Evaluation of Mechanical Heart Valve Size and Function With ECG-Gated 64-MDCT

Transcription

1 Cardiopulmonary Imaging Original Research Laounty et al. 64-MDCT of Mechanical Heart Valve Cardiopulmonary Imaging Original Research Troy M. Laounty 1 Prachi P. garwal 2 amer Chughtai 2 David S. ach 3 Eric Wizauer 2 Ella. Kazerooni 2 Laounty TM, garwal PP, Chughtai, ach DS, Wizauer E, Kazerooni E Keywords: cardiac CT, valvular heart disease DOI: /JR Received November 26, 2008; accepted after revision pril 5, Presented as an abstract at the 2008 annual meeting of the merican Heart ssociation, New Orleans, L. 1 Department of Medicine, Division of Cardiology, Weill Cornell Medical College, 520 E 70th St., Starr Pavilion 4th Floor, New York, NY ddress correspondence to T. M. Laounty (tml9001@med.cornell.edu). 2 Department of Radiology, Division of Cardiothoracic Radiology, University of Michigan Health System, nn rbor, MI. 3 Department of Medicine, Division of Cardiovascular Medicine, University of Michigan Cardiovascular Center, nn rbor, MI. WE This is a Web exclusive article. JR 2009; 193:W389 W X/09/1935 W389 merican Roentgen Ray Society Evaluation of Mechanical Heart Valve Size and Function With ECG-Gated 64-MDCT OJECTIVE. The purpose of our study was to determine whether CT can accurately evaluate mechanical heart valve size and function. MTERILS ND METHODS. Sixty-two patients with mechanical valves (37 singledisc, 27 bileaflet; 59 aortic, 5 mitral) were evaluated with ECG-gated 64-MDCT and transthoracic echocardiography; a subset of 10 patients underwent cinefluoroscopy. Two readers independently interpreted each study. RESULTS. The mean age of the patients was 46.4 ± 14.4 years; 50 were men and 12 were women. There was excellent correlation, and differences between CT readers were absent to small in measuring the opening angle (r = 0.96, p < 0.001; 76.7 ± 9.0 vs 76.8 ± 9.6, p = 0.73), annulus diameter (r = 0.96, p < 0.001; 25.9 ± 3.3 vs 25.9 ± 3.2 mm, p = 0.62), and geometric orifice area (r = 0.98, p < 0.001; 3.8 ± 0.9 vs 3.6 ± 0.8 cm 2, p < 0.001). There was strong correlation without difference in opening angle between CT and cinefluoroscopy (r = 0.77, p < 0.001; 79.2 ± 9.8 vs 77.2 ± 15.5, p = 0.45). Compared with manufacturer specifications, CT reported opening angles that were smaller for single-disc valves (n = 36, 67.4 ± 5.7 vs 75, p < 0.001) and similar for bileaflet valves (n = 42 for 21 valves, 83.8 ± 3.9 vs 85, p = 0.05), valves, with small underestimation with CT versus specifications in annulus diameter (n = 41; r = 0.75, p < 0.001; 26.4 ± 3.0 vs 27.5 ± 3.3 mm, p = 0.003), and geometric orifice area (n = 35; r = 0.90, p < 0.001; 3.7 ± 0.7 vs 3.8 ± 0.8 cm 2, p = 0.04). Each disc closed fully on CT; none had more than mild regurgitation on echocardiography. CONCLUSION. CT can measure the size and function of mechanical valves with high interobserver agreement and results similar to specifications. The opening angle with CT strongly correlates with cinefluoroscopy. CT is promising for the assessment of mechanical valves. C omplications of mechanical heart valves include thrombosis, pannus formation, and infective endocarditis, which can lead to valve obstruction and elevated transvalvular gradients or regurgitation. Transthoracic echo cardiography (TTE) is the initial test of choice because it can measure the transvalvular gradient and assess for regurgitation. However, evaluation of disc excursion is often difficult because of both reverberation artifact and difficulty in appropriately aligning the ultrasound beam. Transesophageal echocardiography (TEE) is better suited to determine valve opening angle and closure than TTE [1], although TEE may still yield diagnostic errors when correlated with surgical findings [2], requires sedation, and is an invasive procedure. Cinefluoroscopy can directly visualize mechanical heart valve components, permits determination of the prosthesis type and valve function in the majority of cases [3], and has been reported to be superior to TEE for the evaluation of disc motion [4]. More recently, ECG-gated MDCT has been evaluated as an alternate technique to evaluate mechanical heart valves. Potential advantages of CT over cinefluoroscopy include the rapid acquisition speed, collection of a 3D dataset that permits reconstruction and visualization of the valve in any plane, and potential for wider availability during off-hours or at facilities lacking cinefluoroscopy. The disadvantages of CT compared with cinefluoroscopy include the lower temporal and spatial resolution of CT as well as the need for potentially higher doses of ionizing radiation. One small study reported that CT could assess the function of bileaflet valves but raised concerns regarding assessment of single-disc valves [5]. Others have described visualization JR:193, November 2009 W389

