Update on radiological imaging of renal cell carcinoma
Article first published online: 13 APR 2007
Volume 99, Issue 5b, pages 1217–1222, May 2007
How to Cite
Coll, D. M. and Smith, R. C. (2007), Update on radiological imaging of renal cell carcinoma. BJU International, 99: 1217–1222. doi: 10.1111/j.1464-410X.2007.06824.x
- Issue published online: 13 APR 2007
- Article first published online: 13 APR 2007
- computed tomography;
- magnetic resonance imaging;
- contrast agents;
USA Food and Drug Administration
field of view
nephrogenic systemic fibrosis.
The rapid growth in the use of cross-sectional imaging has resulted in the increased detection of small asymptomatic renal masses. RCC is the most common primary malignant neoplasm of the kidney; 35 000 patients are diagnosed annually with RCC and there are >12 000 deaths annually in the USA . RCC constitutes 85–90% of all renal tumours; up to 60% might be serendipitously discovered on cross-sectional imaging. These tumours are generally smaller and with an earlier tumour stage, and therefore a better prognosis [2–4]. In turn, this has led to the increased use of minimally invasive techniques, e.g. partial nephrectomy, laparoscopic resection, robotic surgery, radiofrequency ablation and cryotherapy. Increasing sophistication of surgical approach has been mirrored with advances in radiological imaging.
To enable the surgeon to make appropriate preoperative decisions and to operate with confidence in an environment with a restricted field of view, imaging must now offer more accurate detection, localization and characterization of smaller masses, and an exquisite depiction of the arterial, venous and lymphatic anatomy (Fig. 1). In this review we update the latest advances in ultrasonography (US), CT and MRI for RCC.
US is an appropriate initial test for nearly all patients who present with symptoms suggestive of renal disease. It is quick, cost-effective, noninvasive, widely available and involves no radiation. The main limitation of US is the limited ability to detect small renal lesions, particularly if they do not distort the renal parenchyma. US only offers a sensitivity of 67% for tumours that are ≤ 3 cm in diameter . Jamis-Dow et al. showed that CT depicted more renal masses and smaller renal masses than US. In their study of 205 masses with pathological correlation, CT and US detection rates for lesions of 0–30 mm (grouped in 5-mm steps, i.e. 0–5, 5–10 mm, etc. were, respectively, 47% and 0%; 60% and 21%; 75% and 28%; 100% and 58%; 100% and 79%; and 100% and 100%.
The kidney is best evaluated with a 2–3 MHz ultrasound transducer; in larger patients the lower frequency permits greater tissue penetration. In smaller patients, the higher frequency transducers offer improved spatial resolution.
TISSUE HARMONIC IMAGING
Tissue harmonic imaging offers an alternative to conventional US and is now routinely available on modern ultrasound machines. Ultrasound transducers emit sound waves at one frequency. As sound waves pass through the body, there is distortion of the wave that alters the frequency composition such that additional frequency components are generated. These additional components comprise integral multiples of the transmission frequency (i.e. 2f, 3f, 4f, etc.). In this manner, ultrasound energy at frequency f is converted to ultrasound energy at higher frequencies as the beam passes through the patient. The degree of conversion depends on the physical properties of the tissue.
With conventional US, transducers only detect the reflected sound waves at the transmission frequency f (i.e. the first harmonic). With harmonic imaging, the transducer detects the reflected sound waves at higher-order harmonics. Because the amplitude of the harmonics decreases with order, the second harmonic (i.e. 2f) is typically used. Harmonic imaging has several advantages over conventional US, including improved axial and lateral resolution, and less reverberation and side-lobe artefacts, and increased contrast resolution than standard US. It is particularly useful for a more accurate characterization of renal cystic lesions .
REAL-TIME COMPOUND US
In conventional US, real-time images (and static images) are constructed from data obtained at one constant angle. In compound US, a composite (i.e. compound) image is obtained by combining data from several different angles. Some studies showed that compound US images result in fewer image artefacts and better image contrast than conventional images. Limitations include blurring from motion between consecutive data acquisitions .
