2022 WUOF/SIU International Consultation on Urological Diseases: Imaging of Renal Cell Carcinoma

Abstract

Imaging plays a central role in the contemporary multidisciplinary management of renal cell carcinoma. This article provides an overview of the current imaging modalities, including ultrasound, computed tomography, multiparametric magnetic resonance imaging, and molecular imaging, used in the evaluation of renal cell carcinoma. A summary of the imaging strategies for renal cell carcinoma staging and restaging post-treatment is provided.

Introduction

Imaging allows for the detection, characterization, staging, treatment planning and guidance, post-treatment evaluation, and surveillance of renal cell carcinoma (RCC). An understanding of the advantages and limitations of each imaging modality, and the evolving role of imaging in newer management strategies (such as active surveillance, ablation, and embolization) and the utility of newer therapeutics (such as antiangiogenic treatments and immunotherapy) is critical. The purpose of this narrative review is to provide an overview of the current imaging modalities, such as ultrasound (US), multidetector computed tomography (CT), multiparametric magnetic resonance imaging (MRI), and molecular imaging, used in the evaluation of RCC; highlight newer imaging techniques, such as contrast-enhanced US (CEUS) and novel molecular imaging agents, as well as radiomics with artificial intelligence technology; and provide a summary of the imaging strategies for RCC staging and post-treatment restaging. Imaging findings following newer treatment techniques, such as ablation and systemic therapy for advanced RCC, are beyond the scope of this article.

Detection and Diagnosis

Clear cell renal cell carcinoma (ccRCC) (70%–80%), papillary renal cell carcinoma (pRCC) (10%–15%), and chromophobe renal cell carcinoma (chRCC) (5%) are the 3 most common histologic subtypes of RCC [1], which as a group show a broad spectrum of imaging appearances. Multiphase contrast-enhanced CT is the imaging modality most commonly used for the evaluation of RCC [2]. A systematic review found that the median sensitivity and specificity of CT for the diagnosis of RCC were 88% and 75%, respectively [3]. A limitation of CT is the necessity for intravenous contrast agent, which may be contraindicated owing to renal dysfunction or iodine allergy.

Multiparametric MRI is frequently used to further characterize renal masses that are indeterminate on CT, but can be used as the initial study for the evaluation of renal masses, especially in patients with contraindication to iodinated contrast material [2]. A systematic review showed that the median sensitivity and specificity of MR for the diagnosis of RCC were 87.5% and 89%, respectively [3]. A limitation of MRI is the contraindication of intravenous contrast in advanced renal disease owing to potential for nephrogenic systemic fibrosis.

Conventional US can be utilized to triage an indeterminate renal lesion incidentally detected at single-phase CT to determine whether it is a simple or minimally complex cyst or a solid lesion. A systematic review showed that the sensitivity and specificity of conventional US for the diagnosis of RCC were 46% and 12%, respectively [3]. Limitations of conventional US include its reduced sensitivity for the detection of small renal masses and the inability to further characterize solid renal masses. CEUS is superior to CT and MRI for the evaluation of septa and mural enhancement in complex cystic renal lesions [4,5] (Figure 1), and its median sensitivity and specificity for the diagnosis of RCC in these lesions were reported to be 94.5% and 69%, respectively [3]. However, there is no current role for CEUS in the Bosniak classification scheme, but proposals for an adaptation of the scheme incorporating CEUS features have been suggested [6]. Suboptimal image acquisition owing to patient factors, such as obesity, inability to breath-hold, and acoustic shadowing from bowel gas or ribs, can limit the utility of CEUS.

18F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) is not recommended by major societies, including the American Urology Association (AUA), European Association of Urology (EAU), and National Comprehensive Cancer Network (NCCN), for the routine diagnosis or evaluation of RCC because the technique is limited by the physiologic excretion of radiotracer through the kidneys and the baseline FDG uptake in normal renal parenchyma, which can obscure part or all of the renal tumor [7,8,9]. Systematic reviews found that FDG PET/CT for RCC detection showed high specificity but variable sensitivity depending on the size, subtype, and grade of RCC [3,10].