2 Laounty et al. of obstructed valve discs [6, 7], pannus formation [8, 9], and infective endocarditis vegetations [10]. We sought to assess the ability of CT to evaluate mechanical heart valve function in a larger cohort of patients. In addition to evaluating valve function, it can be useful to determine the valve size. When the valve size is not known and the surgical record is not immediately available, diagnosis can be a challenge because TTE may not be able to accurately determine valve size. lthough not previously established, CT may be able to measure valve size, which can assist in evaluation for patient prosthesis mismatch and can be used to determine the expected normal gradient by valve type and size. We hypothesized that CT could determine mechanical heart valve function and size with high interobserver reproducibility and with good agreement with manufacturer specifications and cinefluoroscopy. We assessed the ability of CT to evaluate the opening angle, disc closure, annulus diameter, and geometric orifice area in tilting-disc mechanical heart valves. We compared results between CT readers, CT and cinefluoroscopy, and CT and manufacturer specifications. We further compared valvular regurgitation on TTE and disc closure on CT and compared transvalvular gradients on TTE to the valve size and disc opening angles. Materials and Methods This retrospective study was approved by the institutional review board with a waiver of informed consent and is HIP-compliant. Patients with previous implantation of a mechanical heart valve who underwent both TTE and ECG-gated 64-MDCT of the cardiac or thoracic aorta between January 2005 and February 2008 formed the study group. Inclusion criteria included adult patients (age, 18 years; age range, years), with no more than 3 months between CT and TTE examinations (mean time difference [SD], 8.8 ± 20.0 days; range, 0 90 days). Exclusion criteria included patients with interval cardiac surgery or altered symptoms that could suggest changes in valve function between studies (based on electronic medical records). total of 62 consecutive patients with 64 valves met these criteria. These patients had 37 single-disc valves and 27 bileaflet valves, with 59 in the aortic position and five in the mitral position. The valve manufacturer was known for 36 of the 37 single-disc valves (Hall Valve, Medtronic) and 21 of the 27 bileaflet valves (St. Jude Medical Valve, St. Jude Medical). subset of patients (n = 10) also had cinefluoroscopy evaluation of the valves. To increase the sample size, cinefluoroscopy evaluations within 3 years of the CT study were compared (mean difference, 9.4 ± 8.0 months; range, 0 26 months), provided there was no clinical concern for altered valve function or change in the patient s clinical symptoms according to electronic medical records. ll enrolled ECG-gated CT studies were performed using an aortic protocol and were clinically ordered for assessment of the aorta. The institutional practice included retrospective ECG-gating for all thoracic aorta CT studies for the purpose of performing measurements of the aortic root; the standard institutional practice also included reconstruction of the thoracic portion of the study in multiple phases for routine qualitative assessment of the native aortic valve, which was performed in patients with mechanical valves by default. Mechanical valve function was not assessed at the time of the CT examination. The indication for CT was follow-up of thoracic aorta aneurysm or dissection (with or without repair) in 59 patients and suspected thoracic aorta aneurysm in three patients. TTE indications included assessment of aortic valve replacement or assessment for thoracic aortic aneurysm. Indications for cinefluoroscopy included assessment of suspected valve dysfunction in five cases, with the remainder performed during coronary angiography to evaluate known or suspected coronary artery disease. Two fellowship-trained cardiothoracic radiologists (each with more than 4 years of experience) independently evaluated the CT examinations, and