Prostate-specific membrane antigen (PSMA) fusion imaging in prostate cancer: PET–CT vs PET–MRI

Abstract

Objectives:

To investigate whether PET–CT or PET–MRI is more appropriate for imaging prostate cancer, in terms of primary tumor detection, local staging and recurrence, as well as lymph nodes and distant metastases.

Methods:

A systematic literature search was conducted on Embase, PubMed/MEDLINE, and the Cochrane Library database. Studies evaluating the diagnostic performance of PET–CT vs PET–MRI in prostate cancer patients were emphasized.

Results:

We reviewed 57 original research articles during the period 2016–2021: 14 articles regarding the radiotracer PSMA; 18 articles regarding the primary tumor detection, local tumor staging, managing local recurrence; 17 articles for managing lymph node metastases; and eight articles for managing bone and other distant metastases. PSMA PET could be complementary to mpMRI for primary prostate cancer localization and is particularly valuable for PI-RADS three lesions. PET–MRI is better than PET–CT in local tumor staging due to its specific benefit in predicting extracapsular extension in MRI-occult prostate cancer patients. PET–MRI is likely superior as compared with PET–CT in detecting local recurrence, and has slightly higher detection rates than PET–CT in lymph node recurrence. PET–CT and PET–MRI seem to have equivalent performance in detecting distant bony or visceral metastases.

Conclusion:

In conclusion, PET–MRI is suitable for local and regional disease, either primary staging or restaging, whereas PET–CT is valuable for managing distant bony or visceral metastasis.

Advances in knowledge:

We reviewed the emerging applications of PET–MRI and PET–CT in clinical aspects. Readers will gain an objective overview on the strength and shortfalls of PET–MRI or PET–CT in the management of prostate cancer.

Introduction

Prostate-specific membrane antigen (PSMA) is a 750 amino acid type II membrane glycoprotein abundantly expressed in prostate cancer cells, either in prostate cancer tissue or at other metastatic sites, with activity depending on tumor dedifferentiation. More than 90% of primary prostate cancer displayed moderate-to-high PSMA expression; ¹ therefore, PSMA is considered as a suitable target for prostate cancer positron emission tomography (PET) imaging. ² Several radiopharmaceuticals, which target PSMA for instance ⁶⁸Ga-PSMA-11, ^{3 18}F-DCFPyL ⁴ and ¹⁸F-PSMA-1007, ⁵ have been developed for targeting biological features of prostate cancer that have been widely reviewed. ^{6 68} Ga-PSMA-11 has been extensively adopted and used globally, that is, data from >15,000 patients have been published till now. ¹ The European Association of Nuclear Medicine (EANM) – Society of Nuclear Medicine and Molecular Imaging (SNMMI) guidelines on PSMA PET provide an in-depth information regarding the procedure of ⁶⁸Ga-PSMA-11 PET. ⁷ In late 2020, ⁶⁸Ga-PSMA-11 became the foremost approved PSMA PET imaging agent by the U.S. Food and Drug Administration, and distant suppliers expand the availability of PET radiopharmaceuticals in the early stage of development. However, ⁶⁸Ga still has some drawbacks including shorter half-life, lower positron yield, and higher positron energy. With the increasing demand for PSMA PET, a trend is emerging toward a shift from ⁶⁸Ga- to in-house production of ¹⁸F-labeled PSMA target compounds in recent years. ⁸ Herein, we use PSMA to generally describe all the PSMA-based radiotracers in this article. For readers seeking for a comprehensive guidelines for interpretation, we recommend an outstanding review by Hofman et al ⁹ . Before diving into the scope of prostate cancer, readers might need to consider a number of benign (i.e. hemangioma, Paget disease, and fibrous dysplasia) or malignant (i.e. thyroid cancer, renal cell carcinoma, and hepatocellular carcinoma) lesions also lead to PSMA uptake. ^10,11 Characterizing PSMA-avid lesions needs a comprehensive consideration to include clinical context and morphological imaging information as possible. For example, the ureteral peristalsis can be misunderstood as a nodal disease.

