3-T in-bore MR-guided prostate biopsy based on a scoring system for target lesions characterization

Abstract

Background

To estimate potential malignant lesions within the prostate gland, the usage of a scoring system has recently been proposed by a European consensus meeting.

Purpose

To prospectively investigate a scoring system for functional prostate magnetic resonance imaging (MRI) using in-bore MR-guided prostate biopsy at 3-T.

Material and Methods

Prostate MRI examinations of 59 patients (between February 2011 and May 2012) with no known prostate cancer, elevated prostate specific antigen (PSA) level, and unsuspicious digital rectal examination were included in the study. In each patient up to three lesions were defined and scored using a 5-point scoring system for each MR sequence (T2-weighted images, diffusion-weighted imaging, dynamic contrast-enhanced imaging). Following MRI-guided in-bore biopsy these lesions were correlated to the histopathological findings.

Results

A total number of 144 lesions were defined in 59 patients. In 28 patients (51 lesions) MR-guided in-bore biopsy was positive for tumor (Gleason grade 6 or higher). A cut-off limit of 10 or more points in summation of the individual scores of all three sequences was used, leading to a 90% sensitivity, 63% specificity, 58% positive predictive value, and 92% negative predictive value.

Conclusion

A simple 5-point scoring system of functional prostate MRI achieves excellent sensitivity and moderate specificity for directing 3-T-guided prostate biopsy relative to the histopathological findings.

Keywords

Prostate cancer 3-T MRI scoring in-bore MR-guided biopsy at 3-T

Introduction

Today the diagnosis of prostate cancer includes digital rectal examination, determination of prostate specific antigen (PSA), transrectal ultrasonography (TRUS), systematic TRUS-guided biopsy, and increasingly the use of magnetic resonance imaging (MRI). For the detection of prostate cancer MRI is predominantly used in patients with negative biopsy and remaining suspicion of prostate cancer. MRI is performed in various combinations of T2-weighted images with functional imaging, such as diffusion-weighted imaging (DWI), dynamic contrast-enhanced MRI (DCE-MRI), and spectroscopic images (MRSI). To estimate potential malignant lesions within the prostate gland the usage of a scoring system has recently been proposed by a European consensus meeting (1). The lately published ESUR guidelines of prostate MRI contain a scoring system named PI-RADS (prostate imaging, reporting, and data system) (2). It has previously been found that using a 5-point scoring system based on the recommendations of the consensus comes along with a high inter-reader reliability (3). In a pilot study with a smaller patient population this scoring system has been assessed by our group in clinical routine (4). The aim of this study is to prospectively validate the scoring system using in-bore MR-guided prostate biopsy at 3-T as a reference standard.

Material and Methods

Patients

Fifty-nine consecutive patients with prostate MR examinations and MRI-guided in-bore biopsy between February 2011 and May 2012 were prospectively included in the study. The study was approved by the local ethics committee and written informed consent was given by each patient. For each patient a maximum of three different lesions was selected and scored. If there were more than three lesions detected, those three with the highest score were biopsied. All lesions were described using a localization scheme proposed by the European consensus meeting (1). Therefore the prostate was divided into 27 different regions. All patients had elevated PSA levels (>4 ng/mL). None of the patients had a known tumor and none had a suspicious digital rectal examination. An interval of at least 6 weeks was kept between the functional prostate MRI and previous biopsies. In a second session a transrectal in-bore MR-guided biopsy of the defined lesions was performed. The scores of all biopsied lesions were correlated with the histopathological findings.

MR techniques

Prostate MRI examinations were acquired on a 3-T MR scanner (Magnetom Trio; Siemens Medical Systems, Erlangen, Germany) using a six-channel phased-array body-coil. Before the examination all patients received 20 mg of Butylscopolamine (Buscopan®, Boehringer Ingelheim Pharma, Ingelheim, Germany) each, both intravenously and intramuscularly directly before the examination. In addition 1 mg of Glucagon Hydrochloride (Glucagen®, Novo Nordisk Pharma, Pakuranga, New Zealand) was administered intramuscularly to suppress peristaltic artifacts.