two experienced cardiologists (certified level 3 for echocardiography) independently reviewed TTE and cinefluoroscopy studies. ll readers were blinded to all clinical data, including the observations from other readers. CT was used to measure the opening angle, annulus diameter, and geometric orifice area and to evaluate for complete valve closure. TTE was used to measure the mean pressure gradient, left ventricular ejection fraction, and grade of valve regurgitation. Cinefluoroscopy was used to measure opening angles and assess for complete valve closure. CT Technique and Evaluation CT examinations were designed to evaluate the thoracic or thoracoabdominal aorta. No pharmacologic agents were used for heart rate control. timing bolus of 15 ml was used in all patients to determine the optimal scanning start time. Nonionic iodinated IV contrast material, either 135 ml iopromide (370 mg/ml; Ultravist, ayer HealthCare) or 95 ml iodixanol (320 mg/ml; Visipaque, GE Healthcare), was administered at a rate of 4 ml/s. CT examinations were performed using a 64-MDCT scanner (VCT, GE Healthcare) with retrospective ECG gating in the cranial to caudal direction. Detector collimation width was mm, coverage was 40 mm, and reconstructed slice thickness was 1.25 mm. Gantry rotation time was 0.35 second, and scanning pitch ranged between 0.16 and 0.20 (adjusted for heart rate). Maximum tube current ranged from 450 to 700 m, depending on patient size, with a fixed tube voltage of 120 kvp. Dose modulation was applied for CT examinations of patients without tachyarrhythmia using maximum tube current during left ventricular diastole (between 60% and 80% of the R-R interval) and 20% of the maximum tube current for the remainder of the cardiac cycle. CT datasets with images throughout the cardiac cycle at 10% increments of the R-R interval (with 5% increments in one case) were reviewed on an advanced processing workstation (dvantage Windows Workstation, version 4.3_05 with CardIQ software, GE Healthcare). Valve evaluation was performed using multiplanar reformatted images in a cine mode. The mechanical valve was centered on a sagittal plane with a reformatting plane on the sagittal image used to generate an oblique long-axis view of the valve. n additional reformatting plane was used on the oblique long-axis view to generate a double-oblique short-axis view of the valve. The 3D cursor was placed in the center of the valve orifice on the short-axis view, and then the valve on the long-axis view was rotated while maintaining the annulus plane to provide an optimal on-edge view of the opening and closing of the disc or discs for the opening angle and evaluation of disc closure. In the case of single-disc valves, the image was rotated to incorporate the center arm pivot (Figs. 1 3). Correct valve orientation was confirmed by examining the valve using maximum intensity projections with adequate thickness to visualize the entire valve apparatus (Fig. 4). n electronic protractor was used to measure the opening angle of each disc, using the plane of the valve annulus as the baseline. Complete closure was defined as closure to 0 (the valve plane) for single-disc valves or complete disc apposition for dual-disc valves. Incomplete closure was defined as closure to > 0 for single-disc valves or incomplete disc apposition for dual-disc valves. The short-axis view of the valve was used to measure the geometric orifice area by planimetry of the internal edge of the annulus, and the annulus diameter was determined by measurement from external edge to external edge of the annulus. To visualize the mechanical valves, the default window for vertebrae was selected, with additional adjustment to the window level and width to minimize blooming at the discretion of the reader. The magnification factor for measurement was at the discretion of the reader. W390 JR:193, November 2009