Meanwhile, multiparametric MRI (mpMRI)—combining morphological T ₁-weighted imaging, T ₂-weighted imaging with functional sequences, dynamic contrast-enhanced (DCE) imaging, and diffusion-weighted imaging (DWI)—has become broadly adopted in the past decade for detecting clinically significant cancer and plays a vital role in performing targeted biopsy. ¹² Using mpMRI as a triage test before biopsy, MRI-targeted biopsies are better than systematic biopsies in prostate cancer detection, which helps in avoiding overdiagnosis or underdiagnosis. ¹³ The advancement of hybrid PET with magnetic resonance imaging (PET–MRI) or computed tomography (PET–CT) might potentially transform the approach of imaging prostate cancer. PET–MRI could result in lower exposure (80%) to radiation than PET–CT, but the acquisition time is substantially increased. ¹⁴ This longer scanning time is attributed to the additional mpMRI that significantly improves the resolution of the prostate region. The above-mentioned points raise the question of whether PET–CT or PET–MRI is more suitable for imaging prostate cancer.

Evidence acquisition

We searched three different electronic databases: Embase, PubMed/MEDLINE, and the Cochrane Library, using the keywords or search terms “prostate cancer”, “PSMA”, “PET–CT”, “PET/MRI”, and “PET/MR” and filtered it for full-text journal articles in English between January 2016 and May 2021. We selected studies that had defined cohorts with any accepted definition of prostate cancer with baseline or follow-up PET–CT or PET–MRI utilizing PSMA that assessed their performance in diagnosis, staging, and detection of recurrence. Next, we excluded articles that were from case reports and proceedings, articles that were predominantly based on secondary metastasis, articles that did not involve human subjects, articles that used radiotracers other than PSMA, and articles that concentrated on other malignancies and were not focused on prostate cancer.

Primary tumor detection and local staging

PET–CT vs mpMRI

For primary tumor detection, Kalapara et al ¹⁵ reported a retrospective study of 205 patients, showing non-significant difference in the localization or detection of primary prostate cancer among PET–CT and mpMRI. In another similar retrospective study of 144 patients published by Donato et al ¹⁶ PET–CT detected more secondary cancer foci and smaller lesions than mpMRI, demonstrating a higher sensitivity in lesion detection (90% vs  83%). In a meta-analysis published by Satapathy et al ¹⁷ PET–CT had an overall sensitivity (97%) and negative likelihood ratio (0.05) for the early diagnosis of prostate cancer and has potential utility to exclude the clinical suspicion of prostate cancer, thereby avoiding needless biopsies.

PET–MRI vs mpMRI

PET–MRI demonstrates a better primary tumor detection rate as compared with mpMRI. ¹⁸ Eiber et al ¹⁹ compared the diagnostic performance of PET–MRI for localization of primary prostate cancer against mpMRI and PET alone. From the 53 patients reported, mpMRI, PET, and PET–MRI identified prostate cancer in 66%, 92%, and 98% of the patients, respectively. On a region-based analysis, PET–MRI has higher sensitivity (76%) than mpMRI and PET alone (58 and 64%). In another study, Hicks et al ¹⁸ reported that the sensitivity of PET–MRI to detect prostate cancer is better than mpMRI. These results indicate that PSMA PET may be complementary to mpMRI for primary prostate cancer localization. An example of PET–MRI complementary to mpMRI for primary prostate cancer localization is shown in Figure 1. Current paradigm shift in prostate cancer diagnosis heads for improved detection of clinically significant cancer. Recent studies focused on the detection of clinically significant prostate cancer. ^16,20,21 PET/ MRI demonstrated a higher specificity (96% vs  71%) while remained the same good sensitivity (89%), as compared with PET–CT in a lesion-based analysis. ²⁰ PET–MRI is valuable for PI-RADS three lesions on mpMRI, leading a correct shift toward higher suspicion of malignancy and enabling correct lesion classification for overall ²² and clinically significant prostate cancers. ²⁰ However, it is still unclear of the additional value of PSMA PET to mpMRI in detecting tumors in mpMRI-difficult area, that is, the transitional zone.

