点击显示 收起
1 From the Department of Radiology (E.M.C., R.H.C., M.K., J.F.P., I.R.F., J.H.E.) and Department of Surgery, Urology Surgery Section (G.J.F., J.E.M.), University of Michigan Medical Center, 1500 E Medical Center Dr, Ann Arbor, MI 48109-9723. Received March 26, 2001; revision requested April 23; revision received June 22; accepted July 25.
| ABSTRACT |
|---|
| Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
MATERIALS AND METHODS: Sixty-five patients referred from the urology service, in whom urinary tract abnormalities were strongly suspected, underwent multi-detector row CT urography. The technique included unenhanced, nephrographic, compression, and excretory-phase images through the abdomen and pelvis. Transverse images and three-dimensional reformations were reviewed by one of two radiologists. Findings were retrospectively compared with results of urinalysis, cystoscopy and/or ureteroscopy, and/or surgery.
RESULTS: Multi-detector row CT urography depicted many clinically diagnosed urinary tract abnormalities, including 15 of 16 uroepithelial malignancies, five congenital anomalies, five urinary tract calculi, and 18 calyceal and/or papillary, 30 renal pelvic and/or ureteral, and 25 bladder abnormalities. All abnormalities were detected on transverse images. These abnormalities included diffuse urothelial wall thickening in four patients (three of whom had transitional cell carcinoma), a renal abscess, a colovesical fistula, and incidentally detected extrarenal disease (a liver mass, hepatic metastases, lymph node metastases, an aortic dissection, and a pheochromocytoma; each of these findings was seen in one patient).
CONCLUSION: Multi-detector row CT urography is a useful method for detecting urinary tract abnormalities.
Index terms: Bladder, CT, 83.1211 • Computed tomography (CT), technology, 80.1211 • Kidney, abnormalities, 81.14 • Kidney, CT, 81.12115 • Ureter, CT, 82.12115 • Urography, 80.1211
| INTRODUCTION |
|---|
| Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Experience with CT in the evaluation of the genitourinary system (CT urography) is limited. McNicholas et al (17) compared conventional single-detector helical CT urograms and excretory urograms. Although their study was limited because separate groups of patients were imaged with the two techniques, McNicholas et al found that excretory-phase collecting system opacification obtained during CT urography was comparable to collecting system opacification obtained during excretory urography.
Multi-detector row CT offers greater speed of acquisition and higher resolution images than single-detector helical CT. The more thinly collimated transverse images obtained in a breath hold and the subsequent better quality of reformatted coronal images should further increase the ability of CT to depict the renal collecting systems accurately. Our purpose was to evaluate multi-detector row CT urography for detection of urinary tract abnormalities.
| MATERIALS AND METHODS |
|---|
| Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
All multi-detector row CT urographic examinations were performed with a multi-detector row CT scanner (Lightspeed QX/i version 1.3; GE Medical Systems, Milwaukee, Wis). Unenhanced scans were obtained through the abdomen and pelvis with the following technique: a 4 x 3.75-mm configuration, a table speed of 11.25 mm per rotation, HQ mode, and 150–240 mA. Images were reconstructed at a thickness of 5 mm. CT technologists then applied abdominal compression to the patient with an inflatable compression band placed on the anterior abdominal wall at the level of the iliac crests. The patients were then intravenously injected with 150 mL of iohexol 300 (Omnipaque; Nycomed, New York, NY) with a power injector at a rate of 3 mL/sec. Nephrographic-phase images were then obtained from the diaphragm through the kidneys beginning 100 seconds after initiation of the injection of contrast material with the following technique: a 4 x 2.5-mm configuration, a table speed of 7.5 mm per rotation, HQ mode, and 100–280 mA. Images were reconstructed at a thickness of 5 mm. Compression excretory-phase images were then obtained through the abdomen and pelvis beginning 200 seconds after initiation of the injection of contrast material with the following technique: a 4 x 1.25-mm configuration, a table speed of 7.5 mm per rotation, HS mode, and 150–280 mA. Images were reconstructed at 1.25-mm intervals. At 300 seconds, abdominal compression was released and compression-release excretory-phase images were obtained with a technique identical to that employed at 200 seconds. Digital scout views were also obtained immediately prior to the unenhanced phase, immediately after the compression excretory phase, and immediately after the compression-release excretory phase. The examinations were well tolerated by all patients. The examination time ranged from 15 to 20 minutes.
