MDCT for Suspected Appendicitis: Effect of Reconstruction Section Thickness on Diagnostic Accuracy, Rate of Appendiceal Visualization, and Reader Confidence Using Axial Images
Pamela T. Johnson, Karen M. Horton, Satomi Kawamoto, John Eng, Marchelle J. Bean, Shannon J. Shan, Elliot K. Fishman
Abstract
OBJECTIVE. The purpose of this study was to evaluate interpretative performance with different MDCT reconstruction parameters in adult patients with suspected appendicitis.
MATERIALS AND METHODS. MDCT scans of 212 adult patients obtained in an emergency department with a 64-MDCT scanner were prospectively collected. Acquisition technique included 24 × 1.2 mm detector configuration and IV contrast administration with or without oral contrast administration. Data sets were reconstructed with three techniques: 5 × 5 mm, 3 × 3 mm, and 2 × 1 mm (section thickness × interval). Each of the 212 sets of images (grouped by reconstruction technique) was reviewed retrospectively using axial sections by two independent readers blinded to diagnosis. Medical record review was conducted to identify patients with appendicitis. Visualization of the appendix, confidence in visualization, confidence for presence or absence of specific CT findings, diagnostic accuracy, and diagnostic confidence were compared across reconstruction techniques. Data were analyzed with simple and ordinal logistic regression with adjustment for multiple observations derived from each patient and for reader differences.
RESULTS. Progressively thinner reconstruction section thickness was associated with a significant increase in the rate of visualization of the appendix (p < 0.001 for 5 × 5 vs 3 × 3; p = 0.03 for 3 × 3 vs 2 × 1), visualization confidence (p < 0.001 for 5 × 5 vs 3 × 3 and 3 × 3 vs 2 × 1), and confidence for presence or absence of findings. Seventeen subjects (8%) had appendicitis. Correctness of diagnosis was not significantly associated with reconstruction method. However, for correctly diagnosed cases interpreted as normal, impression confidence increased with progressively thinner section thickness (p < 0.001 for 5 × 5 vs 3 × 3 and 3 × 3 vs 2 × 1).
CONCLUSION. In this investigation of contrast-enhanced MDCT of the appendix, visual ization of the appendix and confidence in interpretation of axial images progressively improved with use of thinner reconstruction sections.
Introduction
Use of CT for suspected appendicitis has increased dramatically since the late 1990s, with 85–93% of patients undergoing preoperative CT in recent years [1]. The improvement in outcome associated with this surge has been attributed to “increased use of high quality preoperative CT, with the results interpreted by experienced abdominal radiologists” [1]. After the implementation of helical CT technology, research studies aimed at protocol optimization have been essential to delineating best practice techniques. In comparisons of data acquisition parameters, investigators have evaluated enteric contrast administration [2–4], the utility of IV contrast enhancement [4–7], the length of anatomic coverage [5], and the role of narrow (5 mm) collimation [8]. With MDCT technology, much attention has been focused on data set reconstruction and display, in particular, multiplanar reconstruction (MPR) [6, 9–13]. Improvements in z-axis resolution have made multiplanar volume interrogation a valuable adjunct to data set evaluation [9].
Whether interpretative display can be limited to coronal reconstructions has been the subject of a few investigations [10–12]; however, axial sections continue to be used in combination with MPRs to evaluate the appendix. Published evidence suggests that appendix protocols for 16- and 64-MDCT continue to include reconstruction sections for axial images similar in thickness to the 5-mm collimation recommended for single-detector CT by Weltman et al. [8] despite the ability to acquire various section thicknesses less than 5 mm. The most common reconstruction section used in a survey of 16-MDCT appendix protocols was 5 mm [14], and 5-mm thickness has been used to reconstruct axial sections in a number of investigations of 16- and 64-MDCT [3, 9–13]. However, because the normal appendix cross-sectional diameter at MDCT can be as small as 3 mm [15, 16], use of reconstruction sections less than 5 mm may be efficacious for visualization of the appendix, particularly in patients without appendicitis. The disadvantages of using thinner reconstruction sections are increased image noise and the need to view a greater number of images [17].
Reconstruction section thickness for appendiceal MDCT was evaluated in a study by Lee et al. [18], who used 16-MDCT. Interpretation of axial images was compared with that of sliding-slab average intensity projection images. The axial images, generated from data sets reconstructed at 5 × 4 mm (section thickness × interval), were viewed in stack mode, whereas the sliding-slab projection images were interpreted with data sets from 2 × 1 mm reconstructions, which could be modified in both slab thickness and viewing angle (from transverse to any other viewing orientation). Use of the thinner-section sliding-slab technique resulted in a trend toward improved diagnostic accuracy in receiver operating characteristic analysis; significantly increased pooled sensitivity and specificity, diagnostic confidence, and visualization of the appendix; and a significant decrease in the rate of indeterminate findings. Various factors potentially contributed to the improved diagnostic capability with the sliding-slab technique in that study, including thinner reconstruction sections, use of average intensity projection, the sliding-slab method of review, and the option of multiplanar viewing [18]. The purpose of our study was to independently analyze the role of reconstruction section thickness on diagnostic performance and confidence in interpretation of axial images acquired with a 64-MDCT scanner.