Figure 1.

PSMA PET–MRI complementary to mpMRI for primary prostate cancer localization. A 80-year-old male with slightly increased PSA level 5.66 ng ml⁻¹. (a) A T2 low signal intensity focus was noted at right posterolateral part of right peripheral zone, the greatest dimension was measured as 1.2 cm. (b, c) The lesion showed hyperintense on high b-value DWI and hypointense on ADC, which was considered a PIRADS four lesion. (d) The fused PET/MR image showed high ⁶⁸Ga-PSMA-11 uptake at same location, and the possibility of extracapuslar extension was excluded. The final biopsy confirmed a Gleason score 4 + 3 prostate cancer lesion.

PET–CT vs PET/MRI

PET/MRI and PET–CT were reported to have a high detection rate for primary prostate cancer (91–98%). ^23–25 The sensitivity and specificity of PET–CT and PET/MRI in the detection of primary prostate cancer from articles with pathologic results as reference standard were reported from 67 to 96%, and from 55.5 to 100%, respectively, as summarized in Table 1. ^{5,16,18,20–22} For primary disease, PSMA PET showed a slightly better detection sensitivity than mpMRI, and the combination may have higher sensitivity than either modality alone. ¹⁹

Table 1.

Diagnostic Performance of PSMA PET–CT and PET–MRI in Detection of Primary Prostate Cancer with Pathologic Results as Reference Standard

Study	Patient	Radiotracer	Data Analysis	Modality	Parameter	Sensitivity (%)	Specificity (%)
Eiber et al 19 e	53	68Ga-PSMA-11	Region-based	PET/MRI	Score 4–5 as positive a	76%	97%
Kesch et al 5 e	10	18F-PSMA-1007	Region-based	PET–CT	Total agreement	71%	81%
					Near-total agreement	93%	92%
				PET/MRI b	Total agreement	81%	81%
					Near-total agreement	89%	99%
Hicks et al 18 e	32	68Ga-PSMA-11	Region-based	PET/MRI	Raw stringent approach	67%	71%
					Neighboring approach	74%	88%
					GEE (population-averaged)	73%	70%
Al-Bayati et al 20 e	22	68Ga-PSMA-11	Lesion-based	PET/MRI	Score 4–5 as positive a	88%	100%
					Score 3–5 as positive a	94%	56%
Chen et al 21 e	54	68Ga-PSMA-11	Lesion-based	PET–CT		89% c	71% c
				PET–MRI b		89% c	96% c
Donato et al 16 e	144	68Ga-PSMA-11	Lesion-based	PET–CT	Equivocal or likely lesions as positive d	95% c	93% c
					Likely lesions as positived	91%c	95%c
Ferraro et al 25 f	42	68Ga-PSMA-11	Patient-based	PET–MRI		96% c	81% c

GEE, Generalized estimating equation.

^aA 5-point Likert-scale

^bPET–CT in combination with MRI

^cDiagnostic performance for clinically significant prostate cancer

^dA 3-point Likert scale

^eRetrospective study including patients with biopsy-proven prostate cancer

^fProspective study including patients with suspected prostate cancer

Besides the detection of primary prostate cancer, the majority of literature discusses how PSMA PET might be a useful tool for local staging of prostate cancer. mpMRI is the imaging modality of choice for pretreatment evaluation of local staging. However, the most updated data showed mpMRI has a sensitivity and specificity of 39 and 56% for detecting extracapsular extension, whereas for seminal vesicle, invasion was 33 and 95%, respectively. ²⁶ Curative surgery is possible when extracapsular extension is absent. Therefore, PET–MRI may improve to define the extracapsular extension or seminal vesicle invasion of prostate cancer, to select the suitable high-risk patients of localized or locally advanced stage for definitive therapy. Based on a prospective study of 80 patients with pretreatment PET–MRI treated with radical prostatectomy published by Grubmüller et al, ²⁵ the accuracy of PET–MRI for T1 stage, T2 stage, T3a stage, and T3b stage were 82.5%, 85%, 79%, and 94%, respectively. In a retrospective study of 40 patients having intermediate- to high-risk prostate cancer who underwent PET–MRI for primary staging followed by radical prostatectomy, Muehlematter et al reported that the diagnostic performance in local staging of PET–MRI and mpMRI was similar, and PET–MRI may improve the sensitivity of MRI to detect extracapsular disease but with reduced specificity. ²⁷ In a similar study published by Brauchli et al, ²⁸ 100 patients who underwent radical prostatectomy after mpMRI and PET–CT imaging were reviewed. The results suggest that tumor-capsule interface measured on PET–CT is comparable to mpMRI criteria for predicting the presence of extracapsular extension. PET–MRI may be of specific advantage in predicting extracapsular extension in patients with MRI-occult prostate cancer.

