OEpsMppdR1�dvsarCP1te

Ttsaspp

FnoP

Tm

5

ADULT UROLOGY

5

INFLUENCE OF TRANSRECTAL ULTRASOUND PROBE ONPROSTATE CANCER DETECTION IN TRANSRECTAL

ULTRASOUND-GUIDED SEXTANT BIOPSY OF PROSTATE

ROGER PAUL, CHRISTIAN KORZINEK, ULRIKE NECKNIG, THOMAS NIESEL,MICHAEL ALSCHIBAJA, HERBERT LEYH, AND RUDOLF HARTUNG

ABSTRACTbjectives. To determine whether end-fire probes increase the prostate cancer (PCa) detection rate.nhancing the PCa detection rate is the main goal of biopsy protocols. Prostate biopsy is limited by side-firerobes to a longitudinal axis, but end-fire probes allow biopsy cores to also be taken in the transverseection.ethods. A total of 2625 patients underwent systematic sextant biopsy in three institutions using the same

rotocol. Three different ultrasound probes were used—the Kretz Combisone and Bruel & Kjaer side-firerobes and the ATL HDI end-fire probe. We retrospectively evaluated the influence of the probe on the PCaetection rate.esults. The Kretz probe was used in 384 men, the Bruel & Kjaer probe in 598 men, and the ATL probe in643 men. Overall, 35.2% had PCa detected. Analyzing all patients, no statistically significant difference (P0.73) was found for the probes, but the subgroup with a prostate-specific antigen level of 4 to 10 ng/mL

emonstrated a statistically significant improvement in the detection rate using the end-fire probe (31.3%ersus 24.5% and 21.5% for the side-fire probes, P � 0.01). Patients with nonpalpable PCa also demon-trated a statistically significant increase in detection with the end-fire probe (P � 0.004). Multivariatenalysis confirmed that the ultrasound probe is an independent parameter to enhance the PCa detectionate.onclusions. Our results showed that end-fire probes provide a statistically significant improvement in theCa detection rate compared with side-fire probes in patients with a prostate-specific antigen level of 4 to0 ng/mL and nonpalpable disease. The reason could be the facilitated sampling in the most lateral part ofhe peripheral zone. Our results suggest that the widespread use of end-fire probes for prostate biopsy couldnhance the PCa detection rate. UROLOGY 64: 532–536, 2004. © 2004 Elsevier Inc.

Hbeap

Ddfmbt

hsg

ransrectal ultrasound (TRUS)-guided biopsyhas become the standard procedure to confirm

he diagnosis of prostate cancer (PCa).1 Systematicextant biopsy of the prostate has been a widelyccepted procedure since the first reports demon-trated that the detection rate was improved com-ared with lesion-directed biopsy.2 The perfectrostate biopsy would offer a sensitivity of 100%.

rom the Department of Urology, Technische Universitat Mu-ich, Klinikum rechts der Isar, Munich, Germany; Departmentf Urology, Klinikum Garmisch-Partenkirchen, Garmisch-artenkirchen, Germany; and Urology, Pfaffenhofen, GermanyReprint requests: Roger Paul, M.D., Department of Urology,

echnische Universitat Munich, Klinikum Rechts der Isar, Is-aninger Strasse 22, Munich 81675, GermanySubmitted: February 19, 2004, accepted (with revisions): April

f, 2004

© 2004 ELSEVIER INC.32 ALL RIGHTS RESERVED

owever, the sensitivity of the systematic sextantiopsy is only about 60%.3 Newer strategies havemphasized an increase in the core numbers4,5

nd/or biopsy sampling of the lateral aspect of theeripheral zone.6,7

A different approach is the use of color flowoppler sonography8 or echographic contrast me-ia.9 Computer analysis of gray-scale informationrom TRUS by artificial neuronal networks, which

ay lead to an enhanced detection of PCa, has alsoeen used.10 These new strategies, however, areime consuming and expensive.The influence of the ultrasound probe, however,

