Biparametric MRI for Local Staging of Prostate Cancer: Current Status and Future Applications

Results and Discussion

Primary staging of prostate cancer. The histologically confirmed T-stage of PCa is important for clinical evaluation and treatment planning based on oncological outcomes. Traditionally, PCa staging has been performed using nomograms such as Partin diagrams based on PSA levels, DRE, Gleason score, and percentage of prostate biopsy core involvement (8). This model often underestimates the true stage of disease and has been shown to be inferior to MRI, with the combination of MRI results and nomograms showing additional importance in predicting negative disease in PCa (6, 8). In addition to improving accuracy, MRI offers information about the location and extent of disease. This helps with surgical planning and decisions about using general surgery to achieve better results or performing nerve-sparing surgery to reduce morbidity (3). Current clinical MRI scanners have similar magnets and high levels of performance. Therefore, users can obtain the best prostate MRI by using different types of MRI scanners. T2-Weighted (T2w) imaging is capable of depicting the zonal anatomy of the prostate. Zonal anatomy is generally not visible on T1-weighted (T1w) images (9); T1w images can be used to evaluate blood products. If the patient has a recent history of prostate biopsy, T1w images should be hyperintense. In general, prostate MRI should not be performed within 8-12 weeks after prostate biopsy to allow sufficient time for resolution of prostate biopsy-related blood products (10). In the presence of blood products, PCa can appear as hypointense area as what some authors call ‘hemorrhage exclusion sign’ which is defined as the presence of a well-defined area of low signal intensity surrounded by areas of high signal intensity on T1w images (11). Although prostate MRI can detect the spread of local PCa, its accuracy is not ideal. Recently Choi et al. (12) observed that the extraprostatic extension grade score may help improve the standard and increase the value of MRI in this setting, just as PI-RADS does in cancer diagnosis. As for bpMRI, the lack of contrast-enhanced imaging does not appear to compromise the accuracy of assessment. This may be related to the fact that most signs of local invasiveness can be better evaluated on T2w images. High-resolution T2w is predictive for the detection of extracapsular growth (13). Prostate hyperplasia is characterized by swelling of the prostate surface, an absence of a high T2w signal on the prostate surface, presumed to be between the prostate and the anus or between the prostate and the seminal vesicles, and a low T2w signal or a large signal in the seminal vesicles (12, 13). Recent advances in positron-emission tomography (PET)/MRI scanner technology have demonstrated the possibility of combining metabolic/receptor information from PET with anatomical and functional imaging from MRI in a multimodal manner (14). While most of the studies on PET/MRI have been on restaging for PCa after treatment, diagnosis of the dominant lesion and characterization using PET/MRI has been an area of increasing interest in recent years, especially with the development of prostate-specific membrane antigen (PSMA) radio ligands (15). Studies show that the use of PSMA PET/MRI improves the diagnosis of clinically significant PCa (csPCa) compared to the use of mpMRI alone. For example, Ferraro and colleagues found sensitivity and specificity of 96% and 81%, respectively (16). Other studies by Park et al. (17) and Hicks et al. (18) found that for bilateral tumors, the positive predictive value of PSMA-PET/MRI was higher than that of mpMRI (70% versus 18%, respectively). PSMA-PET/MRI can accurately detect disease but its impact on patient outcomes and role in changing clinical management has not been established, making its value 50% higher compared to PET/computed tomography (CT) (18).

