Imaging in Prostate Cancer: Magnetic Resonance Imaging and Beyond

Abstract

Imaging is becoming critical for guiding management decisions in prostate cancer (PCa) both at initial diagnosis and at recurrence. Multiparametric magnetic resonance imaging (mpMRI) and positron emission tomography (PET) of the prostate have proven valuable in the detection and localization of aggressive disease. Therefore, understanding the indications for mpMRI, imaging techniques and interpretation, limitations of current imaging approaches, and utility of PET and simultaneous PET/MRI has become increasingly important.

Indications for mpMRI

Screening for PCa is conventionally performed with serum prostate-specific antigen (PSA) testing and digital rectal exam (DRE), followed by transrectal ultrasound (TRUS)-guided biopsy, if indicated. In recent years, mpMRI has played a growing but primarily supportive role. Published guidelines from the European Association of Urology (EAU), American Urological Association (AUA), and National Comprehensive Cancer Network (NCCN) have made similar recommendations regarding the use of mpMRI for the diagnosis, initial staging, and detection of suspected recurrence of PCa. These guidelines are summarized in Table 1.^1–3

Table 1.

Professional society guidelines for use of MRI in prostate cancer.

	Diagnosis	Staging	Recurrence
EAU1 Year: 2016	“Before repeat biopsy, perform mpMRI when clinical suspicion of PCa persists in spite of negative biopsies.” “During repeat biopsy include systematic biopsies and targeting of any mpMRI lesions seen.”	“[In intermediate- and high-risk PCa], use prostate mpMRI for local staging. “In intermediate- and high-risk disease, use multiparametric MRI as a decision tool to select patients for nerve-sparing procedures.”	“[In patients with PSA recurrence after radiation therapy], perform prostate mpMRI only in patients who are considered candidates for local salvage therapy, use mpMRI to localize abnormal areas and guide biopsies.”
AUA3 Year: 2017	“MRI [potentially allows] pre-biopsy risk stratification for individualized decision-making. [….] Its use may be considered in men for whom clinical indications for biopsy are uncertain.” “[MRI] should be strongly considered in any patient with a prior negative biopsy who has persistent clinical suspicion for prostate cancer and who is undergoing a repeat biopsy.”	“Staging patients with prostate cancer using MRI to evaluate for possible lymph node metastasis can be considered in selected patients (T3/T4 and T1/T2 patients with nomograms predicting the risk of lymph node metastasis >10%.)” “mpMRI/TRUS may offer valuable staging information when performed prior to definitive local therapy.”	“It is possible that mpMRI may be useful for follow up evaluation of men treated with radical prostatectomy or in situ ablative therapies.”
NCCN2 Year: 2016	“Consider mpMRI if anterior and/or aggressive cancer is suspected when PSA increases and systematic prostate biopsies are negative [in patients with expected survival >10 years].”	“mpMRI may be used to better risk stratify men who are considering active surveillance.” “[Consider for] initial evaluation of high-risk patients [including those with] T3 or T4 disease [or those with] T1 or T2 disease and nomogram-indicated probability of lymph node involvement >10%”.	“MRI may be considered in patients after RP when [PSA becomes detectable], or after RT for rising PSA or positive DRE if the patient is a candidate for additional local therapy.”

EUA = European Assoication of Urology; AUA = American Urological Assocation; SAR = Society of Abdominal Radiology; NCCN = National Comprehensive Cancer Network; mpMRI = multiparametric magnetic resonance imaging; RT = radiation therapy; DRE = digital rectal examination; PSA = prostate-specific antigen; RP = radical prostatectomy.

