One of the earliest descriptions of using CT to assist in evaluation of the urinary tract involved combining CT and intravenous urography (IVU) (1). Using this approach, patients were evaluated initially with plain radiography and linear tomography. Contrast material was then injected intravenously, during which time a standard intravenous urogram was obtained. Subsequently, the patient was transferred to a CT suite, at which time delayed enhanced CT images were obtained through the kidneys. With this approach, the purpose of the CT images was to identify small renal lesions that would have been missed if only an intravenous urogram was performed.

Although this technique likely improved study sensitivity in detecting renal masses over that of IVU, there were two limitations. First, it is not always easy to transfer a patient quickly from a conventional x-ray room onto a CT scanner. Essentially, both rooms must be vacant at some point for the transfer to occur, which is not always easily done in a busy radiology department. Second, renal masses are not evaluated fully; the delayed enhanced CT images are not obtained during the nephrographic phase of renal enhancement—the phase that is most ideal for detecting renal masses. More important, an unenhanced CT is not obtained. Therefore, renal masses cannot be evaluated for enhancement.

As an alternative, a CT suite may be modified by adding conventional x-ray equipment to a CT scanner (2). This modification allows patients to have conventional radiographs obtained while they remain on the CT table. Preliminary results with this approach were encouraging because now CT evaluation of the renal parenchyma could be obtained at any time (3). This approach is also predicated on the belief that the excretory phase is imaged better with radiography than with CT. Axial CT images are not obtained routinely after excretion of contrast material into the renal collecting systems and ureters has occurred, even though studies have suggested that excretory phase conventional radiography is probably not as accurate as are excretory phase axial images. Also, adopting this hybrid technique requires substantial additional expenditure.

Nevertheless, there have been a number of investigations in which the CT digital (or scan projection) radiograph was used to assess the urinary tract instead of conventional radiographs. Enhanced digital CT radiographs were then developed that appeared similar to conventional radiographs and were as sensitive as radiographs in detecting urinary tract abnormalities, including urolithiasis (4). Thus, the CT scanner could be used to obtain intravenous urography-like images without the need for CT room alterations. As with all projection-type techniques, however, with digital radiography the urinary tract can be obscured by overlying structures, such as gas-filled loops of bowel. As a result, bowel preparation may be helpful when this method of excretory phase urinary tract imaging is employed.

Concurrently, several investigators began to utilize thin section axial images rather than conventional radiographs to evaluate the urinary tract in its entirety (5, 6 and 7). In comparison to conventional or digital radiography, axial images would not be affected by overlying structures, making bowel preparation unnecessary. However, this approach required a paradigm shift. It had been thought that urothelial abnormalities could not be detected on axial CT images due to their decreased resolution compared with IVU (3). However, with the introduction of multidetector row CT, it became possible to obtain thin section images of the entire urinary tract during a single breath-hold, permitting one to create high spatial resolution images of the urothelium in multiple planes. Preliminary results suggested that, in fact, CTU performed with this approach, generally termed multidetector or multislice CTU, was effective in detecting both benign and malignant urinary tract pathology (6). Small abnormalities, such as renal tubular ectasia, papillary necrosis, and even tiny (2- or 3-mm maximal diameter) urothelial cancers, could be identified. Thus, it was suggested that multidetector CTU could replace IVU and the combined IVU–CT hybrid studies (6, 8).

It has become apparent that CTU obtained with thin section excretory phase axial images (generally acquired using a thickness of 2.5 mm or less) is more sensitive in detecting urinary tract pathology than digital scout images or conventional excretory urographic images. The advantages of the axial CT approach are best seen in its ability to detect urothelial cancers and ureteritis. IVU and digital
scout CT images only image the lumina of the renal collecting systems, ureters, and bladder. They cannot be used to visualize their walls. Although it has been assumed that abnormalities of the urinary tract that affect the urothelium nearly always affect the lumen, this is now known not to be the case. Occasional urothelial malignancies and some cases of ureteritis and cystitis can produce pronounced urothelial wall thickening and yet have little or no effect on the diameter or the morphology of the renal collecting system, ureteral, or bladder lumen. Circumferential wall thickening can be identified on axial CT images when it is undetectable on digital CT images, IVU, and even retrograde pyelography (6, 8). For this reason, some early advocates of the combined IVU–CT approach have recently converted to a CTU technique that relies only on axial images to assess the urothelium.

Two different approaches in performing axial image CTU have evolved. The most comprehensive imaging protocol utilizes at least three different image acquisitions: unenhanced scans to assess the kidneys and ureters for calculi and to assist in characterization of any subsequently detected renal masses; early enhanced nephrographic phase scans for optimal detection and characterization of renal masses (commenced after injection of a single bolus of contrast material has been administered); and delayed excretory phase images for evaluation of the renal collecting systems, ureters, and bladder (7, 9). Although some have added more series, such as arterial, corticomedullary phase (10, 11), and additional excretory phase images (6), these are not obtained routinely.