Three-dimensional (3D) reconstructions of the compression excretory-phase scans and the compression-release excretory-phase scans were created at an independent workstation (Advantage Windows 3.1; GE Medical Systems). One of two abdominal radiologists (E.M.C., R.H.C.) performed the postprocessing prior to interpreting the studies. The 3D reconstructions in coronal and bilateral 25° coronal oblique projections were created with maximal intensity projection (MIP), average intensity projection (AIP), and volume-rendering algorithms. The MIP and AIP images were created by electronically selecting an arbitrary section thickness (usually exceeding 50 mm) that included the kidneys and the ureters. Curves for the volume-rendered images were chosen to preferentially show enhanced soft tissue and contrast-opacified structures. Postprocessing time ranged from 20 to 30 minutes for each study.
All images were available on an independent workstation. All unenhanced and nephrographic-phase transverse images, digital scout views, and 3D reformations were also filmed. Only every fourth image of the compression excretory-phase images was filmed in order to conserve film. All filmed CT images were photographed at standardized window width (500 HU) and level (50 HU) settings. Images were photographed in an identical fashion and were printed on the same printer. The compression-release excretory-phase transverse images were not filmed.
One of two abdominal radiologists (E.M.C., R.H.C.), blinded to the patients’ clinical diagnoses, prospectively and independently interpreted the multi-detector row CT urograms. Scans were interpreted after both hard and soft copies of the images were reviewed; however, because the compression excretory-phase images were incompletely filmed and the compression-release excretory-phase images were not filmed, all images obtained in these series were carefully reviewed at the workstation. The CT scout views were subjectively thought to be inferior to the 3D reconstructions, but these were not systematically compared or reviewed during this investigation. All 3D reconstructions (with MIP, AIP, and volume-rendering algorithms), in addition to the transverse data, were evaluated for the presence of urinary tract abnormalities.
One author (E.M.C.) retrospectively compared the results of multi-detector row CT urography with the results of other clinical and imaging examinations. Generally, patients with native bladders or a prior history of urothelial malignancy were evaluated with cystoscopy and cytologic examination. Patients with neobladders were evaluated primarily with cytologic examination of their urine. Patient clinical assessment was determined through a review of patient medical records performed in accordance with the standards of our institutional review board.
| RESULTS |
|---|
| Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
|
|
Congenital Anomalies
The following five urinary tract anomalies were identified in five patients: two partial ureteral duplications, one complete ureteral duplication in which an ectopic ureterocele was also seen (Fig 2), one horseshoe kidney containing a calculus in a renal pelvis, and one pelvic kidney. Of the patients with anomalies, only one underwent another study for confirmation. In the patient with a complete ureteral duplication, results of cystoscopy confirmed ectopic ureteral insertion associated with a ureterocele. The renal pelvic calculus in the patient with the horseshoe kidney was subsequently extracted ureteroscopically. All anomalies were easily identified on both transverse images and 3D reformatted images.
|
Calyceal and/or Papillary Abnormalities
Eighteen calyceal and/or papillary abnormalities were identified at multi-detector row CT urography in 17 patients: 15 cases of diffuse caliectasis (10 of which were unilateral and five of which were bilateral), one case of renal tubular ectasia, one case of papillary necrosis, and one case of focal blunting of upper pole calyces that was thought to be due to reflux because there was associated focal renal parenchymal loss. One patient had two calyceal abnormalities—diffuse bilateral caliectasis and papillary necrosis.
In seven of the 15 patients with diffuse caliectasis, correlative imaging or clinical information was available to help confirm the finding at multi-detector row CT urography. Unilateral caliectasis was confirmed to be present during ureteroscopy or surgery in three patients, excretory urography in one patient, and antegrade pyelography in another. Bilateral caliectasis was identified at surgery in two patients who underwent surgery due to the presence of a bladder mass. Eight patients with diffuse caliectasis underwent no specific clinical follow-up of this radiologic finding. All eight patients had previously undergone radical cystectomy, and the dilatation was considered to be either due to prior obstruction or the result of urinary diversion. Urine cytologic results obtained for six of these eight patients were normal in each. Two of the patients with diffuse caliectasis had previously undergone a multi-detector row CT urographic examination in which a bladder malignancy had been identified.