At present, there is no evidence to support the use of mpMRI for population-based screening, and insufficient evidence to support its routine pre-biopsy use.⁴ There may, however, be a role for mpMRI in risk stratification for biopsy-naïve patients with equivocal indications for biopsy. Indeed, the utility of mpMRI prior to initial biopsy is the focus of two ongoing large-scale clinical trials in the United Kingdom: Prostate MR Imaging Study (PROMIS)⁵ and PRostate Evaluation for Clinically Important Disease: Sampling Using Image-guidance Or Not? (PRECISION).⁶ Furthermore, there is strong evidence to support the use of mpMRI in patients with persistent clinical suspicion for PCa despite a negative systematic biopsy. In these patients, targeting of any suspicious lesion identified on mpMRI at the time of repeat systematic biopsy can improve the rate of diagnosis of clinically significant cancers (Gleason score ≥ 7).⁷

Several approaches have been established for targeting suspicious lesions identified on mpMRI. One approach relies on visual approximation, in which the clinician performing the biopsy reviews the pre-procedural mpMRI data and utilizes TRUS to target the expected location of the lesion.⁸ Although this approach does not provide real-time feedback, it is easy to implement without much modification to standard workflow. Another approach, which is currently utilized at our institution, employs MRI/TRUS fusion. In this technique, suspicious lesions are delineated on anatomic MRI data and projected onto real-time ultrasound images to assist in lesion targeting.⁹ Registration is achieved by using the contour of the prostate gland, which is also delineated on the anatomic MRI data, as an internal fiducial marker. This approach has been shown to be more accurate for sampling smaller lesions and provides significantly more informative histologic results on the presence of cancer compared to visual targeting.⁹ However, this MRI/TRUS fusion approach requires image acquisition and biopsy to be performed at the same institution to ensure accurate co-registration and targeting. Finally, real-time MRI-guided biopsy, which is performed within the MRI bore itself, has the potential to provide the most accurate targeting of suspicious lesions.¹⁰ This approach, however, requires longer procedure times, more expensive equipment, and can be uncomfortable for the patient.

Once a tissue diagnosis has been made, mpMRI has an established role in the initial staging of intermediate-risk and high-risk PCas. The added value of mpMRI in this setting derives from its ability to detect extraprostatic extension, seminal vesicle invasion, and locoregional lymph node involvement. Furthermore, there is moderate evidence to support the use of mpMRI in selecting candidates for nerve-sparing radical prostatectomy and identifying patients suitable for active surveillance.¹¹ In contrast, the role of mpMRI in detecting recurrent disease in treated patients is less clear. mpMRI should be considered for the evaluation of biochemical recurrence (BCR) after radiation therapy, in patients that are candidates for local salvage therapy. Similarly, mpMRI may help differentiate recurrence within the prostatectomy bed from nodal, osseous, or visceral metastatic disease in patients with rising PSA levels after radical prostatectomy, thereby guiding decisions regarding further treatment. Overall, additional research is still needed to define the proper role of mpMRI in the assessment of recurrent disease.

mpMRI Technique

The Prostate Imaging Reporting and Data System (PI-RADS) is a structured reporting system developed by the European Society of Urogenital Radiology (ESUR) in 2012 to standardize prostate MRI acquisition and interpretation. It was subsequently revised in 2015 (PI-RADS v2) by the American College of Radiology (ACR) and ESUR.¹²

PI-RADS v2 specifies that mpMRI should include T1-weighted (T1W) images, high-resolution T2-weighted (T2W) images, diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) maps, and dynamic contrast-enhanced (DCE) imaging.¹² When possible, DWI should include high b-value (≥1400 sec/mm²) acquisitions. T1W images are most helpful for identifying intraprostatic hemorrhage and skeletal lesions. High-resolution T2W images are important for localizing tumors within the prostate gland and detecting extraprostatic extension (including seminal vesicle invasion). DWI is also valuable for tumor identification. Furthermore, the degree of diffusion restriction positively correlates with tumor aggressiveness and Gleason scores.¹³ DCE imaging provides an assessment of lesion vascularity, which may aid in the detection of recurrent tumor after therapy.¹⁴ However, the information provided by DCE is of limited impact in the most current diagnostic algorithm, especially in treatment-naïve patients.

