Multiparametric magnetic resonance imaging and combined prostate biopsy: opportunities, advantages and pitfalls

Multiparametric MRI (mpMRI) has superb sensitivity in prostate cancer detection. mpMRI is increasingly used not only for primary diagnostics, but for location of suspicious lesion before biopsy in case of targeted biopsy (TB). In many recent studies have been shown higher level of TB accuracy in prostate cancer detection in comparison with traditional systemic biopsy. In recent EAU, NICE, ACR recommendations mpMRI is indicated for men with high level of prostate cancer suspicion with previous negative results of systemic biopsy. However, it is not absolutely clear, whether mpMRI is indicated for biopsy-naïve men. This study is dedicated for prostate biopsy planning in the groups of biopsy-naïve men and with the history of previous biopsy.

Критериями исключения из исследования стали признаки распространенного опухолевого процесса: уровень сывороточного ПСА более 50 нг/мл Multiparametric MRI (mpMRI) has superb sensitivity in prostate cancer detection. mpMRI is increasingly used not only for primary diagnostics, but for location of suspicious lesion before biopsy in case of targeted biopsy (TB). In many recent studies have been shown higher level of TB accuracy in prostate cancer detection in comparison with traditional systemic biopsy. In recent EAU, NICE, ACR recommendations mpMRI is indicated for men with high level of prostate cancer suspicion with previous negative results of systemic biopsy. However, it is not absolutely clear, whether mpMRI is indicated for biopsy-naïve men. This study is dedicated for prostate biopsy planning in the groups of biopsy-naïve men and with the history of previous biopsy. и/или распространение опухолевого процесса на соседние органы по данным МРТ, так как в обоих случаях целесообразность выполнения ПБ крайне низкая. Также были исключены пациенты с протезами тазобедренного сустава, снижающими качество исследования и затрудняющими интерпретацию данных в соответствии с рекомендациями PI-RADS.

Introduction
The prostate cancer diagnostic pathway offers transrectal ultrasoundguided (TRUS) systemic biopsy (SB) [1-3]. Overall SB accuracy limitations and necessity to increase the effectiveness clinically significant prostate cancer (csPCa) detection initiated the search for new diagnostic methods [4][5][6]. Excellent mpMRI accuracy in detection of csPCa allows wide implementation in clinical practice. Application of mp-MRI for planning and navigation of "index foci" potentially increases sensitivity of the procedure. The high negative predictive value (NPV) of negative mpMRI result in the detection of csPCa, in a number of cases, might allow abandon prostate biopsy in favor of active surveillance [7]. In 2020, the European Association of Urologists published indications for targeted prostate biopsy [7]. Based on the previous studies, targeted biopsy (TB) demonstrated high efficacy in patients with previous negative SB results with persistent clinical suspicion of prostate cancer [3,7].
However, in order to widely introduce such diagnostic algorithms into clinical practice, it is necessary to clarify the effectiveness of mpMRI in detecting csPCa, analyze the reproducibility of the technique and interpretation of MRI data using a unified system PI-RADS v.2.1 in the conditions of routine clinical practice, and evaluate the diagnostic effectiveness of targeted biopsy performed based on positive MRI results.