US-SPECIFIC I.V. CONTRAST AGENTS
Intravenous ‘microbubble’ contrast agents typically consist of an encapsulating shell (made of phospholipids or albumin) surrounding a gas such as nitrogen or a fluorocarbon. These particles are relatively stable in the bloodstream and provide a highly reflective interface for ultrasound at typical frequencies used in medical imaging. In general, the degree of increased echogenicity in a renal lesion will depend on the relative perfusion of the lesion compared to the renal parenchyma (therefore yielding information similar to a contrast-enhanced CT or MRI). Preliminary studies show that contrast-enhanced US provides more information than colour Doppler imaging alone . These agents have not been approved by the USA Food and Drug Administration (FDA) for use on non-cardiac applications in the USA. However, they have been used in Europe for the last few years. Given the relatively low sensitivity of US for detecting small renal masses these agents offer great potential for improving the performance of US in this area in those patients for whom CT or MRI is contraindicated or difficult .
RCC: US APPEARANCE
RCC on US is generally a solid mass and can be isoechoic, hyperechoic or hypoechoic relative to the surrounding parenchyma. Traditionally RCCs on US were described as hypoechoic or isoechoic to the adjacent renal parenchyma. Angiomyolipomas were described as hyperechoic. However, RCCs can also appear hyperechoic so that there is an overlap in the appearance of these two tumours . The diagnosis of an angiomyolipoma can be usually made by detecting fat within an otherwise solid mass on non-contrast CT (Fig. 2a,b). Small lesions with less necrosis are more likely to be hyperechoic. Isoechoic tumours can only be detected by distortion of the renal contour, focal enlargement of a portion of the kidney, or distortion of the fat in the central sinus.
CT scanner technology has progressed rapidly since the introduction of the first helical (or slip-ring) scanners in the early 1990s. The earliest helical CT scanners allowed continuous data acquisition as a patient was advanced through the scanner. A thin fan-shaped beam was used with a single row of detectors. Slip-ring electrical connections allowed for continuous rotation of the source and detectors as the patient was moved through the scanner. In recent years, additional rows of detectors have been added and the beam shape has been widened (so-called cone-shaped X-ray beams) to simultaneously expose many rows of detectors (up to 256 currently in development) and thereby obtain a true volume scan. This latest CT technology, usually referred to as multidetector CT (MDCT), has reduced acquisition times such that scans can be obtained through nearly the entire body in 10–20 s. In addition, the MDCT scanners allow ultra-thin sections to be obtained (<0.5 mm) with minimal time for motion artefact. Some of these ‘extreme’ MDCT scanners are designed for cardiac imaging, where sub-second imaging for each slice is necessary. Most renal imaging can be done satisfactorily on a basic helical scanner with one to four rows of detectors.
The main advantages of MDCT is that the imaging time is significantly faster, e.g. with a 16-row scanner we can scan a patient from head to toe in <1 min. Therefore we can scan through the entire kidney in <10 s, which allows scanning in multiple phases of enhancement and lessens the problems of respiratory motion. These thinner slices through the kidney allow near-isotropic imaging (i.e. voxels of identical size in all three dimensions) and exquisite three-dimensional (3D) images. Such 3D imaging (Fig. 3a,b) and CT angiography (Fig. 4) are very helpful in identifying lesion location for the surgeon and for vascular mapping, and should be considered for patients who are candidates for partial nephrectomy. CT examination times (5 min with delayed scans) are much shorter than for MRI in general (at least 30 min) and offer better patient acceptance and more reliable image quality with minimal motion artefact.
Non-contrast, arterial (15–25 s), corticomedullary (30–60 s), nephrographic (80–180 s) and delayed images (3–10 min) can be taken through the kidneys. Non-contrast scans detect calcifications and allow quantification of enhancement on the post-contrast scans. Arterial-phase scans are taken to depict the arterial anatomy and perform CT angiography. The corticomedullary phase provides information on the venous and arterial vasculature and renal vascular lesions. The nephrographic phase is the most sensitive for detecting and characterizing lesions. The excretory phase is used to evaluate the renal collecting system and ureters. For patients who have RCC, this is useful for assessing the relationship of the tumour to the collecting system.