Patient Considerations

Because hemorrhage in the post-biopsy setting may confound mpMRI assessment, an interval of at least 6 weeks between biopsy and mpMRI is advisable. Administration of an antiperistaltic agent (e.g., glucagon) prior to the study may reduce artifact related to rectal motility, especially when using an endorectal coil. To minimize artifacts related to gas in the rectum, a pre-procedural enema may be beneficial. Importantly, the use of high field strength MRI systems (i.e. 3 Tesla [3T]) is preferable to maximize the signal-to-noise ratio. However, in circumstances when it is necessary to image the patient at lower field strength (1.5T), use of an endorectal coil is essential. Such circumstances include the presence of hip arthroplasties, which can create significant artifacts, or certain implanted medical devices, which may not be safe for imaging at 3T. Direct discussions between the radiologist and referring clinician are advised in this setting.

Imaging Interpretation

The Prostate Imaging Reporting And Data System, Version 2 (PI-RADS v2) includes five assessment categories based on the probability that an individual lesion corresponds to a clinically significant cancer, defined as a tumor with Gleason score ≥7: PI-RADS 1 – very low; PI-RADS 2 – low; PI-RADS 3 – intermediate; PI-RADS 4 – high; and PI-RADS 5 – very high. The final PI-RADS category is determined by a composite of findings on DWI, DCE, and T2W images, with different emphases placed on specific sequences for the peripheral (PZ) versus transition zones (TZ). Up to four lesions with an assessment category of 3 or greater may be reported, and a standardized sector map should be utilized to assist in biopsy targeting. Reporting lesions with a PI-RADS score of 1 or 2 is optional.

PI-RADS v2 characterizes lesions in the PZ differently from lesions in the TZ. Assessment of lesions in the PZ relies almost entirely on its appearance on DWI. Contrast enhancement is useful for PZ lesions in cases where characterization of the lesion on DWI is equivocal. A PI-RADS 5 lesion in the PZ, which corresponded to histologic Gleason 4+5 prostate adenocarcinoma on biopsy, is shown in Figure 1. Conversely, assessment of lesions in the TZ relies predominantly on the T2W score. Contrast enhancement is not used for evaluation of TZ lesions; however, DWI can be used to further characterize equivocal TZ lesions. A PI-RADS 5 lesion in the TZ, which corresponded to histologic Gleason 4+4 prostate adenocarcinoma on biopsy, is shown in Figure 2. In general, PI-RADS 5 lesions in the PZ or TZ are characterized by either a minimum diameter of 1.5 cm or signs of invasive disease.

Figure 1. Peripheral zone prostate adenocarcinoma.

Within the right lateral peripheral zone in the mid-portion of the gland is an area of focal marked hyperintensity on high b-value DWI, corresponding to marked hypointensity on the ADC map and measuring greater than 1.5 cm in size (DWI score of 5; arrows). On the T2W images, there is a circumscribed, homogenously hypointense focus measuring greater than 1.5 cm and corresponding to the abnormality on DWI (T2W score of 5; arrow). This lesion demonstrates focal early hyperenhancement on DCE images (DCE positive; arrow). The overall PI-RADS score was 5, and biopsy demonstrated Gleason 4+5 adenocarcinoma. The patient was treated with radiation and hormonal therapy.

Figure 2. Transitional zone prostate adenocarcinoma.

Within the central anterior transition zone in the mid-portion of the gland, there is a lentiform, homogenously hypointense focus on T2W images measuring greater than 1.5 cm, (T2W score of 5; arrow). There is a corresponding focal marked hyperintensity on high b-value DWI and a marked hypointensity on the ADC map measuring greater than 1.5 cm (DWI score of 5, arrows). DCE is positive (arrow). The overall PI-RADS score was 5. Biopsy revealed Gleason 4+4 disease, and the patient was treated with transurethral resection of the prostate.

Staging information, including extraprostatic extension, involvement of adjacent structures (i.e., neurovascular bundles, external urethral sphincter, seminal vesicles, and bladder wall), enlargement of locoregional lymph nodes (>8 mm in the short axis), and suspicious osseous lesions, should be reported.

