Hybrid MR and PET Imaging for Prostate Cancer Recurrence

Abstract

Purpose:

Recurrence post definitive local therapy by prostatectomy or radiation therapy is often detected via rise in serum prostate specific antigen (PSA) levels; however, PSA rise does not localize the disease. Distinguishing local versus distant recurrence guides whether to choose subsequent local versus systemic therapy. The purpose of the article is to review imaging for prostate cancer recurrence post local therapy.

Recent Findings:

Among imaging modalities, multiparametric magnetic resonance imaging (mpMRI) is commonly used to assess for local recurrence. New radiopharmaceuticals target prostate cancer cells and enable whole-body imaging. These tend to be more sensitive for lymph node metastases than MRI or computed tomography (CT) and for bone lesions than bone scan at lower PSA levels but can be limited for local prostate cancer recurrence. Given greater soft tissue contrast, similar criteria for lymph nodes, and greater sensitivity for prostate bone metastases, MRI is advantageous to CT. MRI of the whole body and mpMRI are now feasible within a reasonable time frame and complimentary to positron emission tomography (PET) imaging, enabling whole-body and pelvis-focused PET-MRI, which should be advantageous in the setting of recurrent prostate cancer.

Summary:

Hybrid PET-MRI with prostate cancer targeted radiopharmaceuticals and whole body plus local multiparametric MRI can be complimentary for detecting local and distant recurrence to guide treatment planning.

Introduction

Prostate cancer is the second most common solid tumor malignancy in men worldwide [1], and the most common amongst men in the United States [2,3]. While advances in treatment and screening have improved the mortality rate, prostate cancer remains the second-leading cause of death from cancer amongst men in the United States. Treatment for clinically significant locally presenting disease primarily involves radical prostatectomy usually with lymph node dissection, or external beam or brachy-radiation therapy with or without hormone therapy. Post prostatectomy, 20–40% of patients develop biochemical recurrence within 10 years with a median time of 20–38 months [4–6]. After radiation therapy, biochemical recurrence is noted in 15–25% of patients with low-risk disease and 30–70% of patients with high-risk disease [7–10].

Given different approaches that may be taken for local versus distant recurrent disease treatment, and the high morbidity associated with salvage prostatectomy, salvage radiation therapy, and systemic treatment; identifying local recurrence, as well as nodal, bony, or more distant metastases, is critical. Biochemical recurrence refers to rising prostate-specific antigen (PSA) levels post local treatments and is most often the first sign of relapse preceding clinically apparent recurrence [6]. However, rising PSA levels are not always sufficient for assessing recurrence as some patients may have non-PSA- producing tumors, but more importantly, the blood test does not anatomically localize the recurrence for therapeutic decisions [11*]. Clinically, there is a need to distinguish local vs distant recurrent disease and to guide local therapies. Thus, imaging is needed.

CT or MR are the most widely used modalities for cross sectional pelvic and abdominal imaging. For local recurrence, MRI including multiparametric MRI provides superior soft tissue contrast [12,13]. CT and MR may be used to detect lymph nodes, but detection is based on size criteria with limited sensitivity and specificity. Mineralized bone seeking nuclear medicine agents such as ^99mTc-methylene diphosphate (MDP) and ¹⁸F sodium fluoride (NaF) may be used for assessing host response to tumor in the form of bone turnover for whole body evaluation of bone metastases but do not evaluate soft tissues. MR primarily images the soft-tissue component of the tumor and is more sensitive than CT for bone metastases; the latter primarily evaluates mineralized bone. New tumor seeking radiopharmaceuticals have become available for whole body imaging and tend to be more sensitive than CT or MR for lymph node metastases and are still being compared for bone involvement; however, these agents tend to be less sensitive than mpMRI for local recurrence. The sensitivity of these tests is limited at lower serum PSA values. For the newer nuclear medicine agents, imaging is commonly done by fusion with CT. However, given the greater soft tissue contrast in the pelvis, similar criteria for lymph nodes, and greater sensitivity for bone metastases, plus potential for local and whole-body imaging in a reasonable time period, fusion with MR, i.e., targeted PET/MR, should be considered for imaging prostate cancer recurrence.

Local Recurrence

While a variety of modalities have been pursued for evaluation of local recurrence, mpMRI (Fig 1) has emerged as the most accurate and sensitive imaging modality [14*,15]. This includes multiple parameters or sequences including T2-weighted imaging, diffusion weighted imaging (DWI) and dynamic contrast enhanced imaging (DCE). Simplistically, DWI reflects the cellularity of the tissue and DCE reflects abnormal (angiogenic) tumor vasculature. With radiation or hormone therapy, T2 and apparent diffusion coefficient (ADC, a quantitative measure of DWI) signal decreases in the peripheral zone, which can confound interpretation. Among the mpMRI sequences, DCE imaging has been suggested to be the most sensitive for recurrence post prostatectomy [16]. However, each mpMRI sequence may be contributory for detecting recurrence.