Materials and Methods
Subjects
Our institutional review board granted approval for this HIPAA-compliant study and waived the requirement for informed consent. Over an approximately 1-year period, CT data sets from 64-MDCT examinations performed in our emergency department for clinical suspicion of appendicitis were collected and prospectively reconstructed with three protocols. At a later time, each set of reconstructions was retrospectively reviewed and correlated with medical records. Inclusion criteria were age 18 years or older, use of the standard 64-MDCT appendiceal CT acquisition parameters of 1.2-mm detector thickness, and administration of IV contrast material. Although most examinations also were performed with oral contrast material, oral contrast administration was not an absolute requirement for the case to be included. Between June 2006 and June 2007, 250 data sets from 247 patients met the inclusion criteria. Excluded from the 250 were one case with substantial motion artifact, 30 cases in which a complete set of three reconstructions was not available, three cases that were repeat CT examinations in three separate patients, three cases in which the medical records disclosed previous report of appendectomy or right hemicolectomy, and one case in which the patient left the emergency department against medical advice. A total of 212 patients composed the study population of 138 women and 74 men (average age, 35.7 years; range, 18–75 years).
64-MDCT Technique
Our 64-MDCT appendix protocol (Somatom Sensation 64 scanner, Siemens Medical Solutions) includes nonfocused coverage (diaphragm to symphysis) with 24 × 1.2 mm detector configuration, 200 effective mAs, and 120 kVp. For oral contrast enhancement (used in 98% of the cases, 208 of 212), we routinely administer 750–1,000 mL of diluted oral iohexol (Omnipaque 350, GE Healthcare). Approximately 120 mL of IV iohexol is administered at a rate of 2–3 mL/s, and acquisition is initiated 50–60 seconds after start of IV contrast infusion. For the purposes of this investigation, MDCT images were reconstructed with three techniques: 5 × 5 mm, 3 × 3 mm, and 2 × 1 mm (reconstruction section thickness × interval). CT data sets were transferred to a PACS workstation for retrospective review.
Medical Record Review
Two investigators who did not perform the retrospective MDCT readings evaluated the electronic records. Medical record review was conducted more than 6 months after the CT (range, 7.5–19.5 months later). Discharge summaries, operative notes, pathology reports, or a combination of these records were evaluated to identify subjects with appendicitis. Appendicitis was defined as pathologically proven appendicitis or appendiceal perforation in a patient who underwent surgery or perforated appendicitis either drained percutaneously or managed with IV antibiotics. Because these were emergency department patients, records were limited to emergency department charts for many subjects. However, any existing electronic records from visits subsequent to CT were reviewed to determine whether a patient given an alternative or normal diagnosis returned to our institution with appendicitis. In addition, surgical and medical records indicating previous appendectomy or right hemicolectomy were noted.
Retrospective CT Review
This study was modeled after the work of Paulson et al. [9]. The outcome measures were visualization of the appendix, visualization confidence, assessment for predefined CT findings of appendicitis, confidence in presence or absence of CT findings, diagnostic accuracy for appendicitis, and diagnostic confidence.
Two radiologists, each with 10 years of experience as an attending radiologist in body CT, blinded to final diagnosis, independently reviewed the MDCT images using a client–server system (Syngo WebSpace, Siemens Medical Solutions) on a standard Intel PC platform. They viewed axial sections with a scroll technique. Each of the 212 sets of images was grouped according to reconstruction technique and reviewed at separate times (minimum interval, approximately 2 weeks) in the following order: 5 × 5 mm sections, 3 × 3 mm sections, and 2 × 1 mm sections. Readers were not blinded to reconstruction technique, which was apparent from the number of images in the data set and was indicated on the screen.
A data collection sheet was used to record the following: whether the appendix was visualized—no, possibly, probably, or definitely; appendiceal caliber—normal versus enlarged or pathologically distended (for abnormal cases, readers recorded the outer-to-outer wall diameter in millimeters; given the readers' experience, this assessment was subjective with no predefined diameter cutoff); presence or absence of wall thickening, appendicolith, periappendiceal stranding, periappendiceal fluid, and abscess; and final diagnosis—appendicitis or no appendicitis. For identification of findings and the final diagnosis, confidence levels were recorded on a scale of 1–3, 3 being high confidence.
Statistical Analysis
Logistic regression was used to analyze the relation between binary outcomes (e.g., accuracy, presence of a finding) and section thickness. Ordinal logistic regression was used to analyze the relation between confidence ratings and section thickness. Both types of analyses were adjusted for differences between readers and statistical correlation associated with multiple observations per subject. Statistical analysis was performed with the Stata software package (version 10, Stata). To simplify presentation, the results from both readers were pooled in the data tables.