Limitations of Current Imaging Techniques

mpMRI has several important limitations. Notably, the sensitivity for detection of PCa in the TZ is limited by the intrinsically heterogeneous nature of the transitional zone in the setting of benign prostatic hyperplasia (BPH). BPH nodules can exhibit high cellularity, thus mimicking cancerous lesions on MRI sequences, specifically DWI.¹⁵ Hence, morphologic features on T2W images are more reliable than DWI for TZ assessment, a concept reflected in the PI-RADS v2 scoring system. Various other benign and pre-malignant processes (e.g., granulomatous prostatitis, adenosis, prostatic intra-epithelial neoplasia) can also mimic PCa on mpMRI.¹⁵ Furthermore, mpMRI is also limited by inter-observer variability. Although PI-RADS v2 was specifically developed to improve inter-observer reliability, sensitivity, and specificity, further revisions will likely be necessary to optimize agreement among different readers and to limit the use of subjectively applied features.¹⁶ Additionally, the main determinant of the overall PI-RADS score in the PZ is DWI, which is intrinsically susceptible to artifact from rectal gas and metal implants. As previously mentioned, the ability to generate sufficient signal on high b-value DWI requires either imaging at 3T or use of an endorectal coil at 1.5T, both of which may be of limited availability in general radiology practices. Finally, mpMRI has a limited ability to discriminate between post-treatment change and local recurrence following prostatectomy, local ablative therapy, and/or radiation therapy.

New PET Radiopharmaceuticals

Oncologic PET/CT, most commonly performed with the glucose analogue [¹⁸F]-fluorodeoxyglucose (FDG), has become standard of care for the initial staging and subsequent treatment response assessment for many different malignancies. However, FDG-PET/CT has a limited role in PCa imaging. Although FDG-PET/CT may be useful in the initial staging of tumors with aggressive features on pathology, many PCas are not FDG-avid.¹⁷ Consequently, there is considerable interest in non-FDG PET tracers with better diagnostic performance for detecting PCa metastases and localizing sites of disease in patients with biochemical recurrence. Furthermore, such PET tracers may have the ability to interrogate the biological pathways underlying PCa behavior, potentially influencing selection of treatment regimens.

There are three FDA-approved non-FDG PET tracers currently available for PCa imaging at Washington University in St. Louis/Barnes-Jewish Hospital ([¹⁸F]-NaF, [¹¹C]-choline, [¹⁸F]-fluciclovine). Additional PET tracers targeting the prostate specific membrane antigen (PSMA) are likely to enter clinical use in the near future. The mechanisms of targeting, imaging indications, strengths, and weaknesses of these agents are described in detail in Table 2. For imaging osseous metastatic disease, [¹⁸F]-NaF-PET/CT outperforms conventional [^99mTc]-methylene diphosphonate (MDP) SPECT bone scans although [¹¹C]-choline-PET/CT provides a greater degree of specificity for osseous metastatic disease (Figure 3).¹⁸ [¹¹C]-choline-PET/CT and [¹⁸F]-fluciclovine-PET/CT have the advantage of being able to image nodal and visceral metastatic disease in addition to osseous disease (Figure 4), with best diagnostic performance in the setting of biochemical recurrence.^{19, 20} Although not yet FDA-approved, the prostate specific membrane antigen (PSMA)-targeting PET tracers ([⁶⁸Ga]-PSMA-HBED-CC and [¹⁸F]-DCFPyL) are likely more sensitive and specific for metastatic PCa than [¹¹C]-choline or [¹⁸F]-fluciclovine.^{21, 22} These PSMA-targeting PET tracers may also add diagnostic value in the setting of initial staging. Overall, though more research is needed to define their optimal clinical roles, these non-FDG PET tracers have already begun to revolutionize the imaging of PCa.

Table 2.