A variety of different image thicknesses have been utilized. In general, unenhanced images are obtained from the kidneys to the pubic symphysis using a reconstructed image thickness of 2.5 to 5 mm and no overlap. The nephrographic phase images are obtained utilizing the same reconstruction image thickness and interval but are performed either through the kidneys alone or through the entire abdomen. The latter approach allows for assessment of abnormalities outside the urinary tract, although enhancement of the liver is not optimal, since images are obtained after the portal venous phase. Despite this limitation, hepatic (and other nonrenal abdominal organs) enhancement is preferred during the nephrographic phase compared to the excretory phase due to improved sensitivity in detecting visceral organ abnormalities. It is important that multidetector CTU protocols include very thin section excretory phase images. Although an image thickness of 5 mm has occasionally been utilized (10), it is preferable for it to be no greater than 2.5 to 3.0 mm. Many researchers are reconstructing the original data set utilizing an image thickness of 0.625 to 1.25 mm (7, 9). Reconstructions are then performed with no overlap if 0.625-mm-thick images are acquired. A 50% overlap can be used if thicker sections are obtained (5, 9).

Contrast material injection involves administration of 100 to 150 ml of a 300 mg I/ml concentration of nonionic contrast material, which is administered by a mechanical injector at rates varying between 2 and 4 ml/sec (6, 7). Since arterial phase images are not obtained routinely, more rapid injection rates are not required.

The recommended timing for each of the two contrast-enhanced phases has varied widely; however, nephrographic phase images are fairly consistently obtained 100 to 120 seconds after the administration of contrast material begins, provided that the contrast material is injected at the rates described in the previous text (6, 7). There is less of a consensus about the timing of the excretory phase images. Excretory phase scans have been acquired between 3 (12) and 15 minutes (13) after the initiation of the contrast material injection. One reported series demonstrated that urinary tract abnormalities were more frequently detected when excretory phase images were obtained at 450 seconds compared with 300 seconds (14). Still another study found that urinary tract distention improved for up to 15 minutes after contrast injection began but then decreased (15). Based on these data, the optimal timing for excretory phase image acquisition appears to be between 7.5 and 15 minutes.

The three- or four-phase single bolus technique for CTU has been the most widely used (16). There is one important criticism, however: patients are exposed to more radiation in comparison with standard IVU. This criticism could be leveled at many CT protocols since CT scans inherently result in more radiation to the patient than does plain radiography.

In 2001, two groups (5, 17) reported on a different approach to performing CTU, in which administered contrast material was divided into two boluses. A sufficient delay was inserted between the two doses so that the patient would be imaged when the first dose was being excreted into the renal collecting systems, ureters, and bladder, while the second dose was still opacifying the renal parenchyma (5, 17). With this approach, a single contrast-enhanced CT acquisition included both nephrographic and excretory phase images of the urinary tract, effectively eliminating the need for one of the contrast-enhanced CT acquisitions. The result was decreased patient radiation exposure compared with the three- or four-phase CTU protocols.

The split bolus protocol used by Chow and Sommer (5) is summarized as follows: first, 5-mm unenhanced images were obtained through the abdomen and pelvis, and then 40 ml of contrast material (300 mg I/ml) were injected at a rate of 2 ml/sec. After a delay of 2 minutes (during which time no scanning was performed), an additional 80 ml of contrast material (of the same concentration) were injected at the same rate. After a delay of an additional 1.5 minutes, CT scans were obtained of the abdomen and pelvis and reconstructed at 2.5-mm sections and 1.25-mm intervals. However, a net delay of 3.5 minutes between the first contrast material injection and the beginning of excretory
phase image acquisition likely did not allow for maximal opacification and distention of the entire urinary tract. Therefore, a compression device was required. Images were obtained of the upper tract before and the lower tract after the compression device was released.

An alternative split bolus protocol was reported by another group (17). First, unenhanced CT scans were obtained of the abdomen and pelvis and reconstructed using 5-mm-thick sections and 5-mm intervals. Second, 30 ml of contrast material (300 mg I/ml) was administered by infusion rather than a power injector. After 15 minutes, the patient received an additional 100 ml of contrast material administered by a power injector at a rate of 2 ml/sec. After another 100 seconds, thin section images were acquired through the abdomen and pelvis (with images reconstructed contiguously using 5-mm thickness for axial image review but using a thickness of 2.5 mm and reconstruction interval of 1 mm for coronal reformatting). The total delay of more than 16.5 minutes was probably longer than desired for optimal urinary tract distention. Also, it is generally recommended that axial images be reviewed using a section thickness of <5 mm.

Kekelidze et al. (18) reported on their use of a triple bolus protocol whereby contrast material is administered at three different times and the patient is then imaged when the first bolus (30 ml of iodixanol 320) has been excreted, the second (50 ml of iodixanol 320), administered 7 minutes later, is in the renal parenchyma, and the third (65 ml of iodixanol 320), administered 2 minutes after that, opacifies the renal arteries. Using this protocol, the authors found that satisfactory images could be obtained.

The ability of the split bolus approach to eliminate one of the contrast-enhanced CT acquisitions has led to its adoption at many institutions. However, there are several theoretical limitations of the split bolus approach. First, the reduced volume of the first bolus may produce less distention and opacification of both the intrarenal collecting systems and the lower urinary tract, which could limit the ability to detect abnormalities. We have occasionally observed this to be a problem (19). Second, it has been reported that dense excreted contrast material in the renal collecting systems can create beam hardening-induced streak artifact that could limit evaluation of the renal parenchyma (20). Although it is unlikely that such artifact would prevent detection of a renal mass, it might preclude obtaining accurate attenuation measurements of renal masses that are adjacent to the collecting system, limiting characterization.