3 64-MDCT of Mechanical Heart Valve The radiation dose was available for 48 of the 62 study patients and included the cumulative dose for the thorax and abdomen in all patients as well as the pelvis in 19 patients. The effective dose was estimated using the conversion factor from the European Working Group for Guidelines on Quality Criteria in CT [11]. ecause all examinations included the thoracic and abdominal aorta and the D G Fig year-old man with prior thoracic aortic aneurysm and aortic insufficiency, with follow-up testing 1 year after ascending aorta repair and Medtronic Hall aortic valve replacement. and, Mechanical valve is centered on sagittal plane with reformatting plane on sagittal image () used to create oblique long-axis image () of valve. C, dditional reformatting plane is used on oblique long-axis image () to create double-oblique short-axis image of valve. nnulus diameter (D) and geometric orifice area (GO) are measured from short-axis image as shown. D F, On short-axis images, 3D cursor is placed in center of valve orifice, and image on oblique long-axis image () is rotated while maintaining plane of valve annulus to provide on-edge image of disc or discs during both disc closure (D) and opening (E). For single-disc valves, image incorporates center pivot arm. Opening angle (O) (E and F) is measured, and disc is assessed for complete closure to plane of annulus. On corresponding cinefluoroscopy image (F), opening angle is shown. G, Transthoracic echocardiography image shows inability to directly visualize valve (arrow) due to artifact. conversion factor for abdominal CT [0.015 msv (mgy cm) 1 ] is higher than for thoracic CT [0.014 msv (mgy cm) 1 ], the former was used so that doses would not be underestimated. The typical effective radiation dose at our institution for the thoracic aorta alone is approximately 25 msv. TTE Technique and Evaluation TTE studies were performed with commercially available echocardiography equipment (Sequoia 256 or 572, Siemens Healthcare; or Philips 5500 or ie33, Philips Healthcare) and reviewed offline using Prosolv Cardiovascular nalyzer version workstation (Prosolv Cardiovascular E Solutions, FujiFilm US). Each reader measured the transvalvular mean pressure gradient by electronically integrating the area under the continuous-wave Doppler spectral envelope of the highestvelocity signal available. Left ventricular ejection fraction was assessed by qualitative evaluation of the parasternal long, parasternal short, apical fourchamber, and apical two-chamber views. The regurgitation grade was determined using standard recommended Doppler parameters [12]. Elevated transvalvular gradients were defined as a mean gradient of greater than 15 mm Hg for aortic mechanical heart valve or a mean gradient of greater than 5 mm Hg for mitral mechanical heart valve. C F JR:193, November 2009 W391

4 Laounty et al. Fig year-old man who underwent St. Jude Medical aortic valve replacement and aortic root repair 12 years earlier and now presented with chest pain. For bileaflet mechanical valves, oblique longaxis image is rotated while maintaining plane of valve annulus to display view that shows two discs on edge during both opening and closing, with 3D cursor centered on short-axis valve orifice., CT image shows bileaflet discs closed with complete apposition., Opening angle (O) of each disc is measured in comparison with annulus as baseline. C and D, Short-axis image (C) permits measurement of annulus diameter (D) and geometric orifice area (GO). Corresponding cinefluoroscopy image (D) also permits measurement of O. Cinefluoroscopy Technique and Evaluation Cinefluoroscopy images were obtained with multiple fluoroscopy units (biplane or single plane), and reviewed using the Syngo Dynamics Workplace Image Management System version (Siemens Healthcare). Opening angles were measured using the plane of the annulus as the baseline. Complete closure was defined as closure to 0 for single-disc valves or complete disc apposition for bileaflet valves. Statistical nalysis land-ltman tests were performed with Medcalc version because these tests are not available with SPSS. ll other statistical analysis was performed using SPSS version 12.0 for Windows. Comparison between groups was performed with Pearson s correlations, land-ltman plots, and paired Student s t tests. Comparison between observed values and manufacturer specifications for opening angles was performed with one-sample Student s t tests. p value of < 0.05 was considered significant. Results The mean subject age was 46.4 ± 14.4 years; 50 patients were men and 12 were women. The mean left ventricular ejection fraction on TTE was 62.1% ± 8.0%; it was less than 50% in two patients. The mean pressure gradient was 12.9 ± 6.4 mm Hg for aortic valves and 5.4 ± 3.2 mm Hg for mitral valves. Elevated valvular gradients were observed in 22 mechanical heart valves (11 single-disc and 11 bileaflet valves; 20 aortic valves and 2 mitral valves); for these aortic valves the mean gradient was 20.3 ± 4.0 mm Hg (range, mm