1615Vol. 11 | No. 03 | May - June2014 |U R O LO G Y J O U R N A L
Pre-Operative Tumor Localization and Evalu-
ation of Extra-Capsular Extension of Prostate
Cancer: How Misleading Can It Be?
Raymond Wai-man Kan, Chi-fai Kan, Lap-yin Ho, Steve Wai-hee Chan
Corresponding Author:
Raymond Wai-man Kan ,
Department of Surgery, Queen Eliza-
beth Hospital, Kowloon, Hong Kong.
Tel: +852 9814 9642
E-mail: kwm.raymond@yahoo.com
Received June 2013
Accepted March 2014
Department of Surgery, Queen
Elizabeth Hospital, Kowloon,
Hong Kong.
UROLOGICAL ONCOLOGY
Purpose: To verify the accuracy of transrectal ultrasound-guided prostatic biopsy (TRUS
Bx), magnetic resonance imaging (MRI) and their combination in evaluating the laterality of
prostate cancer and to determine the accuracy of MRI in assessing extra-capsular extension
of prostate cancer.
Materials and Methods: We retrospectively reviewed our past 100 consecutive series of radi-
cal prostatectomy performed between February 2010 and April 2012 at our institution. Their
TRUS Bx and MRI results were compared with the pathology of the radical prostatectomy
specimens. For tumor localization, we calculated the accuracies in unilateral diseases, bilateral
diseases, overall accuracies and Cohen Kappa concordance coefficient of TRUS Bx, MRI and
their combination. For the assessment of extra-capsular extension, we calculated the sensitiv-
ity, specificity, positive predictive value, negative predictive value, overall accuracy, likelihood
ratio positive and likelihood ratio negative of MRI.
Results: Eighty-two percent of our radical prostatectomy specimens had bilateral tumor in-
volvement and 32% had extra-capsular extension. The accuracies of TRUS Bx in unilateral
disease, bilateral disease and overall accuracy were 15.2%, 91.4% and 43.6%, respectively.
The accuracies of MRI in unilateral disease, bilateral disease and overall accuracy were
11.1%, 66.7% and 38.9%, respectively. When combining the assessment of TRUS Bx and
MRI, the accuracies in unilateral disease, bilateral disease and overall accuracy were 16.7%,
75% and 55.6%, respectively. The Cohen Kappa concordance co-efficient of TRUS Bx, MRI,
and combination of them were 0.1165, -0.2047 and -0.1084, respectively. The positive pre-
dictive value, negative predictive value, sensitivity, specificity, overall accuracy, likelihood
ratio positive and likelihood ratio negative of MRI in assessing extra-capsular extension were
33.3%, 69.8%, 5.9%, 94.9%, 67.9%, 1.16 and 0.99, respectively.
Conclusion: TRUS Bx, MRI, and their combination had poor concordance and limited ac-
curacies in assessment of the laterality of tumor involvement. The combination of TRUS Bx
and MRI offered a better of accuracy when compared to either modality alone. MRI was a
specific, but not sensitive tool in assessing the presence of extra-capsular extension.
Keywords: neoplasm invasiveness; prostatic neoplasms; image interpretation; neoplasm
staging; predictive value of tests; sensitivity and specificity; magnetic resonance imaging.
1616 |
INTRODUCTION
Since the adoption of prostate-specific antigen (PSA) testing, the incidence of prostate cancer had been gradually climbing up in the ranking of most preva-
lent cancers. In the United States, prostate cancer had be-
come the most common solid-organ cancer diagnosed in
2012.(1) In Hong Kong, prostate cancer ranked fifth with
1,492 new cases registered in 2010.(2) In addition, a larger
proportion of prostate cancer was diagnosed in their early
and low-risk stage.(3) Thus, the demand for definitive treat-
ment, such as radical prostatectomy, was unprecedented.