It is important to keep in mind that there are false-positive and false-negative conditions in PSMA PET studies. Despite the term “prostate-specific”, PSMA expression can also be found in other tissues including sympathetic ganglia, inflammation/infection or epithelial cells, and tumor-associated neovascularization, hence causing the false-positivity. ²⁹ Low-to-moderate PSMA expression can also be seen in osteoblastic areas in bony degeneration, osteoarthritis, or insufficiency fractures. ³⁰ On the other hand, a small proportion of prostate cancers display no or minimal uptake at PSMA PET, possibly reflecting low PSMA expression, ⁹ with an exemplary case shown in Figure 2. In this situation PSMA PET can be false-negative and mpMRI might be helpful for lesion identification.

Figure 2.

Low PSMA expression cancer with nodal metastasis depicted by mpMRI. A 73-year-old male with biopsy-proved prostate cancer, the Gleason score was 5 + 4, and the PSA level was 5.73 ng ml⁻¹. (a) The PSMA maximal intensity projection did not reveal any abnormal ⁶⁸Ga-PSMA-11 uptake. (b) The fused PET–CT image showed low ⁶⁸Ga-PSMA-11 uptake at right portion of prostate gland and right periprostatic node, labelled by yellow arrows. (c, d, e) The MRI showed a large prostate tumor with right extracapsular extension and a periprostatic lymph node. The lesions showed restricted diffusion on DWI.

Management of local recurrence

Evidence from a prospective multicenter study, ³¹ with 312 patients underwent PET–CT for restaging due to biochemical relapse with negative/equivocal conventional imaging (i.e., bone scintigraphy, CT and MRI), PET–CT revealed additional lesions in the prostate bed in 32% of patients. Lütje et al ³² compared PET–MRI with PET–CT for detection of recurring prostate cancer. In 25 patients, PET–MRI and PET–CT identified 14 and 9 local resurrecting cases, respectively, showing that PET–MRI was better as compared with PET–CT in the prostate bed recurrences. The major advantage of PET–MRI for detection of tumor recurrence in the prostate bed was the superior soft tissue contrast from MRI-component of PET–MRI. ³³ Data from 119 patients with biochemical recurrence after radical prostatectomy underwent both PET–CT and PET–MRI, Freitag et al ³⁴ showed that more cases of local recurrence were diagnosed in MRI-component of PET–MRI as compared to PET-component of PET–CT and PET–MRI (18 vs 9). In addition, detection of local recurrence using the PET-component was considerably influenced by its proximity to the bladder. The above studies suggested that PSMA PET is useful for detection of local recurrence, ³¹ and PET–MRI is likely superior as compared with PET–CT. ^32,34 An example of prostate cancer with bladder involvement better delineated by mpMRI is shown in Figure 3. Since PSMA uptake is strongly influenced by the viable cells percentage, ^9,10 and DWI is known to represent cell density, combined PET–MRI might enhance tumor characterization, however not yet been fully evaluated. The ⁶⁸Ga-PSMA is excreted into urinary tract resulting high radioactivity in the bladder, which might mask lesions close proximity to the bladder and limit the PET detection of local recurrence of prostate cancer. By comparison to ⁶⁸Ga-PSMA, ¹⁸F-PSMA is characterized by hepatobiliary clearance resulting in reducing the bladder radioactivity and has the potential to overcome the limitations. ³³

Figure 3.