Turbo spin echo sequences were used to acquire T2-weighted images in standard orthogonal planes: axial (TR, 10,630 ms; TE, 117 ms; field of view [FOV], 12.8 cm; voxel size, 0.5 × 0.5 × 3.0 mm; image matrix, 256 × 256; turbo factor, 23), sagittal HASTE (TE, 90 ms; FOV, 17 cm; voxel size, 1.1 × 0.9 × 3.0 mm; image matrix, 138 × 192; turbo factor, 138), and coronal (TR, 11,330 ms; TE, 103 ms; FOV, 17 cm; voxel size, 0.7 × 0.7 × 3.0 mm; image matrix, 256 × 256; turbo factor, 25). T1-weighted turbo spin echo images were acquired in the axial plane (TR, 650 ms; TE, 13 ms; FOV, 30 cm; voxel size, 1.3 × 0.9 × 5.0 mm; gap, 10%; image matrix, 240 × 320; turbo factor, 3). For axial diffusion-weighted imaging (DWI), a single-shot spin-echo echo-planar sequence was used (TR, 4,600 ms; TE, 90 ms; FOV, 20.4 cm; voxel size, 1.5 × 1.5 × 3.0 mm; image matrix, 136 × 136; using GRAPPA parallel imaging scheme with acceleration factor 2; scan time, 7:12 min). Diffusion-weighted images were acquired using five b-values (0, 250, 500, 750, 1,000 s/mm²) with five averages, applying diffusion gradients in three orthogonal directions for each b-value. Calculation of ADC parameter maps used the standard monoexponential model. Perfusion imaging was performed with a volume-interpolated gradient echo sequence (TR, 5.26 ms; TE, 1.76 ms; FOV, 19.2 cm; voxel size, 1.5 × 1.5 × 3.0 mm; image matrix, 128 × 128; GRAPPA parallel imaging scheme with acceleration factor 2; scans, 31; scan time, 5:05 min; temporal resolution, 10 s). Contrast media injection started after the second measurement using Gadoteric Acid (Dotarem®, Guerbet, Aulnay-sous-Bois, France) in weight-adapted standard dose (0.2 mmol/kg body weight) with an injection rate of 3 mL/s.

For postprocessing of dynamic contrast-enhanced (DCE)-MRI images the software DynaCAD (Invivo, Orlando, FL, USA) was used on an external workstation. A qualitative analysis was performed assessing inflow and wash-out characteristics of contrast media for each lesion. Total scanning duration was approximately 33 min.

Scoring

All lesions were scored by two radiologists in consensus with a minimum of 3 years of experience in prostate MRI. Each lesion was evaluated in the different MR sequences (T2, DWI, DCE-MRI) using a 5-point scale for each sequence as described before (Table 1) (3). Fig. 1 shows the lesions characterization in T2-weighted imaging for the different scores. A summed up total score of all three sequences (T2, DWI, DCE-MRI) of ≥10 was considered to be suspicious for prostate cancer.

Fig. 1.

Examples of lesions characterization in T2-weighted imaging for each score, ranging from score 1 (benign lesion) to score 5 (malignant lesion).

Table 1.

Scoring system for lesions characterization.

Score	T2-weighted images	DCE-MRI	DWI (ADC x 10^-6)
1	Hyperintense	No initial enhancement	>800 mm²/s
2	Inhomogeneous	Type 1 curve	<800 mm²/s
3	Hypointense, well defined	Type 2 curve	<700 mm²/s
4	Hypointense, fuzzy	Type 3 curve	<600 mm²/s
5	Hypointense, bulging	Type 3 curve + asymmetrical	<500 mm²/s

DCE-MRI, perfusion imaging; DWI, diffusion-weighted imaging; type 1 curve, progressive enhancement with a continuous increase in signal intensity; type 2 curve, initial increase in signal intensity followed by a plateau; type 3 curve, initial increase and subsequent decrease in signal intensity (wash-out).