PET Tracers for Prostate Cancer

PET tracer	Indications	Strengths	Weaknesses
[18F]-NaF	Evaluation of known or suspected osseous metastatic disease 18	Faster, higher resolution imaging than conventional bone scans, with better diagnostic accuracy 18	Time intensive interpretation due to high sensitivity for benign processes (e.g., degenerative disease)
[11C]-choline	Identification of recurrent disease in patients with rising PSA after completion of therapy 19	Potential for detecting spread to bones, lymph nodes, and distant organs 19	Prohibitively low disease detection rates for initial staging; uptake by other cancers and benign processes
[18F]-fluciclovine	Same as [11C]-choline	Same as [11C]-choline, though with better recurrence detection rates 20	Same as [11C]-choline
[68Ga]-PSMA-HBED-CC	To be determined; possible roles in initial staging, suspected disease recurrence, and treatment response assessment	Likely greater sensitivity and specificity for prostate cancer than [11C]-choline or [18F]-fluciclovine 21	Not yet FDA-approved
[18F]-DCFPyL	Same as [68Ga]-PSMA-HBED-CC	Same as [68Ga]-PSMA-HBED-CC, though with longer half-life and better spatial resolution 22	Same as [68Ga]-PSMA-HBED-CC

Abbreviations: PSA – prostate specific antigen; PSMA – prostate specific membrane antigen

Figure 3.

An 85-year-old man with prostate cancer treated with radiation therapy several years prior presented for ¹⁸F-NaF-PET/CT (A–D) due to a rising PSA. Transaxial CT images (A, C) with PET fusion (B, D) showed a focus of tracer uptake in the left pedicle of T1 without a definite anatomic correlate (arrowheads). A similar focus of tracer uptake in the right lamina of T1 (arrows) corresponded to sclerosis on CT. These findings were interpreted as consistent with degenerative change, with no definite evidence of osseous metastatic disease. A ¹¹C-choline-PET/CT (E–I) was subsequently performed for localization of disease recurrence. A maximum intensity projection (E) demonstrated normal physiologic uptake in the liver, spleen, pancreas, kidneys, stomach/intestines, and salivary glands. However, a suspicious focus of activity was noted in the upper thoracic spine (arrowhead). Mild uptake in several mediastinal lymph nodes (arrow) was felt to be inflammatory in etiology. Transaxial CT images (F, G) with PET fusion (G, I) showed a focus of tracer uptake in the left pedicle of T1 (arrowhead) corresponding to the site of increased ¹⁸F-NaF uptake, concerning for a prostate cancer metastasis (later confirmed by biopsy). In contrast, there was no abnormal uptake in the right lamina of T1, indicating that the tracer activity seen at this site on ¹⁸F-NaF-PET/CT instead represented degenerative change. This case demonstrates that ¹⁸F-NaF has excellent sensitivity but lower specificity for prostate cancer osseous metastasis and that the differentiation of metastasis from degenerative change can be challenging. Other PET tracers, such as ¹¹C-choline, may provide a higher level of specificity for osseous metastatic disease.

Figure 4.

A 64-year-old man with prostate cancer (Gleason 5 + 4) presented for ¹⁸F-fluciclovine-PET/CT due to rising PSA levels 8 months after radical prostatectomy. A maximum intensity projection (A) showed normal physiologic tracer activity in the salivary glands, liver, and pancreas, with a small amount of activity in the left renal pelvis (arrow) and several suspicious foci in the pelvis (arrowheads). Transaxial CT images (B) with PET fusion (C) near the prostatectomy bed revealed activity in a tissue lesion (arrowheads), consistent with local recurrence. Additionally, more superiorly, transaxial CT images (D) with PET fusion (E) demonstrated an intense focus of tracer avidity in the right sacrum corresponding to a subtle sclerotic lesion (arrows), suspicious for osseous metastatic disease. A conventional 99mTc-MDP bone scan 1 month prior had been normal (images not shown). Percutaneous biopsy of the sacral lesion confirmed metastatic prostate cancer. This case illustrates the benefits of ¹⁸F-fluciclovine for localizing both osseous and non-osseous sites of recurrent disease.