Results
Diagnosis
Seventeen of the 212 patients (8%) had appendicitis (five women, 12 men; mean age, 34 years; range, 19–70 years). Sixteen of the 17 patients had a surgical diagnosis of appendicitis supported by pathologic findings (Table 1). One of these patients had a perforated appendix due to a malignant tumor. The patient who did not have a surgical diagnosis had perforated appendicitis that was managed with IV antibiotics. One patient had an equivocal final diagnosis; two CT studies were suspicious for appendicitis, but the clinical impression was gastritis. The patient was advised to undergo elective appendectomy, which had not been performed as of this writing. This case was excluded from accuracy assessment. In none of the three patients who underwent repeated CT examinations was appendicitis proved at follow-up. Medical records disclosed no other evidence of delayed diagnosis in a patient whose condition was not diagnosed as appendicitis at initial presentation.
VIEW TABLE 1: Surgical and Pathologic Diagnoses in 16 of 17 Patients with Abnormal Findings
Appendiceal Visualization and Confidence According to Technique
Table 2 summarizes the number of cases in which the appendix was visualized and the confidence scores according to reconstruction technique. The rate of appendiceal visualization increased with decreasing section thickness, and a statistically significant association was found between visualization of the appendix and section thickness (p < 0.001 for 5 × 5 vs 3 × 3; p = 0.03 for 3 × 3 vs 2 × 1). When the appendix was visualized, there was a highly significant association between visualization confidence level and section thickness, use of thinner sections improving confidence (p < 0.001 for 5 × 5 vs 3 × 3 and 3 × 3 vs 2 × 1).
VIEW TABLE 2: Rate of Appendiceal Visualization and Visualization Confidence Levels According to Reconstruction Technique
Findings of Appendicitis According to Technique
The frequency of individual true-positive findings is reported in Table 3. There was no statistically significant relation between presence or absence of each of the CT findings and section thickness in any of the correctly diagnosed cases (p = 0.13–0.9), in any of the correctly diagnosed abnormal cases (p = 0.13–0.92), or in any of the correctly diagnosed normal cases (p = 0.07–0.8). Excluded from the analysis of abnormal cases were caliber and wall thickening because these findings had a high frequency (93–96%) across section thicknesses. Normal cases were not compared for the presence of abscess because abscess was not found in any of the normal cases.
VIEW TABLE 3: Frequency of Identification of Individual CT Findings in Cases of True-Positive Diagnoses
In correctly diagnosed cases, there were statistically significant associations between confidence for the presence or absence of each finding and section thickness. Higher confidence was associated with use of thinner sections. One exception was the lack of significant difference between 3 × 3 and 2 × 1 techniques for confidence related to identification of appendicolith. The p values were as follows: caliber confidence (p = 0.006 for 5 × 5 vs 3 × 3; p < 0.001 for 3 × 3 vs 2 × 1), wall confidence (p < 0.001 for 5 × 5 vs 3 × 3 and 3 × 3 vs 2 × 1), appendicolith confidence (p = 0.001 for 5 × 5 vs 3 × 3; p = 0.09 for 3 × 3 vs 2 × 1), periappendiceal stranding confidence (p < 0.001 for 5 × 5 vs 3 × 3 and 3 × 3 vs 2 × 1), periappendiceal fluid confidence (p = 0.008 for 5 × 5 vs 3 × 3; p = 0.001 for 3 × 3 vs 2 × 1), abscess confidence (p < 0.001 for 5 × 5 vs 3 × 3; p = 0.002 for 3 × 3 vs 2 × 1).
 | Fig. 1A —21-year-old man with acute abdominal pain, nausea, and vomiting. Pathologic diagnosis was acute appendicitis with microabscess formation and serositis. Interpretation was difficult owing to paucity of intraabdominal fat. CT finding was no appendicitis in five of six readings. Reader 2 did not visualize appendix with 5 × 5 mm or 3 × 3 mm reconstructions and misdiagnosed findings as normal. CT reconstructions at 2 × 1 mm show enlarged (8 mm) appendix (arrow, A) with wall thickening identified by reader 2, who wrote additional comments about lack of oral contrast opacification and presence of wall enhancement (arrowheads, B) and diagnosed appendicitis with low level of confidence. |
 | Fig. 1B —21-year-old man with acute abdominal pain, nausea, and vomiting. Pathologic diagnosis was acute appendicitis with microabscess formation and serositis. Interpretation was difficult owing to paucity of intraabdominal fat. CT finding was no appendicitis in five of six readings. Reader 2 did not visualize appendix with 5 × 5 mm or 3 × 3 mm reconstructions and misdiagnosed findings as normal. CT reconstructions at 2 × 1 mm show enlarged (8 mm) appendix (arrow, A) with wall thickening identified by reader 2, who wrote additional comments about lack of oral contrast opacification and presence of wall enhancement (arrowheads, B) and diagnosed appendicitis with low level of confidence. |
MDCT Diagnostic Accuracy and Confidence According to Technique
The readers' diagnostic performances according to reconstruction technique are summarized in Table 4. Mean sensitivities were 79.4%, 82.4%, and 82.4% for 5 × 5, 3 × 3, and 2 × 1; specificities were 99.2%, 98.7%, and 98.2% for 5 × 5, 3 × 3, and 2 × 1. Correctness (percentage correct, sensitivity, specificity) was not statistically significantly associated with section thickness (p = 0.42–1.0) With respect to missed diagnoses, both reviewers, using more than one reconstruction technique, incorrectly categorized the same four cases as normal (Figs. 1A, 1B, 2A, 2B, 2C, 2D, 2E, 2F, 3, 4A, and 4B).