Imaging Features of Common Subtypes of RCC

Clear Cell Renal Cell Carcinom

At US, ccRCC has variable appearances. It typically appears as a heterogeneously hypoechoic or isoechoic mass (Figure 2), but may show hyperechoic components [11,12]. Fluid components may be present due to cystic, necrotic, or hemorrhagic change. Doppler flow is readily identified owing to its hypervascular nature. At CEUS, ccRCC shows avid, early enhancement, followed by washout [13] (Figure 3). At CT and MRI, ccRCC is typically exophytic and shows vivid early contrast enhancement [14]. It has low-to-intermediate T1 signal and high T2 signal compared to adjacent renal parenchyma [15]. Internal tumor heterogeneity can occur owing to areas of hemorrhage, necrosis, and/or cystic degeneration, which appear as nonenhancing regions [16]. ccRCC may show reduced signal on opposed-phase chemical shift MR images compared to in-phase images owing to intracellular fat [17]. A peritumoral pseudocapsule may be present, which appears as a regular low or high attenuation rim on CT [18], and low T1 and T2 signal on MR images [19]. Calcifications are uncommon [20] (Figure 4 and Figure 5).

Papillary Renal Cell Carcinoma

At US, pRCC may appear as a solid, well-circumscribed mass (Figure 6), or sometimes may be partially solid with cystic or hemorrhagic components [11]. At CEUS, it shows hypoenhancement with a later and lower peak of enhancement compared to ccRCC [13] (Figure 7). At CT and MRI, pRCC is generally a small peripheral homogeneous mass that has low T2 signal compared to renal cortex, and shows weak enhancement that progressively increases on subsequent phases [21]. pRCC may show loss of signal on in-phase images compared to opposed phase at chemical shift MRI owing to hemosiderin [22]. Some pRCCs appear as hemorrhagic cystic masses with enhancing papillary projections [23]. Calcifications occur in 7% of cases [21] (Figure 8 and Figure 9). Type 1 and 2 pRCCs cannot be reliably differentiated on imaging, but type 2 pRCCs are more likely to be heterogeneous, show infiltrative margins, and contain calcifications [24].

Chromophobe Renal Cell Carcinoma

At imaging, chRCC is typically a solid, well-circum-scribed mass that is more homogeneous than ccRCC. At CEUS, its enhancement is often nearly isoenhancing to renal cortex and can be difficult to discriminate from ccRCC, especially with small tumors [13]. chRCC shows heterogeneous T2 signal on MRI [25]. At CT and MRI, it shows intermediate contrast enhancement in between that of ccRCC and pRCC [14,23]. A central scar, spokewheel enhancement pattern and segmental enhancement inversion may be present, but these features overlap with oncocytoma [25,26,27]. Calcifications occur in 14% to 38% of cases, and perinephric infiltration and venous invasion are uncommon [20,26].

Differentiation of RCC from Benign Renal Tumors

Imaging is unable to reliably discriminate between benign and malignant renal masses owing to overlapping imaging characteristics in 10% to 15% of cases [28]. RCC can be challenging to differentiate from oncocytoma and lipid-poor angiomyolipoma (AML). However, composite imaging features can suggest a likely diagnosis. AML is a typically homogeneous and markedly hyperechoic mass on US, but up to 30% of small RCCs may be hyperechoic, and a definitive diagnosis of AML cannot be established on US appearances alone [12]. At CEUS, AML typically shows homogeneous hypoenhancement relative to renal parenchyma and can be difficult to distinguish from pRCC and chRCC [12,13]. Macroscopic fat within a noncalcified renal mass on CT is almost diagnostic of an AML. Macroscopic fat rarely occurs in RCC [29]. Intracellular fat can be identified in clear cell RCC but this feature in isolation does not allow its differentiation from lipid-poor AML [30]. A renal mass containing fat with calcification or one that shows necrosis is more likely to be an RCC than AML [23,29].

pRCC can be differentiated from a hemorrhagic cyst and lipid-poor AML because it shows weak progressive contrast enhancement. Hemorrhagic cyst shows no contrast enhancement [31], and lipid-poor AML shows avid early contrast enhancement with subsequent contrast washout [32].

chRCC and oncocytoma show multiple overlapping imaging features and are most challenging to differentiate from each other [25,33]. At CEUS, oncocytoma typically shows hyperenhancement and can show persistent delayed enhancement, but the features are inadequate to allow for confident discrimination from chRCC [12]. Quantitative imaging parameters, such as tumor enhancement characteristics [34,35], diffusion-weighted MRI [36], and texture analysis [37], have shown some ability to differentiate between benign and malignant renal masses.

