Advances in medical imaging for the diagnosis and management of common genitourinary cancers

Abstract

Medical imaging of the three most common genitourinary (GU) cancers—prostate adenocarcinoma, clear cell renal cell carcinoma (RCC), and urothelial carcinoma of the bladder—has evolved significantly during the past decades. The most commonly used imaging modalities for the diagnosis, staging, and follow-up of GU cancers are computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET). Multiplanar multidetector CT and multiparametric MRI (mpMRI) with diffusion-weighted imaging are the main imaging modalities for RCC and urothelial carcinoma, and while mpMRI is rapidly becoming the main imaging tool in the evaluation of prostate adenocarcinoma, biopsy is still required for diagnosis. Functional and molecular imaging using 18-fluorodeoxyglucose (FDG)-PET, and sodium fluoride (NaF)-PET are essential for the diagnosis, and especially follow-up, of metastatic GU tumors. This review will provide an overview of the latest advances in the imaging of these three major GU cancers.

1. Introduction

Genitourinary (GU) cancers are some of the most common malignancies. Renal cell carcinoma (RCC) is the most common kidney tumor, with an estimated 62,700 new cases and an estimated 14,240 deaths in 2016 [1, 2] in the United States (US). Urothelial carcinoma of the bladder is more common than RCC, with an estimated 77,000 new cases and 16,390 deaths in 2016 [2] in the US. Adenocarcinoma of the prostate is the second most common cancer in men and the most common GU cancer, with more than 180,000 new cases in the US estimated in 2016 and an estimated 26,120 deaths per year [2].

Computed tomography (CT) is the major imaging modality for RCC and urothelial carcinoma of the bladder. The current recommendation of the National Comprehensive Cancer Network for initial staging of muscle-invasive bladder cancer (MIBC) is CT scan of the chest, abdomen, and pelvis or magnetic resonance imaging (MRI) of the abdomen and pelvis with non-contrast CT scan of the chest.

MRI is helpful in patients allergic to iodinated contrast agents and in patients with renal insufficiency, and may provide more information for the evaluation of locally advanced disease [3, 4]. Advanced MRI techniques are providing better assessment of renal and bladder tumors, while multiparametric MRI (mpMRI) is increasingly being used to diagnose prostate adenocarcinoma.

PET/CT is a powerful noninvasive tool used by major medical centers and many community hospitals for characterizing solid tumors. Fluorodeoxyglucose positron emission tomography (FDG-PET), and sodium fluoride (NaF)-PET are common functional and molecular imaging modalities used for staging and assessing treatment response [3, 5].

In this review, we discuss the latest advances in imaging and their potential impact on the diagnosis and management of RCC, urothelial carcinoma of the bladder, and prostate adenocarcinoma.

2. Renal cell carcinoma

2.1. CT in RCC

CT is the primary modality for the diagnosis, staging, treatment planning, and surveillance of RCC [6-8]. CT can be performed in multiple phases before and after injection of an intravenous (IV) contrast agent. The most reliable phase for the detection of RCC is the nephrographic phase, 70–100 sec after injection of IV contrast [6]. Because papillary carcinoma is hypovascular, the arterial phase of a kidney CT exam can differentiate between clear cell and papillary carcinoma with 95.7% accuracy and a sensitivity and specificity of 98.3%, and 92%, respectively [9]. Clear cell carcinoma, which accounts for 60%–65% of RCCs, appears as a hypervascular mass with a heterogeneous texture due to cystic or necrotic components [10]. Papillary RCC is usually homogeneous, well-defined, and hypovascular [10]. Chromophobic RCC has an intermediate vascularity, is usually well-circumscribed, and measures > 4 cm [10]. Multiphasic multidetector CT scan is useful in differentiating clear cell RCC from other subtypes of RCC and oncocytoma [11].

An important and unique finding in RCC is tumor extension and thrombosis to renal vein and inferior vena cava (IVC) [12]. This has been reported in 4%–10% of all renal neoplasms, having therapeutic and prognostic importance [13, 14]. This invasion can only be limited to the vessel wall. CT is very helpful for venous extension of RCC, best seen in the corticomedullary phase of contrast enhancement [15]. Enhancement of the thrombus will help to differentiate tumor thrombus from bland thrombus [15]. Although CT can be as sensitive as MRI in assessment of tumor thrombus, in some cases it may fail due to extrinsic pressure over the vein or inadequate filling of IVC [16].

Dual-energy CT (DECT) is a new type of CT scanner in which two sets of x-ray beams with two different energy levels are passing through the body, in contrast to traditional CT scans that have an x-ray beam with only one energy level [17-19]. The interaction of these two x-ray beams with body tissues can help in material identification and quantification [18]. These two CT datasets at a low or high polychromatic peak tube energy (kilovoltage, or kV) provide material-specific information that allows for additional characterization of the kidneys and urinary tract [20]. The two varying energy levels can enhance or mute the conspicuity of IV contrast. DECT can be used to interrogate iodine and calcium concentrations and increase the iodine signal to help differentiate pathologic processes and clarify the internal structure of mass lesions [21]. The capability to decompose materials and extract iodine can improve differentiation of lesions such as minimal enhancement in a hypodense kidney tumor (papillary HCC) from a cyst. Quantification of iodine in the lesion can be a biomarker of tumor vascularity [18].

Improving lesion conspicuity can potentially reduce the need for or frequency of follow-up imaging. DECT is helpful in differentiating between clear cell and papillary cell RCC, the latter of which accounts for approximately 15%–20% of all RCCs [6, 21], with an overall accuracy of 95.3% and a sensitivity and specificity of 98.2% and 86.3%, respectively [20]. DECT can also help to assess tumor response, such as changes in tumor vascularity and necrosis [22]. Another clear benefit of DECT is the potential to reduce the radiation dose by applying virtual non-contrast images, negating the need for pre-contrast CT series [23]. This is described in more detail below (CT in urothelial carcinoma).

