Abstract
ARAB JOURNAL OF UROLOGY 2022, VOL. 20, NO. 3, 121–125 https://doi.org/10.1080/2090598X.2022.2067615 Performance of multi-parametric magnetic resonance imaging through PIRADS scoring system in biopsy naïve patients with suspicious prostate cancer a a b a Amr Nowier , Hesham Mazhar , Rasha Salah and Mohamed Shabayek a b Department of Urology, Faculty of Medicine Ain-Shams University, Cairo, Egypt; Department of Radiology, Faculty of Medicine, Ain- Shams University, Cairo, Egypt ABSTRACT ARTICLE HISTORY Received 31 January 2022 Background: Use of multi-parametric magnetic resonance imaging (mp-MRI) and Prostate Accepted 12 April 2022 Imaging Reporting and Data System (PI-RADS) scoring system allowed more precise detection of prostate cancer (PCa). Our study aimed at evaluating the diagnostic performance of mp-MRI KEYWORDS in detection of PCa. Prostate cancer; multi- Methods: Eighty-six patients suspected to have prostate cancer were enrolled. All patients parametric magnetic underwent mp-MRI followed by systematic and targeted trans-rectal ultrasound (TRUS) guided resonance imaging; PI-RADS; prostate biopsies. Sensitivity, specificity, positive predictive value (PPV), negative predictive TRUS guided prostate biopsy value (NPV) and accuracy of mp-MRI were evaluated. Results: Forty-six patients (53.5%) had prostate cancer on targeted and systematic TRUS biopsies. On mp-MRI, 96.6% of lesions with PI-RADS < 3 revealed to be benign by TRUS biopsy, 73.3% of lesions with PI-RADS 4 showed ISUP grades ≥1, whereas all PI-RADS 5 lesions showed high ISUP grades ≥ 3. For PI-RADS 3 lesions, 62.5% of them revealed to be benign and 37.5% showed ISUP grades ≥1 by TRUS biopsy. PI-RADS scores ˃3 had 69.57% sensitivity and 85% specificity for detection of PCa. On adding the equivocal PI-RADS 3 lesions, PI-RADS scores ≥3 had higher sensitivity (97.83%), but at the cost of lower specificity (32.5%). Conclusion: Mp-MRI using PI-RADS V2 scoring system categories ≤3 and >3 could help in detection of PCa. PI-RADS 3 lesions are equivocal. Including PI-RADS lesions ≥3 demonstrated higher sensitivity, but at the cost of lower specificity for mp-MRI in diagnosis for Pca. Abbreviations: CDR: cancer detection rates; DRE: digital rectal examination; ISUP: interna- tional society of urological pathology; mp-MRI: multi-parametric magnetic resonance imaging; NPV: negative predictive value; PCa: prosatate cancer; PI-RADS: Prostate Imaging Reporting and Data System; PPV: Positive predictive value; PSA: prostate specific antigen; TRUS: transrectal ultrasound. Introduction T2 weighted image (T2WI): important to deter- Prostate cancer (PCa) is the most common cancer mine transitional zone (TZ) lesions. among men worldwide [1]. Systematic 10 to 12 cores Dynamic contrast-enhanced (DCE): ineffective in prostate biopsy is the standard method for diagnosis assessment of TZ and low-volume lesions [8]. of PCa [2]. The 10 to 12 core prostate biopsy demon- Magnetic resonance spectroscopy (MRS): time strated cancer detection rates (CDR) of 31–42%, but consuming and expensive [9]. still has the risk of a false-negative prostate biopsy [3,4]. Many studies were performed to improve the Previously, different scores for mp-MRI were used CDR by increasing the number of cores, but this strat- to categorize the level of suspicion of the presence egy proved to be ineffective as it identified more insig- of PCa [10,11]. European Society of Urogenital nificant tumors [5]. Radiology (ESUR) published guidelines based on Multi-parametric magnetic resonance imaging (mp- expert consensus in 2012, to standardize a score for MRI) of the prostate has been studied as an alternative evaluation and reporting of prostate MRI, known as to increase CDR even in patients with a previously the Prostate Imaging Reporting and Data System (PI- negative prostate biopsy [6]. Different aspects of mp- RADS) [12]. Since then, many clinical and research MRI are evaluated with the following characteristics [7]: programs have validated this score. In 