Differentiation of Subtypes of RCC

Imaging is as of yet unable to reliably differentiate between the subtypes of RCC owing to overlapping imaging characteristics. A study showed the performance of CT to predict ccRCC and chRCC on morpho-logic features alone had a positive predictive value of less than 75%, but evaluation of their contrast enhancement profile allowed for differentiation of ccRCC from other subtypes with a sensitivity, specificity, and accuracy of 64%, 87%, and 75%, respectively [34]. Dynamic contrast-enhanced MRI studies found that RCC subtypes showed contrast enhancement profiles concordant with CT findings, but considerable overlap occurs and does not allow for definitive tumor histologic subtyping [14,35]. The application of algorithmic and scoring systems, such as the clear cell likelihood score, would help to achieve greater accuracy [38,39]. Type 1 and 2 pRCCs show overlapping imaging findings that do not permit reliable differentiation between them on morphological features [14,24] or metabolic parameters [40,41]. Early dynamic imaging with FDG PET/CT may be more helpful than traditional static scanning in distinguishing aggressive RCC subtypes. A study showed that the maximum standardized uptake value (SUVmax) from dynamic scans was higher in ccRCC than non-ccRCC [42]. Another study showed that chRCC demonstrated lower SUVmax values than ccRCC and pRCC, but there was no significant difference between ccRCC and pRCC [43].

Grading of RCC

Nuclear grade of RCC correlates with patient survival [44]. Imaging features that act as accurate surrogate markers of histologic grade of RCC would allow for noninvasive prediction of prognosis and triage management. Most studies have attempted to differentiate between low- and high-Fuhrman grade ccRCC. One study showed that the sensitivity and specificity of MRI to diagnose low-grade ccRCC were 50% and 94%, respectively, and to diagnose high-grade ccRCC they were 93% and 75%, respectively [45]. Another study showed no significant correlation between histologic grade and MRI features for pRCC and chRCC [46]. Morphologic imaging features suggestive of higher grade tumor or sarcomatoid dedifferentiation include larger tumors with intratumoral necrosis, calcification, infiltrative margins, increased peritumoral neovascularity, larger peritumoral vessels, and renal vein thrombosis [45,47,48]. An uncommon predominantly cystic appearance of ccRCC has been shown to have low-grade malignant potential [49]. FDG PET/CT studies showed that higher SUVmax and tumor-to-normal reference tissue ratios corresponded to more aggressive RCC features, such as higher TNM stage and Fuhrman grade, as well as presence of venous and lymphatic invasion [42,43]. One FDG PET/CT study showed higher maximum, mean, and peak standardized uptake values in RCC with sarcomatoid differentiation compared to ccRCC [50]. Other metabolic measures, such as metabolic tumor volume and tumor-to-liver ratios, also appear to correlate with RCC grade [51,52]. Quantitative imaging parameters, such as tumor enhancement characteristics [53], diffusion-weighted MR imaging [54], and texture analysis [55,56], have shown some correlation with nuclear grading.

Staging

The 8th edition of American Joint Committee on Cancer TNM staging manual is the most commonly used staging system for RCC [57] (

Table 1. T staging categories.

). The American College of Radiology appropriateness criteria recommend CT or MRI of the abdomen without and with contrast as the most appropriate imaging modalities to stage RCC [58]. CT and MRI have similar accuracy for the staging of the primary tumor [59,60]. However, CT is more commonly utilized owing to its ready availability and rapid acquisition time. MRI is generally utilized when iodinated contrast medium administration is contraindicated. US is generally considered inferior to CT or MRI in staging and post-treatment evaluation for RCC, but it has a role in select patients. CEUS can be helpful in certain situations if CT/MRI remains indeterminate, or if CT/MRI cannot be performed with contrast (Figure 10). CT chest is recommended at initial staging, as the lungs are the most common site of RCC metastases [8,61]. Targeted imaging should be considered in patients with organ-specific symptoms, such as MRI or CT of the brain in patients with neurological symptoms, or bone scintigraphy in patients with bone pain, elevated alkaline phosphatase, or radiographic findings suggestive of bone metastases [61]. Bone metastases from RCC are typically lytic, and the poor osteoblastic response may limit the uptake of radiotracer at bone scintigraphy. One study of patients with stage IV RCC showed that the sensitivity of bone scintigraphy for the detection of osseous metastases was 29% [62].

Evaluation of Primary Tumor

Key imaging features of the primary tumor to be evaluated include the tumor’s size, location, degree of local invasion (into collecting system, perirenal fat, perirenal fascia, and adjacent organs), and renal vascular anatomy [63]. However, both CT and MRI may underestimate tumor size, as well as early urinary collecting system, renal sinus fat, and perinephric fat invasion, compared to pathologic examination, which may result in tumor upstaging [64,65,66,67].