2.2. MRI in localized invasive RCC: morphologic and functional information

MRI offers a wealth of morphologic and functional information for characterizing and staging of RCCs, which makes it a tool for preoperative staging of RCCs when used with chest CT [24, 25]. Inherent T1 and T2 signal intensity of the mass can be used to differentiate subtypes of RCC. For example, macroscopic fat within angiomyolipomas can easily be differentiated by detecting India ink artifacts surrounding the macroscopic fat-containing areas of the mass on in-and out-of-phase images [26]. Similarly, T2-weighted and post-contrast images can delineate the complexity of cystic or necrotic RCCs. Multiple studies have demonstrated close sensitivity and specificity of CT and MRI in detection of the contrast enhancement and complexity of renal masses. In a study of 69 renal masses, Israel et al. [27] were able to demonstrate similar results for categorization of masses into different Bosniak categories in 81% of CT and MRI exams. In the reminder of the cases, MR images may depict additional septa, thickening of the wall and/or septa, or enhancement, which may lead to an upgraded Bosniak cyst classification and can affect case management. Given the higher accuracy of MRI in depicting the fine details of complex cystic lesions, use of MRI is preferable to CT in cases where upgrading or downgrading the lesion would be consequential.

Diffusion-weighted imaging (DWI) exploits the random motion of water molecules. The extent of tissue cellularity and the presence of an intact cell membrane help determine the impedance of water-molecule diffusion, which can be quantitatively assessed using apparent diffusion coefficient (ADC) maps. The lower the ADC value, the more restricted the diffusion of water molecules in the microenvironment, which can be seen in tissues with a high nucleus-to-cytoplasm ratio or dysfunctional cell membranes such as tumor cells [28]. Several studies have investigated the role of this imaging method in characterizing RCC. A meta-analysis of nine studies of DWI's ability to differentiate benign from malignant renal masses [29] demonstrated the significant difference of ADC, with pooled weighted sensitivity and specificity of 88% and 72%, respectively.

The lack of ionizing radiation and high temporal resolution of MRI are beneficial in designing dynamic contrast-enhanced MRI (DCE-MRI) protocols, which typically have four phases: precontrast, arterial, nephrogenic, and excretory. Important information regarding vascular flow within the tumor can be obtained by acquiring images in different phases of contrast administration, which in turn can be helpful in differentiating benign vs. malignant lesions (Figure 1) [30]. In a study of 152 RCCs, Vargas et al. [31] demonstrated significantly higher enhancement of clear cell carcinomas in all phases compared to papillary and chromophobic lesions.

Figure 1.

Dynamic contrast-enhanced MRI in a 27-year-old female with a large, enhancing high-grade RCC within the left kidney, with extension into the left renal vein and IVC on axial (A) and coronal (B) delayed post-contrast images. Patient underwent left nephrectomy in addition to IVC thrombectomy.

Extension of RCC into perinephric fat is associated with increased mortality independent of tumor size [32]. Therefore, determining fat involvement is an important prognostic indicator. MRI is superior to CT for detecting RCC involvement of the kidney capsule, perinephric fat, or Gerota's fascia (Figure 2) [33]. Stranding and enhancement of the perinephric fat are reliable signs of extension beyond the capsule [34]. However, early-stage involvement of Gerota's fascia can be difficult to assess [35]. Both multiphasic MRI and CT have sensitivity and specificity of > 90% and 80%, respectively, for detection of renal vein involvement in locally invasive RCCs [16]. Some studies have reported 100% accuracy of multiphase contrast-enhanced MRI for detection of renal vein thrombosis and extension into the IVC [36]. Enhancement of the thrombus is suggestive of tumor growth into the vessel rather than bland thrombosis. Overall, MRI provides a versatile array of morphologic and functional information for accurate pretreatment lesion characterization and local staging for patients with RCC.

Figure 2.

MRI in a 19-year-old male with a large, locally advanced high-grade RCC with papillary features in the setting of hereditary leiomyomatosis and RCC syndrome within the lower pole of the right kidney. The mass demonstrates heterogeneous signal intensity on T2-weighted images (two panels on left) and heterogeneous enhancement on post-contrast image (second from the right) and restricted diffusion on DWI image (right panel). There is extension of the mass into the mesonephric fat beyond the capsule and involvement of Gerota's fascia and right psoas muscle.

2.3. PET/CT for metastatic RCC

PET's utility in evaluating RCC stems primarily from the poor prognosis for patients with advanced disease, as indicated by the presence of metastases. In conjunction with anatomic imaging with CT and MRI, PET offers potentially high levels of visual conspicuity at the most common sites of metastatic disease. In order of frequency, these sites include the lungs, bones, lymph nodes, liver, adrenal glands, and brain [37]. Currently, clinical PET imaging for metastatic RCC is primarily achieved through metabolic imaging with FDG. Table 1 lists selected studies evaluating recurrent or metastatic RCC currently available in PubMed. Reviews, case reports, meta-analyses, and studies with < 10 subjects were excluded from our search. Studies listed calculate sensitivity and specificity but do not evaluate treatment response [38-56].

Table 1.

Selected studies evaluating recurrent or metastatic RCC.