Continence and erectile function are two major concerns
affecting the post-operative quality of life. Nerve-sparing
technique was shown to offer early return of continence(4)
and preservation of erectile function,(5) hence nerve-sparing
prostatectomy had become the standard of care for organ-
confined prostate cancer.(6) Pre-operative risk assessment
and tumor localization are important for operative decisions.
Tools for pre-operative tumor localization, however, had ar-
guable accuracy despite technological advancement. Tran-
srectal ultrasound-guided 12-core prostatic biopsy (TRUS
Bx) was used to diagnose and to determine the laterality of
prostate cancer. Magnetic resonance imaging (MRI) of the
prostate, besides the laterality of tumor, also provides infor-
mation regarding extra-capsular extension of the tumor.(7)
Based on our experience, we conducted this review with the
primary objective to determine the accuracy of TRUS Bx,
MRI and their combination in evaluating the laterality of
prostate cancer. Our second objective was to determine the
reliability of MRI in assessing extra-capsular extension of
prostate cancer.
MATERIALS AND METHODS
We retrospectively reviewed our past 100 consecutive se-
ries of radical prostatectomy performed between February
2010 and April 2012 at our institution. Their TRUS Bx and
MRI results were compared with the final pathology of the
radical prostatectomy specimens. All radical prostatectomy
specimens were submitted for standardized slicing and pro-
cessing,(8) and were reported by experienced pathologists.
TRUS Bx with at least 12 cores, performed at our institution
and elsewhere with retrievable pathology reports were in-
cluded for analysis. Our institution adopted the use of Aloka
Prosound 6 ultrasound machine (Prosound Alpha 7, Aloka,
Tokyo, Japan) and Pajunk Delta Cut Biopsy System.
In order to minimize the confounding factor of using dif-
ferent MRI scanners and sequences, only those scans per-
formed at our institution were included. We used the Sie-
mens Magnetom Avanto 1.5-Tesla MRI system (Siemens
Magnetom Avanto, Erlangen, Germany). With a pelvis-
phased body array coil system, our scanner produced T1 and
T2 images with contrast phase, as well as diffusion weight
images. Our MRI scans were reported by at least one experi-
enced radiologist. The above specifications were in accord-
ance with the minimal requirement established by the Euro-
pean Consensus Meeting in 2009.(7) The key information we
retrieved from the MRI reports were laterality of the tumor,
and whether there was extra-capsular extension. Regarding
the evaluation of tumor laterality, we calculated the accu-
racy in unilateral disease, accuracy in bilateral disease, and
overall accuracy for TRUS Bx, MRI and their combination.
Bilateral disease in combination was defined as bilateral dis-
ease in either TRUS Bx or MRI, or when TRUS Bx and MRI
indicated unilateral disease of opposite sides. We calculated
the Cohen Kappa to indicate the concordance of TRUS Bx,
MRI and their combination with the final pathology.
Statistical Analysis
Statistical analysis was performed using the Statistical Pack-
Urological Oncology
Table 1. Tumor characteristics (n = 100).
Variables %
PSA level ng/mL
< 10 64
10-20 29
> 20 7
Gleason score
3 + 3 79
3 + 4 / 4 + 3 19
4 + 4 or above 2
Laterality on final pathology
Bilateral 82
Left only 11
Right only 7
Extra-capsular extension
Absent 68
Present 32
Key: PSA, prostate-specific antigen.
1617Vol. 11 | No. 03 | May - June2014 |U R O LO G Y J O U R N A L
Evaluation of Extra-Capsular Extension of Prostate Cancer | Kan et al
age for the Social Science (SPSS Inc, Chicago, Illinois,
USA) version 10.0. Regarding extra-capsular extension, we
computed the sensitivity, specificity, positive predicted val-
ue, negative predictive value, overall accuracy, likelihood
ratio positive and likelihood ratio negative of MRI in as-
sessing the presence of extra-capsular extension in the final
pathology.