VIEW TABLE 4: Diagnostic Performance According to Reconstruction Technique (n = 211)
In correctly diagnosed cases, there was a statistically significant association between impression confidence level and section thickness (p < 0.001 for 5 × 5 vs 3 × 3 and 3 × 3 vs 2 × 1). Higher confidence levels were associated with use of thinner sections (Table 5). Subgroup analysis of abnormal and normal cases revealed that the association was significant for normal cases (p < 0.001 for 5 × 5 vs 3 × 3 and 3 × 3 vs 2 × 1) but not for abnormal cases (p = 0.6 for 5 × 5 vs 3 × 3 and 3 × 3 vs 2 × 1).
TABLE 5: Diagnostic Confidence in Correctly Diagnosed Cases According to Reconstruction Technique
Discussion
The increase in the use of CT for suspected appendicitis [1, 19] results from a number of factors, including increased availability, many emergency departments now housing a dedicated MDCT scanner; improved resolution due to advancing CT technology [9, 20]; growing radiologist experience [1, 20]; and data that suggest lower negative appendectomy rates after the introduction of helical CT [21]. One measurable advantage of advancing technology pertaining to radiologic interpretation relates to definitive diagnosis. Daly et al. [20] compared the rate of equivocal CT interpretations between 1998 and 2002 and found a decline from 23.3% to 9.5%. This reduction was ascribed to growing radiologist experience in conjunction with the narrow reconstruction sections afforded by MDCT technology. The value of the increased resolution is made apparent by the results of a study in which 132 patients with suspected appendicitis underwent 16-MDCT with indeterminate findings in 0.8% of the cases [22].
Contributing to improved diagnostic confidence is MPR, made feasible by the spatial resolution of 16- and 64-MDCT. Paulson et al. [9], using isotropic data sets obtained with 16 × 0.625 detector configuration in CT of 100 subjects, found that supplementation of transverse sections with coronal reconstructions significantly improved reader agreement, confidence in visualization of the appendix, confidence in identification or exclusion of individual CT findings, and diagnostic confidence. In other work [10], supplementation of axial sections with coronal reconstructions significantly increased radiologists' confidence in visualization of the appendix in true-positive cases. That study, conducted with IV contrast-enhanced 16 × 1.5 mm MDCT on 110 subjects, also revealed that combining coronal reconstructions with axial sections resulted in a trend toward improved sensitivity and diagnostic performance.
 | Fig. 2A —55-year-old man with right lower quadrant pain. Pathologic diagnosis was appendicitis with periappendicitis. CT findings were interpreted as no appendicitis in five of six readings. (A and B, 2 × 1 mm; C and D, 3 × 3 mm; E and F, 5 × 5 mm). CT reconstructions at 2 × 1 mm (A and B) and 3 × 3 mm (C and D) show appendix (arrow); however, appendicitis was not diagnosed by either reader using these reconstructions. Arrowhead indicates collapsed small-bowel loop adjacent to appendix. |
 | Fig. 2B —55-year-old man with right lower quadrant pain. Pathologic diagnosis was appendicitis with periappendicitis. CT findings were interpreted as no appendicitis in five of six readings. (A and B, 2 × 1 mm; C and D, 3 × 3 mm; E and F, 5 × 5 mm). CT reconstructions at 2 × 1 mm (A and B) and 3 × 3 mm (C and D) show appendix (arrow); however, appendicitis was not diagnosed by either reader using these reconstructions. Arrowhead indicates collapsed small-bowel loop adjacent to appendix. |
 | Fig. 2C —55-year-old man with right lower quadrant pain. Pathologic diagnosis was appendicitis with periappendicitis. CT findings were interpreted as no appendicitis in five of six readings. (A and B, 2 × 1 mm; C and D, 3 × 3 mm; E and F, 5 × 5 mm). CT reconstructions at 2 × 1 mm (A and B) and 3 × 3 mm (C and D) show appendix (arrow); however, appendicitis was not diagnosed by either reader using these reconstructions. Arrowhead indicates collapsed small-bowel loop adjacent to appendix. |
 | Fig. 2D —55-year-old man with right lower quadrant pain. Pathologic diagnosis was appendicitis with periappendicitis. CT findings were interpreted as no appendicitis in five of six readings. (A and B, 2 × 1 mm; C and D, 3 × 3 mm; E and F, 5 × 5 mm). CT reconstructions at 2 × 1 mm (A and B) and 3 × 3 mm (C and D) show appendix (arrow); however, appendicitis was not diagnosed by either reader using these reconstructions. Arrowhead indicates collapsed small-bowel loop adjacent to appendix. |
 | Fig. 2E —55-year-old man with right lower quadrant pain. Pathologic diagnosis was appendicitis with periappendicitis. CT findings were interpreted as no appendicitis in five of six readings. (A and B, 2 × 1 mm; C and D, 3 × 3 mm; E and F, 5 × 5 mm). CT reconstructions at 2 × 1 mm (A and B) and 3 × 3 mm (C and D) show appendix (arrow); however, appendicitis was not diagnosed by either reader using these reconstructions. Arrowhead indicates collapsed small-bowel loop adjacent to appendix. |