2015, the PI- RADS Steering Committee developed an updated Diffusion-weighted imaging (DWI): important to version (PI-RADS V2) to overcome some of the determine the peripheral zone (PZ) lesions. CONTACT Hesham Mazhar heshammazhar777@yahoo.com Department of Urology, Faculty of Medicine, Ain-Shams University © 2022 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 122 A. NOWIER ET AL. limitations of PI-RADS V1 [13]. Purpose of our study coronal T2-weighted HASTE without FS; axial 3D T1- twist dynamic study in free breath for about 50 frames was to evaluate the diagnostic performance of in 3–4 minutes. Gadolinium-based contrast was given PIRADS V2 scores in patients with suspicious PCa. intravenously by means of a power injector (Ulrich Medical® Tennessee TM, Germany) at an infusion rate Materials and methods of 1 ml/s. Then, pre-contrast T1- mapping with two flip angles were obtained. Subtracted images were com- Study subjects puted if needed. Retrospective analysis of prospectively collected data of 86 patients with suspicious prostate cancer TRUS-guided prostate biopsy recruited from the Urology department, Faculty of Systematic and targeted TRUS-guided prostate biop- Medicine, Ain Shams University between sies using the cognitive fusion technique were per- August 2019, and December 2020 was evaluated. All formed by an experienced senior urologist with more participants were Egyptians. All patients were sus- than 5 years experience. All patients had 12 cores pected to have prostate cancer based on elevated systematic TRUS-guided prostate biopsy in addition prostate specific antigen (PSA) and/or positive digital to 2–3 targeted biopsies from PIRADS ≥3 on mp-MRI. rectal examination (DRE). Patients with metallic pros- Biopsy specimens were then assessed by two different thesis, pacemaker, or abnormal kidney function or pre- expert pathologists in urologic oncology, who were vious negative biopsies were excluded. Our study was blinded to MRI examination results, before reporting carried out after local ethical committee approval the International Society of Urological Pathology (ISUP) (FMASU M S 65/2019) and obtaining informed consent 2014 updated Gleason score grading system [14,15]. from all patients included in our study. Statistical analysis Methods The collected data was revised, coded, tabulated, and All patients were subjected to full history taking, care- evaluated using Statistical Package for Social Science ful digital rectal examination, basic laboratory investi- (SPSS 25). Mean, Standard deviation (± SD) and range gations, serum PSA, mp-MRI. PI-RADS v.2 scores were were reported for parametric numerical data, whereas reported by senior radiologist for all patients in 8 sites frequency and percentage of non-numerical data. of the prostate; right base, right mid zone, right apex, Sensitivity, specificity, positive predictive value (PPV) left base, left mid zone, left apex, right and left anterior and negative predictive value (NPV) of mp-MRI were fibromuscular stroma. This was followed by systematic evaluated. Kappa statistics was used to evaluate the and targeted transrectal ultrasound (TRUS) guided agreement between two investigational methods. prostate biopsy using the cognitive fusion technique. Kappa’s value 0.4–0.75 meant fair to good agreement, whereas Kappa’s value below 0.4 meant poor Mp-MRI protocol agreement. The study was performed on a 3.0-T MRI system (MAGNETOM Skyra; Siemens Healthcare, Erlangen, Germany) with an 18-element body phased array coil Results and a 32-element spine array coil. Before contrast injection, anatomical MRI was performed including Our current study included 86 patients with suspicious sagittal and axial T2-weighted (T2W) HASTE (half- PCa. The basic characteristics of the patients included Fourier acquisition single shot turbo spin-echo) with were described in (Table 1). Mp-MRI revealed that 72 controlled respiration, without fat-suppression (FS); patients (83.7%) had one or more PI-RADS lesions ≥3, Table 1. Basic Characteristics