Author	Year	N	Study type	Instrument	Reference Std	Met Location	Sensitivity	Specificity	PPV	NPV	Acc
Safaei, et al.	2002	36	Retrospective	PET	Histology, clinical outcome	-	87	100	-	-	89
Wu, et al.	2003	18	Not Reported	PET	Histology, clinical outcome	Bone	100	100	-	-	-
Aide, et al.	2003	53	Prospective	PET	Histology, clinical outcome	-	100	93	NA	NA	94
Chang, et al.	2003	15	Retrospective	PET	Histology	Lung	90	80	-	-	87
Jadvar, et al.	2003	25	Retrospective	PET	Histology, clinical outcome	-	87	75	94	33	72
Majhail, et al.	2004	24	Retrospective	PET	Histology	-	63.6	100	100	-	-
Kang, et al.	2004	66	Retrospective	PET	Histology, clinical outcome	Retroperitoneum	75	100	-	-	-
Kang, et al.	2004	66	Retrospective	PET	Histology, clinical outcome	Lung	75	97	-	-	-
Kang, et al.	2004	66	Retrospective	PET	Histology, clinical outcome	Bone	77	100	-	-	-
Dilhuydy, et al.	2007	24	Retrospective	PET	Histology, clinical outcome	-	75	50	92.3	33.3	-
Park, et al.	2009	63	Retrospective	PET/CT	Histology, clinical outcome	-	87	83.3	77.3	92.6	85.7
de Llano, et al.	2010	58	Retrospective	PET	Histology, clinical outcome	-	80.56	86.36	90.6	73.8	58.7
Kumar, et al.	2010	63	Retrospective	PET/CT	Histology, clinical outcome	-	90	91	-	-	90
Nakatani, et al.	2011	23	Retrospective	PET	Histology, clinical outcome	-	80	67	94	33	78
Bertagna, et al.	2013	68	Retrospective	PET/CT	Histology, clinical outcome	-	82	100	100	66.7	86.8
Fuccio, et al.	2014	69	Retrospective	PET/CT	Histology, clinical outcome	-	90	92	95	85	91
Öztürk	2015	132	Retrospective	PET/CT	Histology, clinical outcome	-	93.8	88.2	92.6	88.2	91.6
Win, Aparici	2015	315	Retrospective	PET/CT	Histology	-	100	100	-	-	-
Alongi, et al.	2016	104	Retrospective	PET/CT	Histology, clinical outcome	-	74	80	83	70	84

Positron emission tomography (PET), computed tomography (CT), positive predictive value (PPV), negative predictive value (NPV), accuracy (Acc)

RCC can present with limited uptake of FDG, likely due to the dominant incidence of the clear cell subtype of the disease. There is emerging evidence that the rarer papillary variant of RCC may have high avidity for FDG [56, 57]. For the clear cell subtype, there is growing interest in non-FDG molecular imaging agents that may have greater sensitivity and specificity and may potentially add phenotypic information for patient/tumor-specific treatment strategies.

Image acquisition and interpretation of RCC may benefit from more advanced techniques such as FDG-PET/CT. Because of the potentially low levels of FDG uptake in RCC, suspicion must remain high for lesions with mild FDG uptake that are concerning on CT. Additionally, foci of abnormal FDG uptake that may not be clearly identified as suspicious lesions on CT must also be considered important until proven otherwise (Figure 3). Thus, the combination of IV and oral contrast-enhanced CT as part of the CT acquired with PET, or as a separate examination that can be registered to PET, is highly valuable. The spatial combination of sensitivity and specificity of the respective modalities can not only aid in diagnosis and staging of multiple tumor types, but can also help to visually and spatially communicate radiologic findings among the subspecialties involved in patient care. For instance, biopsy and surgical planning with a PET/CT combination (Figure 4) can show the spatial coexistence of complex CT structures and elevated activity [58]. Structures with a large component of nonviable cells, necrosis, and fluid can be approached in favor of regions with viable tumor cells with intact cell membranes, as indicated by higher FDG activity. These regions do not always have a discriminatory appearance on CT. However, a certain level of comfort with FDG-PET is required due to the common confounding occurrence of structures with normal activity, such as urinary elimination in the renal calyces/pelvis, ureters, and bladder, as well as normal FDG activity in the gastrointestinal tract.

Figure 3.

Pre- (left) and post-treatment (right) maximum-intensity FDG-PET projections of the whole body of a patient with RCC. The extent of the new sites of lung metastases are clearly observed, in addition to new sites of osseous metastatic disease (white arrows) in the right iliac wing and left femur, which were not detected on CT.

Figure 4.

FDG-PET/CT of a patient with RCC. From left to right, maximum-intensity projection PET, coronal PET, CT, and PET/CT fusion are shown. Clearly visualized by PET, the left supraclavicular cluster of nodal metastasis was not detected on CT.

Another application of FDG in RCC is in venous tumor thrombus. Some investigators indicate that high FDG uptake is related to tumor thrombus and provide quantitative cutoffs to help differentiate tumor from bland thrombus [59]. However, prospective trials are currently limited, and we avoid quantitative cutoffs for this purpose because of the known high degree of variability among PET/CT scanners and methods of image reconstruction. The same concerns apply to the use of cutoffs for primary and distant metastatic RCC.

A recent meta-analysis reported an overall sensitivity and specificity for FDG-PET in identifying extrarenal lesions as 79% and 90%, respectively [38]. However, most of the studies included in the analysis were of PET without CT. A multimodality approach might achieve higher sensitivity and specificity. However, the clinical utility of FDG extends beyond traditional measures of diagnostic accuracy. Alongi et al. [51] observed that despite modest sensitivity and specificity for RCC (74% and 80%, respectively), imaging with FDG influenced the therapeutic strategy for a high proportion of patients (43%), and FDG-positive cases resulted in significantly lower five-year overall survival rates compared to FDG-negative cases (69% vs. 19%, respectively).

The high sensitivity of PET, the use of positron-emitting radiotracer agents with various potential tumor targets (glycolysis, hypoxia, amino acids, malignant cell proliferation and amino acid transport, Ab-Ag complex and receptors), and hybrid imaging technologies such as PET-MRI allow us to image malignant lesions more successfully than ever, which results in better patient care and outcomes [60].

3. Urothelial carcinoma of the bladder

3.1. CT scan in urothelial carcinoma

In the last few years, a major advance in cancer imaging has been improved multidetector CT acquisition quality with reduced radiation exposure, thanks to model-based iterative reconstruction and other dose-reduction initiatives. Processing speeds finally allow for real-time noise reduction to make up for reduced exposures. CT can detect UC in the bladder with a sensitivity of 79%–89.7% and specificity of 91%–94.7% [61]. Before patients undergo radical cystectomy, a contrast-enhanced CT or MRI scan is taken to assess local/regional tumor invasion and distant metastases. In MIBC and high-grade non-muscle invasive bladder cancer (NMIBC), CT of the abdomen and pelvis can help to stage and differentiate tumors with transmural spread to the extravesical space from tumors localized to the bladder wall. Furthermore, CT is relatively low-cost, more time-efficient, is often easier and faster to schedule than MRI or PET, and can assess nodal and metastatic stage. Unfortunately, CT is not optimal for accurately determining local extension and early lymph node involvement [5]. Specificity ranges from 68%–100% and does not consistently correlate with pathology because lymph node enlargement can be secondary to other etiologies [5].