Combined PET/MRI

Simultaneous PET/MRI systems combine the respective advantages of PET and MRI into a single imaging study. These systems have been in use in the United States since 2011 and are currently available for clinical use at Washington University in St. Louis. In the context of PCa, PET is ideal for identifying locoregional nodal and distant metastatic disease, whereas mpMRI provides high-resolution images with excellent soft tissue contrast for lesion localization within the prostate and staging. The results of early studies evaluating simultaneous PET/MRI suggest benefits for preoperative tumor identification, treatment planning, and localizing suspected recurrence (Figure 5). Simultaneous [⁶⁸Ga]-PSMA-PET/MRI has been shown to improve both pre-prostatectomy tumor localization and diagnostic accuracy compared with either modality alone.²³ Likewise, simultaneous [¹⁸F]-choline-PET/mpMRI outperformed mpMRI alone with respect to sensitivity and accuracy for the initial diagnosis of PCa.²⁴ In the setting of biochemical recurrence, [⁶⁸Ga]-PSMA-PET/MRI outperformed [⁶⁸Ga]-PSMA-PET/CT for the detection of local recurrence in the prostatectomy bed.²⁵ Bone-specific PET tracers (e.g., [¹⁸F]-NaF) may improve the sensitivity of whole-body MRI for osseous metastatic disease. Overall, further work is needed to make novel PET tracers more readily available and to optimize PET/MRI workflows.

Figure 5.

A 63-year-old man with prostate cancer (Gleason 3 + 4) status-post radical prostatectomy with subsequent recurrence in the surgical bed (treated with salvage radiation therapy) presented for a restaging ¹¹C-choline- PET due to a rising PSA. Given its superior soft tissue contrast and the patient’s prior recurrence in the surgical bed, PET/MRI was selected instead of PET/CT. Contrast-enhanced transaxial T1-weight images (A, D) with PET fusion (B, E) and transaxial diffusion-weighted images (C, F) are provided. There was no abnormal enhancement, diffusion restriction, or tracer activity in the surgical bed (images not shown). A mildly enlarged right common iliac lymph node (arrowheads) and a prominent right external iliac lymph node (arrows) had similar enhancement (A, D) and diffusion (C, F) characteristics. However, the right common iliac lymph node demonstrated intense tracer uptake, whereas tracer activity in the right external iliac lymph node was similar to background. These findings were consistent with a right common iliac nodal recurrence, and the patient was empirically treated with radiation therapy to this area. This case demonstrates the value of the amino acid PET tracers for identifying nodal sites of recurrence, which can be challenging to differentiate from normal lymph nodes on MRI alone.

Conclusions

Prostate mpMRI plays an important role in the management of patients with known or clinically suspected PCa, and its strength lies in its ability to detect and localize clinically significant cancers. Current indications for mpMRI include persistent clinical suspicion for PCa despite negative systematic biopsy and initial staging of intermediate-risk and high-risk PCas. Multiple clinical trials are underway to assess the utility of mpMRI prior to initial prostate biopsy. Furthermore, mpMRI may play a role in identifying patients suitable for active surveillance. Although the role of mpMRI in biochemical recurrence is less clear, PET/CT and PET/MRI both show significant potential in localizing sites of active disease after initial treatment. Multiple promising PET tracers are already FDA-approved, though further work is needed to define their optimal roles in guiding clinical management.

Biography

Daniel R. Ludwig, MD, Radiology Resident, Tyler J. Fraum, MD, Radiology Resident, Kathryn J. Fowler, MD, Director, Abdominal and Pelvic MRI, and Joseph E. Ippolito, MD, PhD, (above), Medical Director, Center for Clinical Imaging Research, are at Washington University School of Medicine, St. Louis, Missouri.

Contact: ippolitoj@mir.wustl.edu