Prostate cancer with bladder involvement better delineated by mpMRI. A 60-year-old male with prostate cancer initial clinical stage IVB. (a) The maximal intensity projection of PSMA PET showed prostate tumor with pelvic and para-aortic lymph nodes, and left proximal femur bone metastasis. (b) The fused PET–CT image showed high ⁶⁸Ga-PSMA-11 uptake at prostate gland. However, high radioactivity in the bladder limit the ability of ⁶⁸Ga-PSMA PET–CT to detect bladder neck invasion by prostate tumor. Prostate cancer with bladder involvement (arrow) and seminal vesicle invasion was clearly delineated on T ₂-weighted image (c), ADC map (d), and DCE MRI (e).

Detection of lymph nodes metastases

PET–CT vs mpMRI

A recent meta-analysis compared the diagnostic performance of PET–CT and mpMRI in identifying lymph nodes (LN) metastasis for the intermediate- to high-risk prostate cancer. ³⁵ The per-patient sensitivity and specificity of PET–CT were 65 and 94%, superior to the corresponding values of mpMRI were 41 and 92%, respectively. The area under the receiver-operating characteristic curve for PET–CT and mpMRI were 0.92 and 0.83, respectively. Malaspina et al prospectively compared PET–CT and whole-body MRI in 79 patients having intermediate- or high-risk prostate cancer at primary staging. ³⁶ Pelvic LN metastases were found in 31 (39%) patients. At the patient level, sensitivity/specificity values for PET–CT and whole-body MRI were 87%/98 and 37%/98%, respectively. From the above studies, it can be supported that PET–CT has a significant higher sensitivity than mpMRI in the detection of LN metastasis in the primary staging. We had prospectively enrolled 34 patients with biochemically recurrent prostate cancer followed by robot-assisted radical prostatectomy, in which PET–CT identified more regional LN metastases than the mpMRI (10 vs 8). ³⁷

PET–CT vs conventional CT/MRI

For primary nodal staging before definite therapy, studies have been conducted to assess PSMA studies in detecting regional LN metastasis with histopathological results from pelvic LN dissection as the gold standard. A retrospective study of 130 patients with intermediate- or high-risk prostate cancer revealed PET–CT to outperform CT or MRI in the detection of LN metastasis, with a per-patient sensitivity of 66vs 44% and a specificity of 99vs 85%. ¹ Another study of 30 patients having intermediate- or high-risk patients showed PET–CT to have a per-patient sensitivity of 64% and specificity of 95%. ³⁸ The “ProPSMA” study, a prospective randomized study for 302 prostate cancer patients with high-risk features considered for curative therapies from ten centers in Australia revealed PET–CT to provide superior per-patient sensitivity (83% vs  23%) and similar specificity (99% vs  96%) to CT for pelvic LN metastasis. ³⁹

However, the sensitivity of PET–CT for detecting LN metastasis is only moderate, in contrast to the specificity which is higher than 90%. Klingenberg et al. evaluated 691 high-risk patients primary staged by PET–CT. LN metastasis was detected in 31% (217/691). ⁴⁰ In 117 patients who undertook radical prostatectomy with concomitant pelvic LN dissection, per-patient sensitivity, specificity, positive and negative predictive values, and accuracy for LN metastasis detection on PET–CT were 31, 97, 69, 85, and 83%, respectively. Undetected LN metastases were either micro-metastases located in the LN border or without PSMA expression. Due to the limited resolution of PET, small LN metastases might undergo undetected. ^38,41 LN metastases can also be missed due to the lack of PSMA expression in a minority of prostate cancers. ⁴² It is clear that PET–CT imaging still cannot replace pelvic LN dissection to exclude pelvic LN metastases.