Among the remaining three of the 18 calyceal and/or papillary abnormalities identified, correlative imaging was performed in only one instance. Results of excretory urography confirmed the presence of renal tubular ectasia (Fig 3). No imaging or surgical follow-up was conducted to confirm the presence of papillary necrosis and focal upper pole caliectasis.
|
|
|
Renal Pelvic or Ureteral Abnormalities Other than Urolithiasis
Thirty renal pelvic or ureteral abnormalities were identified in 23 patients. Pelviectasis and ureterectasis were present in 14 patients (nine findings were unilateral and five bilateral), and isolated unilateral pelviectasis was present in one patient. Findings also included parapelvic cysts in four patients, renal pelvic masses in two patients, concentric urothelial wall thickening in four patients, a ureteral mass in one patient, a ureteral filling defect in one patient, and ureteral narrowing in three patients. The 15 patients with dilatation of the renal pelvis also had caliectasis and are also described in the section on calyceal and/or papillary abnormalities.
There was more than one renal pelvic and/or ureteral finding in five patients. Three patients with unilateral pelviectasis also had ipsilateral ureteral narrowing; one of these patients also had parapelvic cysts. One patient with a renal pelvic mass also had a ureteral mass (Fig 4). Another patient with a renal pelvic mass had diffuse urothelial wall thickening and a ureteral filling defect.
|
|
|
|
|
|
|
Of the three instances in which CT demonstrated ureteral narrowing, two proved to be benign strictures at endoscopy (one was due to ureteritis cystica and one had no known etiology). Renal collecting system obstruction (as determined by the presence of proximal collecting system dilatation and delay in the appearance of the ipsilateral nephrogram) was present in these two patients. There was one false-positive case in which endoscopy revealed kinking of the ureter rather than a true stricture. Renal collecting system obstruction was not present in this patient and, in fact, had not been suggested at multi-detector row CT urography (as there was no delay in the appearance of the ipsilateral nephrogram). No renal pelvic or ureteral abnormalities were subsequently identified in any of the patients in whom multi-detector row CT urography failed to depict a renal pelvic or ureteral abnormality.
Renal pelvic and ureteral abnormalities were best visualized on the compression-release excretory-phase images. Transverse scans allowed detection of all abnormalities. Pelviectasis, ureterectasis, and ureteral narrowing could be seen on 3D reformatted images if the collecting system lumen was opacified at the time of image acquisition. Although easily diagnosed on transverse images, parapelvic cysts could be only indirectly suspected on the 3D reformations due to their displacement of the renal pelvis and the intrarenal collecting system. The renal pelvic masses, ureteral mass, and ureteral filling defect were identified on both transverse imaging and 3D images; however, concentric ureteral wall thickening was not seen on the 3D images.
Bladder Abnormalities Other than Urolithiasis
Twenty-five bladder abnormalities were detected at multi-detector row CT urography in 23 patients, 22 of whom also underwent cystoscopy with or without biopsy or surgery to enable a diagnosis. The remaining patient was followed up with cytologic examinations of urine. Focal or asymmetric bladder wall thickening was demonstrated at multi-detector row CT urography in 10 patients, diffuse bladder wall thickening in 10, bladder hematomas in three, a bladder diverticulum in one, and a colovesical fistula secondary to sigmoid diverticulitis in one. Two patients had two bladder abnormalities. One patient had focal bladder wall thickening adjacent to a bladder diverticulum. The other had asymmetric bladder wall thickening and a bladder hematoma.
Bladder malignancy was subsequently diagnosed in eight of the 10 patients with focal or asymmetric bladder wall thickening (Fig 6). Radiation cystitis and scarring were seen at cystoscopy in the two other patients. Among the 10 patients in whom diffuse bladder wall thickening was identified at multi-detector row CT urography, bladder malignancy was subsequently diagnosed in only one. These latter 10 patients underwent cystoscopy and were given the diagnosis of obstructive uropathy due to benign prostatic hyperplasia (n = 6), normal bladder walls (n = 2; these were counted as false-positive results), erosive cystitis (n = 1), or diffuse, infiltrative transitional cell carcinoma (n = 1). When the patients with diffuse bladder wall thickening as a result of benign disease were compared with the patient with bladder wall thickening as a result of malignant disease, no differences in appearance at multi-detector row CT urography that might facilitate differentiation could be detected. The severity and morphology of the bladder wall thickening were similar.
|
|
In two patients, multi-detector row CT urographic results were initially false-negative for detecting pathologic conditions in the bladder; in both cases, however, the bladder abnormalities could be identified in retrospect. Cystoscopy revealed a focal transitional cell carcinoma at the base of the bladder in one patient. Review of the multi-detector row CT urograms following cystoscopy revealed a corresponding area of focal bladder wall thickening adjacent to the prostate. In another patient in whom the multi-detector row CT urograms were prospectively read as negative, cystoscopy subsequently revealed radiation cystitis. Review of the multi-detector row CT urographic results after cystoscopy revealed mildly prominent wall thickening on the transverse excretory-phase images.