of studied population. n = 86 Age Mean ±SD 63.07 ± 7.28 Range 49–79 DRE NAD 40 (46.5%) Firm 33 (38.4%) Hard 13 (15.1%) PSA Mean ±SD 10.07 ± 4.62 Range 3.5–34 PSA density Mean ±SD 0.19 ± 0.09 Range 0.05–0.69 Prostate volume Mean ±SD 55.37 ± 16.51 Range 26–105 SD: standard deviation, DRE: digital rectal examination, NAD: no abnormality detected, PSA: prostate specific antigen. ARAB JOURNAL OF UROLOGY 123 Table 2. Correlation between PIRADS score & ISUP grade results in all studied lesions. Mp-MRI (PIRADS score) < 3 3 4 5 Agreement ISUP grade(GS) n (%) n (%) n (%) n (%) Kappa P value Benign 535 (96.6) 40(62.5) 16(26.7) 0 0.345 <0.001 1 (3 + 3) 7 (1.3) 12 (18.8) 9 (15) 0 2 (3 + 4) 9 (1.6) 8 (12.5) 17(28.3) 0 3 (3 + 5, 4 + 4, 5 + 3) 2 (0.4) 2(3.1) 4(6.7) 1 (10) 4 (4 + 5, 5 + 4) 1 (0.2) 2 (3.1) 14(23.3) 9 (90) 5 (5 + 5) 0 0 0 0 Total (n = 688) 554 (80.5%) 64 (9.3%) 60 (8.7%) 10 (1.5%) TRUS: transrectal ultrasound, ISUP: international society of urological pathology, mp-MRI: multiparametric magnetic resonance imaging, PIRADS: prostate imaging-reporting and data system score, GS:Gleason score Table 3. Correlation between PI-RADS score >3 lesions on mp-MRI and results of TRUS Biopsy. TRUS Bx Agreement PI-RADS Negative Positive Total Kappa p value Negative (≤3) 34 (85%) 14 (30.43%) 48 (55.81%) 0.539 <0.001 Positive (>3) 6 (15%) 32 (69.57%) 38 (44.19%) Total 40 (100%) 46 (100%) 86 (100%) TRUS: transrectal ultrasound, mp-MRI: multiparametric magnetic resonance imaging, PIRADS: prostate imaging-reporting and data system score. whereas 14 patients had PIRADS score <3 in all low specificity of 36%, so elevated PSA does not reported sites. After targeted and systematic biopsies, necessarily mean the presence of a malignant only 46 patients (53.5%) were proved to have PCa. lesion, and TRUS-guided biopsy may underestimate 688 sites were evaluated in mp-MRI of 86 patients extent and grade of prostate cancer [12]. Recently, and PIRADS V.2 scores were reported. 591 (85.9%) sites mp-MRI has been widely used for the diagnosis of were proved to be benign by TRUS biopsies, whereas PCa and demonstrated high sensitivity in detection 97 (14.1%) sites harbored PCa. 134 PI-RADS ≥3 lesions of PCa [1]. Adding a standardized reporting method were detected by mp-MRI, of which 64 lesions were through PI-RADS scoring system to mp-MRI classified as PI-RADS 3, whereas 60 lesions were classi- increased its ability to detect PCa [16]. fied as PI-RADS 4, and only 10 lesions were classified as In our study mp-MRI done for all cases with suspi- PI-RADS 5. cious prostate cancer before TRUS guided biopsy In lesions with PI-RADS score <3, there was a very to evaluate the diagnostic performance of mp- low likelihood of presence of any PCa (3.4%). However, MRI (PI-RADS ≥3 score) to detect PCa using TRUS- in patients with PIRADS score 3, 4 and 5, PCa was guided biopsy as a reference standard. Our study detected in 37.5%, 73.3% and 100% respectively. We demonstrated statistically significant correlation also found statistically significant correlation between between PI-RADS score and ISUP grade, where lesions PIRADS score and ISUP grade (P < 0.001). (Table 2) with PI-RADS ˃3 demonstrated high percentage of We found good agreement between PI-RADS score cancer by TRUS biopsy (73.3% and 100% respectively). ˃3 on mp-MRI and detection of prostate cancer Most of lesions with PI-RADS < 3 were benign by TRUS (kappa = 0.539) and mp-MRI achieved sensitivity biopsy (96.6%) and this may help to avoid unnecessary of 69.57%, specificity of 85%, PPV of 84.21%, NPV of TRUS guided biopsied in these patients. These results 70.83% and accuracy reached 76.74% in detection of were in concordance with the study of Junker and prostate cancer compared to TRUS biopsy (Table 3, colleagues who showed that 92% and 100% of lesions Figure 1). It is worth mentioning that adding the equi- with PI-RADS score ˃ 3 were found as high grade PCa vocal PI-RADS score of 3, performance of mp-MRI with by TRUS biopsy and 97% of lesions with PI-RADS score PI-RADS score of ≥3 demonstrated higher sensitivity < 3 were benign by biopsy [17]. (97.83%), but lower specificity (32.50%). PPV and NPV Regarding the equivocal PI-RADS 3 lesions; our were 62.50% and 92.86% respectively and accuracy study included 34 patients with 64 PI-RADS 3 lesions, reached 67.44% (Table 4, Figure 1). 