Detection of Renal Stones on Portal Venous Phase CT: Comparison of Thin Axial and Coronal Maximum- Intensity-Projection Images

Transcription

1 Genitourinary Imaging Original Research Corwin et al. Detection of Renal Stones on Portal Venous Phase CT Genitourinary Imaging Original Research Michael T. Corwin 1 Justin S. Lee 1 Ghaneh Fananapazir 1 Machelle Wilson 2 Ramit Lamba 1 Corwin MT, Lee JS, Fananapazir G, Wilson M, Lamba R Keywords: contrast material, CT, nephrolithiasis, portal venous phase, renal stone DOI: /AJR Received January 6, 2016; accepted after revision May 24, Supported by the National Center for Advancing Translational Sciences, National Institutes of Health, through grant UL1 TR Department of Radiology, University of California, Davis Medical Center, 4860 Y St, Ste 3100, Sacramento, CA Address correspondence to M. T. Corwin (mtcorwin@ucdavis.edu). 2 Department of Public Health Sciences, University of California, Davis, Davis, CA. AJR 2016; 207: X/16/ American Roentgen Ray Society Detection of Renal Stones on Portal Venous Phase CT: Comparison of Thin Axial and Coronal Maximum- Intensity-Projection Images OBJECTIVE. The objective of our study was to determine the sensitivity of thin axial and coronal maximum-intensity-projection (MIP) images for the detection of renal stones on contrast-enhanced CT performed in the portal venous phase. MATERIALS AND METHODS. This retrospective study included 72 patients, 59 of whom had at least one renal stone, who underwent unenhanced CT immediately followed by contrast-enhanced CT in the portal venous phase. Two abdominal imaging fellowship-trained radiologists independently recorded the number of stones on both thin ( mm) axial and 5-mm coronal MIP images in the portal venous phase. The reference standard was determined by consensus review of the thin axial unenhanced images. Reviewer sensitivity was calculated and categorized by stone diameter. RESULTS. One hundred forty-eight stones were present; the mean number of stones per patient was 2.5 (SD, 2.7). The mean stone size was 2.5 mm (SD, 2.7). The sensitivity of thin axial images was 89.9%, 99.0%, and 100.0% for reviewer 1 and 83.1%, 98.0%, and 100.0% for reviewer 2 for all stones, stones 2 mm, and stones 3 mm, respectively. The sensitivity of coronal MIP images was 86.5%, 96.2%, and 100.0% for reviewer 1 and 79.0%, 91.4%, and 96.6% for reviewer 2 for all stones, stones 2 mm, and stones 3 mm, respectively. CONCLUSION. Thin axial images are highly sensitive for the detection of renal stones 2 mm on portal venous phase CT. Coronal MIP images do not improve renal stone detection over thin axial images. U nenhanced CT is the imaging modality of choice for patients presenting with acute flank pain and suspicion of urolithiasis [1]. However, not all patients in this setting have stone disease; some patients have an alternative diagnosis for acute pain such as diverticulitis, appendicitis, or even pancreatitis [2 5]. In contradistinction to urolithiasis, these entities are better evaluated with contrast-enhanced CT (CECT) [6]. The clinical distinction between these disorders and stone disease may often be difficult, and the decision to perform CT with or without contrast material may be unclear in some patients [7, 8]. Furthermore, unenhanced CT is limited in the detection of pyelonephritis and renal abscess, which could either mimic or complicate symptomatic urolithiasis [9]. Although the presence of IV contrast material has traditionally been thought to reduce the sensitivity of CT for the detection of renal stones, few studies have directly examined this phenomenon. In a recent study, Dym et al. [10] reported an overall sensitivity of 81% for renal stone detection using portal venous phase contrast-enhanced CT. In that study, 2.5-mm axial images were reviewed. However, it is clear that renal stone detection is improved on unenhanced CT with the use of a thinner slice thickness [11 13]. Therefore, it is possible that the sensitivity of portal venous phase CT could be increased using even thinner slices. Investigators have also recently shown that coronal maximum-intensity-projection (MIP) images are as sensitive as thin (1.25-mm) axial images for renal stone detection on unenhanced CT [14]. Therefore, the purpose of this study was to determine the sensitivity of thin axial and coronal MIP images for the detection of renal stones on portal venous phase CECT. Materials and Methods Study Group This study was HIPAA-compliant and was approved by our institutional review board. Informed consent was waived owing to the retrospective nature of the study. A search of the radiology database was performed to identify patients with renal stones 1200 AJR:207, December 2016