Bladder abnormalities were best visualized on the compression-release excretory-phase images. Transverse scans allowed detection of all abnormalities. Focal bladder wall thickening could be seen on 3D reformatted images; however, diffuse bladder wall thickening was appreciated only on transverse images. The bladder hematomas and the bladder diverticulum were well seen on both the 3D reformatted images and the transverse images. The colovesical fistula was detected only on transverse images.
Comparison of Multi–Detector Row CT Urography and Confirmatory Studies
Because our study population did not undergo a standardized clinical follow-up, not all patients underwent confirmatory studies. Therefore, our determination of the numbers of true- and false-positive and true- and false-negative diagnoses that occur when multi-detector row CT urography is performed should be viewed with caution. In addition, some of our patients with normal multi-detector row CT urographic studies underwent no further evaluation. Given these limitations, our findings in those patients in whom some confirmation was available are as follows. Five false-positive findings were identified: a urinary calculus (n = 1), ureteral wall thickening (n = 1), a ureteral stricture (n = 1), and diffuse bladder wall thickening (n = 2). Two false-negative findings were seen: radiation cystitis (n = 1) and bladder malignancy (n = 1). These findings are all described in detail in prior sections. In 21 patients, no urinary tract abnormality was found both at multi-detector row CT urography and during clinical evaluation, which included cystoscopy (n = 10), cytologic examination of urine without cystoscopy (n = 9), or clinical follow-up (n = 2); these results were considered true-negative.
Other Pathologic Results
Five patients had important findings apart from those identified in the urinary tract. These included metastatic abdominal lymph node involvement from transitional cell carcinoma of the bladder, hepatic metastases, an unsuspected liver mass in a cirrhotic liver, an unsuspected abdominal aortic dissection, and an adrenal pheochromocytoma (each of these findings was seen in one patient).
| DISCUSSION |
|---|
| Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
To our knowledge, the sensitivity of multi-detector row CT urography in detection of other subtle renal collecting system abnormalities that may produce symptoms such as calyceal diverticula, renal tubular ectasia, and papillary necrosis is unknown. In our study, two patients with these abnormalities had diagnostic findings that were easily identified on the 3D reformatted images and, in retrospect, on the transverse images.
We believe that our CT urographic technique has some advantages over other potential CT urographic protocols. As an alternative to our approach, it is possible to perform hybrid studies, in which initial transverse CT images are obtained before the administration of contrast material and during the nephrographic phase after contrast material administration. These CT images are then followed with radiographs obtained after excretion of contrast material into the renal collecting system has occurred. Such radiographs are obtained either with a modified radiographic unit installed in the room containing the CT scanner or through removing the patient from the CT scanner and moving him or her to a radiographic room. These approaches require either that an expensive modification to a CT scanner be made or that there be coordination between CT and general radiology rooms. In our protocol, all the information is acquired with the CT scanner. No room modifications or patient transfers are needed.
There is another, more important potential advantage of our multi-detector row CT urographic technique over hybrid CT urography. In our series, many of the detected abnormalities in the renal collecting systems and the bladder were most conspicuous on transverse excretory-phase images obtained with thin collimation and intraluminal contrast material. In fact, some urothelial wall and bladder abnormalities were seen only on the transverse excretory-phase scans. Concentric urothelial wall thickening (which represented malignancy in three of four patients in whom it was identified) could be identified only on transverse images. If only 3D reformatted images or radiographs had been obtained during the excretory phase in these patients, we believe that the ureteral wall thickening would not have been detected, because, similar to excretory urograms, the 3D images are limited to depiction of the opacified lumen. The inability to visualize the urothelium may, in part, explain the relatively low urothelial malignancy detection rates previously reported with excretory urography.
Although our series is small, we successfully detected six (100%) of the six foci of upper tract malignancy. We detected nine (90%) of the 10 bladder malignancies. However, most patients with a high risk of urinary tract disease will be evaluated with cystoscopy, and thus the ability of multi-detector row CT urography to identify bladder masses will not affect clinical practice as much as will its ability to detect upper tract disease.