Since the development of CT urography (CTU), IV pyelogram is no longer used in the diagnosis and surveillance of localized urothelial carcinoma of the bladder, ureter, and renal pelvis. Instead, a CT scan of the abdomen with and without contrast with delayed/excretory phase images can carefully assess the collecting system, ureters, and bladder, while evaluating the kidney for parenchymal lesions. The specificity and sensitivity of CTU in identifying the source of hematuria range from 83%–99% and 79%–95%, respectively [62-64]. Unfortunately, CTU cannot adequately stage the depth of invasion of the primary bladder tumor because it doesn't have the resolution to distinguish between individual layers of the bladder, nor can it distinguish high-risk MIBC from low-grade NMIBC [65]. In addition, CTU can miss lesions of < 1 cm and carcinoma in situ [66]. When CTU is employed following transurethral resection of bladder tumor (TURBT), inflammatory changes are indistinguishable from tumor extension, reducing the accuracy of local staging to approximately 60% [67]. Also, in patients on surveillance for NMIBC, upper urinary tract surveillance by CTU was inadequate, as only 29% of patients with upper urinary tract disease were diagnosed by routine CTU [68].

Although CTU is an excellent imaging modality for screening for hematuria and surveillance after NMIBC, it is of limited use in staging primary bladder tumors. However, an increase in bladder wall thickness of > 150% is associated with all-cause and urothelial carcinoma mortality [69]. Furthermore, CTU is only 70%–90% accurate in determining lymph node involvement, and false negative rates range from 25%–40%, likely due to CT's inability to distinguish between inflammatory and tumor-positive lymph nodes [68]. Nor is CT useful in detecting small (< 1 cm) lymph nodes with microscopic tumors. Overall, the accuracy of pelvic lymph node staging in localized bladder urothelial carcinoma by CT scan is 78%, with sensitivity and specificity ranging from 30%–53% and 94%–98%, respectively [70]. Thus, while CTU has replaced IV pyelogram for diagnosing and staging of urothelial carcinoma of the bladder, it has limitations (Figure 5).

Figure 5.

CT urogram multiplanar volume-rendered (MPVR) coronal reformation showing ureters in white, with normal peristalsis, hence incompletely seen the moment this CT was acquired. This was a 10-min delay for a hematuria workup.

DECT is a new imaging approach that is also being investigated in assessing metastatic urothelial carcinoma and RCC, as described above. DECT, which uses two merged CT datasets, provides additional characterization of the urinary tract and abdominal organs. In addition to technological iterative reconstruction solutions to reduced dose, dual energy and virtual non-contrast allow for what is called split-dose CT urogram. The subtraction of IV contrast from CT-acquired data is called virtual non-contrast CT (VNC). In this split-dose CT urogram, 1/3 of the total IV dose is given 8 mins prior to scan; the last 2/3 of IV contrast are injected 2 mins prior to scan. Giving the IV contrast 8 mins prior to scan effectively provides for a delay superimposed on the venous-phase CT images. Then VNC is applied to that dataset to subtract all IV contrast, providing the non-contrast images. This effective triple-phase exam (non-contrast, 2-min, and 8-min delay) is completed in one scan acquisition at 1/3 the radiation dose, since the delay and non-contrast pass are not needed [71]. In addition to the benefits of reduced radiation exposure, iodine mapping and other material characterization post-processing analysis have several benefits.

For patients undergoing treatment for metastatic urothelial carcinoma, there has been a concerted effort to improve tumor quantification with automated volumetric measurements, and to assess tumor necrosis using CT density measurements as a response to therapy (Figure 6) [72]. Many ongoing imaging studies are embedded in clinical trials of antiangiogenic agents, chemotherapy, and immunotherapy with checkpoint inhibition (NCT02788201, NCT01688999 and NCT02496208), with the goal of improving tumor burden and treatment response assessment in metastatic urothelial carcinoma.

Figure 6.

Demonstration of viable tumor volume (VTV) where a metastatic urothelial carcinoma lesion became less dense (less IV contrast enhancement) visually; however, the left shift of the histogram objectively represents likely tumor necrosis from an antiangiogenic treatment.

3.2. MRI in localized muscle-invasive bladder cancer

Routine MRI protocol includes mpMRI sequences, i.e., T1, T2, diffusion-weighted precontrast; T1-weighted gradient echo pre- and contrast-enhanced MRI sequences targeted to the urinary bladder/pelvis; and, when appropriate, a single non-contrast, T1-weighted sequence of the lumbar spine in the axial plane for further evaluation of possible bone metastasis. The mpMRI bladder/pelvis and lumbosacral sequences can be accomplished in a single session without moving the patient [73].

For MIBC and high-grade NMIBC, MRI can detect and stage tumors with relatively high sensitivity and specificity, with the advantage of not having the ionizing radiation associated with CT [5]. Though not commonly used, MRI of the pelvis can stage the bladder tumor after cystoscopic diagnosis. Although TURBT is the preferred, and most accurate, technique for staging NMIBC and MIBC tumors, retrospective analyses have found that TURBT has understaged cases by 42% [74]. MRI, on the other hand, provides extensive soft-tissue resolution. It has even been shown to detect T3 and T4 disease and has proven superior to CT in distinguishing between stages T2a and T2b by identifying bladder-wall invasion [75].

Traditional T1-weighted imaging can identify extravesical fat infiltration, pelvic lymphadenopathy, and suspected bone metastases, but is of limited use in local staging of urothelial carcinoma because these tumors appear similar to the normal detrusor muscle of the bladder wall due to similar signal intensities [76, 77]. On the other hand, T2-weighted imaging can differentiate between tumor edges and the bladder wall because urothelial carcinomas have increased T2-weighted signaling compared to normal detrusor muscle [78].