Fig. 1.

Recurrence (arrow) at the anastomosis site post prostatectomy seen as an area of increased signal on T2-weighted imaging (A) and on diffusion-weighted imaging (B), as well as enhancement on dynamic contrast enhanced MR imaging (C).

A systematic review by Sandgren et al. noted 90% sensitivity for detecting local recurrence using mpMRI in patients with mean PSA of <1 ng/ml when biopsy and salvage radiation therapy were used for follow-up [17]. Comparing mpMRI and ¹¹C-choline PET/CT, Kitajima et al. reported 52.4% of local recurrences were correctly diagnosed by both modalities whereas 36.1% were correctly diagnosed by only mpMRI; they reported mpMRI sensitivity, specificity, and accuracy for diagnosing local recurrence were 88.5%, 84.6%, and 87.4%, respectively [18]. Venkatesan et al. found in a clinical setting that mpMRI is very likely to be negative if PSA is <0.5 ng/ml for low Gleason grade tumors but not high grade, whereas, for high grade tumors mpMRI is almost always positive if PSA >1.5 ng/ml, suggesting that between these values mpMRI is more likely to be diagnostic; and, in the latter scenario, mpMRI is more likely to be of value for guiding biopsy [19].

Recently, a consensus guideline in the vein of PIRADS was published for mpMRI in the recurrence setting [14*]. It suggested that post radiation therapy, DCE or DWI be used as the primary sequence, whichever is higher and the other be used to upgrade from 4 to 5. Laterality on the same side versus opposite side as the original lesion and diffuse versus focal shape are used in the grading categorization. Post prostatectomy, the consensus guideline suggested using DCE as the primary sequence and DWI for upgrading. Of note, both DWI and DCE can be degraded by susceptibility artifact such as clips to a greater extent than T2-weighted imaging.

Other functional MRI methods that may also find value in the future for detecting prostate cancer recurrence, especially post radiation, are spectroscopy and hyperpolarized ¹³C-pyruvate MR imaging, which look to evaluate metabolic activity.

Imaging also usually includes large fields of view of the pelvis for evaluation of pelvic lymph nodes and bones. In addition to T1-weighted imaging, DWI imaging can help identify lymph nodes and bone lesions. Post lymph node dissection or pelvic irradiation, it is important to also look outside (i.e., superior to) the treatment field for lymph node recurrence.

PET Imaging

Several prostate tumor targeting radiopharmaceuticals have been evaluated for imaging prostate carcinomas by PET/CT. These have different mechanisms for targeting but improve detection compared to more traditional PET agents. For example, ¹⁸F-choline had a higher sensitivity (73%) than ¹⁸F-FDG (53%) in detecting prostate cancer preoperatively [20,21]. Such agents have been tested for imaging prostate cancer recurrence. Utilization of such prostate cancer-targeting agents have primarily been restricted to certain institutions due to availability issues, but this should start to change given now that FDA approved agents are coming online.

¹¹C-Choline and ¹⁸F-Fluciclovine PET Imaging

¹¹C-choline is phosphorylated by choline kinase overexpressed in prostate cancer cells [22] and subsequently is incorporated into the cell membrane as phosphatidyl choline. Over the past decade, ¹¹C-choline PET/CT has been evaluated as an additional tool for initial diagnosis of prostate cancer [23] and for recurrence. ¹¹C-choline has been available for a decade under IND but is limited by its short half-life requiring a nearby cyclotron. This has been addressed by ¹⁸F-choline. An advantage of using ¹⁸F-choline is that uptake of radiolabeled choline reflects a basic metabolic process not directly regulated by androgen signaling. In contrast, prostate specific membrane antigen (PSMA) expression is negatively regulated by androgens, which may limit its utility for following therapy but not necessarily for detecting recurrence (24). Although a meta-analysis has suggested sensitivity of 85–89% and specificity of 88–93% for ¹¹C-choline detecting recurrence, sensitivity decreases in patients with biochemically recurrent prostate cancer after prostatectomy with PSA levels less than 2 ng/ml, and particularly less than 1 ng/ml [23,25–30].

¹¹C-choline PET/CT is inferior to mpMRI for evaluation of local recurrence especially at low PSA levels <2.0 ng/ml. In a direct comparison study, ¹¹C-choline PET/CT was found to be superior to mpMRI for pelvic lymph node metastases (0.92 vs 0.81 ROC), inferior for local metastases (0.76 vs 0.91 ROC), and equivalent for pelvic bone metastases [18].