 | Fig. 2F —55-year-old man with right lower quadrant pain. Pathologic diagnosis was appendicitis with periappendicitis. CT findings were interpreted as no appendicitis in five of six readings. (A and B, 2 × 1 mm; C and D, 3 × 3 mm; E and F, 5 × 5 mm). CT reconstructions at 5 × 5 mm, only data set correctly interpreted as appendicitis with a high confidence level by reader 2. Appendix (arrow) is enlarged (16 mm) with wall thickening. Arrowhead indicates collapsed small-bowel loop adjacent to appendix. |
 | Fig. 3 —20-year-old man with right lower quadrant pain and guarding. Pathologic diagnosis was appendix with chronic inflammation. With exception of reader 2, who noted wall thickening on 3 × 3 mm reconstructions, appendix was interpreted as normal in appearance in all six readings with all CT diagnoses of no appendicitis. Axial CT image from 2 × 1 mm reconstructions shows air-distended appendix (arrows). |
Beyond investigations showing the utility of MPR [9–12], few published data exist to direct 16-MDCT and 64-MDCT acquisition and reconstruction parameters for suspected appendicitis. Accordingly, some components of protocol design continue to be guided by research conducted on single-detector helical CT. With respect to section thickness, Weltman et al. [8] revealed the importance of using 5-mm collimation and reconstruction sections for interpretation of axial images obtained with single-detector helical CT. Compared with the use of 10-mm collimation, the use of 5-mm collimation significantly improved visualization of normal appendices, identification of abnormal appendices and periappendiceal inflammation, diagnostic confidence, sensitivity, and diagnostic accuracy for appendicitis. As a result, 5-mm collimation became widely used for evaluating the appendix with single-detector helical CT.
 | Fig. 4A —33-year-old woman with right lower quadrant pain. Pathologic diagnosis was acute transmural appendicitis with serositis. CT findings were interpreted as no appendicitis in three of six readings. Reader 1 made correct diagnosis with low level of confidence using only 3 × 3 mm reconstructions after identifying enlarged (10 mm) appendix with wall thickening and periappendiceal inflammation. Reader 2 misinterpreted 5 × 5 mm reconstructions as nonvisualization with no appendicitis. CT reconstructions using 2 × 1 mm section × interval show enlarged (9 mm) appendix with wall thickening (arrow) identified by reader 2, who made the correct diagnosis using both 3 × 3 mm and 2 × 1 mm reconstructions. |
 | Fig. 4B —33-year-old woman with right lower quadrant pain. Pathologic diagnosis was acute transmural appendicitis with serositis. CT findings were interpreted as no appendicitis in three of six readings. Reader 1 made correct diagnosis with low level of confidence using only 3 × 3 mm reconstructions after identifying enlarged (10 mm) appendix with wall thickening and periappendiceal inflammation. Reader 2 misinterpreted 5 × 5 mm reconstructions as nonvisualization with no appendicitis. CT reconstructions using 2 × 1 mm section × interval show enlarged (9 mm) appendix with wall thickening (arrow) identified by reader 2, who made the correct diagnosis using both 3 × 3 mm and 2 × 1 mm reconstructions. |
After the introduction of 4-MDCT in 1998, scanning the abdomen and pelvis with effective detector width and reconstruction section thickness less than 5 mm became a reality, but optimal parameters were not definitively elucidated for appendiceal imaging. Increased resolution resulting from narrow detector configuration (4 × 1 mm vs 4 × 5 mm) was constrained by an increase in radiation dose to the patient. In the absence of data showing measurable improvement in diagnostic performance or outcome with narrow detectors, evidence to support the use of the thinnest detectors with 4-MDCT was lacking.
With advancing MDCT technology, detector section thickness, which dictates minimum reconstruction section thickness, has decreased concomitantly with increases in z-axis resolution and radiation dose. The range of detector configurations becomes more limited with later-generation models. For example, three 4-MDCT scanners have three or four detector configurations (4 × 1.0 or 4 × 1.25 mm, 4 × 2.5 mm, 4 × 3.75 mm, and 4 × 5 mm) [23, 24], but only two configurations are available on 16-MDCT scanners from the same three manufacturers (16 × 0.625 mm and 16 × 1.25 mm or 16 × 0.75 mm and 16 × 1.5 mm) [25]. For 64-MDCT, the smallest detectors (all detectors in matrix array or centrally located detectors in hybrid array) are less than 1 mm in effective thickness. These detectors can be used individually (0.5, 0.6, or 0.625 mm) or in pairs (1.0, 1.2, or 1.25 mm) [24, 26]. Accordingly, there is little variability in detector sections with 64-MDCT. Although Dalrymple et al. [24] pointed out that use of the narrowest 64-MDCT collimation does not increase radiation dose when exposure parameters are held constant, the increased noise from the use of narrow detectors often necessitates use of a higher tube current–time product setting to obtain an adequate signal-to-noise ratio [26]. For this reason, we use 1.2-mm detectors, which are adequate for generating MPRs, as shown by several publications on appendiceal CT [6, 10, 13].