RESULTS
Table 1 showed the tumor characteristics of our cohort of
100 patients who underwent radical prostatectomy. Sixty-
four patients had a pre-operative PSA level of less than 10
ng/mL, 29 patients had a level between 10 and 20 ng/mL,
while 7 patients had a level greater than 20 ng/mL. Regard-
ing Gleason score, 79 patients had a score of 3 + 3 in the
final pathology, 19 patients had a score of 7, while 2 patients
had a score of 8 or above. Regarding laterality of tumor in
the prostatectomy specimen, 82 patients had bilateral tumor
involvement, while 18 had unilateral disease. Among these
100 radical prostatectomy specimens, 32 had extra-capsular
extension.
Table 2 showed our results in pre-operative evaluation of
tumor laterality. There were 94 patients whose TRUS Bx re-
ports were available for analysis. Fifty six patients had their
MRI scans performed at our institution, among which 20
scans did not visualize the biopsy-proven prostate cancer.
Therefore we performed our analysis based on the remain-
ing 36 MRI scans. The accuracies of TRUS Bx in unilateral
disease, bilateral disease and overall accuracy were 15.2%,
91.4% and 43.6%, respectively. The accuracies of MRI in
unilateral disease, bilateral disease and overall accuracy
were 11.1%, 66.7% and 38.9% respectively. When combin-
ing the assessment of TRUS Bx and MRI, the accuracies
in unilateral disease, bilateral disease and overall accuracy
were 16.7%, 75% and 55.6%, respectively. The concordance
of TRUS Bx, MRI and their combination, as indicated by
their Cohen Kappa co-efficient, were 0.1165, -0.2047 and
-0.1084, respectively.
Table 3 demonstrated our analysis regarding the assessment
of extra-capsular extension. Among these 56 patients, 17
(30.4%) had extra-capsular extension on the prostatectomy
specimens. The positive predictive value, negative predic-
tive value, sensitivity, specificity, overall accuracy, likeli-
hood ratio positive and likelihood ratio negative of MRI
in assessing extra-capsular extension were 33.3%, 69.8%,
5.9%, 94.9%, 67.9%, 1.16 and 0.99, respectively.
DISCUSSION
TRUS Bx is one of the most important pre-operative inves-
tigations to determine the tumor laterality, however, our re-
sults fell short of satisfying. Due to the multifocal nature of
prostate cancer, 82% of all radical prostatectomy specimens
in our series were bilaterally involved. As these foci were
microscopically present, they easily succumbed to sampling
Table 2. Accuracy in prediction of tumor laterality.
Variables
Final Pathology Cohen Kappa (к) Data Analysis (%)
Right only Left only Bilateral Total
Accuracy of unilateral disease = 15.2
Accuracy of bilateral disease = 91.4
Overall accuracy = 43.6TRUS Bx (n = 94)
Right only 4 4 19 27
0.1165
95% CI = 0.0066 – 0.2265
Left only 1 5 26 32
Bilateral 2 1 32 35
Total 7 10 77 94
MRI (n = 56)
Visible (n = 36)
Not seen (n = 20)
Right only 0 0 8 8
-0.2047
95% CI = -0.4397 – 0.0302
Accuracy of unilateral disease = 11.1
Accuracy of bilateral disease = 66.7
Overall accuracy = 38.9
Left only 0 2 8 10
Bilateral 4 2 12 18
Total 4 4 28 36
Combination (n = 36)
Right only 0 0 5 5
-0.1084
95% CI = -0.3767 – 0.1600
Accuracy of unilateral disease = 16.7
Accuracy of bilateral disease = 75
Overall accuracy = 55.6
Left only 0 2 5 7
Bilateral 4 2 18 24
Total 4 4 28 36
Keys: CI, confidence interval; TRUS Bx, transrectal ultrasound-guided prostatic biopsy; MRI, magnetic resonance imaging.