2 Detection of Renal Stones on Portal Venous Phase CT seen on CT of the abdomen performed with and without IV contrast material from August 1, 2009, to November 1, Patients were included if one or more renal stones were visualized in the kidneys and if the CT examination had both an unenhanced phase and a portal venous phase. This search yielded examinations performed using a four-phase liver protocol that were not performed specifically to assess for urolithiasis. Patients with medullary nephrocalcinosis (n = 3 patients) were excluded because accurate assessment of the number of stones in these patients is precluded; in addition, patients with CT images showing motion artifact (n = 1) and patients with polycystic kidney disease (n = 1) were excluded. The search yielded 59 patients (mean age, 58.6 years; age range, years; 55 men and four women) with renal stones; 13 patients without stones were also included during the image analysis to minimize bias during image review. CT Technique CT examinations were performed on either a GE Healthcare 64-MDCT scanner (Volume CT) or a Siemens Healthcare 64- or 128-MDCT scanner (Definition DS or Definition AS Plus, respectively). On the GE scanner, scans were obtained using a detector configuration of , beam collimation of 40 mm, pitch of 1.375, and gantry rotation speed of 0.5 second. The tube voltage was 120 kv, and the tube current exposure time product was variable and was based on automated dose modulation with a noise index ranging from 30 or 36 (prescribed for 1.25-mm reconstructed images). On the Siemens scanners, scans were obtained using a detector configuration of either or , beam collimation of 28.8 or 38.4 mm, pitch of 0.8, and gantry rotation speed of 0.5 second. The tube voltage was 120 kv, and the tube current exposure time product was variable and was based on automated dose modulation with a reference value ranging from 180 to 220 mas. Images were reconstructed at a thickness and interval of 1.25 mm on the GE 64-MDCT scanner, 1.0 mm on the Siemens 128-MDCT scanner, and 1.5 mm on the Siemens 64-MDCT scanner using a standard body filter on the GE scanner and a B40f kernel on the Siemens scanners. Coronal MIP reformatted images with a thickness of 5 mm and interval of 3 mm were generated for the portal venous phase. Portal venous phase images were acquired after a fixed delay of 80 seconds after IV injection of 125 ml of iodinated contrast material (iohexol [Omnipaque 350, GE Healthcare]) at a rate of ml/s. Image Analysis Portal venous phase thin axial and 5-mm coronal MIP images were reviewed independently by two radiologists with subspecialty training in abdominal imaging with 4 and 2 years of postfellowship experience. The reviewers were blinded to clinical information, the radiology reports, and the reference standard. They were aware that patients with no stones were included in the review but did not know the number of such cases. The review sessions were split so that each reviewer viewed the thin axial images first for one-half of the patients and the coronal MIP images first for the other half. Reviewers were instructed to record the number of stones in each kidney but did not review the ureters or bladder. Reference Standard The reference standard was obtained by consensus review of thin axial unenhanced CT images by the same two reviewers. There was a 1-month separation between the independent review and the consensus review. All kidney stones and the sizes of all stones were recorded in each kidney; punctate stones were recorded as being 1 mm in size. Care was taken to identify and exclude vascular calcifications, which were identified by linear and branching morphology corresponding to the course of renal artery branches, particularly in the setting of renal artery and aortic calcifications. Parenchymal calcifications such as dystrophic calcifications were not considered stones. Statistical Analysis Sensitivities were calculated for each reviewer for the detection of all stones and stones stratified by size, and 95% CIs were determined. We used the Spearman correlation coefficient to assess interrater reliability. Analyses were performed using the CORR procedure in SAS software (version 9.4, SAS Institute). Results A total of 148 stones were present with a mean of 2.5 (SD, 2.7) stones per patient and a range of 1 19 stones per patient. There was a mean of 1.3 (SD, 1.7) stones per kidney with a range of 1 12 stones per kidney. The mean stone size was 2.5 mm (SD, 2.7), and the number of stones by size is presented in Table 1. The sensitivities for each reviewer using thin axial and coronal MIP images are presented in Table 2. Figure 1 shows an example of a small stone seen on thin axial images but missed on MIP images. Reviewer 1 had one false-positive finding on the thin axial images, and reviewer 2 had two false-positive findings