Unlike the limited selection of detector widths for data acquisition, a wide variety of reconstruction section thicknesses are available with 64-MDCT. For example, numerous options between 0.6 or 1.5 mm (depending on the detector width selected) and 10 mm are available with the hybrid-array model [27].
Data set resolution is improved by the use of narrow reconstruction sections and overlapping reconstruction intervals. Several studies of MDCT have shown advantages to using narrow reconstruction sections (1–3 mm), in comparison with sections ranging from 3 to 6 mm, for other applications dependent on high resolution [17, 28–32]. In separate investigations, thinner sections have been found to improve diagnostic efficacy in the identification of segmental and subsegmental pulmonary emboli [17, 31], pulmonary nodules [28], renal calculi [29], and endoleaks after endovascular stent placement [30]. These results combined with the findings made by Lee et al. [18], which showed improved pooled sensitivity and specificity with a sliding-slab display generated from 2 × 1 mm reconstructions, support our data suggesting a role for narrow reconstructions in the interpretation of axial images of patients with suspected appendicitis.
The compromises in selecting narrow reconstruction sections are an increase in image noise and an inverse relation to the number of images, which are disadvantageous to workflow. For example, a comparison of reconstruction section thicknesses for MDCT with 16 × 0.75 mm detector configuration in the evaluation of pulmonary embolism showed the number of images generated with various reconstruction parameters. Compared with 6 × 5 mm section × interval, use of 4 × 3 mm reconstructions yielded an average of 1.7 times as many axial images per case; 2 × 1.5 mm reconstructions generated 3.3 times as many images per case; and 0.75 × 0.4 mm reconstructions resulted in 12.8 times as many images per case [17]. For abdominal CT with the reconstruction techniques compared in our study, 450 mm of coverage would result in 90 images at 5 × 5 mm, 150 images at 3 × 3 mm, and 450 images at 2 × 1 mm.
Our results nonetheless revealed that use of increasingly narrow reconstructions less than 5 mm significantly increased the rate of visualization of the appendix, confidence in visualization of the appendix, confidence in identification or exclusion of individual CT findings, and diagnostic confidence in exclusion of appendicitis. Ability to confidently exclude the diagnosis becomes increasingly important in the setting of increased use of CT because the percentage of patients undergoing imaging who actually have appendicitis decreases as a consequence of a declining threshold for imaging.
We did not identify any effect on diagnostic accuracy associated with reconstruction section thickness, owing in part to the small number of subjects with appendicitis, a limitation of this study. However, investigations of the utility of MPR that included a higher percentage of abnormal cases (24–42%) have had similar results [9, 10]. Studies by Paulson et al. [9] and Lee et al. [10] showed no significant differences in sensitivity, specificity [9, 10], or results of receiver operating characteristic analysis [10] (with a trend toward increased Az values and improved sensitivity for radiologists in the study by Lee et al.) but showed that use of MPR significantly increased diagnostic confidence and appendiceal visualization [9, 10]. Whether use of thinner reconstructions can positively affect diagnostic accuracy will require investigations that include a larger number of patients with appendicitis.
The limitation of our investigation that the percentage of abnormal cases was small reflects the increasing use of CT. This is within the range of incidences reported in single-institution studies of CT. In a population imaged between 2002 and 2004, 4% (21 of 525) of patients had appendicitis [33], and among 400 patients imaged in 2003, 20% (80 of 400) had appendicitis [34]. Use of CT can lead to alternative diagnoses in some patients who do not have appendicitis (12–46% in series described between 2005 and 2008) [22, 33, 34]. Because they expedite decisions about abnormal, normal, and alternative diagnoses, CT findings facilitate the determination of whether hospitalization and surgery are warranted, reduce the rate of unnecessary admissions, and can alter surgical management decisions [35]. The risk of ionizing radiation, however, warrants development of reduced radiation dose protocols and algorithms for decreasing the percentage of true-negative findings at CT, particularly in light of the mean age (fourth decade) of patients undergoing imaging for suspected appendicitis. Petrosyan et al. [36] reviewed the records of 1,630 patients with suspected appendicitis, 905 of whom were imaged with CT. The use of CT increased from 2005 to 2006, and record review revealed that CT was ordered for patients for whom clinical suspicion of the diagnosis was low, intermediate, and high. Identifying improvement in the rate of negative appendectomy associated with CT only among patients with intermediate clinical suspicion (Alvarado score, 5–7), the authors recommended limiting CT to patients in that clinical category. Alternative triage protocols to avert CT of some patients include use of sonography as the initial imaging tool, which was associated with a 29% rate of appendicitis in patients who underwent CT after indeterminate sonographic findings or discrepancy between sonographic findings and clinical assessment [22].
This study had several additional limitations. The retrospective collection of clinical data from medical record review is less accurate than prospective patient interview and follow-up in confirming that no patient assigned a normal diagnosis went to another hospital with a delayed diagnosis (i.e., potential underestimation of false-negative diagnosis) and that all subjects had an appendix. In addition, some patients may have had appendicitis that resolved [36, 37], accounting for their lack of return visit (i.e., potential overestimation of false-positive diagnoses or underestimation of false-negative diagnoses). The readers' knowledge of section thickness might have introduced bias in favor of thinner sections; however, this phenomenon was unavoidable because experienced readers can deduce section thickness on the basis of the number of slices. Interpretation of axial sections in isolation is not consistent with current practice; MPRs are used routinely and for challenging cases. However, because the standard of care for appendiceal CT includes interpretation of axial sections, we sought to investigate axial reconstructions independent of MPRs.