Lymph node (LN) staging. EAU guidelines recommend N-staging with preliminary diagnostic MRI in patients in all risk groups. PCa spreads mainly to the four pelvic regional LNs: obturator, internal iliac, external iliac, and presacral LNs. Involvement of one region is classified as N1 stage, whereas involvement of non-regional nodes (para-aortic or paracaval LNs) represents M1 disease (19). According to the risk stratification system, the probability of LN metastasis depends on the T-stage, histology, and Gleason score of patients with PCa (20). Similar to CT, traditional evaluation of MRI relies on size-based morphological interpretation, often leading to false-positive and false-negative LN staging results. It is recommended that LNs with a minor axis diameter greater than 8 mm be classified as suspicious (19, 21). Studies have shown that in the early stages of metastatic disease, small tumors are often affected and these cannot be detected radiologically. Almost 80% of LN metastases occur in LNs less than 1 cm in diameter. Hövels et al. (22) in a systematic review of 24 CT studies found a pooled sensitivity of 42% [95% confidence interval (CI)=0.26-0.56] and specificity of 82% (95% CI=80-83%) for the staging of pelvic LN metastases in PCa by CT. MRI data showed a similar, non-statistically significantly different accuracy, with overall sensitivity of 39% (95% CI=22-56%) and specificity of 82% (95% CI=79-83%). Additionally, the identification of benign cases of enlarged lymphadenopathy makes the size-based model questionable. The authors subsequently concluded that CT and MRI were equally effective in detecting LN metastasis in PCa staging. In addition to size, shape and internal structure are essential and enhancement can be used for assessment. DWI is not accurate, there were false-negative findings in small LNs and false-positives even in larger LNs due to inflammatory changes (22). The newest and most advanced method in MRI is the use of ultra-small superparamagnetic iron oxide particles for LN staging. Ferumoxtran-10 is a popular iron oxide particle that is unfortunately not yet commercially available and has not been approved by the European Medicines Agency and Food and Drug Administration, but it has shown very good results in research, with a correct diagnosis of 93% of all metastases in a study with histologically proven PCa (23). Different PET radiopharmaceuticals, such as ¹⁸F-choline, 11C-choline, and ¹⁸F-fluciclovine, have been evaluated and show similar specificity but lower sensitivity to MRI/CT. Evangelista et al. (24) reported a concordance of 49.2% (95% CI=39-58%) and a specificity of 95% (95% CI=92-970) for ¹⁸F/¹¹C-choline. However, heterogeneity varies between 22.7% and 78.4%. Selnaes et al. (25) found a sensitivity of 40% and a specificity of 87.5% for ¹⁸F-fluciclovine. However, EAU guidelines clearly state that choline PET/CT should not be used to evaluate LN metastasis due to its low sensitivity and inability to provide clinical evidence for the diagnosis of LN metastasis (26). The use of PSMA has experienced rapid growth since its development. Various PSMA radiopharmaceuticals are currently used. Different studies have evaluated the use of PSMA in nodal staging of PCa. The ProPSMA test showed an overall sensitivity of 85% (95% CI=74-96%) and a specificity of 98% (95% CI=95-100%) for ⁶⁸Ga-PSMA-11 in the preoperative phase of high risk. It is superior to normal staging in terms of LN and distant metastasis (27). ¹⁸F-PSMA-1007 has been used more frequently since its launch in 2017, but baseline evidence for its use in the early stage of PCa is still lacking. In summary, PSMA PET/CT was compared to imaging modalities, such as mpMRI, abdominal contrast-enhanced CT, or choline PET/CT; however, the resolution of PET (approximately 5 mm) limits the detection of small LN metastases (28).