¹⁸F-Fluciclovine (Fig 2) is a radiotracer predominantly transported via ASCT2 and LAT1 amino acid transporters [31,32]. It was approved for clinical use in 2016. ¹⁸F-Fluciclovine has been reported to have a sensitivity and specificity of ~89% [27]. Nanni et al. reported a higher detection rate for recurrent prostate cancer using ¹⁸F-fluciclovine versus ¹¹C-choline PET/CT, with 60% additional lesions detected in 20% more patients [33]. It also has the advantage of lower background activity. A review found an average of 38%, 65% and 78% positivity rates for recurrence detection using ¹⁸F-fluciclovine PET/CT (across several publications), with PSA levels of <1.0, 1.0–2.0 and >2.0 ng/ml, respectively [34]. Comparing several studies, its detection rate was greater than ¹¹C-choline for PSA<2 ng/ml but less than for ⁶⁸Ga-PSMA-11. Limitations of ¹⁸F-fluciclovine PET/CT include intense physiological uptake in the liver and pancreas; urinary excretion that can limit evaluating the prostate/surgical bed region; and as with choline, low uptake in sclerotic bone metastases [14*,35,36]. The latter is thought to be due to low cellularity, but such lesions can be seen by CT as dense lesions and by MR as low signal on T1- and T2-weighted imaging.

Fig. 2.

Lymph node metastasis post prostatectomy. The lymph node (arrow) does not appear enlarged on T1-weighted MR (A) or CT (B), but has increased uptake on ¹⁸F-fluciclovine PET.

PSMA

⁶⁸Ga-PSMA-11 is an emerging radiopharmaceutical used in conjunction with PET/CT as a non-invasive technique to image prostate cancer with increased PSMA expression [37]. Another PSMA-targeted agent was approved in 2021, ¹⁸F-DCFPyl (Fig 3). Such agents bind the type II cell surface glycoprotein PSMA, and detect more prostate cancer recurrence at lower PSA values than ¹¹C-choline (PSA<2 ng/ml) or ¹⁸F-fluciclovine (PSA<1 ng/ml) [18,34]. At <1 ng/ml PSA values, recurrence detection of ~51–57% has been reported with these tracers [34,38]. Limitations of PSMA-targeting agents include that 5–10% of prostate cancers do not express PSMA [39]; and PSMA-targeting therapies may result in down-regulation of PSMA or selection of non-PSMA-producing tumors. Other radiotracers with different mechanisms of action may find utility in low or no PSMA-producing tumors such as ¹⁸F-fluciclovine [40*] and such lesions may also be detectable by MRI.

Fig. 3.

Recurrence (arrow) on the right side post brachytherapy (high attenuation dot-like appearing structures within the prostate). The lesion is not seen on CT (A) but is seen as an area of increased uptake on PSMA-targeted (¹⁸F-DCFPyL) PET-CT (B).

Bone metastases

Bone metastases may be visualized by MDP bone scan, Na¹⁸F, or prostate tumor targeted agents (Fig 4), all of which can be utilized for whole-body evaluation. Bone scan is rarely positive with PSA<5 ng/ml after radical prostatectomy or radiation therapy [41]. In the setting of biochemical recurrence, Jadvar et al. reported 43% sensitivity and 83% specificity for Na¹⁸F, but no lesion was seen with PSA <2 ng/ml [42].

Fig. 4.

Bone metastases (arrows). More lesions are seen in the ¹⁸F-fluciclovine PET than bone scan.

MR is more sensitive for bone metastases than CT but has traditionally been used for local imaging. A meta-analysis of 45 studies found ⁶⁸Ga-PSMA PET/CT had sensitivity of .91 and specificity of .96, whereas 3T MR which had sensitivity of .94 and specificity of .94 on a per patient basis [43*]. Thus, these methods can be confirmatory and may be complimentary such as if there is a PSMA negative metastasis. Network meta-analysis in the same paper suggested that on a per lesion basis ¹¹C-choline PET/CT had the highest diagnostic value with the highest superiority index, followed closely by ⁶⁸Ga-PSMA PET/CT, ¹⁸F-choline PET/CT, MRI, and ¹⁸F-NaF PET/CT [43*].

Whole body DWI to create an index of disease activity has been proposed but multisequence imaging has been limited by time on the scanner. Amorim et al. suggested that MR is superior to ¹⁸F-fluciclovine alone in detecting castration resistant bone metastases, but the combination is better [44]. Multisequence, multiplane, contrast enhanced whole body imaging that can be performed in a reasonable time frame of about an hour is now achievable using moving table techniques. This includes whole body DWI, which can aide not only bone lesion detection, but also highlights lymph nodes. DWI tends to be a sensitive but not specific sequence. Other sequences such as T1-weighted and contrast enhancement can aide lesion characterization and additional images can be derived without increased scan time.