Hg), with mean gradients of 6.9 and 10.3 mm Hg for the mitral valves. Using CT, the opening angle, annulus diameter, and assessment of closure could be determined in all cases; the geometric orifice area could be measured in 58 cases, with motion artifact preventing planimetry of the geometric orifice area in six cases. The mean radiation dose was 44.0 ± 8.3 msv, which included the thoracic and abdominal aorta in all patients as well as the pelvis in 19 patients. Interobserver greement on CT There was excellent correlation and no significant difference between CT readers in the opening angle for both single-disc valves (n = 37; r = 0.85, p < 0.001) and bileaflet valves (n = 54 discs from 27 valves; r = 0.86, p < 0.001). When combined, the findings were similar (n = 91 discs from 64 valves; r = 0.96, p < 0.001) (Table 1). land-ltman plot showed clinically acceptable agreement between readers (mean difference, 0.1, ± 1.96 = 5.4 ) (Fig. 5). ll discs closed fully on CT, with closure to 0 in all single-disc valves and complete coaptation of the discs with all bileaflet valves. There was excellent correlation with absent to small differences between CT readers for annulus diameter (n = 64; r = 0.96, p < C 0.001) and geometric orifice area (n = 58; r = 0.98, p < 0.001); there was clinically acceptable agreement between readers (annulus diameter; mean difference, 0.1 ± 1.96 = 1.8 mm; geometric orifice area: mean difference, = 0.17 ± 1.96 = 0.31 cm 2 ) (Fig. 6). CT Compared With Cinefluoroscopy There was a strong correlation and no difference between the mean opening angle on CT and cinefluoroscopy (n = 16 discs from four single-disc and six bileaflet valves; r = 0.77, p < 0.001) (Fig. 7). With a land-ltman plot, there were clinically relevant differences between the two methods (mean difference, 2.0 ± 1.96 = 19.9 ), which was driven by two outliers. In one outlier, the opening angle was 99.5 with cinefluoroscopy and 82.5 with CT; according to manufacturer specifications, the normal opening angle is 85. In the other case (Fig. 3), the opening angle was 42.5 with cinefluoroscopy and 76.2 with CT; according to manufacturer specifications, the normal opening angle is 75. For this latter patient, the D W392 JR:193, November 2009

5 64-MDCT of Mechanical Heart Valve D Fig year-old man with follow-up CT 1 year after Medtronic Hall aortic valve replacement and aorta repair for thoracic aortic aneurysm. and, This case represents disagreement in opening angle (O) between CT () and cinefluoroscopy (). With CT, opening angle was 76.2, which is close to normal O of 75 defined by manufacturer specifications. With cinefluoroscopy, O appears to be 42.5, which would represent significant decrease consistent with partial obstruction. On echocardiography, patient had normal ejection fraction and mean transvalvular gradient of 9.2 mm Hg, which are consistent with normally functioning valve, suggesting that cinefluoroscopy image was not obtained from appropriate orientation and is likely inaccurate. Lines are offset from image to improve visualization. Fig. 4 Impact of valve orientation on opening angles in 41-year-old man with follow-up study 1 year after thoracic aneurysm repair and Meditronic-Hall aortic valve replacement. F, Medtronic Hall aortic valve replacement is shown with thick maximum-intensity-projection (MIP) images that permit 3D orientation of valve, with corresponding thin images that can be used to measure valve opening. These images are helpful to confirm accurate valve orientation. Visualization of valve with these thick MIP images can confirm appropriate orientation. Same valve is shown with appropriate orientation () and opening angle of 76 (), off-axis orientation (C) that results in artificially low opening angle measurement of 62 (D), and additional off-axis orientation (E) that results in artificially high opening angle of 98 (F). E C F mean pressure gradient on TTE was 9.2 mm Hg. oth cases suggest errors with cinefluoroscopy and not CT (see Discussion). With these two outliers excluded, the correlation was very strong, with no difference between CT and cinefluoroscopy (n = 14 discs; r = 0.96, p < 0.001; 79.2 ± 10.4 vs 78.1 ± 12.1, p = 0.27), and clinically acceptable agreement between techniques with a land-ltman plot (mean difference, 1.1 ± 1.96 = 6.9 ). ll discs had complete closure on cinefluoroscopy in agreement with CT. JR:193, November 2009 W393