62.5% of them revealed to be benign and 37.5% showed ISUP grades ≥1 by TRUS biopsy. This also agreed with Aslam and colleagues who demonstrated Discussion that 56% of patients with PI-RADS 3 had non- malignant findings and 43% of patients had malignant PCa is the most diagnosed cancer in men [1]. prostatic adenocarcinoma on transrectal/transperineal However, the available diagnostic modalities have ultrasound-guided biopsy [18]. unfavorable sensitivities and specificities. PSA has 124 A. NOWIER ET AL. 97.83% 100% 92.86% 90% 85.00% 84.21% 76.74% 80% 70.83% 69.57% 67.44% 70% 62.50% 60% PI-RADS ≥ 3 50% PI-RADS >3 40% 32.50% 30% 20% 10% 0% sensitivity specificity PPV NPV Accuracy Figure 1. Diagnostic performance of PIRADS score ˃ 3 versus PIRADS score ≥3 on mp-MRI and detection of prostate cancer. Table 4. Correlation between PI-RADS ≥3 on mp-MRI and results of TRUS Biopsy. PI-RADS TRUS Bx Total Agreement Negative Positive Kappa p value Negative (<3) 13 (32.5%) 1 (2.17%) 14 (16.28%) 0.317 <0.001 Positive (≥3) 27 (67.5%) 45 (97.83%) 72 (83.72%) Total 40 (100%) 46 (100%) 86 (100%) TRUS: transrectal ultrasound, mp-MRI: multiparametric magnetic resonance imaging, PIRADS: prostate imaging-reporting and data system score. PI-RADS scores ˃3 demonstrated acceptable ability to Conclusion detect PCa compared to TRUS biopsy with good sensi- Mp-MRI using PI-RADS V2 scoring system categories tivity (69.57%), specificity (85%), PPV (84.21%), and NPV ≤3 and >3 could help in diagnosis of PCa. We can (70.83%) and accuracy reaching 76.74%. However, safely refrain patients with PI-RADS lesions ˂3 including lesions with PI-RADS score ≥3 for detection of lesions from prostate biopsies. PIRADS 3 lesions PCa raised sensitivity of mp-MRI to 97.83%, but at the are equivocal. Evaluating PIRADS lesions ˃3, cost of lower specificity of 32.5%. Patel and colleagues achieves high specificity for dianosis of Pca. demonstrated that including lesions with PI-RADS ≥3 However including PI-RADS lesions ≥3 demon- had high sensitivity of 81.25%, but low specificity of strated higher sensitivity, but at the cost of lower 32.26%. They also reported that PI-RADS ˃ 3 lesions specificity for mp-MRI in diagnosis of PCa. were associated with low sensitivity (43.7% and 37.5% respectively), but higher specificity (64.5% and 100% respectively) [19]. This was also confirmed by Youn and Author contributions colleagues who demonstrated increased sensitivity but Amr Nowier: project development, manuscript editing; decreased specificity with the use of PI-RADS score of ≥3 Hesham Mazhar: data collection, data analysis, manuscript compared to PI-RADS score of ˃ 3. They reported that writing Rasha Salah: data analysis, manuscript editing; sensitivity, specificity, PPV, NPV, and accuracy for PI-RADS Mohamed Shabayek: data analysis, manuscript editing. score of ≥3 were 92.3%, 58%, 62.3%, 90.9% and 72.7% respectively compared to 84.6%, 81.2%, 77.2%, 87.5% and 82.6% respectively for PI-RADS score of ˃ 3 [20]. Disclosure statement Main limitation of our study was the small popu- No potential conflict of interest was reported by the author(s). lation, so large multicentric studies are recom- mended to ascertain our results. Another limitation was lack of comparing results of TRUS biopsy to Funding final histological findings of radical prostatectomy The authors have no funding to report. specimen. ARAB JOURNAL OF UROLOGY 125 [10] Lin WC, Muglia VF, Silva GEB, et al. Multiparametric MRI References of the prostate: diagnostic performance and interreader [1] Mäkelä P, Anttinen M, Suomi V, et al. Acute and agreement of two scoring systems. Br J Radiol. 