as not been studied in detail. Earlier reportshowed that the ultrasound frequency only mar-inally affected the PCa detection rate.11 Two dif-

erent transrectal ultrasound probes are currently

0090-4295/04/$30.00doi:10.1016/j.urology.2004.04.005

aspeov

et

tdtpooczttnlnp

C&t9t

assbspbstbattss

uuu

ufKcS

P

radntapnAfiTbnpoaT5avs

P

oit032

FsvtpdtW(

U

vailable—the side-fire and end-fire probes. Inide-fire probes with a longitudinal transducer, therostate biopsy is limited to a sagittal axis, but withnd-fire probes with a curved array probe, the bi-psy can be performed in the sagittal and trans-erse sections.The aim of our study was to evaluate the differ-

nce in the PCa detection rate using these twoypes of ultrasound probes.

MATERIAL AND METHODS

Between 1990 and 2003, 2625 patients underwent a first-ime, TRUS-guided sextant prostate biopsy at one of threeifferent institutions (university hospital, community hospi-al, and private practice) using the same biopsy protocol. Allatients received local anesthesia (lidocaine gel) and antibi-tic prophylaxis (gyrase inhibitor). A systematic sextant bi-psy was performed. The cores were sampled at the base,enter, and apex of the prostate on both sides of the peripheralone. The cores were also sampled from the lateral aspect ofhe peripheral zone as determined on the TRUS image. Addi-ional biopsies from suspicious areas on digital rectal exami-ation (DRE) and/or TRUS outside the sextant area were al-

owed. Sampling was performed with an 18-gauge, single-useeedle. The biopsy strategy remained unchanged during theeriod of this analysis.Three different ultrasound probes were used: the Kretz

ombisone 320 S side-fire probe (Kretz, 7.5 MHz), the BruelKjaer Medical side-fire probe (B&K, type 8551 and 8808, 6

o 10 MHz), and the ATL HDI 1000 end-fire probe (ATL, 5 toMHz). The differences in the ultrasound probes are illus-

IGURE 1. Ultrasound sections of prostate. With use ofide-fire probe in (A) transverse section, biopsy needleisualized as spot only and sampling in transverse sec-ion is not possible; in (B) longitudinal section of lateraleripheral zone, biopsy sampling is possible, but bor-ers of peripheral zone are difficult to define; (C) longi-udinal section of midline of prostate and biopsy line.ith use of end-fire probe biopsy sampling is possible in

D) transverse and (E) longitudinal sections.

rated in Figure 1. The curved endorectal array (Kretz probe) t

ROLOGY 64 (3), 2004

llows sagittal imaging of the prostate by sending the ultra-ound waves from a longitudinal transducer in the probe. Aecond transverse section is possible by a rotating transducer,ut biopsy sampling is restricted to the sagittal section. Theecond endorectal biplanar curved array (B&K probe) is com-arable to the Kretz system. The only difference is the possi-ility of continuously changing the angle of scanning from theagittal to the transverse section by turning the transducer inhe probe mechanically. Again, biopsy sampling is only possi-le in the sagittal section. The endorectal multiplanar curvedrray (ATL probe) sends the ultrasound waves from the tip ofhe probe; therefore, turning and rotating the probe allows oneo change the scanning angle liberally. Biopsy sampling is pos-ible in any possible section of the prostate, especially in theagittal and transverse sections (Fig. 1).

The Kretz-Probe was used in 384 patients by a total of 5rologists. The B&K probe was used in 598 patients by 8rologists, and the ATL probe was used in 1643 men by 18rologists.Statistical analysis of this retrospective study was performed

sing Statview, version 4.0, software. To assess the differencesor the different probes, we used the Mann-Whitney U test andruskal-Wallis test for continuous variables and the Pearsonhi-square test or Fisher’s exact test for nominal variables.tatistical significance was accepted at 5% (P �0.05).