PET–MRI vs mpMRI

Park et al ³⁸ prospectively enrolled 33 patients with intermediate- or high-risk prostate cancer who were scheduled for radical prostatectomy with pelvic LN dissection, ⁴³ with 12 out of 382 dissected nodes confirmed to harbor metastatic disease in three patients. PSMA PET showed six true-positive and six false-positive LNs while mpMRI showed no true-positive lesion and nine false-positive LNs. It is thus suggested that PET–MRI may better inform the need and extent for pelvic node dissection, yet more evidence is needed.

Management of nodal recurrence

For patients with biochemical failure and planning for salvage lymphadenectomy, Rauscher et al ⁴⁴ reported data with mixed PET–CT and PET–MRI, detecting LN metastasis in 78% metastatic fields while CT or MRI was positive in only 27%. Specificity of PSMA PET and CT or MRI was 97 and 99%, respectively. The mean size of PET-positive LNs measured by MRI or CT was 8.3 mm (range 4–25 mm). Mandel et al ⁴⁵ retrospectively analyzed 23 patients with nodal recurrence and found 29/109 (27%) resected fields harbored histologically confirmed LN metastases. The sensitivity and specificity of PSMA PET studies were 76 and 88% in region-based analysis. Linxweiler et al ⁴⁶ identified 25 patients who underwent through robotic salvage lymphadenectomy for nodal recurrence detected by PET–CT. In total, 43 suspicious spots were detected on PET–CT and 66 histologically positive LNs were removed. In 23 LN-positive patients, PET–CT was correct in 13 (57%) patients, whereas nodal metastatic spread was more extensive in 10 (43%) patients than suggested by imaging. Siriwardana et al ⁴⁷ retrospectively analyzed 35 patients who underwent robotic salvage lymphadenectomy for nodal recurrence detected by PET–CT. A total of 58 lesions suspicious for LN metastasis were detected and 32 patients (91%) were histopathologically proven. Jilg et al ⁴² retrospectively analyzed 30 patients with suspicion of exclusively nodal relapse who underwent a template pelvic/retroperitoneal salvage lymphadenectomy after PET–CT. ⁴¹ The sensitivity and specificity for main region (pelvic left/right, retroperitoneal) were 93.2 and 100%, and for subregion (common iliac, internal iliac, external iliac, obturator, aortal, vena cava, presacral, and aortic-bifurcation) were 81.2 and 99.5%, respectively. Based on anatomical sub regions containing just one LN metastasis, the essential short diameter of tumor deposits in LN metastasis required to reach a detection rate of 50 and 90% was estimated to be ≥2.3 mm and ≥4.5 mm, respectively.

PET–CT vs PET–MRI

There are few head-to-head comparisons of PET–CT vs PET–MRI in this regard. Domachevsky et al prospectively enrolled 140 successive patients with biopsy-proven prostate cancer to undergo prostate and pelvic PET/MR immediately after ⁶⁸Ga-PSMA injection followed by whole-body PET–CT at 60 min after injection. ⁴⁸ PSMA-avid pelvic LNs were noted on PET–CT in 33 patients (24%) and on PET/MR in 32 patients (23%), although no pathological approval was obtained. We have identified four studies with retrievable information for comparing PET–CT and PET–MRI imaging performed on the same day in patients with biochemical failure, with PET–CT acquisition performed around 1 h after PSMA injection and PET–MRI acquisition performed around 3 h after injection. ^32,49–51 The patient-based and lesion-based detection rates for LN metastasis are listed in Table 2. An example of PET–MRI demonstrated local, regional, and distant tumor recurrence is shown in Figure 4. Overall PET–MRI seems to have slightly higher detection rates than PET–CT. This may be attributed to the delayed imaging time of PET–MRI and also to the better soft tissue contrast of MRI.

Table 2.

Detection of lymph node metastases in patients by ⁶⁸Ga-PSMA PET–CT and PET–MRI with biochemically recurrent prostate cancer

Study	Patient-based Analysis				Lesion-based Analysis
	Patient number	LN positive patient	PET–CT a	PET–MRI a	LN positive lesion	PET–CT b	PET–MRI b
Afshar-Oromieh et al 49	20	11	100%	100%	49	100%	100%
Freitag et al 48	26	20	100%	100%	64	98%	100%
Lütje et al 32	25	14	93%	100%	24	83%	96%
Guberina et al 50	93				103	95%	98%

LN, Lymph node.