In-bore MR-guided biopsy

In-bore MR-guided biopsy of all lesions was performed in a second session on the same 3-T MR scanner (Magnetom Trio; Siemens Medical Systems, Erlangen, Germany). All lesions independent of their score were biopsied. All patients took levofloxacin 500 mg (Tavanic, Sanofi-Aventis, Frankfurt, Germany) for oral antibiotic prophylaxis 2 days before, on the morning of the biopsy, and for 2 days after biopsy. Coagulation levels were checked before biopsy. The biopsies were performed transrectally with the patient in prone position. Targeting the predefined lesions was done with the dynatrim biopsy device (Invivo, Orlando, FL, USA) and the corresponding software DynaCAD (Invivo, Orlando, FL, USA). A six-channel phased-array body-coil was placed on the back of the patient gaining fast T2-weighted HASTE images in sagittal (TE, 76 ms; FOV, 28 cm; voxel size, 1.4 × 1.1 × 3.0 mm) and axial (TE, 76 ms; FOV, 28 cm; voxel size, 1.4 × 1.1 × 3.0 mm) planes for biopsy planning. The 3D-tool of the scanner software (Magnetom Trio; Siemens Medical Systems, Erlangen, Germany) was used for cross-checking the needle position. Two tissue samples of each lesion were obtained either with a 150 mm or 175 mm 18-G needle biopsy gun (Invivo, Orlando, FL, USA).

Statistics

The statistical analysis was performed using IBM® SPSS® Statistics 19 for Windows (SPSS Inc., Chicago, IL, USA). Data are expressed as mean ± standard division. Correlation coefficients were calculated when possible.

Results

Patients and biopsy procedure

In 59 patients (median, 65 years; range, 52–83 years) with a median PSA of 8 ng/mL (range, 4–49 ng/mL), 144 lesions were defined. The lesions were distributed as follows: 88 central lesions, 74 peripheral lesions, and 14 lesions in the anterior fibrosmuscular stroma. Thirteen of the peripheral lesions were located anteriorly, and 61 posteriorly. Twenty-eight lesions covered more than one region and were counted in each region. The median size of all lesions was 11 mm (range, 5–33 mm). Thirty-three patients underwent prior negative random TRUS-guided biopsy. The other 26 did not receive a prior biopsy.

Scoring: different sequences

The mean score of all lesions for T2-weighted images was 3.56 ± 0.86 (range, 1–5). For DWI the mean score was 3.10 ± 1.38 (range, 1–5), and for DCE-MRI the mean score was 3.34 ± 1.29 (range, 1–5). Gleason grade and lesion scores are shown in Table 2. Considering a score of 4 or 5 as suspicious for prostate cancer resulted in a sensitivity and specificity of 96% and 71% for T2-weighted images, 51% and 67% for DWI, and 63% and 58% for DCE-MRI, respectively. The score of T2-weighted images has the closest correlation (0.91) to the probability of malignancy compared to the score of the DWI (0.79) or the DCE-MRI (0.71).

Table 2.

Number of malignant lesions, Gleason grade, and corresponding score in the different sequences.

Gleason 6	Gleason 3 + 4	Gleason 4 + 3	Gleason 8	Gleason 9	Sum
T2
Score 1
Score 2
Score 3	1	1	2
Score 4	13	15	3	2	33
Score 5	3	7	3	3	16
DWI
Score 1
Score 2	1	1
Score 3	8	12	1	3	24
Score 4	3	6	2	11
Score 5	4	5	6	15
DCE-MRI
Score 1	2	2
Score 2	1	1	1	3
Score 3	6	5	1	2	14
Score 4	2	3	1	6
Score 5	5	14	4	3	26

DCE-MRI, dynamic contrast-enhanced MRI; DWI, diffusion-weighted imaging; T2, T2-weighted imaging.