6 Laounty et al. TLE 1: Comparison of CT Variables Variable CT Interobserver greement CT Mean vs Cinefluoroscopy (CF) CT Compared With Manufacturer Specifications Manufacturer specifications for opening angle were available for 57 valves; the opening angle was greater than 10 below specification for 14 valves, and no valve on CT had an opening angle greater than 20 below specification. In comparison with available manufacturer specifications, the opening angle with CT was lower for single-disc valves (n = 36) and similar for bileaflet valves (n = 42 discs for 21 valves). n underestimation was observed with CT compared with avail- able specifications for annulus diameter (n = 41; r = 0.75, p < 0.001) and geometric orifice area (n = 35; r = 0.90, p < 0.001). Echocardiography On echocardiography, all valves had absent to mild regurgitation, with no cases of mild to moderate, moderate, or severe regurgitation. This is consistent with the complete closure of all valves observed on CT because a small amount of regurgitation occurs with most mechanical valves and is thought to prevent thrombus formation on the upstream side of CT Mean vs Normal Manufacturer Specifications (MS) Reader 1 Reader 2 p CT CF p CT MS p Single-disc opening angle ( ) 67.5 ± ± ± < ileaflet opening angle ( ) 83.0 ± ± ± Overall opening angle ( ) 76.7 ± ± ± ± nnulus diameter (mm) 25.9 ± ± ± ± Geometric orifice area (cm 2 ) 3.8 ± ± 0.8 < ± ± Note Data are mean ± SD. Dash indicates not applicable. Opening ngle CT, Reader 2 ( ) nnulus Diameter, CT, Reader 2 (mm) Opening ngle, CT Reader 1 ( ) nnulus Diameter, CT Reader 1 (mm) Difference in CT Opening ngles ( ) Difference in CT nnular Diameter (mm) SD = 5.3 Mean SD = Mean CT Opening ngles ( ) 1.96 SD = 1.9 mm Mean = 0.1 mm 1.96 SD = 1.7 mm Mean CT nnular Diameter (mm) Fig. 5 Comparison of valve opening angles between CT readers., Graph shows strong correlation between readers., land-ltman plot shows clinically acceptable agreement. Fig. 6 Comparison of annulus diameter between CT readers., Graph shows strong correlation between readers., land-ltman plot shows clinically acceptable agreement. the valve [13]. The mean transvalvular pressure gradient for aortic valves inversely correlated to the annulus diameter (n = 59; r = 0.65, p < 0.001) and geometric orifice area (n = 54; r = 0.63, p < 0.001) on CT. Elevated valvular gradients were not associated with decreased opening angles (p = 0.17). Discussion ecause many patients with a mechanical heart valve have multiple medical comorbidities, it often can be difficult to determine the cause of symptomatic complaints such as dyspnea, and a thorough evaluation of mechanical heart valve function is often needed. In addition to TTE, further testing may be necessary, particularly in cases of elevated valve gradients. CT has high interobserver agreement and has good agreement with both cinefluoroscopy and manufacturer specifications for disc opening and closing. The opening angle with CT may be slightly smaller than manufacturer specifications; this is similar to observations W394 JR:193, November 2009

7 64-MDCT of Mechanical Heart Valve Mean Opening ngle With Fluoroscopy ( ) Mean Opening ngle With CT ( ) CT Fluoroscopy Opening ngle ( ) SD = 21.9 Mean = SD = Mean Opening ngle With CT and Fluoroscopy (mm) 100 in normally functioning Medtronic Hall and St. Jude Medical aortic valves using cinefluoroscopy, which reported opening angles of 72.0 and 80.2 respectively [4]. These differences from manufacturer specifications may be due to small amounts of pannus or scar formation that prevent complete opening or potentially normal in vivo parameters. Normally functioning mechanical tilting disc valves open and close very rapidly at the start of systole and diastole, which is the reason for the opening and closing clicks heard on auscultation and seen on transvalvular Doppler signals, although these clicks may be decreased with abnormally functioning valves [13], suggesting less rapid opening or closure. In this cohort of patients with normally functioning valves, the valves were observed by the readers to rapidly alternate between fully open and fully closed, and the temporal resolution of the scanner did not prevent assessment of disc opening and closure in any case. The impact of the limited temporal resolution of CT on the evaluation of mechanical heart valves with abnormal disc excursion may require additional study. The elevated SD between the opening angles with CT and cinefluoroscopy (Fig. 7) was driven by two