^aPatient-based detection rate.

^bLesion-based detection rate.

Figure 4.

PSMA PET–MRI demonstrated local, regional and distant tumor recurrence. A 65-year-old male underwent radical prostatectomy 7 years ago, initial staging pT3bN0M0, Gleason score 4 + 3, PSA was slow raising to 4.20 ng ml⁻¹. (a) The maximum intensity projection of PSMA PET shows abnormally-increased ⁶⁸Ga-PSMA-11 uptake at (b) left prostate fossa, (c) left obturator lymph node, and (d) T12 spine. The yellow arrows on the corresponding multisequence MR imaging demonstrates (e) a focal nodule at left anastomotic site, (f) a left obturator lymph node with short axis diameter measured as 0.6 cm, and (g) a focal hyperintensity on DWI at T12 vertebral body. These lesions were considered as recurrent and metastatic prostate cancer.

Detection of distant metastases

PET–CT vs mpMRI

Although mpMRI may achieve a sensitivity of 95% to detect bone metastasis for the purpose of primary staging, the routine MRI protocol usually limits the scanning field to the pelvifemoral area and lumbar spine and cannot assess the global burden of bone metastasis. ⁵² For detecting distant metastases at primary staging, the prospective “proPSMA” study revealed PET–CT to provide superior per-patient sensitivity (92% vs  54%) and specificity (99% vs  93%) to conventional imaging consisting of bone scintigraphy and CT. ³⁹ Klingenberg et al ⁴¹ evaluated 691 consecutive newly diagnosed high-risk prostate cancer patients with PET–CT. ⁴⁰ Bone metastases were identified in 17% of males (116/691) and 78% of these patients (90/116) had concurrent LN metastasis. Bone metastases were correlated with increasing PSA levels, clinical stages, and tumor grades, and around 2% of patients (13/691) also presented with visceral metastases. ⁴⁰ Damjanovic et al retrospectively evaluated PET–CT scans of 739 prostate cancer patients to detect lung metastases and non-solid focal pulmonary opacities. ⁵³ Ninety-one pulmonary metastases, of which 66 (72.5%) were PSMA-positive and 25 (27.5%) were PSMA-negative, and 14 opacities were recognized in 34 patients. Pulmonary opacities revealed a moderate tracer uptake, significantly lower than PSMA-positive lung metastases yet higher than PSMA-negative metastases. The study showed that PET–CT was not able to differentiate between pulmonary opacities and pulmonary metastases. Damjanovic et al also identified 123 liver metastases in 18 patients in these scans, ⁵⁴ with 8 (78%) metastases being PSMA-positive while 23 (22%) metastases being PSMA-negative. Whole-body PSMA PET can be valuable in the detection of occult distant metastases and help identifying these patients to receive individualized multimodal treatment. Special attention should be taken on interpreting the lesions in the lung and liver.

PET–MRI vs mpMRI

As mpMRI can be simultaneously acquired during a PET–MRI study, PET–MRI is expected to be the one-stop-shop modality for prostate cancer imaging. The combined interpretation of distant lesions with both PSMA PET and mpMRI may decrease the false-positive and false-negative findings. An exploratory cost-effectiveness analysis for prostate cancer patients with biochemical failure suggested the use of PET/MRI to be cost-effective comparative to normal care. ⁵⁵ Further studies are wanted to determine if PET–MRI can be cost-effective in comparison with PET–CT with mpMRI.

PET–CT vs PET–MRI

In the study conducted by Domachevsky et al of 140 newly diagnosed patients undergoing prostate and pelvic PET/MR followed by whole-body PET–CT, 11 patients (7.8%) were found to have bone metastases. ⁴⁸ In three patients, bone metastases were detected on PET–CT outside the PET/MR field of view. Two patients had other clinically significant findings on PET–CT (1.4%), including one patient with PSMA-avid cervical and supra clavicular LNs and a second patient with PSMA-avid lung nodules. From the studies for comparing PET–CT and PET–MRI performed on the same day in patients with biochemical failure, the patient-based and lesion-based detection rates for bone metastases are listed in Table 3 injection. ^32,49–51 Both modalities seem to have equivalent performance. In the study by Guberina et al ⁵⁰ , 15 visceral metastases were also detected by both modalities. ⁵¹

Table 3.