RCC: CT APPEARANCE AND ROLE OF CT IN STAGING
On the non-contrast scan the tumour is generally a solid lesion with an attenuation value of ≥ 20 Hounsfield Units. Lesions that are <3 cm are usually homogeneous. Larger lesions have a variable appearance often showing central necrosis and or haemorrhage. RCCs have a rich vascular supply; if there is an increase in attenuation of the lesion of >15 Hounsfield Units from the baseline scan, this is considered significant enhancement.
Recent studies have shown that clear cell RCC is generally vascular and heterogeneous, papillary RCC tends to be hypovascular and homogenous (Fig. 5a,b), and that chromophobe RCCs and oncocytomas tend to be vascular and homogenous [11–13]. The importance of suggesting the subtype before surgery means that the surgeon can try to plan the approach accordingly. It can offer an additional reason to propose that the tumour can be removed by minimally invasive techniques.
Studies have shown that CT has a staging accuracy of up to 91%[14,15] making it the imaging method of choice for most patients. One of the main limitations in staging with CT occurs with correct identification of perirenal fat invasion (Fig. 6), to distinguish T2 and T3a lesions. This remains a problem even with 3D CT and newer multislice technology. Perinephric stranding is a non-specific sign and can be seen secondary to inflammation, oedema (e.g. related to previous obstruction) and vascular engorgement.
Another remaining problem for MDCT is the imaging of the superior extent of tumour thrombus. On CT, it is difficult to accurately obtain optimal venous opacification and differentiate bland from tumour thrombus. In addition, radiation exposure considerations on CT limits the number of imaging passes through the kidney. Therefore it would seem at present that MRI remains the imaging test of choice for delineating tumour thrombus [16,17].
MRI of the kidney is now used routinely with a torso phased-array surface coil. MRI surface coils receive the MRI signal after an excitation. When using a single surface coil, the effective field of view (FOV) is proportional to the diameter of the coil, while the signal-to-noise ratio (SNR) is inversely proportional to the coil size. Therefore, a small surface coil provides the best SNR but has a limited effective FOV. Phased-array coils contain several small surface coils that are used simultaneously and independently to receive the MRI signal and reconstruct a composite MR image. Phased-array coils provide the effective FOV of a much larger coil, but the SNR of a small surface coil. The improved SNR of a phased-array coil allows thinner sections to be obtained with higher in-plane spatial resolution and reduced imaging times.
High field-strength MRI (usually defined as ≥ 1.0 T) using a phased-array surface coil and improved software allows imaging of the entire kidney in one breath-hold. A further reduction of imaging time can be obtained using phased-array coils with so-called ‘parallel imaging’, a major advance in MR technology. It reduces the required number of phase-encoding steps (and hence reduces imaging time because imaging time is directly proportional to the number of phase-encoding steps) by exploiting low-resolution data (from the individual coils) that reflects coil sensitivity, to combine aliased images from the individual coils into a single composite image with the aliasing removed. These techniques are known by various acronyms including SENSE (sensitivity encoding), IPAT (integrated parallel acquisition techniques), SMASH (simultaneous acquisition of spatial harmonics), SPEEDER and ASSET (array spatial sensitivity encoding technique) by different vendors .
Higher field-strength 3 T magnets have now become available commercially. They offer the potential for improved image quality and resolution due to their increased SNR. The SNR correlates approximately with field strength so it is roughly twice as great at 3 T than at 1.5 T. However, at these higher field strengths, susceptibility and flow artefacts can cause image degradation. There are also safety concerns when operating at 3 T, related to heating effects and induced currents in peripheral nerves. While 3 T scanners have shown much promise in the field of neuroradiology and bone imaging, more research needs to be done to define their role in renal imaging.