Scoring: overall

The summed up mean score for all lesions was 9.97 ± 2.53 (range, 4–15). In 64 out of 144 lesions we found a total score of < 10 points and therefore these lesions were considered to be benign. Histopathologic examination of the biopsy specimen revealed prostate cancer in five of these lesions, and benign tissue in 59 cases. The five false-negative findings were scored as follows: 4 × total score: 9 (2 × Gleason 6, 1 × Gleason 3 + 4, 1 × Gleason 4 + 3), 1 × total score: 8 (Gleason 6). For the other 80 lesions the total score was between 10 and 15 points and they were called suspicious for malignancy. Histopathologic findings were positive in 46 of these lesions (true-positive findings) and negative in 34 lesions (false-positives). Gleason distribution in the positive findings was as follows: 13 × Gleason 6, 22 × Gleason 3 + 4, 6 × Gleason 4 + 3, 2 × Gleason 8, 3 × Gleason 9, respectively. Therefore, the detection rate using a cut-off limit of 10 was 58% in patients with elevated PSA and positive MR. A cut-off limit of 10 or more for a combined scoring of all three MR sequences led to a sensitivity of 90%, a specificity of 63%, a positive predictive value of 58%, and a negative predictive value of 92% (Table 3).

Table 3.

Cut-off limits of the total scores.

Discussion

The application of a simple 5-point scoring system based on the recommendations of a European consensus meeting in functional prostate MRI is feasible and leads to high sensitivity and high negative predictive value of 92%. It has been shown previously that this method has high inter-observer reliability when scoring MR data-sets (3). Thus, the prostate MR scoring system provides a reliable tool to diagnose prostate cancer and must be considered an essential part for MR-based prostate biopsy.

In patients with elevated PSA and negative TRUS-guided biopsy the highest published detection rate using a saturation needle biopsy technique is 34% (5). In this study the detection rate is in the upper range of published detection rates of MR-guided biopsy ranging from 38% to 59% (6 –11). Differences in detection rates between studies may be a result of different inclusion criteria and differences in the imaging technique, such as the use of an endorectal vs. a surface body coil or 1.5 T vs. 3-T. While modern surface coil technology allows a resolution similar to an endorectal coil, the field strength of the magnet must be considered crucial (12). The main advantages of a 3-T MR system are higher anatomic detail imaging and consecutive faster re-identification of suspicious lesions during in-bore MR-guided biopsy (9). Compared to initial results from the pilot study (4) the sensitivity was similar while the specificity was lower in the pilot study. This difference within the same imaging group may be contributed to a certain training effect of in-bore MR-guided biopsy.

This study uses a cut-off value of 10 to define malignancy on MR of the prostate. Using higher cut-off values than 10 for the total score leads to higher specificity and higher positive predictive value, but the sensitivity and the negative predictive value decrease. Using lower cut-off values enables sensitivities up to 100% with a decrease in specificity. However, the aim of lesion scoring should be to achieve a high negative predictive value with reasonable specificity to exclude prostate cancer in patients with negative MRI. This must be considered a substantial advantage over TRUS-guided biopsy with reported negative predictive values ranging from 36% to 89% (13).

Comparing the different sequences, T2-weighted imaging shows the highest sensitivity and specificity compared to DWI and DCE-MRI, with nearly equal results using a cut-off value of 4 or more for potential malignant lesions. In the literature values for sensitivity are ranging from 54% to 91% and for specificity from 27% to 91% for T2-weighted imaging (8). Both functional imaging sequences are accepted to further increase both sensitivity and specificity (14,15). Therefore, diagnosis of malignancy is typically based on the overall score including T2 and functional data-sets, rather than the score based on only one sequence (16).

In line with other MRI studies more than half of the lesions were located in the transitional zone (9). This is in contrast to the literature, which locates about 25 % of prostate cancers in the transitional zone (17). The large number of suspect lesions in the transitional zone may be attributed to a preselection effect, due to the prior TRUS-biopsy, which more than half of the patients received before. The transitional zone is frequently missed on TRUS-biopsies.