Detection of local recurrence. Detection of local recurrence of PCa is crucial for determining prognosis, guiding treatment decisions, and optimizing treatment. PSA remains the most widely used test to assess response to treatment, including radiation therapy (RT); however, PSA values alone do not necessarily indicate metastatic disease and delay treatment decisions (29). 68Ga-PSMA PET/CT appears promising as a diagnostic tool in biochemical relapse of PCa but, to date, there is no clear conclusion regarding the correct diagnosis of local recurrence of PCa after RT and radical surgery. Carvalho et al. (30) detected pelvic recurrence with Ga-PSMA PET/CT after radical prostatectomy (RP) (pre-PET PSA ≥0.8 ng/ml) and RT (pre-PET PSA ≥2.3 ng/ml). The rates were 31% and 63.6%, respectively. In the setting of patients underwent RP, urinary excretion of choline- or Ga-PSMA obscures local recurrence in the prostatic fossa due to the superimposed activity in the urinary bladder. In the setting of post-RT, mpMRI is a good method for diagnosing local recurrence of PCa. The average sensitivity for detecting local recurrence of PCa was reported as 84% with specificity of 92%. Sensitivity and specificity after RT ranged from 72% to 100% and 85% to 100%, respectively. Recently, the Prostate Imaging for Relapse Reporting (PI-RR) test was developed to standardize mpMRI acquisition and reporting and to provide a reproducible tool for detecting PCa local recurrence in patients with biochemical relapse after RT or RP (31). PI-RR provides a 5-point scoring system for RT and RP, and the actual probability of recurrence for each. The PI-RR system, validated through face-to-face and online interviews, is not based on advanced scientific evidence. It evaluated DCE as a routine for local recurrence of PCa clinical trials but did not report the site of recurrence after RP. This is important for choosing the best treatment (32). In addition, the generalizability of results to the community setting is limited by the experience of radiologists. bpMRI-based PI-RADS gives DWI/apparent diffusion coefficient (ADC) an important role in detecting recurrence in the prostate gland both in peripheral and in transitional zones after RT and in the prostatectomy bed after RP (32, 33). PI-RRADS provides four diagnostic categories and assigns a probability of occurrence to each category, ranging from very low to very high. PI-RRADS groups 1 and 2 (negative for recurrence in the prostate bed and prostatectomy after RT and RP, respectively) have a very low and low recurrence rate, respectively. Areas with low signal intensity on the ADC map were assigned to PI-RRADS category 3 (possible recurrence), while areas with low signal intensity on the ADC map were assigned to PI-RRADS category 4 (high probability of recurrence) (32). Quantitative analysis of grayscale ADC images improves the discrimination between class 3 and class 4, making the system reliable and reproducible. After RT, T2w allows the detection of suspicious areas in the prostate using a 41-sector map and assesses capsular involvement. After RP, T2w enables localization of the lesion using the vesicourethral clock in the axial plane and the inferior border of the pubic symphysis in the sagittal plane as landmarks. The exact location of the wound is important to guide treatment decisions and the success of salvage therapy (34).

The importance of mpMRI in PCa management. Kumar et al. observed that mpMRI plays an important role in the evaluation of PCa using tissue contrast of prostate sequences (35). PCa can be localised, detected and isolated using these tests, which provide information about anatomy and function. The authors also noted that use of this device reduced unnecessary biopsies and prevented adverse outcomes, which was also supported by the studies by Stabile et al. (36) and Bosen et al. (37). In their review, they concluded that mpMRI has a high accuracy for PCa. The American Urological Association recommends using mpMRI before biopsy because it is more sensitive and specific than PSA in screening for PCa. Before mpMRI, PSA was used to determine whether patients needed a biopsy; some patients were not unnecessarily exposed to biopsy risks, such as hematuria, sepsis, and urine (38). This imaging technique can also be used to analyze biopsies; In this case, mpMRI can be used in three types: Cognitive fusion, ultrasound-MRI fusion, and MRI-MRI fusion (also known as core biopsy). Compared with random biopsy, mpMRI targeted biopsy increases the PCa detection rate (from 21% to 43% in some studies) (36, 38). According to the PROMIS study, mpMRI has higher sensitivity (93% vs. 48%) and higher negative predictive value (89% vs. 74%) than transrectal ultrasound-guided biopsy (39). Another advantage of using mpMRI biopsy is that it reduces the number of biopsy cores needed. In patients who need to be re-evaluated after biopsy for any reason, it is recommended that the next mpMRI be performed at least 6 weeks after the biopsy because swelling and bleeding will be considered artifacts and will lead to negative results (10). Regarding AS, Kumar et al. (35) stated that mpMRI is not included in the clinical process but using mpMRI for prostate evaluation before blood tests may allow patients to receive additional treatment for AS. Some studies have shown that it can effectively detect PCa and is less painful, so it can be used as a tool for monitoring AS (35, 36). Researchers have stated that the use of mpMRI in follow-up will reduce the number of biopsies required annually by patients if the disease is suspicious, without increasing PI-RADS (10). Another application of mpMRI is local treatment, which provides a good understanding of the area where an abnormality is located. Another important role worth mentioning is the management of RP (40). mpMRI is indicative of relapse when vascularization or new changes occur in previously operated areas. One of the major limitations of mpMRI is that MRI may miss some clinically important prostatic lesions; therefore, concerns remain regarding the use of mpMRI as a PCa screening tool (31, 37). For patients, absolute/relative contraindications to mpMRI include incompatible implants, severe claustrophobia, and previous severe reactions to gadolinium-containing contrast agents (37).