Detection of bone metastases by ⁶⁸Ga-PSMA PET–CT and PET–MRI with biochemically recurrent prostate cancer

Study	Patient-based Analysis				Lesion-based Analysis
	Patient number	Bone positive patient	PET–CT a	PET/MRI a	Bone positive lesion	PET–CT b	PET/MRI b
Afshar-Oromieh et al 49	20	6	100%	100%	23	100%	100%
Freitag et al 48	26	8	100%	100%	28	100%	100%
Lütje et al 32	25	2	100%	100%	4	100%	100%
Guberina et al 50	93	10	100%	100%	23	100%	100%

LN, Lymph node.

^aPatient-based detection rate.

^bLesion-based detection rate.

Management of distant recurrence

For patients with extensive metastases, PSMA PET may not provide additional impact for initial therapy but may serve as a baseline study to assess the global burden for later utility. In our prospective study of 34 patients with biochemically recurrent prostate cancer, ⁶⁸Ga-PSMA-11 PET–CT identified non-regional LN metastases and bone metastases in 1 and 5 patients while mpMRI spotted these lesions in 1 and 4 patients, respectively. ³⁷ With recurrent prostate cancer, Verburg et al retrospectively studied 155 patients receiving PET–CT studies. ⁵⁶ PET–CT was positive in 44%, 79 and 89% patients and detected bone metastasis in 15%, 16 and 39% of patients having PSA levels of ≤1, 1–2 and ≥2 ng ml⁻¹, respectively. Pyka et al retrospectively analyzed 126 patients who underwent either PET–CT or PET–MRI studies and bone scintigraphy with a median interval of 20 days. ⁵⁷ In the patient-based analysis, with ambiguous scans considered negative, PSMA PET showed a sensitivity and specificity of 100vs57 and 100% vs  96%, respectively, as compared with bone scintigraphy. PSMA PET outperforms bone scintigraphy for the detection of bone metastasis for the primary staging and biochemical recurrence. PSMA PET and bone scintigraphy have equivalent performance even in the subset of metastatic castration-resistant prostate cancer (mCRPC) subgroup, ⁵⁷ shown as in Figure 5. It can be concluded that additional bone scintigraphy is not needed if PSMA PET has been performed. To date no direct comparison of PET–CT vs PET–MRI has been reported on this aspect.

Figure 5.

PSMA PET–CT for managing castration-resistant prostate cancer (mCRPC). A 73-year-old male with initial staging T3N1M1b with androgen deprivation therapy since 7 years ago, PSA keep progression even after change treatment to new antiandrogen drug, castration-resistant prostate cancer (mCRPC) status was confirmed, and current PSA level was 675 ng ml⁻¹. (a) The bone scintigraphy showed multiple hot spots involving spines, left posterior ribs, and pelvic bones. (b) The PSMA PET showed more extensively bony metastasis, (c) primary prostate cancer lesion labelled by yellow arrow, bony metastasis by arrowheads and (d) liver metastasis.

Conclusions

PSMA PET may be complementary to mpMRI for primary prostate cancer localization and is particularly valuable for PI-RADS three lesions. PET–MRI is superior to PET–CT in local tumor staging, may be of particular benefit in predicting extracapsular extension in patients with MRI-occult prostate cancer. PET–MRI is likely superior as compared with PET–CT in detecting local recurrence, and have slightly higher detection rates than PET–CT in lymph node recurrence. PET–CT and PET–MRI seem to have equivalent performance in detecting distant bony or visceral metastases. PET–MRI is useful for local and regional disease, either primary staging, or restaging. PET–CT is useful for managing distant bony or visceral metastasis. The cost-effective comparison of both modalities is still to be exploited.