RESULTS

ATIENT COHORT

A total of 2625 patients were included in thisetrospective analysis (Table I). The mean patientge was 66.2 years (range 40.2 to 93.1). The me-ian prostate-specific antigen (PSA) level was 7.5g/mL for all patients. For the Kretz probe group,he median PSA level was 8.3 ng/mL; for the B&Knd ATL groups, it was 7.4 ng/mL. Of the 2625atients, 52.0% had a PSA value between 4 and 10g/mL (Kretz 35.7%, B&K 58.1%, ATL 53.6%).lso, 52.2% of all patients had suspicious DREndings (Kretz 63.7%, B&K 37.3%, ATL 47.6%).he differences in the median PSA and DRE distri-ution with increased PSA levels and an increasedumber of suspicious DRE results for the Kretzrobe can be explained by different times of bi-psy. The Kretz probe was used from 1990 to 1994nd the B&K and ATL probes from 1994 to 2002.he prostate size was 62.2 cm3, on average (Kretz8.1 cm3, B&K 57.2 cm3, ATL 66.1 cm3). When wenalyzed only the subgroup of patients with a PSAalue between 4 and 10 ng/mL, the findings wereimilar (Table I).

Ca DETECTION RATE

Overall, the PCa detection rate of first-time bi-psy was 35.2% (n � 920). No statistically signif-cant difference was found among all patients inhe detection rate for the different probes (P �.73). Overall, the PCa detection rate (Fig. 2) was4.6% (n � 134) for the Kretz probe, 34.1% (n �02) for the B&K probe, and 35.8% (n � 584) for

he ATL probe.

533

vsl

nlhd2p

cccwsPa1tnPPa1aswp

stPapcwu

Fd(

5

When patients were stratified according to PSAalues (Table II), we could not detect a statisticallyignificant difference in patients with a PSA level

TABLE I. Patie

ParameterKretz Com

320

Mean age (yr)All patients 66.5PSA subgroup* 65.7

Median PSA (ng/mL)All patients 8.3PSA subgroup* 6.7

PSA (%)�4 ng/mL 22.84–10 ng/mL 35.7�10 ng/mL 41.5

DRE suspicious (%)All patients 63.7PSA subgroup* 51.1

Mean prostate size (cm3)All patients 58.1PSA subgroup* 62.6

KEY: PSA � prostate-specific antigen; DRE � digital recta* Subgroup with PSA level of 4–10 ng/mL.

IGURE 2. PCa detection rate for first-time biopsy byifferent ultrasound probes for (A) overall patients andB) patients with PSA level between 4 and 10 ng/mL.

ess than 4 ng/mL (P � 0.48) and greater than 10 �

34

g/mL (P � 0.11). However, in patients with a PSAevel between 4 and 10 ng/mL, the difference wasighly statistically significant (P � 0.01). The PCaetection rate was 21.5% for the Kretz probe,4.5% for the B&K probe, and 31.3% for the ATLrobe (Fig. 2).In considering patients with and without suspi-

ious DRE findings, we could not detect a statisti-ally significant difference for patients with suspi-ious DRE findings (P � 0.062), but in patientsith nonsuspicious DRE, the difference was highly

tatistically significant (P � 0.004). The observedCa detection rate was greatest for the ATL probet 26.7% (223 of 835) compared with 14.5% (20 of18) for the Kretz probe and 22.0% (78 of 355) forhe B&K probe. The difference was even more pro-ounced for patients with nonpalpable PCa with aSA level between 4 and 10 ng/mL (Table II). TheCa detection rate was 25.3% for the ATL probend only 12.3% for the Kretz probe (P � 0.02) and7.6% for the B&K probe (P � 0.019). We alsonalyzed the influence of prostate size and the in-titution on the detection rate, but the differenceas not statistically significant for the differentrobes (data not shown).We performed a binary logistic regression analy-