bpMRI accuracy and prostate cancer detection rates. Most research articles on this topic reach the same conclusion: bpMRI can replace mpMRI in diagnosing csPCa with nearly the same accuracy in detecting abnormalities, with the exception that any PI-RADS 3 therapy that must be followed by DCE (7, 12). A DCE sequence is required because a PI-RADS 3 lesion correlates equally with the likelihood of developing csPCa. Authors note that use of DCE for PI-RADS 3 may change the PI-RADS 4 score if improvement is evident. DCE sequences are also required for extracapsular expansion, as described by Caglic et al. (41) in their study. A 2018 meta-analysis also sought to determine the overall accuracy of bpMRI in diagnosing PCa. Niu et al. (5) described a total of 33 studies evaluating bpMRI and PCa. In selected published studies, overall bpMRI sensitivity and specificity for each Gleason grade group were 81% and 77%, respectively. Bass et al. (42) updated this meta-analysis in 2020 to strengthen our understanding by providing information on csPCa. Among 44 studies, the sensitivity and specificity of bpMRI for clinically significant csPCa were 87% and 72%, respectively. Recently, Cuocolo et al. (43) conducted their own meta-analysis of 17 studies including 3,964 patients. These meta-analyses demonstrate the high accuracy of bpMRI for PCa and clinically significant cancer. Kuhl et al. (44) analyzed PCa rates in 542 men who underwent bpMRI for elevated PSA. The distribution of PI-RADS 3-5 versus PI-RADS 1-2 lesions in their groups was 36.7% and 66.3%, respectively. PI-RADS 3-5 lesions had true-positive, false-negative, and false-positive rates of 77%, 13%, and 9.5%, respectively. De Visschere et al. (45) took this work one step further by evaluating PCa rates for each PI-RADS v2 score. PI-RADS 3, 4, and 5 showed a 40%, 78.8%, and 93% increase in PCa risk, respectively. However, these figures may be overestimated, as the average PSA increased to 9.2 ng/ml in their group. Boesen et al. (37) published field reports examining whether bpMRI is a good screening tool. This study included 1,020 biopsy-naïve men with PSA ≥4 ng/ml and/or abnormal DRE. The study found PCa in 64% of the cohort across all biopsies. These researchers evaluated all biopsy samples based on bpMRI PI-RADS scores. A total of 2.6% of these patients were diagnosed with csPCa in biopsies with PI-RADS scores of 1 and 2. PI-RADS 3, 4, and 5 were detected more frequently, with significant cancer diagnoses occurring in 13%, 39%, and 77%, respectively. Jambor et al. (46) also wanted to evaluate whether bpMRI could be used as a diagnostic tool; 175 men participated in their study, where two targeted biopsies were performed per region of interest. Biopsy decisions for men with relevant bpMRI findings resulted in 24% of patients receiving an inappropriate diagnosis. When using bpMRI as a secondary tool in patients with elevated PSA, this study urges physicians to diagnose only regions of interest with a PI-RADS score ≥3. Reducing the number of unnecessary biopsies allows patients to avoid complications while also saving costs associated with the procedure.