PET/MR

mpMRI is most often used to evaluate locally recurrent prostate cancer and has utility particularly in the post-surgical bed or post-radiation therapy prostate [18]. Tumor targeting agents such as for PSMA have been used primarily for locating metastases to lymph nodes and bone. These agents are commonly combined with CT, however, CT has less soft tissue resolution than MR, particularly for pelvic imaging. In the skeleton, CT looks primarily at the mineralized bone response of host to the tumor whereas MRI sees the soft tissue component, i.e., tumor itself, and has greater sensitivity than CT, particularly in the axial skeleton for prostate bone metastases. MR can be superior to CT in evaluating bone lesions particularly when employing T1 weighted imaging with DWI [43*,45]. With the advent of faster sequences, multiplanar, multisequence MR can now be performed in a clinically reasonable time [46*,47]. This now enables lesion localization and characterization of not only lymph node and bone metastases but also other visceral lesions from prostate cancer metastases. CT is most often performed in conjunction with PET agents without intravenous contrast, further limiting lesion detection and characterization.

Increased tracer can be masked by the radiotracer accumulating in the bladder; however, MRI improves localization of focal radiopharmaceutical uptake identified on PET imaging. Using MR instead of CT reduces total irradiation of the hybrid PET study. It also increases patient convenience in that they have one integrated exam rather than two separate exams. Simultaneous acquisition should also improve registration of the PET and MR images. Thus, PET and MR should be complimentary in localizing prostate cancer recurrence locally and systemically. Both MR and CT depend on size criteria primarily to characterize lymph nodes, and both have similarly poor sensitivity and specificity for lymph node metastases. PSMA-targeting agents can detect metastatic disease, and targeted agents outperform CT and MR for lymph node metastases but are relatively limited for local recurrence compared to MRI [18] (Fig 5). This enables assessment for the key therapeutic decision of whether local therapy alone is needed or if systemic therapy should be used in the setting of biochemically recurrent prostate cancer.

Fig. 5.

Recurrence (arrow) at the anastomosis site post prostatectomy seen as an area of enhancement on dynamic contrast enhanced imaging (A) and increased signal on T2-weighted MR imaging (B). This lesion was not seen on PSMA-targeted (¹⁸F-DCFPyL) PET-CT imaging. The region of the anastomosis on PET is obscured by excreted radiopharmaceutical in the bladder.

Small studies have compared PSMA-targeted PET/CT versus PET/MRI. Due to small numbers of patients, statistical power has been limited in these studies. In 34 patients, Jentjens et al. found that ⁶⁸Ga-PSMA-11 PET/MRI performed at least as well as PET/CT [48**]), and in this small study although not statistically significant, PET/MRI found more patients with positive lesions and more lesions over all (88 vs 81), including lymph nodes (57 vs 51) and bone metastases (16 vs 15), than by PET/CT despite not including mpMRI sequences; further, PET/MRI was more sensitive for lymph node metastases than MRI or CT alone. In a study of 62 patients with recurrence and 150 lesions, ⁶⁸Ga-PSMA-11 PET/MR detected 11 lesions not seen by PET/CT, including 5 lymph nodes and 6 local recurrences, with the latter being statistically significant; further, interpreting PET/MR resulted in higher diagnostic confidence than PET/CT [49**]. In a separate study with a local recurrence subgroup, 18 patients were positive for recurrence by mpMRI compared to 9 by ⁶⁸Ga-PSMA-11 PET/CT or PET/MR without mpMRI [50], emphasizing the importance of the MR sequences chosen and complementarity with PSMA-targeted PET. PET with MR imaging can provide diagnostic confidence. It has been suggested that in patients with questionable recurrent prostate cancer post radiation therapy, if both PSMA PET and MRI are positive, given nearly 100% PPV, biopsy confirmation may not be needed [51*].

Conclusion

Combined whole-body MRI with mpMRI and a targeted PET agent, such as against PSMA for detecting recurrence at low PSA levels, should be advantageous and should better enable detection of locally recurrent and metastatic prostate cancer. Detecting and distinguishing local versus metastatic disease via imaging is crucial to salvage treatment decisions such as whether to use local versus systemic salvage therapy.

Key Points:

• PET-MR should be advantageous for imaging recurrent prostate cancer and guiding local and/or systemic therapy.

• Detectability of prostate cancer recurrence by imaging is affected by imaging technique and PSA level.

• mpMR mechanistically assesses anatomic and functional changes and is more useful for evaluating the post-treatment bed.

• Whole body PET-MR sensitivity is dependent on the radiopharmaceutical used and aids evaluation for lymph node and distant metastases.