Triplanar anatomical T1- and T2-weighted sequences are used for local tumor staging of urothelial carcinoma, and have been used to make 3D virtual MR cystoscopies which have detection rates of > 90% [5]. In cases of conventional visualization obscured by significant hematuria, urothelial stricture disease, or lesions within the diverticula, virtual MR cystoscopy can provide an alternative to, and perhaps better visualization than, conventional modalities.

Other forms of MRI, including DWI-MRI and DCE-MRI, are commonly used in RCC and provide functional sequences that can differentiate between tumor and normal detrusor muscle. DWI-MRI can characterize tissue by measuring the Brownian motion of water molecules [79]. In DWI-MRI, malignant tissue has lower apparent diffusion coefficient values than normal tissue. Sensitivity and specificity have been shown to be 80% and 79%, respectively, but when used in combination with anatomic T2-weighted MRI, sensitivity and positive predictive value were 100% in a small cohort (n = 15) of patients. This combination has been shown to provide more accurate staging in patients in clinical stage ≤ T2 [80, 81]. On the other hand, DCE-MRI has been found to be superior to traditional unenhanced T1- and T2-weighted anatomic imaging owing to its ability to distinguish urothelial carcinoma from normal bladder wall. The hypervascularity of tumor tissue leads to earlier and better contrast enhancement compared to normal bladder wall. One study revealed 84% accuracy for DCE-MRI in tumor staging compared to 67% accuracy for unenhanced T1- and T2-weighted imaging [82]. A prospective study investigating the role of DCE-MRI in 122 patients demonstrated an 87.5% sensitivity for differentiation of lymph node-negative organ-confined UC from nonorgan-confined urothelial carcinomaand accuracy of 74%. However, the specificity of MRI findings were mediocre at 47.6% [83].

MR urography (MRU) is used to evaluate the urinary system while avoiding radiation exposure. However, the technology is time-intensive and costly, and requires technical expertise to produce adequate images. As with CTU, the use of MRU is predicated on adequate renal function, because IV gadolinium contrast is required to evaluate the urinary system. In the imaging of bladder tumors, MRU does not offer any improved ability to determine local stage of invasion [84]. However, T1- and T2-weighted imaging might confirm the presence of a bladder tumor. With T1-weighted imaging, a bladder tumor will likely appear as an area of intermediate intensity, while with T2-weighted imaging, a bladder tumor will appear hypointense to the surrounding urine. The Mayo Clinic described a large retrospective analysis of MRU, specifically examining differences between the parenchymal and pyelographic phases [84]. They found that the parenchymal phase was superior for lesion detection in the bladder. Overall accuracy ranged from 78%–87%; sensitivity and specificity ranged from 73%–85% and 75%– 91%, respectively [84]. Recently, a Danish group prospectively compared split-bolus CTU and MRU to flexible cystoscopy, and found that MRU had a slightly superior sensitivity to CTU with similar specificity (sensitivity: 76.9% vs. 61.5%; specificity: 93.4% vs. 94.9%) [85]. Although MRU provides no additional diagnostic information with or without CTU, it can serve as an alternative imaging modality in urothelial carcinoma.

3.3. FDG-PET/CT in metastatic UC

At diagnosis, 10%–15% of patients with urothelial carcinomaare found to already have distant metastatic disease, and 50% of those with true localized disease eventually develop metastasis within two years, even with aggressive therapy [86]. Molecular imaging is critical in accurately staging and evaluating response to treatment, and in developing a therapeutic strategy that avoids unnecessarily aggressive interventions and maximizes quality of life. FDG is the most established radiotracer, but is not widely used in urothelial carcinoma because it is excreted through the urinary tract and bladder. While FDG is not ideal for detecting primary disease, its greatest value is in assessing metastatic lesions and disease recurrence.

FDG-PET/CT has been studied in a few trials, most of which were limited by small patient numbers and lack of histopathologic correlation. Nevertheless, FDG-PET/CT has shown promise in evaluating metastatic disease. A systematic review and meta-analysis in 2012 found only six studies that fit standardized criteria, but pooled data on sensitivity and specificity of FDG-PET/CT for staging or restaging of metastatic disease was 82% and 89%, respectively. Its accuracy for diagnosing metastatic urothelial carcinoma lesions was deemed satisfactory [87].

Early studies have reported superior sensitivity and specificity for FDG-PET/CT in diagnosing urothelial carcinoma metastases. More lesions were detected with FDG-PET/CT than with CT or MRI, which substantially affected clinical management [60]. In evaluating response to therapy in metastatic urothelial carcinoma, Ozturk et al. found FDG-PET/CT and European Organisation for Research and Treatment of Cancer criteria more accurate in determining remission and overall response rate than CT and RECIST criteria [88]. FDG-PET/CT also appears useful in metastatic urothelial carcinoma, with an associated survival advantage in those with a favorable response on imaging after two cycles of first-line chemotherapy (Figure 7) [89].

Figure 7.

A 62-year-old female with metastatic urothelial carcinoma. Focal abnormal FDG metabolism in supraclavicular, mediastinal, hepatic, retroperitoneal, and osseous metastases.

There are conflicting viewpoints on preoperative staging. Scanning with FDG-PET/CT before radical cystectomy in MIBC has detected metastatic lesions not seen on conventional CT or bone scintigraphy, which significantly correlated with overall survival [90]. Another study demonstrated that FDG-PET/CT was more sensitive than CT at lymph node staging prior to radical cystectomy and had histopathologic validation from lymph node dissection [91]. However, another study asserted that FDG-PET/CT added only a small benefit in detecting lymph node metastasis outside the pelvis compared to CT, with 69% sensitivity and 95% specificity for FDG-PET/CT compared to 41% sensitivity and 98% specificity for CT, which was not considered enough of an advantage to justify the use of FDG PET/CT [92]. A recent study agreed that FDG is no better than CT in detecting metastatic lymph nodes in initial staging of MIBC prior to radical cystectomy [93].

Using FDG-PET/CT for restaging after surgical treatment appears more promising. With a sensitivity of 87.1% and specificity of 89.7%, it outperformed CT, ultrasound, and MRI in changing management and correctly restaging urothelial carcinoma after surgery [94]. Ozturk similarly demonstrated good sensitivity, specificity, positive predictive value, and negative predictive value for FDG-PET/CT in identifying local recurrence and distant metastases in restaging scans after radical cystectomy for MIBC [95].