outliers. In one case, cinefluoroscopy reported an opening angle that was 14.5 beyond the maximum opening angle defined by manufacturer specifications. In the other case, the opening angle on cinefluoroscopy was 32.5 below the normal manufacturer-reported opening angle (suggestive of a partial obstruction) in a patient with gradients suggesting normal valve function. oth cases suggest that the discs were not imaged in the appropriate orientation at cinefluoroscopy, which requires visualization in the plane of the valve annulus and rotation to visualize the opening and closing of tilting discs on edge. If the valve is not properly oriented, with the annulus aligned in the line of sight and with the valve rotated in this plane for on-edge visualization of the disc, it is possible to overestimate or underestimate the valve opening on cinefluoroscopy or CT (Fig. 4). Unlike CT datasets, which have the ability to reorient the valve in any plane, cinefluoroscopy images are cine clips that cannot be further adjusted and are dependent on accurate orientation during image acquisition. When these two cases of probable error with cinefluoroscopy are excluded, we found excellent and clinically acceptable agreement between CT and cinefluoroscopy. In addition to the evaluation of disc opening and closing, CT can measure the annulus diameter and geometric orifice area with high interobserver agreement and with very good agreement with known manufacturer specifications. There was a small underestimation of annulus diameter with CT compared with manufacturer specifications. To minimize blooming artifact, the vertebrae window preset was used, with further adjustment to the window level and width at the discretion of the reader. It is possible that excessive correction for blooming led to a small underestimation of the annulus diameter with CT, although the observed difference is small enough that clinically adequate approximation within one valve size (usually available in 2-mm increments) is possible. common cause of elevated gradients in the setting of normal disc function is patient prosthesis mismatch, in which prosthesis size is small relative to the patient s body size and expected normal cardiac output. This diagnosis is often made when other diagnoses are excluded and can prompt repeat surgery in Fig. 7 Comparison of valve opening angles between CT and cinefluoroscopy., Graph shows strong correlation between techniques., land-ltman plot shows two significant outliers. selected symptomatic patients. Determination of valve size and function is important in diagnosing patient prosthesis mismatch. CT can determine the valve type (number of discs) and size when these data are not available historically. The obtained Doppler velocities and gradients can then be compared with anticipated normal values, which vary by the size and type of valve [14, 15]. Limitations of this study include the retrospective design. This study evaluates two of the most common contemporary mechanical valves, and the applicability of these findings to other mechanical valves is not known. The CT data were limited to 10 axial datasets across the cardiac cycle reconstructed previously, with no opportunity to reconstruct additional images during other time points from the raw data. This raises the possibility that true systole or diastole could be missed in some patients. Cinefluoroscopy images were limited to cine clips saved during the initial acquisition, with no ability to acquire additional views. Furthermore, all available cinefluoroscopy images were compared with CT to provide a more complete comparison between CT and cinefluoroscopy, even in cases of probable error with cinefluoroscopy. To include the largest number of comparisons between CT and cinefluoroscopy, studies up to 3 years apart were evaluated, provided there were no clinical concerns for changes in valve function between studies. Finally, cases with significant valve dysfunction or obstruction were not available in this cohort, which is not surprising because all CT studies were performed for follow-up or evaluation of suspected thoracic aortic aneurysm or dissection. ecause CT is not part of the routine assessment of patients with suspected valve dysfunction (and currently JR:193, November 2009 W395

