum PSA level (P < 0.001), as well as between the PZ volume and serum PSA level (r2=0.19; P= 0.002); whereas there was a negative correlation between the PZ : CG ratio and serum PSA level (r2=0.30; P < 0.001; Fig. 3). Additionally, there was a weak correlation between serum PSA level and age (r2=0.08; P= 0.048).
For logistic regression model analysis, the IPSS of patients were divided into three groups (group 1 [mild], 0–7; group 2 [moderate], 8–19; group 3 [severe], ≥20). IPSS severity positively correlated with the WP and CG volumes (P < 0.001); whereas it negatively correlated with the PZ : CG ratio (P < 0.001). Although no threshold WP or CG volume could be assessed to predict severity of micturition symptoms, severe symptoms were not encountered in patients with WP and CG volumes of <20 mL and <8 mL, respectively (Supplement Data S2).
SHIM scores of patients were divided into five groups: group 1, 22–25; group 2, 17–21; group 3, 12–16; group 4, 8–11; group 5, 0–7. SHIM score severity positively correlated with the WP (P= 0.015) and CG (P= 0.023) volumes; whereas it negatively correlated with the PZ : CG ratio (P < 0.001). There was no correlation of PZ volume with IPSS and SHIM scores.
When comparing patients who had contour-deforming tumours with patients without contour-deforming tumours, the mean WP volume was, interestingly, significantly larger in the latter group (49.2 vs 35.3 mL, P < 0.001). In other words, smaller prostate glands were associated with more pronounced growth of cancers, perhaps because such growth was more apparent. Consistent with this finding, the mean CG volume was also larger in the unaffected contour group (29.2 vs 16.1 mL, P < 0.001) and the mean IPSS score was also higher (9.5 vs 6.6, P= 0.021); on the other hand PZ volume, age, serum PSA level and SHIM scores were not different.
Multivariate analysis to predict the WP, CG, PZ volumes and PZ : CG volume ratio by using patient age, serum PSA levels, IPSS and SHIM scores showed that patient age, serum PSA level and IPSS were significant predictors (Table 1). Multivariate analysis to predict IPSS, SHIM scores by using age, serum PSA level, WP, CG, PZ volumes and PZ : CG volume ratio showed that the most predictive factors were patient age and CG volume (Table 2).
Table 1. Result of multivariate analysis to predict WP, CG, PZ volumes and PZ : CG volume ratio by using patient age, serum PSA, IPSS, SHIM scores. Patient age, serum PSA and IPSS score were found as significant predictors of WP, CG volumes and PZ : CG volume ratio
PZ : CG ratio
Table 2. Result of multivariate analysis to predict IPSS, SHIM scores by using age, serum PSA, WP, CG, and PZ : CG volume ratio. Patient age was found to predict both IPSS and SHIM scores, whereas CG volume was found to predict IPSS score
Interaction (WP vs CG)
Comparison of the IPSS and SHIM scores with first and tenth decile age-specific prostate zonal volumes was performed for WP and CG volumes, but not for the PZ volume, as no correlation was present at the logistic regression analysis. The IPSS of patients with WP and CG volumes in the age-specific first decile (adjusted for age) were significantly lower than those of patients in the tenth decile (Figs 4–6).
The present results indicate that high resolution T2W-MRI can be used to determine whole gland and prostate zonal anatomy, which in turn, can be used to study the effects of prostate zonal volumes on clinical symptoms. Not surprisingly, the CG and WP volumes increased with age and serum PSA level; however, the PZ volume had no correlation with age and only a weak correlation with PSA level. These findings are consistent with the concept that the CG is the major determinant factor in BPH and elevated serum PSA levels [11,12]. Although few previous studies reported longitudinal changes of prostate volume with age in small to medium sized cohorts by using serial MRI measurements (e.g. Williams et al.  in 64 patients, Loeb et al.  in 278 patients), to the best of our knowledge, the present study is the largest aiming to correlate baseline MRI-derived zonal prostate volumes with age, serum PSA level and symptoms. The present data suggests that the WP volume peaks in the sixth and seventh decades of life and is mainly driven by changes in the CG volume, while the PZ volume remains stable or slightly decreases during aging. Rhodes et al.  reported a similar increase in prostate volume with advancing age in a cohort of 631 men, who underwent prostate volume determinations using TRUS; however, in that study the relative contribution of the CG and PZ could not be discerned.
Serum PSA levels increase proportionately with the volume of BPH tissue, which is characterised by variable stromal and epithelial hyperplasia . In the present study, there was a positive correlation between serum PSA levels and the WP and CG volumes; whereas there was no correlation between PSA level and the PZ volume. The present results are similar to those in previous studies reporting a positive correlation between serum PSA levels and prostate volume [2,17]; however, correlation in the present trial could be skewed by the presence of variable amounts of cancer in some patients [1,18].
An important consequence of BPH is its effect on LUTS and sexual health. Logistic regression showed a positive correlation between IPSS and the WP and CG volumes. Similarly, the SHIM score had a positive correlation with WP and CG volumes, but not with the PZ volume. Using a multivariable analysis model (which included patient age, serum PSA level, WP, CG, PZ volumes, and PZ : CG volume ratio) revealed that patient age and CG volume were significant predictors for IPSS, whereas patient age was the only predictor for SHIM score. Kaplan et al.  reported that CG volume may serve as a useful parameter for evaluating patients with worsening urinary outlet obstruction in a cohort of 61 men, who had their prostate volumes measured using ultrasonography. Girman et al.  reported a modest positive correlation between prostate volume and IPSS in 2115 men, who had TRUS-measured volumes. However, this was contradicted by Sauver et al.  who reported only a weak correlation between the CG volume and urinary symptoms in a cohort of 336 patients. The present overall findings are similar, showing a modest correlation between prostate volumes and IPSS. Interestingly, a more detailed analysis of patients in the first decile of volume for age compared with patients in the tenth decile of volume for age, revealed that IPSS scores from the first decile of WP and CG volumes were significantly lower than those of patients in the tenth decile (WP, 7.7 (first) vs 13.5 (tenth); CG, 6.6 (first) vs 13 (tenth)).
The present study has several limitations. First, many of the patients in the present study had prostate cancer. Indeed, it is difficult to find a cohort of men in this age cohort in whom prostate cancer can reliably be excluded. As a result it is likely that the PSA measurements are somewhat higher than a population consisting only of patients with BPH. Another limitation is that we were unable to follow patients longitudinally and had to rely on a cross-sectional design. Although we did not perform a longitudinal study, the present findings are similar to those of Williams et al.  who also reported a peak increase in volume between the mid-fifth and mid-sixth decades. Longitudinal data may be a by-product of the growing number of patients who have low-risk prostate cancer, and who are undergoing active surveillance with annual MRI and semi-annual biopsy. This longitudinal data will become available in the next few years. Finally, we did not include quantitative laboratory values, e.g. urine flow studies (for IPSS) or serum testosterone (for SHIM score), which may have provided insight into the correlations we observed.
In conclusion, high-resolution T2W-MRI can provide accurate measurements of prostate zonal volumes and total volumes that correlate with prostate weight. Because of its ability to differentiate the central zone and PZ of the prostate, MRI can be used to identify the main driver of prostate volume, BPH. Changes in WP and CG volumes correlated with LUTS; whereas changes in the PZ volumes correlated with changes in sexual health. Overall prostate and prostate zonal volumes can be used to better understand the aetiology of LUTS in men. Longitudinal studies with MRI, which will be a by-product of the increasing number of men followed during active surveillance, will reveal the kinetics of prostate volume changes with age.