1618 | Urological Oncology
error in TRUS Bx. As a result, most of the apparently uni-
lateral disease in TRUS Bx turned out to be bilaterally in-
volved in the prostatectomy specimens. This resulted in a
very disappointing accuracy in unilateral disease of 15.2%.
Notwithstanding, the accuracy in bilateral disease was a
reassuring 91.4%. The Cohen Kappa coefficient of TRUS
Bx was 0.1165, indicating only slight agreement between
TRUS Bx and final pathology. It was evident that TRUS Bx
had limited reliability in evaluating unilateral disease.
A handful of cases where TRUS Bx indicated unilateral dis-
ease turned out to be unilaterally involved on the opposite
side in the prostatectomy specimens. This situation was also
present in other similar studies.(9,10) This “unilateral vanish-
ing cancer syndrome” was another proof of the multifocal
nature of prostate cancer.
Advances in MRI had allowed a combination of modern
MRI sequences into a more informative multi-parametric
MRI scanning. To address the diversity in techniques and
image interpretation, the European Consensus Meeting had
established a set of guidelines regarding the multi-paramet-
ric MRI scanning.(7) The evidence regarding multi-paramet-
ric MRI scanning was conflicting. Although in general the
performance of multi-parametric MRI was reckoned prom-
ising,(11,12) there existed conflicting opinion regarding its ac-
curacy and usefulness.(13)
We routinely recommended MRI scanning for all patients
diagnosed with prostate cancer who opted to undergo radi-
cal prostatectomy. The scan was scheduled 8-12 weeks after
TRUS Bx. Among the 56 MRI scans performed at our insti-
tution, 20 scans could not visualize a biopsy-proven prostate
cancer. For the remaining 36 scans, the accuracy in unilat-
eral disease was 11.1% and the accuracy in bilateral disease
was 66.7%, both of which were worse than those of TRUS
Bx. This resulted in an overall accuracy of 38.9%. The Co-
hen Kappa coefficient of MRI was negative, which indicated
no agreement between MRI evaluation and final pathology.
Our results clearly showed that the use of our MRI scanning
sequences, which met the minimal requirement as suggested
by the European Consensus Meeting(7) was suboptimal. The
adoption of higher magnetic field power and endorectal coil,
as well as the addition of dynamic contrast enhancement and
spectroscopy could arguably increase the sensitivity and ac-
curacy in tumor localization.(11)
The complementary combination of TRUS Bx and MRI
was able to improve the accuracy in evaluating unilateral
disease, as well as the overall accuracy, when compared to
either modality alone. The overall accuracy of their combi-
nation reached 55.6%.
Regarding the assessment of extra-capsular extension, our
results showed that MRI was a highly specific, but not sensi-
tive tool. The specificity was as high as 94.9%, but the sen-
sitivity was unacceptably low at 5.9%. This might be partly
explained by the difference between macroscopic and mi-
croscopic extra-capsular extension. The other explanation
might be the inter-observer variability in deciphering the
MRI images.
The inter-observer variability, and the differences in imag-
ing criteria used for a positive MRI finding contributed to
the wide range of accuracy in MRI performance.(11) In an
attempt to standardize the interpretation and reporting of
MRI scanning, the European Society of Urogenital Radiol-
ogy had proposed the Prostate Imaging Reporting and Data
System (PI-RADS) scoring system.(14) The adoption of this
objective and structured reporting system, together with des-
ignating an experienced group of urogenital radiologists for
the interpretation, might aid to reduce the inter-observer var-
iability and to enable comparison among different patients.
As new technologies regarding MRI scanning was devel-
Table 3. Accuracy in prediction of extra-capsular extension.