8 Laounty et al. lacks a clinical indication for this), only five patients had suspected valve dysfunction and all patients were clinically deemed to have normal valve function after cinefluoroscopy. Larger prospective studies may be needed to assess the role of CT more completely in patients with suspected valve dysfunction. Radiation exposure is another limitation of examining mechanical heart valves with CT. lthough cinefluoroscopy is associated with lower doses of radiation than in the cohort in this study, CT assessment limited to the mechanical heart valve could permit only smaller z-axis coverage and hence reduce the dose. dditionally, that all valves were adequately visualized even in the setting of the reduced current present during systole with dose modulation suggests that reduced current could be used without dose modulation, which could further reduce radiation exposure. It is also possible that IV contrast material would not be needed for mechanical valve assessment, although that was not evaluated in this study. Future prospective studies could determine the ability of CT to assess mechanical valves with lower radiation doses and without iodinated contrast material. CT can measure the size and function of mechanical valves with high interobserver agreement and with results similar to specifications. Furthermore, CT assessment of disc excursion has clinically acceptable agreement with cinefluoroscopy. Compared with cinefluoroscopy, CT may be associated with increased radiation exposure and potentially require contrast administration, suggesting that this approach should be reserved for patients in whom information on valve size and function is clinically important but cannot be obtained by conventional means. CT is promising for the assessment of mechanical tilting-disc valves in appropriately selected patients and may merit further study. References 1. Muratori M, Montorsi P, Teruzzi G, et al. Feasibility and diagnostic accuracy of quantitative assessment of mechanical prostheses leaflet motion by transthoracic and transesophageal echocardiography in suspected prosthetic valve dysfunction. m J Cardiol 2006; 97: Faletra F, Constantin C, De Chiara F, et al. Incorrect echocardiographic diagnosis in patients with mechanical prosthetic valve dysfunction: correlation with surgical findings. m J Med 2000; 108: Montorsi P, Cavoretto D, Repossini, artorelli L, Guazzi MD. Valve design characteristics and cine-fluoroscopic appearance of five currently available bileaflet prosthetic heart valves. m J Card Imaging 1996; 10: Cianciulli TE, Lax J, eck M, et al. Cinefluoroscopic assessment of mechanical disc prostheses: its value as a complementary method to echocardiography. J Heart Valve Dis 2005; 14: Konen E, Goitein O, Feinberg MS, et al. The role of ECG-gated MDCT in the evaluation of aortic and mitral mechanical valves: initial experience. JR 2008; 191: Faletra FF, lain M, Moccetti T. lockage of bileaflet mitral valve prosthesis imaged by computed tomography virtual endoscopy. Heart 2007; 93: Leborgne L, Renard C, Tribouilloy C. Usefulness of ECG-gated multi-detector computed tomography for the diagnosis of mechanical prosthetic valve dysfunction. Eur Heart J 2006; 27: Teshima H, Hayashida N, Fukunaga S, et al. Usefulness of a multidetector-row computed tomography scanner for detecting pannus formation. nn Thorac Surg 2004; 77: oyagi S, Fukunaga S, rinaga K, Shojima T, Ueda T. Prosthetic valve obstruction: diagnostic usefulness of cineradiography and multidetectorrow computed tomography. Thorac Cardiovasc Surg 2007; 55: Kim RJ, Weinsaft JW, Callister TQ, Min JK. Evaluation of prosthetic valve endocarditis by 64- row multidetector computed tomography. Int J Cardiol 2007; 120:e27 e Shrimpton PC. European Guidelines for Multislice Computed Tomography: ppendix C. Funded by the European Commission; March Contract No. FIGM-CT CT-TIP. PDF_FILES/EC%20C%20Report%20D5%2- %20Dosimetry.pdf. ccessed ugust 31, Zoghbi W, Enriquez-Sarano M, Foster E, et al. Recommendations for evaluation of the severity of native valvular regurgitation with two-dimensional and Doppler echocardiography. J m Soc Echocardiogr 2003; 16: Feigenbaum H. Prosthetic valves. In: Feigenbaum H, ed. Feigenbaum s echocardiography, 6th ed. Philadelphia, P: Lippincott Williams & Wilkins, 2005: Rosenhek R, inder T, Maurer G, aumgartner H. Normal values for Doppler echocardiographic assessment of heart valve prostheses. J m Soc Echocardiogr 2003; 16: Rajani R, Mukherjee D, Chambers J. Doppler echocardiography in normally functioning replacement aortic valves: a review of 129 studies. J Heart Valve Dis 2007; 16: W396 JR:193, November 2009