on the thin axial images. Reviewer 1 had two false-positive findings on the coronal MIP images, and reviewer 2 had nine false-positive findings on the coronal MIP images. Using thin axial images, reviewer 1 detected no stones in one patient with at least one stone, yielding a per-patient sensitivity of 98.3% (95% CI, %), and reviewer 2 TABLE 2: Sensitivities for Each Reviewer for the Detection of Renal Stones Using Thin Axial or Coronal Maximum- Intensity-Projection (MIP) Images Stones Sensitivity (%) TABLE 1: Number of Stones by Stone Size Stone Size (mm) Reviewer 1 Reviewer 2 No. of Stones Thin Axial Images Coronal MIP Images Thin Axial Images Coronal MIP Images All stones 89.9 ( ) 86.5 ( ) 83.1 ( ) 79.0 ( ) Stones 2 mm 99.0 ( ) 96.2 ( ) 98.0 ( ) 91.4 ( ) Stones 3 mm ( ) ( ) ( ) 96.6 ( ) Note Numbers in parentheses are 95% CIs. AJR:207, December

3 Corwin et al. detected no stones in five patients with at least one stone, yielding a per-patient sensitivity of 91.5% (95% CI, %). All of these patients except one patient had only one 1-mm stone in one kidney. The remaining patient had one 1-mm stone in each kidney. Correlations were high within each imaging technique between reviewers, with R values of 0.89 for both thin axial and coronal MIP images (both p values < 0.001). Discussion The results of our study show that thin images on portal venous phase CECT have a high sensitivity for the detection of renal stones 2 mm. Unenhanced CT is the recommended imaging test for patients presenting with acute flank pain and suspicion of renal stone disease [15]. CECT has traditionally been thought to be limited in detecting renal stones and has not been recommended in this setting. In fact, the American College of Radiology appropriateness criteria give CT of the abdomen and pelvis with contrast material the lowest rating level (2, usually not appropriate) in this situation [1]. Although the decision to perform unenhanced CT can be straightforward in some patients with suspected stone disease, urolithiasis can mimic other causes of acute abdominal pain and the clinical differential diagnosis may be broad. Non urinary tract A Fig year-old man with nephrolithiasis. A and B, Thin axial unenhanced (A) and thin axial portal venous phase (B) CT images show 2-mm renal stone (arrows). C, Coronal maximum-intensity-projection (MIP) CT image from portal venous phase shows faint density (arrow) in region of stone, but this finding is less conspicuous than on A and B. Stone was identified by both reviewers on thin axial portal venous phase images and was missed by both reviewers on MIP images. causes of acute flank pain have been found in up to 20 35% of unenhanced CT examinations performed for the assessment of flank pain [8, 16]. Additionally, other renal processes such as pyelonephritis, renal infarcts, and even renal neoplasms can present with acute flank pain and hematuria similar to urolithiasis [16]. Therefore, a dilemma may arise when choosing the most appropriate protocol because most of the common alternative diagnoses are best evaluated with CECT during the portal venous phase. In our practice, we commonly encounter referrals of patients with clinical histories that request evaluation for both urolithiasis and other processes such as appendicitis or diverticulitis. If contrast material is chosen to ensure optimal evaluation of these processes, it is important to know the sensitivity of CECT for the detection of renal stones. The reduced sensitivity of CECT for renal stone detection is theoretically because of the avid enhancement of the kidneys, which decreases the contrast between the radiodense stones and the renal parenchyma. However, this phenomenon has not been formally studied until recently. Dym et al. [10] found an overall sensitivity of 81% for the detection of renal stones on portal venous phase CECT. Not surprisingly, large stones were more easily identified than small stones with sensitivities of 88%, 95%, and 99% for stones B measuring 2, 3, and 4 mm, respectively. The authors concluded that contrast material can be given to patients when urolithiasis and other inflammatory conditions are being considered because most large renal stones will be visualized. However, the overall sensitivity for renal stones was only moderate, and the sensitivity for renal stones < 2 mm was only 61%. Kawamoto et al. [17] evaluated the detection of renal stones during arterial phase CT and found that 75% of renal calculi and all stones > 5 mm were detected. The discrepancy in sensitivities between these studies and ours is likely related to the slice thickness of the reviewed images. Whereas we used thin images ( mm thickness), the slice thickness was 2.5 mm and 3 5 mm for the studies by Dym et al. and Kawamoto et al., respectively. It is well established that renal stone detection is improved with decreasing slice thickness on unenhanced CT [11 13]. This effect may be even more important during the portal venous phase because of the decreased contrast between stones and kidney. In our practice, we routinely evaluate thin axial reconstructed images or coronal MIP images to optimize renal stone detection on unenhanced CT. The results of our study show that thin axial images improve the sensitivity of stone detection during portal venous phase CECT to % for all stones. It C 1202 AJR:207, December 2016