Despite some disagreement on its value in staging, FDG-PET/CT appears to provide constructive information to guide treatment in this challenging malignancy. The technology is widespread, costs for the scanning procedure are reasonable, and the radiation exposure is acceptable for cancer patients. Imaging metastatic urothelial carcinoma with FDG-PET/CT can efficiently lead to advanced treatment options that may improve patient outcomes and pave the way for innovative therapies.

3.4. PET/MR in metastatic urothelial carcinoma

Both anatomical and nuclear medicine imaging techniques play an important role in the detection of primary and/or local relapse of UC, as well as lymph node and skeletal metastases. Most centers rely on mpMRI for this purpose. The main advantage of MRI is its excellent anatomic resolution, which makes it highly accurate in detecting local recurrence. The advantage of PET/CT is its ability to delineate biochemical or physiologic phenomena, making it easier to differentiate between benign and malignant lesions [96]. MRI-PET, FDA-approved in 2011, simultaneously performs a PET scan and an MRI scan, combining the advantages of PET and MRI, which may be even more beneficial in managing patients with cancer [97].

mpMRI of the bladder and pelvis combined with whole-body or targeted PET imaging may serve as a single staging study, or “one-stop” imaging tool, for patients with intermediate and high-risk urothelial carcinoma(Figure 8). In the PET/CT modality, CT does not provide the excellent soft-tissue contrast that MRI does, yet it gives patients a significant dose of radiation. MRI has the advantage of having no ionizing radiation [98]. Also, PET/CT imaging does not allow for true simultaneous imaging, as does PET/MRI [97]. MRI's anatomical resolution is superior to CT's, while PET provides metabolic and membrane receptor information. In addition to FDG, which is the most commonly used radiotracer in PET/CT, several other radiotracers are available for PET/MRI imaging in the management of urothelial carcinoma (Table 2). Based on preliminary studies, hybrid PET/MRI appears to provide superior sensitivity and specificity for tumor detection and characterization. Further studies are required to determine whether simultaneous acquisition of MRI and PET provide more accurate imaging than subsequent sequencing (Figure 9). In addition, PET/MRI's diagnostic accuracy, influence on therapeutic management, and related economic factors need to be carefully considered and evidenced by larger prospective clinical studies [73]. We found that PET/MRI allowed us to detect malignant lesions more accurately and confidently, potentially allowing physicians to improve treatment planning and disease monitoring (Figure 10) [97].

Figure 8.

mpMRI of the bladder combined with targeted PET imaging. Moderately distended urinary bladder activity. Adequate image magnification during the interpretation, and diluted radioactive urine by adequate hydration of the patient during and prior to image acquisition are simple factors that improve our ability to properly evaluate the wall of the urinary bladder. A moderately distended urinary bladder is more optimally imaged compared to a non-distended or excessively distended bladder, both of which may hinder accurate visualization of a bladder wall tumor.

Table 2.

Radiotracers used for PET/MRI imaging of urothelial carcinoma.

Metabolic Tracers	18F-FDG
	11C-choline
	18F-FM choline
	11C-acetate
Amino Acids	11C-methionine
	18F-FLT
	Anti-18F-FACBC
Receptor-Specific	18F-FDHT (androgen receptors)
	18F-FMISO (tumor hypoxia, ischemic tumors)
	124I-cG250 (Ab-CaIX- carbonic anhydrase IX clear cell carcinomas) identification of RCC*
	Bombesin – neuropeptide, 4 subtype
	18F-DCFPyL
	68Ga-PSMA (HBED-CC)

^*Potential agent for local radioimmunotherapy: 131I-cG250 (Ab-CaIX- carbonic anhydrase IX RCC)

fluorine-18-fluorodeoxyglucose (¹⁸F-FDG), fluorine-18- Fluoromethylcholine (¹⁸F-FM Choline), 3′-deoxy-3′ [(18)F]-fluorothymidine (¹⁸F-FLT), anti-1-amino-3-18F-fluorocyclobutyl-1-carboxylic acid (Anti-¹⁸F-FACBC), 18F-fluoro-5α-dihydrotestosterone (18F-FDHT), 18fluorine-fluoromisonidazole (18F-FMISO), 24I-labeled chimeric monoclonal antibody G250 (124I-cG250), 18F[fluoro-pyridine-3-carbonyl)-amino]-pentyl)-ureido)-pentanedioic acid (¹⁸F-DCFPyL), ⁶⁸Gallium-labelled ligand of the prostate-specific membrane antigen Glu-NH- CO-NH-Lys-(Ahx)-[68Ga(HBED-CC)] (⁶⁸Ga-PSMA (HBED-CC)), clear cell renal cell carcinoma (RCC)

Figure 9.

Urothelial carcinoma FDG-PET/MRI. Focal wall thickening and increased FDG activity of the posterior inferior bladder wall in its neck and within the superior portion of the prostate are present.

Figure 10.

Urothelial carcinoma FDG-PET/MRI. A 59-year-old male with high-grade non-muscle-invasive bladder cancer post-TURBT and intravesical chemotherapy. PET/MRI showed activity at or near the expected segment of intraprostatic urethra and retroperitoneal lymph nodes, upgrading the stage.

In summary, although current evidence shows a good correlation between PET/CT and PET/MRI in lesion detection of most oncologic diseases, including urothelial carcinoma, there are not enough clearly defined prospective studies with enough patients to validate these findings. Initial results are encouraging, particularly the reduction in radiation exposure. However, there is as yet no approved indication by which PET/CT is replaced by PET/MRI. PET/MRI may replace PET/CT for certain indications, such as a need for higher soft tissue resolution. For some tumors, the clear benefits of PET/MRI are its one-stop ease of use, comprehensive tumor staging, and potentially improved lymph node staging in the abdomen and pelvis. An ongoing study is testing PET/MRI in metastatic urothelial carcinoma in addition to conventional CT (NCT02788201).