Guidelines regarding bpMRI. mpMRI recommended by current PI-RADS v2.1 guidelines includes multiplanar T1w and T2w images as well as sequential DWI and DCE images (47). Due to variability, the time required for mpMRI is long, approximately 30-45 minutes. Moreover, the use of gadolinium-based contrast agents brings with it the problem of fragility of blood vessels and the risk of various problems, such as nephrogenic systemic fibrosis or gadolinium accumulation in the brain (48). In PI-RADS v2.1, unlike previous guidelines, DCE-MRI has a smaller role in the evaluation of cancer (in the evaluation of peripheral disease, but no role in transition zone). To overcome these shortcomings, prostate bpMRI (T2w imaging and DWI only) is considered an option for PCa detection and evaluation; some studies have reported positive results (48, 49). However, there are currently no guidelines for using bpMRI over mpMRI. Some European studies have shown adverse outcomes of PCa when bpMRI is used. Clinicians may prefer bpMRI to mpMRI for a variety of reasons, including sensitivity differences, cost, or inability to obtain contrast for other reasons (e.g., end-stage renal disease) (49, 50). Internationally, bpMRI has had great usage in the European Union; most of the data on the sensitivity and specificity of bpMRI compared with mpMRI come from European researchers. One of the landmark publications regarding the accuracy of bpMRI comes from Denmark and concluded that bpMRI has excellent PCa detection rates (37).

Benefits of omitting DCE. In contrast to mpMRI protocols, bpMRI protocols do not include use of DCE. Therefore, it has three major advantages: Examination times are shorter, costs are lower, and the risk of adverse outcomes with contrast agents is removed (51).

Examination time: Omitting a MRI sequence reduces the examination time. There are wide differences in the literature on how much time can be saved by using a biparametric approach. Obmann et al. found a scanning time of 11.9 min for bpMRI and 45 min for mpMRI (52). This was confirmed by Lee et al., who found times of 15 min for bpMRI and 45 min for mpMRI. Other authors have found differences in short-term imaging between mpMRI and bpMRI (53). Junker et al. described a reduction of 12 min in examination time using bpMRI, whereas another group found a reduction of only 2.30 min (54). Kuhl and colleagues reported a very rapid bp MRI technique that took only 8.45 min (44).

Costs: DCE processing includes the direct costs of different media and physical equipment, as well as indirect labor costs and longer printing times. Data for costs diverge in the literature. Junker et al. showed additional costs of about $56 for a 70-kg patient for the contrast agent only (54). A Korean group estimated the cost of mpMRI at $600, while bpRI cost only $300. The cost potential of bpMRI compared to mpMRI was modeled to define what cost savings could be achieved if DCE were eliminated from these studies. Considering a 45-min period (allowing for one full mpMRI or three full bpMRI runs), Porter et al. found the total cost of a 45-min bpMRI was $1,531.32. Compared to mpMRI for PCa detection, total revenue increased by $892.58 for a 45-min period or $10,710.98 for a 9-h workday when performing bpMRI. The authors demonstrated that bpMRI can provide the best results compared with conventional mpMRI, increasing the use of MRI in the diagnosis and risk stratification of patients being assessed for PCa (53).

Risks using contrast agents. The contrast medium for DCE in mpMRI is gadolinium-based. Gadolinium-based contrast agents differ from iodinated and barium-based contrast agents in that they boost the signal intensity of biological tissues during imaging. This is achieved by reducing the alignment time of water protons with the magnetic field generated by the imaging machine (55). Additionally, the chelating agents in gadolinium-based contrast agents enable the substance to stay in circulatory vessels for a more extended period before extravasation compared to radiographic contrast agents. The immediate risk of hypersensitivity reactions to gadolinium-based contrast agents is low but recent discoveries indicate potential brain deposition, challenging the previous perception of gadolinium as being safe. This calls for caution in gadolinium management, prompting a reconsideration of the necessity for DCE procedures (56).