Interpretative performance relies on high-quality MDCT studies for convincing identification of the appendix and findings associated with appendicitis. Some investigations have shown that lack of visualization of an appendix when CT findings are otherwise normal is associated with a very low rate of appendicitis [34, 38]. Nonetheless, confidence in excluding the diagnosis is bolstered by definitive visualization of a normal appendix [9, 13]. In this study, use of narrow reconstruction sections improved radiologists' capability to confidently visualize the appendix and increased diagnostic confidence in exclusion of appendicitis on axial sections, although reconstruction section thickness was not associated with diagnostic accuracy. Our results and conclusions pertain to the specific 64-MDCT acquisition and display parameters used in this investigation, including 24 × 1.2 mm detector configuration, IV and oral contrast enhancement, and interpretation of axial sections. Of the three reconstruction parameters compared, appendiceal visualization and interpretative confidence were optimized with use of 2 × 1 mm sections. Although we did not evaluate interpretation with MPRs, these reconstruction parameters will support interactive multiplanar volume interrogation, either routinely or in problematic cases, according to individual practice. If workstation and network constraints limit a department's capacity to manage 2 × 1 mm sections, the study results indicate that the thinnest feasible sections (e.g., 3 × 3 mm rather than 5 × 5 mm) should be used for the best possible interpretative confidence.
References
- Raman SS, Osuagwu FC, Kadell B, Cryer H, Sayre J, Lu DS. Effect of CT on false positive diagnosis of appendicitis and perforation. N Engl J Med 2008; 358:972 -973
Medline - Hershko DD, Awad N, Fischer D, et al. Focused helical CT using rectal contrast material only as the preferred technique for the diagnosis of suspected acute appendicitis: a prospective, randomized, controlled study comparing three different techniques. Dis Colon Rectum2007 ; 50:1223 -1229
Medline - Hebert JJ, Taylor AJ, Winter TC. Comparison of colonic transit between polyethylene glycol and water as oral contrast vehicles in the CT evaluation of acute appendicitis. AJR2006 ; 187:1188 -1191
Abstract - Tamburrini S, Brunetti A, Brown M, Sirlin C, Casola G. Acute appendicitis: diagnostic value of nonenhanced CT with selective use of contrast in routine clinical settings. Eur Radiol2007 ; 17:2055 -2061
Medline - Jacobs JE, Birnbaum BA, Macari M, et al. Acute appendicitis: comparison of helical CT diagnosis focused technique with oral contrast material versus nonfocused technique with oral and intravenous contrast material. Radiology 2001;220 : 683-690
Medline - Jan YT, Yang FS, Huang JK. Visualization rate and pattern of normal appendix on multidetector computed tomography by using multiplanar reformation display. J Comput Assist Tomogr 2005;29 : 446-451
Medline - Keyzer C, Pargov S, Tack D, et al. Normal appendix in adults: reproducibility of detection with unenhanced and contrast-enhanced MDCT. AJR 2008; 191:507 -514
Abstract - Weltman DI, Yu J, Krumenacker J Jr, Huang S, Moh P. Diagnosis of acute appendicitis: comparison of 5- and 10-mm CT sections in the same patient. Radiology 2000;216 : 172-177
Medline - Paulson EK, Harris JP, Jaffe TA, Haugan PA, Nelson RC. Acute appendicitis: added diagnostic value of coronal reformations from isotropic voxels at multi-detector row CT. Radiology2005 ; 235:879 -885
Medline - Lee KH, Kim YH, Hahn S, et al. Added value of coronal reformations for duty radiologists and for referring physicians or surgeons in the CT diagnosis of acute appendicitis. Korean J Radiol2006 ; 7:87 -96
Medline - Yaghmai V, Brandwein WM, Hammond N, Nikolaidis P. MDCT diagnosis of appendicitis using only coronal reformations. Emerg Radiol 2007; 14:167 -172
Medline - Sebastian S, Kalra MK, Mittal P, Saini S, Small WC. Can independent coronal multiplanar reformatted images obtained using state-of-the-art MDCT scanners be used for primary interpretation of MDCT of the abdomen and pelvis? A feasibility study. Eur J Radiol 2007;64 : 439-446
Medline - Kim HC, Yang DM, Jin W, Park SJ. Added diagnostic value of multiplanar reformation of multidetector CT data in patients with suspected appendicitis. RadioGraphics 2008;28 : 393-405
Medline - Johnson PT, Horton KM, Mahesh M, Fishman EK. Multidetector computed tomography for suspected appendicitis: multi-institutional survey of 16-MDCT data acquisition protocols and review of pertinent literature. J Comput Assist Tomogr 2006;30 : 758-764