4 Detection of Renal Stones on Portal Venous Phase CT is also important to note that stones < 1 mm were excluded from the Dym et al. study. Exclusion of stones < 1 mm would falsely increase the reported sensitivity because it is common to see punctate stones measuring < 1 mm in the kidneys, and these stones would be the ones most likely to be missed. We included stones < 1 mm in our study, which may be one reason that the reported sensitivities for all stones was only slightly higher than that in the prior study. However, we believe that including these stones provides a more accurate assessment of the ability of CECT to detect nephrolithiasis. Comparing the sensitivities between the two studies for stones 2 mm, our sensitivity of % was much higher than the 88% seen in the prior study [10]. Therefore, we conclude that thin axial images during portal venous phase CECT depict renal stones 2 mm with high sensitivity. If CECT is performed when there is some concern for renal stones, review of thin images should be performed to optimize stone detection. On the basis of our results, some stones 1 mm will be missed with CECT. The clinical importance of small nonobstructing renal stones in the setting of acute flank pain is controversial. Multiple studies have reported improvement in pain after successful treatment of these stones [18 20]. However, these studies were performed before the advent of high-resolution thin-section CT and thus stones measuring < 1 mm would not have been assessed. A more recent study by Furlan et al. [21] suggested that small nonobstructing stones seen on unenhanced CT in patients presenting with renal colic were a source of pain. The authors hypothesized that the pain is secondary to uroepithelial irritation or to intermittent obstruction at the ureteropelvic junction or infundibulum-calyx junction. However, that study evaluated unenhanced CT using 5-mm section thickness, and it is very likely that the small 1 mm stones seen in our study group would have been missed. Thus, the clinical relevance of these small stones remains unclear. However, the mean sensitivity of the two reviewers for detecting at least one stone on a per-patient basis was 94.9%. This result suggests that the diagnosis of nephrolithiasis in a given patient would still almost always be made even if one small stone was missed. It has recently been shown that coronal MIP images are as sensitive as thin axial images for renal stone detection on unenhanced CT [14]. However, our study found that stone detection is poorer with coronal MIP images than with thin axial images during the portal venous phase. Without contrast material, the MIP technique accentuates the contrast between the bright stone and relatively hypodense renal parenchyma. However, with contrast material, the kidney avidly enhances, and the MIP technique will also heighten the brightness of the kidney, thus minimizing the differences in contrast between stones and kidney. Therefore, coronal MIP images are unlikely to be of added value for renal stone detection during portal venous phase CT. On the basis of the slightly inferior sensitivity of CECT for small renal stones compared with unenhanced CT, we emphasize that unenhanced CT is still the test of choice for clear cases of suspected nephrolithiasis. There are disadvantages to using IV contrast material such as the risks of contrast reactions and the higher radiation dose [22, 23]. Contrast administration would also not be appropriate in patients with acute or chronic renal failure. We consider CECT to be an appropriate choice when there is a reasonably high pretest probability of an alternative diagnosis that would justify the use of IV contrast material. We did not evaluate for stones in the ureters or bladder. One would not expect CECT to have any lower sensitivity for ureteral stones than unenhanced CT, although this issue has never been formally studied to our knowledge. In fact, the presence of IV contrast material may better delineate the course of the ureters and help make the distinction from vessels as long as the images were obtained before contrast excretion into the ureters. Our study has limitations. Although the reviewers were aware that negative cases were