One disadvantage of our protocol is the time required to create the 3D reformatted images. In our experience, this ranged between 20 and 30 minutes. Although the reformatted images in this study were all created by physicians, we have since educated our technologists so that they can create these images. The 3D reconstructions are useful in conveying information to our clinicians who are more familiar with excretory urography, because a large amount of transverse data can be summarized in one coronal image and because our urology colleagues are most comfortable visualizing the renal collecting systems in the same planes that are used during excretory urography. In addition, we found 3D images to be particularly helpful in the diagnosis of papillary abnormalities such as renal tubular ectasia and papillary necrosis. These abnormalities were seen on the transverse scans only after they were first detected on the reformatted images. This suggests that until all pathologic conditions of the collecting system are more easily recognized in the transverse plane, 3D reformatted images will also be useful to radiologists as a bridge between excretory urography data and transverse CT data. At least for the time being, both methods of display will likely be beneficial. However, we believe that as additional experience is gained, the 3D reformats may become unnecessary. In no instance in our series did the 3D reformatted images reveal an abnormality that was not identifiable on the transverse images.
A theoretical concern with multi-detector row CT urography is the possibility of incremental radiation dose in comparison with excretory urography. The patient radiation dose from multi-detector row CT urography can be estimated from the dose-length product, a measurement of radiation exposure that takes into account the volume of irradiation (18). This number is generated at the CT computer console after each examination. An effective dose (measured in millisieverts) can be estimated from the dose-length product. Calculated with the formula and normalized coefficients recommended by Jessen et al (18), the effective dose for multi-detector row CT urography for an average man is approximately 25–35 mSv. The effective dose for an abdominopelvic CT examination is approximately 10–15 mSv. The standard excretory urography protocol at our institution consists of four anteroposterior views requiring a 14 x 17-inch screen-film combination, two to four anteroposterior views requiring a 14 x 11-inch screen-film combination, and four to six tomograms at 70 kVp and 64 mAs. The typical effective dose for a man of average size is 5–10 mSv. A large number of our patients who had a history of previously treated urothelial malignancy would probably have undergone an abdominopelvic CT examination to detect metastatic disease and excretory urography to detect recurrent disease. The additional radiation received during the compression and excretory phases of multi-detector row CT urography is similar to or slightly greater than that of excretory urography. Therefore, we estimate that performing multi-detector row CT urography exposes a patient to an amount of radiation similar to what he or she would experience during a combination of excretory urography and standard abdominal and pelvic CT. As experience with multi-detector row CT urography increases, it may even become possible to reduce the radiation dose by further tailoring the study (by eliminating one of two excretory phases, for example).
Our study had several limitations. First, although a variety of lesions were seen, our study population consisted of small numbers of patients with each of these abnormalities. With respect to detecting neoplastic disease in the upper collecting tracts, a large number of patients would be needed to assess the effectiveness of multi-detector row CT urography because the frequency of ureteral tumors is small. Second, patients with several findings detected at multi-detector row CT urography (particularly findings in the renal calyces) did not undergo definitive follow-up. Third, there were no direct comparisons between multi-detector row CT urography and excretory urography. The relative effectiveness of each examination in the detection of each urinary tract abnormality we evaluated could not be determined. Last, the interpretations of the multi-detector row CT urographic data were often used to guide subsequent patient care. This introduces some bias into our study, because patients with positive multi-detector row CT urographic examinations were more likely to undergo additional evaluation than those with negative studies. In particular, such bias would be likely to spuriously lower the number of false-negative results that would be encountered. Such an algorithm, leading to close follow-up in some patients, also favors an increased rate of detection of abnormalities.
In summary, multi-detector row CT urographic examinations in which both transverse and 3D images are evaluated are useful for revealing urinary tract anatomy and pathologic conditions in the selected group of patients in whom urinary tract abnormalities are strongly suspected. Multi-detector row CT urography has the potential to reveal even small or subtle collecting system abnormalities. Combining the known utility of CT as an aid in detecting and characterizing renal masses and evaluating patients suspected of having urolithiasis with its now-demonstrated ability to depict urothelial and bladder lesions, multi-detector row CT urography has the potential to enable a comprehensive evaluation of the genitourinary system in a single examination. Given these attributes, we believe multi-detector row CT urography has potential as an alternative to excretory urography for patient evaluation.
| REFERENCES |
|---|
| Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|