The scoring system used in this study is similar to PI-RADS, which has recently been published as part of the ESUR guidelines of prostate MRI (2). In comparison both scoring systems use a likert-like five-grade scale ranging from score 1 (benign finding) to score 5 (most malignant finding), while score 3 is equivocal. Other authors use similar scoring (18). One main difference of our score as compared with PI-RADS is that the ADC cut-off values are an integral part of our score, while PI-RADS does not use ADC values at all. ADC thresholds used in our institution are based on previous experience (4). However, published ADC thresholds are different in between publications (19,20). Technical parameters, especially the b-values and the mathematical model to calculate the ADC affect the proposed cut-off values (21). Thus, ADC-values may vary from scanner to scanner and cannot be standardized well. This may be the reason, why PI-RADS does not quantitatively include the ADC. In contrast to PI-RADS our score does not differentiate between central or peripheral lesion in the T2-weighted imaging. The spectroscopy can be used with the PI-RADS approach but MRSI is not part of our scoring system. Due to poor performance compared to T2-weighted imaging, long acquisition time, and often extensive postprocessing, MRSI is not used in our institution (22).

Aim of this study was not to establish a new scoring system apart from the PI-RADS scoring system. This study included patients before the PI-RADS system was released and therefore PI-RADS could not have been used additionally. Due to the similarity of both scoring system the present results should be comparable to the PI-RADS scoring system but further studies are needed to evaluate the PI-RADS scoring system in clinical routine.

The study has some limitations. The given sensitivity, specificity, negative, positive predictive value, and detection rate in this study are based on the scoring system used. To investigate the scoring system was the main aim of the study. These values can be different in other studies primarily focusing on the detection rate of MRI or MR-guided in-bore biopsy for several reasons: tumors can be inapparent in the MRI, tumors could have been missed by the readers in the MRI, and tumors can be missed in the MR-guided in-bore biopsy due to technical factors from the targeting. For these studies a gold standard those as a prostatectomy or a systematic random biopsy as a surrogate is essential, which is not available in the present study. The defined maximum of scored lesions was limited to three. This was chosen to prevent clinically unacceptable biopsy duration and consecutive patient discomfort. Due to the increasing movement of the patient with increasing duration of biopsy procedure we accept this as a compromise between the intention to perform a biopsy in every definable lesion and the duration of the procedure. This study does only distinguish between malignant and benign lesions. Other pathologies like prostatitis in benign lesions were not investigated. Since this study is focusing on the scoring system parameters like tumor size or percentage of tumor in the core are not investigated.

In conclusion, a simple 5-point scoring system of functional prostate MRI has an excellent sensitivity and moderate specificity of 63% using as reference standard in-bore MR-guided prostate biopsy at 3-T.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Dickinson

Ahmed

Allen

. Magnetic resonance imaging for the detection, localisation, and characterisation of prostate cancer: recommendations from a European consensus meeting. Eur Urol 2011; 59: 477–494.

Barentsz

Richenberg

Clements

. ESUR prostate MR guidelines 2012. Eur Radiol 2012; 22: 746–757.

Quentin

Arsov

Röhlen

. Inter-Reader Agreement of Multi-Parametric MR Imaging for the Detection of Prostate Cancer: Evaluation of a Scoring System. Rofo 2012; 184: 925–929.

Arsov

Blondin

Rabenalt

. [Standardised scoring of a multi-parametric 3-T MRI for a targeted MRI-guided prostate biopsy.]. Urologe A 2012; 51: 848–856.

Stewart

Leibovich

Weaver

. Prostate cancer diagnosis using a saturation needle biopsy technique after previous negative sextant biopsies. J Urol 2001; 166: 86–91.

Roethke

Anastasiadis

Lichy

. MRI-guided prostate biopsy detects clinically significant cancer: analysis of a cohort of 100 patients after previous negative TRUS biopsy. World J Urol 2012; 30: 213–218.