2016;89 subacute prostate MRI findings after MRI-guided (1062):1–10. [11] Rastinehad AR, Waingankar N, Turkbe B, et al. transurethral ultrasound ablation of prostate Comparison of multi parametric MRI scoring systems cancer. Acta Radiol. 2021;62(12):1687–1695. and the impact on cancer detection in patients under- [2] Heidenreich A, Bastian PJ, Bellmunt J, et al. European going MR US fusion guided prostate biopsies. PLoS One. Association of U. EAU guidelines on prostate cancer. 2015;10(11):1–16. Part 1: screening, diagnosis, and local treatment with [12] Barentsz JO, Richenberg J, Clements R, et al. ESUR pros- curative intent-update 2013. Eur Urol. tate. MR guidelines. Eur Radiol. 2012;22(4):746–757. 2014;65:124–137. [13] Weinreb JC, Barentsz JO, Choyke PL, et al. PI-RADS pros- [3] Irani J, Blanchet P, Salomon L, et al. Is an extended tate imaging – reporting and data system: 2015, version 20-core prostate biopsy protocol more efficient than 2. Eur Urol. 2016;69:16–40. the standard 12-core? A randomized multicenter trial. [14] Schlenker B, Apfelbeck M, Buchner A, et al. MRI-TRUS J Urol . 2013;190:77–83. fusion biopsy of the prostate: quality of image fusion [4] Campos-Fernandes JL, Bastien L, Nicolaiew N, et al. in a clinical setting. Clin Hemorheol Microcirc. Prostate cancer detection rate in patients with 2018;70:433–440. repeated extended 21-sample needle biopsy. Eur [15] Epstein JI, Egevad L, Amin MB, et al. The 2014 Urol. 2009;55:600–606. International Society of Urological Pathology ISUP) con- [5] Ploussard G, Nicolaiew N, Marchand C, et al. sensus conference on gleason grading of prostatic carci- Prospective evaluation of an extended 21-core noma: definition of grading patterns and proposal for biopsy scheme as initial prostate cancer diagnostic a new grading system. Am J Surg Pathol. strategy. Eur Urol. 2014;65:154–161. 2016;40:244–252. [6] Pokorny MR, de Rooij M, Duncan E, et al. [16] Rosenkrantz AB, Kim S, Lim RP, et al. Prostate cancer Prospective study of diagnostic accuracy compar- localization using multiparametric MR imaging: compar- ing prostate cancer detection by transrectal ison of Prostate Imaging Reporting and Data System ultrasound-guided biopsy versus magnetic reso- (PI-RADS) and Likert scales. Radiology. 2013;269 nance (MR) imaging with subsequent MR-guided (2):482–492. biopsy in men without previous prostate biopsies. [17] Junker D, Quentin M, Nagele U, et al. Evaluation of the Eur Urol. 2014;66:22–29. PI-RADS scoring system for mpMRI of the prostate: a [7] Renard-Penna R, Mozer P, Cornud F, et al. Prostate whole-mount step-section analysis. World J Urol. imaging reporting and data system and Likert scor- 2015;33(7):1023–1030. ing system: multiparametric MR imaging validation [18] Aslam S, Tsang J, Bickle I, et al. Multiparametric MRI study to screen patients for initial biopsy. Radiology. prostate PI-RAD scoring in a district general hospital: correlating PI-RADS 3 results with histological findings. 2015;275(2):458–468. Br J Radiol. 2022 [cited 2022 Jan 7];20210804. [8] Roethke MC, Kuru TH, Schultze S, et al. Evaluation of DOI:10.1259/bjr.20210804. the ESUR PI-RADS scoring system for multiparametric [19] Patel NU, Lind KE, Garg K, et al. Assessment of PI-RADS v2 MRI of the prostate with targeted MR/TRUS categories ≥ 3 for diagnosis of clinically significant pros- fusion-guided biopsy at 3.0 Tesla. Eur Radiol. 2014;24 tate cancer. Abdom Radiol (NY). 2019;44(2):705–712. (2):344–352. [20] Youn SY, Choi MH, Kim DH, et al. Detection and PI-RADS [9] Schimmoller L, Quentin M, Arsov C, et al. classification of focal lesions in prostate MRI: perfor- Predictive power of the ESUR scoring system for mance comparison between a deep learning-based prostate cancer diagnosis verified with targeted algorithm (DLA) and radiologists with various levels of MR-guided in-bore biopsy. Eur J Radiol. 2014;83 experience. Eur J Radiol. 2021;142:109894. (12):2103–2108.
Journal
Arab Journal of Urology
– Taylor & Francis
Published: Jul 3, 2022
Keywords: Prostate cancer; multi-parametric magnetic resonance imaging; PI-RADS; TRUS guided prostate biopsy