is with forward selection of parameters. We iden-ified three independent parameters predictive ofCa: PSA (P �0.0001), DRE findings (P �0.0001),nd TRUS findings (P �0.0001). The ultrasoundrobe was not an independent predictor of prostateancer. However, when we analyzed the subgroupith a PSA level of 4 to 10 ng/mL, the differentltrasound probes were statistically significant (P

characteristicsne B&K

MedicalATL HDI

1000 Total

67.1 65.7 66.266.1 65.0 65.4

7.4 7.4 7.56.4 6.5 6.5

11.2 15.4 15.558.1 53.6 52.030.7 31.0 32.5

37.3 47.6 52.225.2 38.0 36.1

57.2 66.1 62.257.2 65.6 62.5

ination.

ntbisoS

l exam

0.007) parameters for the prostate biopsy results

UROLOGY 64 (3), 2004

t�

wrtTdogistpDw

errTsTat

gttittopptctPp

fit

btttw(bedprtzTitemolcoa

tsToi1fsigpmea

P

�4

�

AD

U

ogether with the DRE (P �0.0001) and TRUS (P0.013) findings.

COMMENT

The random systematic sextant biopsy is aidely used and accepted protocol. Hodge et al.2

eported in 1989 that this procedure was superioro directed biopsies to hypoechoic lesions onRUS. However, the real detection rate of the ran-om systematic biopsy is only 60% at the first bi-psy.3 Therefore, new strategies are under investi-ation to improve the PCa detection rate. Mostnvestigators have focused on different biopsytrategies, such as increasing the number of coresaken12,13 or sampling from distinct regions of therostate.6,7,14 Another approach is to use color flowoppler sonography8,9 or artificial neuronal net-orks10 to enhance the PCa detection rate.Only a little information is available on the influ-

nce of the ultrasound equipment on the detectionate of PCa. This is because of the disappointingesults when trying to visualize PCa usingRUS.15,16 The sensitivity and specificity of TRUScanning is low. The limited potential of gray-scaleRUS is due to the variable sonographic appear-nce of PCa and other benign prostate abnormali-ies.17,18

Although it is widely accepted that ultrasounduidance of prostatic biopsy is necessary,19 the ul-rasound equipment probably does not influencehe limitations regarding sensitivity and specific-ty. In this respect, Vleeming et al.11 reported thathe frequency of the ultrasound transducer had lit-le influence on the detection rate of prostate bi-psy. The first reports of biplanar ultrasoundrobes date back to 1989.20 Biplanar, side-firerobes enable the investigator to biopsy in a longi-udinal section. The second plane can be used toontrol the placement of the biopsy needle in theransverse section.20 The first reports from Torp-edersen et al.21 demonstrated that biopsy of the

TABLE II. Prostate cancer detection rate straprobe used for

SA (ng/mL)

Kretz Combisone320 S

Bru

Total PCa Total

4 86 19 (22.1) 65–10Total 135 29 (21.5) 339DRE suspicious 68 20 (29.4) 83DRE not suspicious 65 8 (12.3) 24410 157 83 (52.9) 179

bbreviations as in Table I.ata presented as number of patients, with percentage in parentheses.

rostate was possible with both side-fire and end- p

ROLOGY 64 (3), 2004

re probes, but they did not analyze whether one ofhese probes is superior to the other.Our hypothesis was that end-fire probes would

e superior because sampling in the transverse sec-ion may detect more smaller tumors. The reason ishat the biopsy needle can be directed more towardhe lateral part of the peripheral zone, the area inhich most of the peripheral zone tumors arise