4. Prostate adenocarcinoma

4.1. CT scan in staging and restaging of prostate adenocarcinoma

CT has a limited role in prostate adenocarcinoma staging, since many patients have metastases predominantly to bone [99]. CT is often used in the initial staging of advanced disease to follow nodal involvement and bone metastasis, and is especially useful in excluding extensive osseous metastasis or evaluating pathologic fractures. Although nodal metastasis on CT is specific, there is low sensitivity [100].

4.2. MRI in localized prostate adenocarcinoma assessment and biopsy

mpMRI of the prostate has been extensively studied in the last two decades. In prostate adenocarcinoma, it was initially used for localization and staging [101]; however, recently it has been increasingly employed to detect intraprostatic lesions for biopsy purposes. This shift in approach is a result of the limitations of transrectal ultrasound (TRUS)-guided systematic biopsies, which could lead to underdiagnoses of clinically significant prostate adenocarcinomas [102]. mpMRI can be employed for cognitive fusion-targeted, in-bore MRI guidance and TRUS/MRI fusion biopsies. The TRUS/MRI fusion method has been extensively studied [103]. In this technique, mpMRI's superior lesion detection is united with TRUS's real-time imaging and intervention capability to accurately guide a biopsy needle within the fused images by electromagnetic tracking or an angle-sensing encoder-loaded mechanical arm [103]. In several large-scale prospective studies, TRUS/MRI fusion-guided biopsy detected more clinically significant prostate adenocarcinomas than TRUS-guided systematic biopsy (Figures 11 and 12) [104-106]. In addition to lesion detection and cancer diagnosis by targeted biopsy approach, mpMRI-guided biopsy can also establish tumor heterogeneity, which can influence patient management and treatment outcomes [107]. Although it is not fully established in clinical routine, mpMRI targeted biopsy can potentially enable sampling of the worst portion of primary prostate adenocarcinoma lesions and aid accurate treatment management [108].

Figure 11.

A 68-year-old man with serum PSA of 4.88 ng/mL and history of 2 prior negative TRUS-guided biopsies. Axial T2W MRI shows a hypointense lesion in the left apical anterior transition zone (arrow) with capsular bulge (A). The lesion shows restricted diffusion on ADC maps (B) and b2000 DW-MRI (C) (arrows) and hypervascularity on DCE-MRI (D) (arrow). This PI-RADS 5 lesion underwent TRUS/MRI fusion-guided biopsy, which revealed Gleason 4+5 prostate adenocarcinoma.

Figure 12.

A 72-year-old man with serum PSA of 4.03 ng/mL with no prior biopsy. Axial T2W MRI shows a hypointense lesion in the left apical peripheral zone (arrow) (A). The lesion shows restricted diffusion on ADC maps (B) and b2000 DW-MRI (C) (arrows) and hypervascularity on DCE-MRI (D) (arrow). This PI-RADS 4 lesion underwent TRUS/MRI fusion-guided biopsy, which revealed Gleason 3+4 prostate adenocarcinoma.

Although mpMRI has been reported to accurately detect and stage prostate adenocarcinoma lesions, it still has certain limitations, mainly related to heterogeneity in image acquisition and interpretation and data processing, all of which strongly depend on the radiologist's level of experience. To address this, in early 2015, Prostate Imaging Reporting and Data System (v.2) guidelines were established by the American College of Radiology and the European Society of Urogenital Radiology [109] that aim to standardize all aspects of mpMRI. Initial prospective studies have revealed promising findings on its use in improving prostate adenocarcinoma care [110] and its impact on inter-observer agreement and related cancer detection [111, 112]. A multi-reader study with 34 patients revealed average sensitivity values of 91% and 63% for detecting index and all lesions across all readers, respectively. Index of specific agreement among readers was 93% for the detection of index lesions, 74% for the detection of all lesions, and 85% for scoring index lesions, and 58% for scoring all lesions [111]. However, more prospective studies are needed to better understand this new system. Several research groups are also working on developing new pulse sequences and signal-processing techniques to further improve the detection capability of mpMRI. New research developments include a) motion correction techniques like BLADE, a T2-weighted MRI sequence, b) use of calculated high b value diffusion-weighted images, and c) clinical use of hyperpolarized 13C MR, a new spectroscopic technique for investigating the metabolism of prostate adenocarcinoma [113-115]. These novel imaging techniques require further evaluation in larger studies.

In summary, while mpMRI has had a significant impact on prostate adenocarcinoma care through its use in biopsy guidance, it needs to be standardized and improved for better cancer detection and staging.

5. Nuclear medicine imaging in prostate adenocarcinoma

Prostate adenocarcinoma imaging is an area of vigorous molecular PET research, and could potentially affect hundreds of patients. Detection of tumor and accurate staging is critical in managing appropriate treatment strategies for patients with prostate adenocarcinoma. ¹⁸F FDG is not a new tracer but is the most common PET radiotracer commercially available for oncological use. However, its utility in prostate adenocarcinoma is confined to aggressive, poorly differentiated disease (Figure 13) [116, 117]. Several novel imaging agents have been developed in the last decade that show a better ability to identify malignancy. It is beyond the scope of this paper to discuss all available radiotracers under investigation or go into great depth, but we will highlight a few notable agents.

Figure 13.

A 53-year-old man with history of prostate adenocarcinoma. NaF-PET/CT demonstrates focal abnormal uptake in multiple bone metastases throughout the skeleton.

¹¹C choline was approved by the FDA in 2012 for evaluating biochemical recurrence in prostate adenocarcinoma patients. The mechanism of action is based on the increased use of choline for the accelerated production of prostate adenocarcinoma cell membranes. Many studies have assessed its value in prostate adenocarcinoma identification, with low sensitivity for primary disease and mixed reviews, demonstrating overall a moderate ability to detect nodal and metastatic tumors after definitive treatment by radical prostatectomy or radiation therapy, and a fair amount of false positive results [118-121]. A fluorinated version has also been developed and is currently available in Europe. Essentially, it is equally effective except for variations in physiologic distribution and a longer half-life radioligand that allows for wider transportation and delivery access.