Evaluation of Nodes and Distant Metastases

The most common sites of RCC metastases, in descending order of frequency, are the lungs, bones, liver, lymph nodes, adrenal glands, and brain [68,69]. However, metastases to any organ can occur. Cross-sectional imaging criteria for the diagnosis of metastatic lymph nodes rely on size larger than 1 cm in short-axis diameter, abnormal shape, disruption of the normal lymph node architecture, and abnormal contrast enhancement characteristics mirroring those of the primary tumor [70]. The accuracy of CT and MRI for lymph node staging is 83% to 88% [69]. Both CT and MRI are unable to differentiate enlarged, reactive nodes from metastatic lymph nodes or identify micrometastases in normal-sized lymph nodes [69,71]. A review showed that FDG PET had a sensitivity and specificity of 75% and 100%, respectively, for the detection of lymph node metastases in RCC [71] (Figure 11).

Imaging in Follow-Up

In RCC, 80% to 85% of tumor recurrence occurs within the first 3 years following surgery [69,72]. The incidence of local recurrence at the surgical bed following surgery for localized RCC is about 2% [69]. Risk for tumor recurrence following surgery depends on the pathological size, stage, grade, and histologic subtype of the primary tumor [69]. Pathological stage and grade of the primary tumor enable risk stratification of surgical candidates [57,61]. Patients with positive surgical margin are considered to be in at least one higher level of risk category than that based upon their surgical specimen [61].

There is no consensus on the surveillance program following treatment. A risk-based postoperative surveillance schedule has been recommended by the AUA [61] as well as EAU [8]. Contrast-enhanced CT or MRI of the abdomen as well as chest imaging are suggested with each follow-up visit [61]. Chest radiograph is recommended for those in low-risk and intermediate-risk categories, and chest CT is recommended for those in high-risk and very high-risk categories [61]. US alternating with CT or MRI may be considered in low-risk and intermediate-risk groups after the initial 2 years of follow-up after surgery or ablation, and in active surveillance of localized renal masses [61]. Patients managed with ablative treatments are recommended to follow an intermediate-risk category surveillance schedule [61]. Patients with relapse, stage IV disease, and surgically unresectable disease are recommended to undergo CT or MRI every 6 to 16 weeks at the physician’s discretion and patient’s clinical status [9].

A detailed discussion of the imaging manifestations following ablation and systemic therapy with targeted agents, such as antiangiogenic agents and immunotherapy, is beyond the scope of this article. Traditional evaluation of tumor size to determine therapy response may be inadequate in these settings. Imaging findings supportive of favorable response include development of marked necrosis, decrease in tumor attenuation, and change in pattern of enhancement [73].

Currently, AUA recommends that PET/CT should not be routinely obtained but may be considered in select cases [61]. A meta-analysis showed that the pooled sensitivity and specificity were 86% and 88%, respectively, of 18F-FDG PET/CT for the detection of metastatic disease in RCC [74]. Another study showed that PET/CT was comparable to CT for the detection of metastatic disease after surgery [75]. PET/CT may have prognostic benefit and can influence clinical decision. A study showed that positive PET/CT scan correlated with lower progression-free survival at 3 years and lower overall survival by 5 years, which affected management decision in 43% of patients [76]. Another study showed that the high number of FDG-positive RCC metastases or metastases with high SUVmax at baseline PET/CT were linked to shorter overall survival [77]. Furthermore, the study showed that disease progression on PET/CT at 16 weeks after start of treatment correlated with decreased overall survival and progression-free survival. Qualitative metrics, such as total lesion glycolysis and metabolic tumor volume, were also shown to be predictive of overall survival and progression-free survival [78,79].

A summary of the currently recommended imaging modalities to be utilized according to the stage of the disease is included (

Table 2. Recommended imaging modalities for evaluation at each clinical stage of disease.

).

Imaging-Assisted Interventions

Renal mass biopsy, percutaneous tumor ablation, and intraoperative surgery can be assisted by real-time imaging. Image-guided percutaneous renal mass biopsy has become more commonly performed and can be guided by US or CT. Biopsy may help to avoid surgery by demonstrating benign pathology, or if showing malignancy can help guide management decision to surgery, ablation, or active surveillance. Image-guided percutaneous thermal ablation of a renal mass is an alternative to surgery in select patients with localized tumors and can be potentially curative. Intraoperative US can be utilized as an adjunct to assist surgery by increasing confidence in selection of the site of parenchymal transection, aid in evaluation of the relationship of the mass to renal vessels and renal collecting system, verify extent of inferior vena caval tumor, and aid in detection of additional lesions [80,81]. Three-dimensional (3D) imaging technology, such as 3D printing model, augmented reality, and mixed reality technology, is a novel application of CT or MR imaging dataset to produce a visually concise representation of a renal tumor to improve its localization within the kidney and understand its relationship to relevant anatomical structures. Three-dimensional printing models and augmented reality have been utilized for preoperative surgical planning in complex cases [82] and for patient counseling [83].