Franiel

Stephan

Erbersdobler

. Areas suspicious for prostate cancer: MR-guided biopsy in patients with at least one transrectal US-guided biopsy with a negative finding–multiparametric MR imaging for detection and biopsy planning. Radiology 2011; 259: 162–172.

Hoeks

Schouten

Bomers

. Three-Tesla Magnetic Resonance-Guided Prostate Biopsy in Men With Increased Prostate-Specific Antigen and Repeated, Negative, Random, Systematic, Transrectal Ultrasound Biopsies: Detection of Clinically Significant Prostate Cancers. Eur Urol 2012; 62: 902–909.

Hambrock

Fütterer

Huisman

. Thirty-two-channel coil 3T magnetic resonance-guided biopsies of prostate tumor suspicious regions identified on multimodality 3T magnetic resonance imaging: technique and feasibility. Invest Radiol 2008; 43: 686–694.

10.

Anastasiadis

Lichy

Nagele

. MRI-guided biopsy of the prostate increases diagnostic performance in men with elevated or increasing PSA levels after previous negative TRUS biopsies. Eur Urol 2006; 50: 738–748.

11.

Hambrock

Somford

Hoeks

. Magnetic resonance imaging guided prostate biopsy in men with repeat negative biopsies and increased prostate specific antigen. J Urol 2010; 183: 520–527.

12.

Kim

Kwon

. Comparison of pelvic phased-array versus endorectal coil magnetic resonance imaging at 3 Tesla for local staging of prostate cancer. Yonsei Med J 2012; 53: 550–556.

13.

Heijmink

van Moerkerk

Kiemeney

. A comparison of the diagnostic performance of systematic versus ultrasound-guided biopsies of prostate cancer. Eur Radiol 2006; 16: 927–938.

14.

Tanimoto

Nakashima

Kohno

. Prostate cancer screening: the clinical value of diffusion-weighted imaging and dynamic MR imaging in combination with T2-weighted imaging. J Magn Reson Imaging 2007; 25: 146–152.

15.

Kitajima

Kaji

Fukabori

. Prostate cancer detection with 3 T MRI: comparison of diffusion-weighted imaging and dynamic contrast-enhanced MRI in combination with T2-weighted imaging. J Magn Reson Imaging 2010; 31: 625–631.

16.

Dickinson

Ahmed

Allen

. Scoring systems used for the interpretation and reporting of multiparametric MRI for prostate cancer detection, localization, and characterization: Could standardization lead to improved utilization of imaging within the diagnostic pathway? J Magn Reson Imaging 2013; 37: 48–58.

17.

Sakai

Harada

Hara

. A comparison of the biological features between prostate cancers arising in the transition and peripheral zones. BJU Int 2005; 96: 528–532.

18.

Jung

Coakley

Vigneron

. Prostate depiction at endorectal MR spectroscopic imaging: investigation of a standardized evaluation system. Radiology 2004; 233: 701–708.

19.

Delongchamps

Rouanne

Flam

. Multiparametric magnetic resonance imaging for the detection and localization of prostate cancer: combination of T2-weighted, dynamic contrast-enhanced and diffusion-weighted imaging. BJU Int 2011; 107: 1411–1418.

20.

Kumar

Jagannathan

Kumar

. Apparent diffusion coefficient of the prostate in men prior to biopsy: determination of a cut-off value to predict malignancy of the peripheral zone. NMR Biomed 2007; 20: 505–511.

21.

Quentin

Blondin

Klasen

. Comparison of different mathematical models of diffusion-weighted prostate MR imaging. Magn Reson Imaging 2012; 30: 1468–1474.

22.

Weinreb

Blume

Coakley

. Prostate cancer: sextant localization at MR imaging and MR spectroscopic imaging before prostatectomy–results of ACRIN prospective multi-institutional clinicopathologic study. Radiology 2009; 251: 122–133.