Table 3: Summary of articles measuring the PGV by MRI in chronological order.

First author,
Year of publication,
Country
Number of patients,
Age
MRI Imaging details,
Median volume
Reference method,
Median volume
Reference detailsCorrelation
data
Concordance
Data and over/under estimation
Other
comments
Scores for Bias (0-2) and
Quality (0-4)

Sosna [30] 2003
USA
11 pts, Mean 59 yrsEC, PC, and ex vivo PC
Mean volumes 26-31 EC, 37 PC, 34 mL ex vivo,3T MRI,
no ERC
Fresh specimen weighed, SGF applied, Mean 40mLSV removedLinear regression
R=0.32-0.75 for EC, 0.65 for PC in vivo, 0.86 for PC ex vivo
6 combinations of various axes used for EC, best was sagittal for AP and SI, axial for RLB2
Q2

Lee [20]
2006
Korea
73 pts, age NSEC, Mean 38cc
3T or ERC NS
Fresh weight within 1 hr, SGF applied, Mean 37ccSV removedPearson R=0.96
P< 0.001
Overestimated if < 35cc, underestimated if >35ccAlso tested TRUS, but MRI more accurateB2
Q4

Jeong [22] 2008 Korea21 pts, Mean 66 yrsEC and PC, Means 41-51 mL,
ERC used,
3T NS
displacement method, Mean 40mLFresh specimen within 1 hr, SV removedLinear regression, R=0.84-92Students t-test
P=0.03-0.70
PC most accurateB2
Q3

Kwon [31] 2010 Korea579, Mean 64 yrsEC,
Mean 32 mL, 1.5T MRI,
ERC NS
Fresh weight, Mean NSSV removedPearson R=0.69
P< 0.001
NSB2
Q2

Bulman [32]
2012
USA
91 pts, Mean 59 yrsmpMRI EC, PC (manual and MFA).
ERC, 3T, Mean 41-45 mL
Freshly weighed, Mean 50 mLAverage weight of SVs subtractedWilcoxon signed rank test and linear regression 0.78-0.90Bland-Altman plots, 92-97% within limits of agreement. All of the MRI methods underestimated the volume by around 15%Multiple readers used, MFA similar in accuracy to manual planimetry, both more accurate than ECB2
Q4

Turkbey [33]
2012
USA
98 pts, Median 61 yrsEC, PC and automated PC, Means 29-48 cc
3T, ERC
Fresh specimen weight, Mean 52 ccIncluded SVsPearson r=0.86-0.91
P<0.0001
Partial and full Dice similarity coefficient 0.85-0.92Autosegmentation faster than manual PCB2
Q3

Karademir [34] 2013,
USA
61 pts, Median 64 yrsAutomated volume calculation,
Mean 46 cc,
1.5T mpMRI, ERC
Weight from pathology reports, mean 50ccStandard SV weight subtractedPearson r=0.94
P<0.0001
Underestimated by 10% on averageB2
Q2

Hong [35],
2014,
USA
1756 pts, Median 59 yrsEC,
Median 31mL
3T NS,
ERC NS
Weight from pathology report,
Mean NS
Details NSPearson R=0.82,
p< 0.0001
NSHigher grade cancer associated with smaller volumeB2
Q1

Le Nobin [36], 2014,
USA
37 pts, Mean 60 yrsPC
Mean 47 mL
3T mpMRI,
ERC NS
Post fixation,
Mean 47 mL
Details NSNSBland Altman 95% limits -7 to +8 mLB2
Q2

Chernyak [37], 2015,
USA
49 pts, Mean 59 yrsEC, either 1.5T or 3T
With and without ERC, Means 46 and 51 cc
Weight retrospectively collected from report, Mean 55 gmIncluded SVICC improved with ERC 0.90-0.96, mainly due to AP measurementMRI underestimated weight, more so with ERC (9 vs 4 gm)B1
Q2

Mazaheri [38], 2015,
USA
195 pts, Median 62 yrsEC and PC, Median 42 cc for both,
3T MRI, ERC
Fresh weight from pathology report, Median 52 ccStandard SV weight subtracted, applied SGFLin CCC used to assess correlation and concordanceLin CCC = 0.85 (EC) and 0.87 (PC), both underestimated by approx. 10mLB2
Q3

Paterson [29], 2016,
Canada
318 pts
Mean 63 yrs
EC
Mean 39cc
Fluid displacement method. Mean vol 37ccAlso prostate weight
(ICC=0.96)
ICC 0.83Overestimation more common when a median lobe was presentAlso used TRUS, MRI slightly more accurate,B1
Q3

Bezinque [39], 2018,
USA
99 pts, Median 63 yrsVarious EC and PC methods, Medians 35 to 49,
3T mpMRI,
No ERC
Specimen wt and volume, Medians 37-54 mLDetails NSICC 0.66-0.73NSMRI with segmentation was considered the referenceB1
Q2

Pts: patients, Yrs: years of age, TRUS: transrectal ultrasound, EC: ellipsoid calculation, PC: planimetric calculation, NS: not stated, SVs: seminal vesicles, TZ: transitional zone, mpMRI: multiparametric magnetic resonance imaging, MFA: multifeature active shape model, ERC: endorectal coil, 3T: 3-tesla, AP: anteroposterior, ICC: intraclass correlation coefficient, SGF: specific gravity factor (1.05 g/mL), CC: craniocaudal, SP: specific gravity, SEARCH: shared equal access regional cancer hospital, and Lin CCC: Lin’s concordance correlation coefficient.