Medline - Johnson PT, Eng J, Moore CJ, Horton KM, Fishman EK. Multidetector-row CT of the appendix in healthy adults. Emerg Radiol 2006; 12:248 -253
Medline - Tamburrini S, Brunetti A, Brown M, Sirlin CB, Casola G. CT appearance of the normal appendix in adults. Eur Radiol 2005; 15:2096 -2103
Medline - Heuschmid M, Mann C, Luz O, et al. Detection of pulmonary embolism using 16-slice multidetector-row computed tomography: evaluation of different image reconstruction parameters. J Comput Assist Tomogr 2006; 30:77 -82
Medline - Lee KH, Kim YH, Hahn S, et al. Computed tomography diagnosis of acute appendicitis: advantages of reviewing thin-section datasets using sliding slab average intensity projection technique. Invest Radiol 2006; 41:579 -585
Medline - Frei SP, Bond WF, Bazuro RK, Richardson DM, Sierzega GM, Reed JF. Appendicitis outcomes with increasing computed tomographic scanning. Am J Emerg Med 2008;26 : 39-44
Medline - Daly CP, Cohan RH, Francis IR, Caoili EM, Ellis JH, Nan B. Incidence of acute appendicitis in patients with equivocal CT findings. AJR 2005; 184:1813 -1820
Abstract - Rao PM, Rhea JT, Rattner DW, Venus LG, Novelline RA. Introduction of appendiceal CT: impact on negative appendectomy and appendiceal perforation rates. Ann Surg 1999;229 : 344-349
Medline - Gaitini D, Beck-Razi N, Mor-Yosef D, et al. Diagnosing acute appendicitis in adults: accuracy of color Doppler sonography and MDCT compared with surgery and clinical follow-up. AJR2008 ; 190:1300 -1306
Abstract - Flohr TG, Schaller S, Stierstorfer K, Bruder H, Ohnesorge BM, Schoepf UJ. Multi-detector row CT systems and image-reconstruction techniques. Radiology 2005;235 : 756-773
Medline - Dalrymple NC, Prasad SR, El-Merhi FM, Chintapalli KN. Price of isotropy in multidetector CT. RadioGraphics2007 ; 27:49 -62
Medline - Dalrymple NC, Prasad SR, Freckleton MW, Chintapalli KN. Informatics in radiology (info-RAD): introduction to the language of three-dimensional imaging with multidetector CT. RadioGraphics2005 ; 25:1409 -1428
Medline - Prokop M. MDCT imaging: new challenges for scan and contrast optimization. Eur Radiol Suppl 2005;15 [suppl 5]:E35 -E45
CrossRef - Flohr TG, Stierstorfer K, Ulzheimer S, Bruder H, Primak AN, McCollough CH. Image reconstruction and image quality evaluation for a 64-slice CT scanner with z-flying focal spot. Med Phys2005 ; 32:2536 -2547
Medline - Fischbach F, Knollmann F, Griesshaber V, Freund T, Akkol E, Felix R. Detection of pulmonary nodules by multislice computed tomography: improved detection rate with reduced slice thickness. Eur Radiol 2003; 13:2378 -2383
Medline - Memarsadeghi M, Heinz-Peer G, Helbich TH, et al. Unenhanced multi-detector row CT in patients suspected of having urinary stone disease: effect of section width on diagnosis. Radiology2005 ; 235:530 -536
Medline - Iezzi R, Cotroneo AR, Filippone A, Di Fabio F, Santoro M, Storto ML. MDCT angiography in abdominal aortic aneurysm treated with endovascular repair: diagnostic impact of slice thickness on detection of endoleaks. AJR 2007; 189:1414 -1420
Abstract - Schoepf UJ, Holzknecht N, Helmberger TK, et al. Subsegmental pulmonary emboli: improved detection with thin-collimation multi-detector row spiral CT. Radiology 2002;222 : 483-490
Medline - Parrish FJ. Volume CT: state-of-the-art reporting. AJR 2007; 189:528 -534
Abstract - Giuliano V, Giuliano C, Pinto F, Scaglione M. CT method for visualization of the appendix using a fixed oral dosage of diatrizoate: clinical experience in 525 cases. Emerg Radiol2005 ; 11:281 -285
Medline - Ganguli S, Raptopoulos V, Komlos F, Siewert B, Kruskal JB. Right lower quadrant pain: value of the nonvisualized appendix in patients at multidetector CT. Radiology 2006;241 : 175-180
Medline - Rosen MP, Siewart B, Sands DZ, Bromberg R, Edlow J, Raptopoulos V. Value of abdominal CT in emergency department patients with abdominal pain. Eur Radiol 2003;13 : 418-424
Medline - Petrosyan M, Estrada J, Chan S, et al. CT scan in patients with suspected appendicitis: clinical implications for the acute care surgeon. Eur Surg Res 2008;40 : 211-219
Medline - Kirshenbaum M, Mishra V, Kuo D, Kaplan G. Resolving appendicitis: role of CT. Abdom Imaging 2003;28 : 276-279
Medline - Nikolaidis P, Hwang CM, Miller FH, Papanicolaou N. The nonvisualized appendix: incidence of acute appendicitis when secondary inflammatory changes are absent. AJR2004 ; 183:889 -892
Abstract