Several sequences are used in a routine renal MRI; these sequences should include T1- and T2-weighted, and fast gradient echo before and after dynamic contrast-enhanced images. The sequences before and after contrast enhancement can be done with either standard 2D axial sequences or a 3D volume acquisition (Fig. 7).
MRI offers the advantages of no radiation and the use of less nephrotoxic contrast material. MR images have better contrast resolution but decreased spatial resolution than CT. All MRI examinations being done to evaluate a renal mass will require administration of contrast medium. Recent reports raised the question of an association between gadolinium administration and nephrogenic systemic fibrosis (NSF), previously known as nephrogenic fibrosing dermopathy [19,20]. NSF has generally occurred in patients with renal insufficiency. It can be a progressively debilitating and eventually fatal disease; there is no acknowledged treatment. The FDA recommended that caution be used when considering gadolinium administration to patients who have severe renal compromise (GFR <15 mL/min). However, it must be acknowledged that ≈ 30 million patients have received gadodiamide (the contrast agent that has been identified in most of the studies so far) and only 200 cases of NSF have been described . Until more definitive data are available, caution is advised when giving paramagnetic contrast media for MRI in renally compromised patients. Double-dose studies should be avoided in this group, and if paramagnetic contrast medium is given to a renally compromised patient, dialysis should be done soon after as per the nephrology recommendation.
RCC: MRI Appearance
Most RCCs have a signal intensity that is lower than the renal cortex on T1-weighted images. RCC is variable in appearance on T2-weighted images. As in CT, the presence of contrast enhancement is the most critical factor for the diagnosis of a solid tumour (Fig. 7). However, as signal intensity units for MRI are arbitrary and can vary substantially with different sequences, scanners and patients, a quantitative assessment can be difficult. In our practice, a 15% increase in signal intensity on post-contrast enhanced images is generally considered significant. Studies show that this yields 100% sensitivity and >94% specificity in helping to distinguish between cysts and solid renal masses . It is often very helpful to use subtraction of the sequence before contrast from that after contrast enhancement to accurately assess any subtle enhancement.
As with CT, image features have emerged that help to predict the histological type of tumour. Clear cell RCC is hyper-intense and heterogeneous on T2-weighted sequences. Papillary tumours have a low signal intensity with a homogeneous signal on T2-weighted images. Enhancement was lower and delayed in the papillary type than in the clear cell type .
MRI appears to offer a similar accuracy to CT for staging RCC ; as with CT, the correct identification of tumoral spread into perinephric tissues remains problematic. MRI remains the method of choice for determining caval extension of tumour thrombus, due to the limitations described with CT above (Fig. 8a,b).
US is frequently used as the initial imaging in patients with suspected renal disease. US contrast agents offer the potential to increase the sensitivity in patients who cannot undergo CT or MRI. Harmonic imaging has helped with the evaluation of renal cystic lesions.
CT remains the reference standard for staging and lesion characterization. The rapid technological advances mean that ultra-thin slices and 3D imaging have now become more widely available.
MRI has also greatly advanced in speed and image quality, but at present, imaging time and scanner availability mean that MRI is generally used as a problem-solving tool. It is particularly helpful for smaller lesions and complex cystic lesions, where subtraction can be used to accurately identify the presence of enhancement. MRI is used as the primary diagnostic tool for patients with radiation concerns and those with renal failure. Caution must now also be taken in these patients, given the recent studies raising the association of gadolinium-based compounds and NSF.
There have been dramatic improvements in renal imaging over the last decade, offering better resolution, shorter imaging times and better patient acceptance. The progress in minimally invasive techniques has driven the need to provide better preoperative information to the surgeon. The future of renal imaging is an exciting field; perhaps with fluorodeoxyglucose-based positron emission tomography we will be able to predict the biological behaviour of a tumour, and molecular imaging agents will become available to identify and hopefully treat specific tumour types [25,26].
CONFLICT OF INTEREST