Fig. 1). This strategy was based on the observationy Stamey22 who reported that sampling in the lat-ral aspect of the peripheral zone increased theetection rate. Newer biopsy strategies have em-hasized sampling in this region. Chang et al.23

eported the benefit of four additional biopsy coresaken from the most lateral edge of the peripheralone at the central and basal part of the prostate.hey noted a statistically significant, but moderate,

ncrease in the detection rate of 14%. In a recentrial,24 they stated that an eight-core biopsy strat-gy would be best. Using that strategy, the twoedian biopsy cores at the prostatic base can be

mitted, demonstrating the importance of the mostateral biopsy. Gore et al.25 also stated that an in-reased detection rate may be attributed to the usef end-fire multiplanar probes but did not performcomparative analysis.The results of our investigation have confirmed

his hypothesis. None of the ultrasound probes wasuperior when used in large tumors. However, in1c PCa with a PSA value of 4 to 10 ng/mL, the usef the end-fire probe was highly statistically signif-cantly better. The detection rate increased from2.3% and 17.3% for the side-fire probes to 25.2%or the multiplanar probe in this subgroup analy-is, nearly a twofold increase. We attributed thisncrease to the ability to biopsy specifically the re-ion in which most PCa arise—the most-lateraleripheral zone. We found that detecting large tu-ors and/or palpable tumors was identical with

ither ultrasound system, because the tumor canlso be detected in the more medial parts of the

d by PSA value, DRE findings, and ultrasoundt-time biopsyKjaer

icalATL HDI

1000 P Value(Chi-square Test)PCa Total PCa

11 (16.9) 246 59 (24.0) 0.48

83 (24.5) 857 268 (31.3) 0.0137 (44.6) 322 132 (41.0) 0.1343 (17.6) 522 132 (25.3) 0.00805 (58.7) 496 246 (49.6) 0.11

tifiefirsel &

Med

1

eripheral zone. Such tumors can be defined as

535

tnpuDPalbei

fdcititgwewo

pb

as

sU

p9

pm

tJ

rn

D

tJ

er1

t(A

dt

ut4

tc

p1

rt

tU

n

duC

g1

ge

ba

t

et

tb1

t

5

hose with either high PSA values (greater than 10g/mL) or with palpable disease. In this respect,atients with a PSA value of less than 4 ng/mL aresually scheduled for biopsy because of suspiciousRE findings. However, small-volume, early-stageCa will be missed more often if the biopsy coresre sampled in the midline of the prostatic sideobes, but may be detected more frequently if theiopsy needle can be selectively guided to the lat-ral edge of the peripheral zone, as demonstratedn our study.

CONCLUSIONS

The results of our study have elucidated the ef-ect of the ultrasound probe on PCa detection. Ourata emphasize that sampling of the sextant biopsyores in the most lateral part of the peripheral zones associated with the greatest yield of PCa in pa-ients with small tumors. This sampling was facil-tated by end-fire probes because of the freedom ofhe biopsy plane, which can be chosen in any an-le. Therefore, sampling in the transverse sectionith an optimal visualization of the lateral periph-

ral zone is possible. Our analysis suggests that theidespread use of end-fire probes for prostate bi-psy could enhance the PCa detection rate.

REFERENCES1. Djavan B, Remzi M, Ghawidel K, et al: Diagnosis of

rostate cancer: the clinical use of transrectal ultrasound andiopsy. EAU Update Series 1: 9–15, 2003.

2. Hodge KK, McNeal JE, Terris M, et al: Random system-tic versus directed ultrasound guided transrectal core biop-ies of the prostate. J Urol 142: 71–74, 1989.

3. Terris MK: Sensitivity and specificity of sextant biop-ies in the detection of prostate cancer: preliminary report.rology 54: 486–489, 1999.

4. Chen ME, Troncoso P, Tang K, et al: Comparison ofrostate biopsy schemes by computer simulation. Urology 53:51–960, 1999.

5. Stricker HJ, Ruddock LJ, Wan J, et al: Detection of non-alpable prostate cancer: a mathematical and laboratoryodel. Br J Urol 71: 43–46, 1997.

6. Chen ME, Troncoso P, Johnston DA, et al: Optimiza-ion of prostate biopsy strategy using computer based analysis.Urol 158: 2168–2175, 1997.