Materials and methods
The study was done in Sverdlovsk Region Oncology Center, Sverdlovsk Regional Clinical Hospital No. 1 and "European medical center UGMK-Health" Ltd. 113 participants were recruited by local urologist. Patient age was 47-75 years (mean age -63 years), all patient presented clinical suspicion of PCa with el-evated serum prostate specific antigen (PSA) level or abnormal digital rectal examination (DRE) or both. The study included 13 (11.5%) patients with a history of SB and negative histological results. In view of persisting PSA growth and continued suspicion of csPCa, they perform mpMRI before repeated biopsy.
All patients underwent mpMRI with consequent transperineal targeted biopsy (TB) and systemic biopsy (SB) in cases 3-5 categories changes detection according to PI-RADS, or transperineal SB only in 1-2 category changes. The study was conducted strictly according to the developed PI-RADS v.2.1 criteria [10]. Time gap between mpMRI and the biopsy did not exceed 4 weeks.
Exclusion criteria of the study were signs of advanced tumor: the level of serum PSA more than 50 ng/ml and/or the spread of the tumor to adjacent organs on MRI, since in both cases the probability of advance cancer is extremely high and there is no reason performing TB. Patients with hip prostheses that reduced study quality and subsequent difficultness in correct interpretation according to PI-RADS criteria were also excluded.
All MRI data and study interpretation were performed in accordance with PI-RADS v.2.1 [10]. Multiparametric MRI was done at 3T Siemens Skyra with an external 20-element coil. The MRI protocol included T2W performed in 3x orthogonal planes, TR 4200, TE 102, matrix 448 × 448, slice thickness 3mm, slice gap 0 mm, FOV 200 mm; axial DWI obtained with multiple b-values (50.800.1000) matrix 128 × 128, slice thickness 3 mm, T1W gradient echo with dynamic bolus contrast enhancement (Gadovist 7.5), injection rate 2 ml/s, repetition 35, time resolution 7 sec, matrix 256 × 256. Interpretation of the data was performed by two independent radiologists. All discrepancies, if any, were analyzed, based on the results of the discussion, a collegial opinion was issued. PI-RADS v.2 category 3-5 sites were placed at high resolution T2W. For navigation during transperineal biopsy, the GE Logic E9 expert class ultrasound system (Milwaukee USA) equipped with a hardware and software complex for fusion biopsy MedCom GmbH BiopSee 2.2 (Darmstadt, Germany) was used. The biopsy was performed by a urologist.
The study used a combined biopsy algorithm involving sequential execution of SB and TB in one session. At the first stage, a systematic biopsy of 10-12 points was performed in accordance with the standard template. Further, according to the MRI findings, foci were marked for targeted biopsy with their subsequent biopsy (3-6 biopsy columns from each marked focus). On average, the total time spent on biopsy was 15-25 minutes. All patients were biopsied under anesthesia. The combination of inhalation (Sevoran) and intravenous (Propofol, Phentanyl) anesthetics was used. In case of 1-2 category of changes on PI-RADS, only transperineal SB from 12-14 points was performed.
SB in the study was performed with transperineal access. According to published studies, there were no significant differences in diagnostic accuracy compared to classical SB with TRUS navigation [11].
The total number of biopsy cores in both patient groups was 12 to 22. All foci of both TB and SB were numbered and entered into the diagnostic biopsy protocol, according to which, in retrospective data analysis, it is possible to obtain information from which segment of the prostate biopsy core was performed (Fig. 1, 2). When comparing MRI data, diagnostic protocol and obtained histological examination results, analysis of positive and negative results was performed for each biopsy site. For statistical analysis, the McNimar test for dependent variable was applied, since the both methods were simultaneously performed in the same patients, which ensures the correctness of the selected method. Test results are presented in the form of a combined table.
Histological evaluation was done in accordance recommendations of the international committee of pathomorphologists specializing in urogenital pathology (ISUP) 2018 [12].

Results
In all patients, an increase in serum PSA level of 2.4 ng/ml to 49.48 ng/ml (mean 15.10 ng/ml) was detected. In 26 (23%) of 113 patients DRE abnormalities were detected. 100 (88.5%) patients were biopsy naïve men; 13 (11.5%) patients had history of negative SB.
The results of the mpMRI were divided into positive and negative groups according to the category of PI-RADS: the categories 3-5 were recognized as positive, 1-2 were recognized as negative. Analysis of mpMRI results is shown in Table 1.
On mpMRI data, 26 (23%) patients had a 1-2 category of changes established by PI-RADS v2.1 corresponding to a very low or low probability of csPCa. In 14 (12%) cases, category 3 corresponding to the intermediate probability of csPCa is defined; Category 4 -in 30 patients (27%), corresponding to a high probability of csPCa and category 5 -in 43 patients (38%), corresponding to a very high probability of csPCa.
26 (23%) patients with 1-2 category PI-RADS findings underwent TB due to high clinical suspicion of PCa (high serum PSA level combined with low, less than 15% free PSA/total PSA ratio), 4 of them had csPCa detected. In retrospective evaluation of MRI data, the previously identified 2nd category on PI-RADS was unchanged.
Based on the obtained results, the number of true positive mpMRI cases was 62 of 113, true negative -22, false negative -4, false positive -25. Thus, the sensitivity of mpMRI was 93.9%, the specificity -46.8%. The positive predictive value (PPV) was 71.2%, NPV was 84.6%. The number of true positive cases of SB was 41 out of 113, true negative -47, false negative -25. There were no false positive SB cases if the ISUP class differences in post-biopsy and post-operative in the case of radical prostatectomy (RP) histology were not taken into account, which was not part of the objectives of the present study. Sensitivity of SB was 62.1%, specificity -100%, NPV -65.2%. Table 2 shows the results of two prostate biopsy techniques.
Thus, in 37 patients (32.5%), csPCa was detected on TB and SB. In 25 (22%) cases, csPCa was detec ted on TB only and not diagnosed on SB. In 4 (3.5%) patients, csPCa was detected on SB data and in case of 47 patients (42%) no csPCa was detected on TB or SB.
In 13 patients with previously negative SB and persistent PSA level growth according to mpMRI, foci with category 4-5 by PI-RADS were identified, 9 of them in the anterior fibromuscular stroma and 4 in the transitional zone. According to the histological results, TB from the detected mpMRI foci was positive. Thus, the sensitivity of mpMRI and TB in this selected category was 100%.
47 patients with negative histological data both after TB or SB were observed for 14-19 months; of these, 25, after anti-inflammatory treatment, showed serum PSA level decreasing, due to which repeated prostate mpMRI and biopsies were not performed. A control study was performed on 2 out of 25 patients. With repeated prostate MRI, in one case, at 15 months there was a decrease in the estimated gradation of PI-RADS v.2.1 from category 4 to 2nd, clinically this was accompanied by a decrease in PSA levels from 6.3 ng/ml to 4.41 ng/ml and was regarded as an outcome of the inflammatory process. In the second case, after 16 months, in the absence of significant dynamics in the MR picture and stable PSA levels, a repeated combined biopsy was performed with negative histological data, which did not confirm the presence of PCa. Thus, in all 47 patients in the dynamic follow-up group, the absence of csPCa at the time of study mp-MRI and subsequent biopsy can be concluded.