Future Directions

A number of novel imaging techniques are being investigated to further characterize indeterminate renal masses including elastography, dual-energy spectral CT and perfusion CT, novel PET radiotracers, 99m Technetium sestamibi, and the utility of radiomics with artificial intelligence [84,85].

Advanced US techniques, such as elastography, are being studied for their potential to differentiate between benign and malignant renal masses [86]. Advanced CT techniques, such as dual-energy spectral CT and perfusion CT, are being studied but their exact role in renal mass CT protocol is unclear. Studies have shown mixed results in the ability of dual-energy spectral CT and perfusion CT to differentiate between benign and malignant renal masses [87], RCC subtypes [88], and RCC histologic grade [89]. The higher radiation dose penalty and more challenging technique of perfusion CT may limit the technique’s wider utility in comparison to dual-energy spectral CT [90].

Novel PET radiotracers linked to specific proteins, such as prostate-specific membrane antigen (PSMA) and carbonic anhydrase IX (CAIX), are under current investigation for the evaluation of RCC.

A systemic review showed that PSMA PET/CT has a potential role in staging, restaging, and predicting treatment response, but not for primary tumor evaluation [91]. It appears superior to FDG PET/CT for detection of local recurrence and bone metastases [92]. CAIX is a cell-surface antigen that is highly expressed in ccRCC but not found in other RCC subtypes or benign renal tissue (Figure 12). Girentuximab is an anti-CAIX monoclonal antibody. Preliminary studies showed that 89Zr-girentuximab PET/CT was able to differentiate between ccRCC and non-ccRCC [93], and improved detection of RCC metastases compared to CT alone or CT in combination with FDG PET/CT [94]. Theragnostic applications directed at PSMA and CAIX are being explored [95].

99m Technetium sestamibi is a radiotracer that accumulates in mitochondria-rich cells, and is commonly utilized in myocardial and parathyroid scintigraphy. Renal oncocytoma has high mitochondrial content compared to chRCC. A meta-analysis showed that 99m Technetium sestamibi scintigraphy had a pooled sensitivity and specificity of 92% and 88%, respectively, for detecting renal oncocytomas versus other renal lesions, and 89% and 67%, respectively, for detecting renal oncocytoma versus chRCC [96]. Novel application of this radiotracer to further characterize indeterminate renal masses would allow for triage of suspected oncocytomas to active surveillance.

Radiomics with artificial intelligence is an emerging field that uses computational methods to extract quantitative metrics, such as shape, size, and texture, from any standard clinical image dataset, such as CT, MRI or PET/CT, which can then be used to help differentiate between benign and malignant renal masses, predict nuclear grade, and evaluate gene expression profile [97]. Preliminary studies have shown that radiomics allows for differentiation of benign from malignant renal masses with CT [98,99] and MRI [100,101]. One CT study showed that sensitivity and accuracy were 85.8% and 74.4%, respectively, in differentiating ccRCC from oncocytoma [98]. Another CT study of 127 patients with RCC showed a sensitivity, specificity, and accuracy of 89%, 92%, and 87%, respectively, for differentiating ccRCC from non-ccRCC, and 87%, 92%, and 78%, respectively, for differentiating pRCC from chRCC [102]. A further CT study of 62 patients with pRCC showed 84% accuracy in differentiating between type 1 and type 2 pRCC [103]. An MRI study found a sensitivity, specificity, and accuracy of 92%, 41%, and 70%, respectively, for distinguishing benign from malignant renal masses when utilizing deep learning algorithms [100]. Studies have shown feasibility of radiomics to differentiate between low- and high-grade RCC [104,105]. One study of 53 patients showed a sensitivity, specificity, and accuracy of 91.3%, 80.6%, and 85.1%, respectively, for predicting high-grade from low-grade clear cell RCC [106].

Radiogenomic studies have shown that BRCA1-associated protein 1 mutation is associated with ill-defined tumor margins and presence of calcification, and more commonly seen with higher grade RCC [107]. Mutation of mucin 4 is found to be associated with exophytic tumor growth and reduced survival [107], while mutation of lysine demethylase 5C is found to be associated with renal vein invasion and reduced survival [108]. However, a systemic review of 57 studies found that translation of radiomics into clinical practice remains technically challenging owing to several factors including heterogeneous image acquisition protocols, reproducibility of radiomics signature, and big data sharing [109].

Conclusion

Imaging plays a central role in the clinical detection, staging, and follow-up of patients with RCC. Contemporary management of RCC has emphasized the role of imaging in the multidisciplinary care of these patients. Clinicians should be cognizant of the strengths and limitations of the different imaging techniques. Newer imaging techniques and the nascent role of artificial intelligence may translate into future clinical practice.