7. Eskew LA, Bare RL, and McCullough DL: Systematic 5egion prostate biopsy is superior to sextant method for diag-osing carcinoma of the prostate. J Urol 157: 199–202, 1997.

8. Roy C, Buy X, Lang H, et al: Contrast enhanced color

oppler endorectal sonography of prostate: efficiency for de- t

36

ecting peripheral zone tumors and role for biopsy procedure.Urol 170: 69–72, 2003.

9. Pepe P, Patane D, Panella P, et al: Does the adjunct ofcographic contrast medium Levovist improve the detectionate of prostate cancer. Prostate Cancer Prostatic Dis 6: 159–62, 2003.10. Loch T, Leuschner I, Genberg C, et al: Improvement of

ransrectal ultrasound: artificial neural network analysisANNA) in detection and staging of prostatic carcinoma. Urol39: 341–347, 2000.11. Vleeming R, Noordzij JW, de Reijke TM, et al: The

etection of prostatic carcinoma: 4- or 7-MHz transrectal ul-rasonography. World J Urol 11: 66–68, 1993.

12. Levine MA, Ittmann M, Melamed J, et al: Two consec-tive sets of transrectal ultrasound guided sextant biopsies ofhe prostate for detection of prostate cancer. J Urol 159: 471–75, 1998.13. Naughton CK, Miller DC, Mager DE, et al: A prospec-

ive randomized trial comparing 6 versus 12 prostate biopsyores: impact on cancer detection. J Urol 164: 388–392, 2000.

14. Ravery V, Goldblatt L, Royer B, et al: Extensive biopsyrotocol improves the detection rate of prostate cancer. J Urol64: 393–396, 2000.15. Sedelaar JP, Vijverberg PL, de Reijke TM, et al: Trans-

ectal ultrasound in the diagnosis of prostate cancer: state ofhe art and perspectives. Eur Urol 40: 275–284, 2001.

16. Scherr DS, Eastham J, Ohori M, et al: Prostate biopsyechniques and indications: when, where, and how. Seminrol Oncol 20: 18–31, 2002.17. Clements R: The role of transrectal ultrasound in diag-

osing prostate cancer. Curr Urol Rep 3: 194–200, 2002.18. Durkan GC, and Greene DR: Diagnostic dilemmas in

etection of prostate cancer in patients undergoing transrectalltrasound-guided needle biopsy of the prostate. Prostateancer Prostatic Dis 3: 13–20, 2000.19. Renfer LG, Schow D, Thompson IM, et al: Is ultrasound

uidance necessary for transrectal prostate biopsy? J Urol 154:390–1391, 1995.20. Bock E, D’Ascenzo R, Valenti P, et al: Ultrasound-

uided prostatic biopsy using biplanar transrectal probe: ourxperience. Radiol Med (Torino) 78: 514–515, 1989.

21. Torp-Pedersen S, Lee F, Littrup PJ, et al: Transrectaliopsy of the prostate guided with transrectal US: longitudinalnd multiplanar scanning. Radiology 170: 23–27, 1989.

22. Stamey TA: Making the most out of six systematic sex-ant biopsies. Urology 45: 2–12, 1995.

23. Chang JJ, Shinohara K, Bhargava V, et al: Prospectivevaluation of lateral biopsies of the peripheral zone for pros-ate cancer detection. J Urol 160: 2111–2114, 1998.

24. Presti JC, Chang JJ, Bhargava V, et al: The optimal sys-ematic prostate biopsy scheme should include 8 rather than 6iopsies: results of a prospective clinical trial. J Urol 163:63–167, 2000.25. Gore JL, Shariat SF, Miles BJ, et al: Optimal combina-

ions of systematic sextant and laterally directed biopsies for

he detection of prostate cancer. J Urol 165: 1554–1559, 2001.

UROLOGY 64 (3), 2004