Discussion
The obtained results of the combined biopsy including TB and SB demonstrate the value of each method in PCa diagnosis. The simultaneous execution of TB and SB makes it possible to compare the results of each technique correctly. Analysis of patients groups with negative mpMRI results undergoing transperineal SB revealed false negative mpMRI data, suggesting a continuing need to perform biopsies in the case of negative mpMRI data. In 22% of observations, csPCa was only detected from mpMRI and TB data, confirming the high feasibility of performing mpMRI in all patients with suspected prostate cancer. Detection from TB csPCa data in all biopsy-na ve men suggests the obligation to perform mpMRI before repeated biopsy.
Our data clearly demonstrate the limitations of mp-MRI specificity, in cases of 3-4 categories of changes according to PI-RADS. This fact is observed in wide range previously published study: the specificity of MRI for 4th category according to PI-RADS was 60%, and for the 3rd category -12% [16]. The low mpMRI specificity became a reason to refuse mpMRI performance for the men without PCa suspicion, according to EAU recommendations [3].
The high NPV mpMRI is a consequence of it high sensitivity. The NPV in our study was 84%, which correlates with P. Moldovan meta-analysis data (2017) [15], where the mean NPV mpMRI was 82.4% (interval 69.0-92.4%). High NPV mpMRI is likely to be useful for urologist physicians in patient management especially with considering mpMRI findings. In the present study, the sensitivity of mpMRI (84%) was significantly higher than the sensitivity of SB (65.2%), which may be the reason to perform mpMRI for all patients with suspected csPCa.
In our study, TB was performed after SB by the same specialist urologist. Sequential execution of SB on the approved template, followed by MPI data import minimizes the effect of biopsy focus positioning correction in SB depending on MRI results. For ethical reasons, the possibility of performing 2 different stages of biopsies (separate execution of TB and separate execution of SB) was not approved, as a result of which patients were offered a single-step combined biopsy. The latter option is dominant in nowadays clinical practice and an increasing number of leading centers are abandoning the pre-existing technique of performing SB with TRUS-navigation immediately or after the TB performed the day before. The effect of mpMRI results on SB performance when TB and SB are performed simultaneously by one person is a well known problem when it is necessary to compare the sensitivity of each method: if a specialist is aware of the mpMRI data, then the part of SB samples can be taken not from the template position, but from the zone of interest according to the mpMRI data, which will increase the SB sensitivity and cause incorrect evaluation of the results. In the O. Rouviere (2018) study, SB was performed after TB by a clinician, who was not aware of the results of mpMRI; the authors indicate that artifacts and traces of hemorrhagic im-pregnation may have influenced site selection for SB and increased it sensitivity.
In contrast, in randomized trials with different patient samples, where one part of the patients undergo only TB and the other part -only SB, both the difficulty of correct randomization still remains (moreover for ethical reasons, it was not possible in our study to offer patients only TB) and the errors of randomization itself exist, which could potentially reduce accuracy of compared data [4,13].
The absence of tumor signs in 47 patients with negative histological examination data after biopsy at dynamic observation for 14-18 months indicates the true absence of csPCa at the time of mpMRI and biopsy, which in turn confirms the accuracy of our data. In the V. Panebianco's (2018) study analyzed patient monitoring with negative MRI and negative SB data; low probability (less than 5% for at least 2 years) of csPCa was demonstrated, it is assumed that all detected tumors developed during dynamic observation and were absent at the time of mpMRI or biopsy [14].

Conclusions
The obtained results of the study indicate a high sensitivity of the mpMRI and TB performed according to its results. The high proportions of patients with detected csPCa on TB solely indicates the feasibility of applying this technique to all patients regardless of previous biopsies history. In patient with previous negative SB results and persistent clinical PCa suspicion, performing of mpMRI before biopsy is mandatory. The number of false negative mpMRI results indicates the impossibility of using the method as a triage test before biopsy. Currently, combined biopsy, which includes TB and SB, can be recognized as the most rational method of PCa detection.

Authors' participation
Gulin G.A. -conducting research, collection and analysis of data, statistical analysis, writing text, analysis and interpretation of the obtained data, participation in scientific design, preparation and creation of the published work.
Zyryanov A.V. -concept and design of the study, participation in scientific design, participation in scientific design, preparation and creation of the published work, approval of the final version of the article.
Rubtsova N.A. -concept and design of the study, participation in scientific design, text preparation and editing, preparation and creation of the published work, approval of the final version of the article.
Artemov V.M. -collection and analysis of data, analysis and interpretation of the obtained data.
Zamyatin A.V. -collection and analysis of data, statistical analysis, analysis and interpretation of the obtained data.
Grebenev E.А. -collection and analysis of data, analysis and interpretation of the obtained data.