Variables
Pathology PPV = 1/3 (33.3%)
NPV = 37/53 (69.8%)
Sensitivity = 1/17 (5.9%)
Specificity = 37/39 (94.9%)
Overall accuracy = 38/56 (67.9%)
LR (+) = 1.16
LR (-) = 0.99
ECE No ECE Total
MRI (n = 56)
ECE 1 2 3
No ECE 16 37 53
Total 17 39 56
Keys: PPV, positive predictive value; NPV, negative predictive value; LR (+), likelihood ratio positive; LR (-), likelihood ratio negative;
MRI, magnetic resonance imaging; ECE, extra-capsular extension.
1619Vol. 11 | No. 03 | May - June2014 |U R O LO G Y J O U R N A L
Evaluation of Extra-Capsular Extension of Prostate Cancer | Kan et al
REFERENCES
1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer
J Clin. 2012;62:10-29.
2. Hong Kong Cancer Registry, Hospital Authority, Hong Kong. Ac-
cessed on: www.ha.org.hk/cancereg.
3. Cooperberg MR, Broering JM, Kantoff PW, Carroll PR. Contemporary
trends in low risk prostate cancer: risk assessment and treatment. J
Urol. 2007;178:S14-9.
4. Srivastava A, Chopra S, Pham A, et al. Effect of a risk-stratified
grade of nerve-sparing technique on early return of continence
after robot-assisted laparoscopic radical prostatectomy. Eur Urol.
2013;63:438-44.
5. Meuleman EJ, Mulders PF. Erectile function after radical prostatec-
tomy: a review. Eur Urol. 2003;43:95-101.
6. Montorsi F, Wilson TG, Rosen RC, et al. Best practice in robot-assist-
ed radical prostatectomy: recommendations of the Pasadena Con-
sensus Panel. Eur Urol. 2012;62:368-381.
7. Dickinson L, Ahmed HU, Allen C, et al. Magnetic resonance imag-
ing for the detection, localisation, and characterisation of prostate
cancer: recommendations from a European Consensus Meeting.
Eur Urol. 2011;59:477-94.
8. Samaratunga H, Montironi R, True L, et al. International Society of
Urological Pathology (ISUP) consensus conference on handling
and staging of radical prostatectomy specimens. Working group 1:
specimen handling. Mod Pathol. 2011;24:6-15.
9. Frota R, Stein RJ, Turna B, et al. Are prostate needle biopsies predic-
tive of the laterality of significant cancer and positive surgical mar-
gins? BJU Int. 2009;104:1599-603.
oping inexorably, we believed the role of MRI in prostate
cancer could be potentially pivotal when making important
clinical decisions in the future.
CONCLUSION
TRUS Bx, MRI and their combination had poor concord-
ance and limited accuracies in evaluating the laterality of
tumor involvement. The combination of TRUS Bx and MRI
offered a better overall accuracy when compared to either
modality alone. MRI was a specific, but not sensitive tool
in assessing the presence of extra-capsular extension. When
planning for nerve-sparing radical prostatectomy, urologists
should recognize the limitations of each pre-operative in-
vestigation in terms of tumor localization and assessment of
extra-capsular extension.
CONFLICT OF INTEREST
None declared.
10. Jeong CW, Ku JH, Moon KC, et al. Can conventional magnetic reso-
nance imaging, prostate needle biopsy, or their combination pre-
dict the laterality of clinically localized prostate cancer? Urology.
2012;79:1322-7.
11. Kirkham AP, Emberton M, Allen C. How good is MRI at detecting and
characterizing cancer within the prostate? Eur Urol. 2006;50:1163-74.
12. Puech P, Huglo D, Petyt G, Lemaitre L, Villers A. Imaging of organ-
confirmed prostate cancer: functional ultrasound, MRI and PET/
computed tomography. Curr Opin Urol. 2009;19:168-76.
13. Kelloff GJ, Choyke P, Coffey DS, Prostate Cancer Imaging Working
Group. Challenges in clinical prostate cancer: role of imaging. Am J
Roentgenol. 2009;192:1455-70.
14. Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR
guidelines 2012. Eur Radiol. 2012;22:746-57.