¹¹C acetate also targets molecular pathways involved in prostate adenocarcinoma cell membrane assembly through upregulated fatty acid synthase, which preferentially converts acetate to fatty acids for incorporation into malignant cell membranes. It also is ineffective in primary disease detection, and is only marginally successful in detecting biochemically recurrent disease [122-124]. Radiolabeled choline and ¹¹C acetate generally perform similarly, but in patients with low PSA values they are not as effective in identifying tumor recurrence [125]. Moreover, with a short (20-min) half-life, ¹¹C acetate's use outside of large research institutions is limited.

¹⁸F fluciclovine (also known as ¹⁸F FACBC) is the newest agent to gain FDA approval (2016) for biochemical recurrence in prostate adenocarcinoma. An amino acid analog to leucine, it appears to have high uptake in prostate adenocarcinoma cells and performs comparably to ¹¹C acetate and radiolabeled choline in primary, recurrent, and advanced disease [126, 127]. In a recent multicenter trial with almost 600 patients, ¹⁸F FACBC identified tumor in only 68% of patients with biochemically recurrent disease [128]. With its recent approval, more experience will be gained to further assess its capabilities in detecting prostate adenocarcinoma in all phases.

Radiotracers under the most aggressive investigation are those that pursue prostate-specific membrane antigen (PSMA), a transmembrane receptor overexpressed on most prostate adenocarcinoma cells. Interestingly, the only other FDA-approved agent for detecting recurrent prostate adenocarcinoma is ¹¹¹In-capromab pendetide (ProstaScint®), an older radiolabeled PSMA antibody that has fallen out of favor due to poor performance. Next-generation compounds have been much more successful by targeting an external epitope to PSMA, unlike ProstaScint, which has an affinity for an internal epitope. Improved antibodies to PSMA are still being studied, but the small molecule receptor antagonists are rapidly influencing prostate adenocarcinoma imaging. ⁶⁸Ga-PSMA-HBED-CC PSMA (⁶⁸Ga PSMA), the most studied tracer in Europe, is slowly gaining attention in the US. Despite its brief history, a recent meta-analysis found that ⁶⁸Ga PSMA had similar or better sensitivities and specificities for detecting prostate adenocarcinoma than choline compounds [129]. Whereas ¹¹C acetate, radiolabeled choline, and ¹⁸F fluciclovine are best used in progressive prostate adenocarcinoma, PSMA tracers appear promising in all stages of the disease and could be helpful in distinguishing challenging benign prostatic tissue from primary tumor [130, 131]. Limited data available for ⁶⁸Ga PSMA indicate a considerable false negative rate, but low false positive findings [132, 133]. ¹⁸F DCFPyL, one of the newest low molecular weight inhibitors to join the field, has also shown similarly encouraging results in detecting primary, recurrent, and metastatic prostate adenocarcinoma and is logistically appealing, with a 110-min half-life and cyclotron radionuclide synthesis [134-136]. Further trials are underway exploring the ability of alpha- and beta-emitter PSMA inhibitors to therapeutically target prostate adenocarcinoma [137, 138].

¹⁸F NaF is the oldest and one of the most accurate radiotracers available for detecting bone metastasis. However, there is reluctance to abandon planar bone scintigraphy despite many publications demonstrating the superiority of ¹⁸F NaF [139-143]. Imaging prostate adenocarcinoma bone metastasis has traditionally relied on 99Tc bone scintigraphy, which is widely available in most communities but requires a 3- to 4-hour time commitment for patients and has only 47% sensitivity and 88% specificity. Moreover, it is not reliable in detecting early metastasis. If single photon emission computed tomography (SPECT) is added, the sensitivity improves to 82% and specificity to 99%. However, the field of view is limited and requires more scanning time, (usually an extra 45 minutes or more) and is therefore not routinely used [144].

An important distinction regarding ¹⁸F NaF is that it detects osteoblastic activity that could be in response to tumor, trauma, or infection and not direct cancer cells; therefore, specificity is low. Especially in the prostate adenocarcinoma patient population, degenerative bone changes due to osteoarthritis can sometimes be difficult to distinguish from malignant disease. These limitations are also found with planar bone scintigraphy. All the tracers previously discussed bind to prostate tumor cells and have the advantage of demonstrating focal uptake in soft tissue as well as bone lesions. Nevertheless, as a marker for bone metastasis, ¹⁸F NaF is a potent tool with better image quality and detection rate than bone scintigraphy. Recent research showed that ¹⁸F NaF PET/CT findings at baseline and over time correlated with clinical outcomes and survival in advanced prostate adenocarcinoma [145]. Against diffusion-weighted MRI, ¹⁸F NaF had better sensitivity but lower specificity for prostate adenocarcinoma metastasis to bone [146]. Research by the National Oncologic PET Registry found that as a staging mechanism for osseous metastasis, ¹⁸F NaF scans highly influenced patient management in hundreds of prostate adenocarcinoma cases [147]. Clinicians relied on ¹⁸F NaF over other advanced imaging techniques such as CT and MRI, and avoided invasive procedures. Further indications point to ¹⁸F NaF's value for monitoring treatment response and prognosis in novel bone-specific therapies, such as ²²³Ra and dasatinib [148-150]. ¹⁸F NaF is an excellent imaging agent for detecting or eliminating bone metastasis in prostate adenocarcinoma. It is emerging as a practical method for evaluating the effects of therapy on bone disease.

6. Conclusion

Multidetector CT, morphologic and functional MRI, and PET/CT can provide valuable information regarding diagnosis, staging, treatment planning, and follow-up of RCC and urothelial carcinoma. mpMRI of the prostate has had a significant impact on prostate adenocarcinoma care, but needs standardization and improved accuracy in detection and staging. PET, with its inherently high sensitivity, and PET/CT functional and molecular imaging are very useful for detecting distant metastases. These and other technologies and approaches, such as the use of positron-emitting radiotracer agents and hybrid imaging technologies such as PET/MRI, enable us to image malignant lesions more successfully than ever, which results in better patient care and outcomes [60]. These newer imaging modalities are increasingly being incorporated into therapeutic clinical trials of GU cancers, which will aid in further determining their clinical impact.