included in the review, they were specifically looking for renal stones, which may have increased sensitivity compared with routine clinical practice. We also did not have a reference standard other than unenhanced CT, and it is possible that small foci of increased density due to noise could have been mistaken for punctate stones. In conclusion, thin axial images are highly sensitive for the detection of renal stones 2 mm on portal venous phase CT. Coronal MIP images of the portal venous phase do not improve renal stone detection over thin images using CECT. In patients presenting with abdominal pain and a broad differential diagnosis that includes nephrolithiasis and other acute inflammatory conditions, CECT using thin images can be performed without significantly limiting the detection of renal stones. References 1. American College of Radiology website. Moreno CC, Beland M, Goldfarb S, et al. ACR appropriateness criteria: acute onset flank pain suspicion of stone disease. Narrative/. Accessed January 6, Patatas K, Panditaratne N, Wah TM, Weston MJ, Irving HC. Emergency department imaging protocol for suspected acute renal colic: re-evaluating our service. Br J Radiol 2012; 85: Hoppe H, Studer R, Kessler TM, Vock P, Studer UE, Thoeny HC. Alternate or additional findings to stone disease on unenhanced computerized tomography for acute flank pain can impact management. J Urol 2006; 175: ; discussion, Dalrymple NC, Verga M, Anderson KR, et al. The value of unenhanced helical computerized tomography in the management of acute flank pain. J Urol 1998; 159: Chen JH, Chern CH, Chen JD, How CK, Wang LM, Lee CH. Left flank pain as the sole manifestation of acute pancreatitis: a report of a case with an initial misdiagnosis. Emerg Med J 2005; 22: American College of Radiology website. ACR appropriateness criteria. Accessed January 6, Ather MH, Faizullah K, Achakzai I, Siwani R, Irani F. Alternate and incidental diagnoses on noncontrast-enhanced spiral computed tomography for acute flank pain. Urol J 2009; 6: Rucker CM, Menias CO, Bhalla S. Mimics of renal colic: alternative diagnoses at unenhanced helical CT. RadioGraphics 2004; 24(suppl 1):S11 S28; discussion, S28 S33 9. Tamm EP, Silverman PM, Shuman WP. Evaluation of the patient with flank pain and possible ureteral calculus. Radiology 2003; 228: Dym RJ, Duncan DR, Spektor M, Cohen HW, Scheinfeld MH. Renal stones on portal venous phase contrast-enhanced CT: does intravenous contrast interfere with detection? Abdom Imaging 2014; 39: Ketelslegers E, Van Beers BE. Urinary calculi: improved detection and characterization with thin-slice multidetector CT. Eur Radiol 2006; 16: Kambadakone AR, Eisner BH, Catalano OA, Sahani DV. New and evolving concepts in the imaging and management of urolithiasis: urologists perspective. RadioGraphics 2010; 30: Memarsadeghi M, Heinz-Peer G, Helbich TH, et al. Unenhanced multi-detector row CT in patients AJR:207, December

5 Corwin et al. suspected of having urinary stone disease: effect of section width on diagnosis. Radiology 2005; 235: Corwin MT, Hsu M, McGahan JP, Wilson M, Lamba R. Unenhanced MDCT in suspected urolithiasis: improved stone detection and density measurements using coronal maximum-intensityprojection images. AJR 2013; 201: Eisner BH, McQuaid JW, Hyams E, Matlaga BR. Nephrolithiasis: what surgeons need to know. AJR 2011; 196: Talner L, Vaughan M. Nonobstructive renal causes of flank pain: findings on noncontrast helical CT (CT KUB). Abdom Imaging 2003; 28: Kawamoto S, Horton KM, Fishman EK. Detection of renal calculi on late arterial phase computed tomography images: are noncontrast scans always needed to detect renal calculi? J Comput Assist Tomogr 2008; 32: Andersson L, Sylven M. Small renal caliceal calculi as a cause of pain. J Urol 1983; 130: Mee SL, Thuroff JW. Small caliceal stones: is extracorporeal shock wave lithotripsy justified? J Urol 1988; 139: Coury TA, Sonda LP, Lingeman JE, Kahnoski RJ. Treatment of painful caliceal stones. Urology 1988; 32: Furlan A, Federle MP, Yealy DM, Averch TD, Pealer K. Nonobstructing renal stones on unenhanced CT: a real cause for renal colic? AJR 2008; 190:[web]W125 W Amato E, Salamone I, Naso S, Bottari A, Gaeta M, Blandino A. Can contrast media increase organ doses in CT examinations? A clinical study. AJR 2013; 200: Pasternak JJ, Williamson EE. Clinical pharmacology, uses, and adverse reactions of iodinated contrast agents: a primer for the non-radiologist. Mayo Clin Proc 2012; 87: FOR YOUR INFORMATION The AJR has made getting the articles you really want really easy with an online tool, Really Simple Syndication, available at It s simple. Click the RSS button located in the menu on the right side of the page. You ll be on your way to syndicating your AJR content in no time AJR:207, December 2016