cock orspsaten Diagnostic Accuracy of Noncontrast CT in Detecting Acute Appendicitis: A Meta-analysis of Prospective Studies ING XIONG, M.D, BAISHU ZHONG, M.D,* ZHENWEL LI, M.D,* FENG ZHOU, M.D, RUYING HU, M.D.t ZHAN FENG, M.D," SHUNLIANG XU, M.D," FENG CHEN, M.D* From *Department of Radiology, The First Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, Zhejiang Province, China; tDepartment of Gastrointestinal Surgery, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou, of Surgical Oncology, University, Hangzhou, Zhejiang Province, China ijiang Province, China and +Department The Second Affiliated Hospital of Zhejiang Chinese Medical The aim of the study is to evaluate the diagnostic accuracy of noncontrast CT in detecting acute appendicitis. Prospective studies in which noncontrast CT was performed to evaluate acute appendicitis were found on PubMed, EMBASE, and Cochrane Library. Pooled sensitivity, speci- ficity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio were assessed. ‘The summary receiver-operating characteristic curve was conducted and the area under the curve was calculated. Seven original studies investigating a total of 845 patients were included in thi meta-analysis. The pooled sensitivity and specificity were 0.90 (95% Cl: 0.86-0.92) and 0.94 ( Ci: 0.92-0.97), respectively. The pooled positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio was 12.90 (95% Cl: 4.80-34.67), 0.09 (85% Cl: 0.04-0.20), and 162.76 (95% ‘CI: 31.05-853.26), respectively. The summary receiver-operating characteristic curve was symmetrical and the area under the curve was 0.97 (95% Cl: 0,95-0.98). In conclusion, noncontzast CT has high diagnostic accuracy in detecting acute appendicitis, which is adequate for clinical decision making. Ay Siz strmrems sone ofthe most common sete abdomen in the emergency department. Once the diagnosis is clear, appendectomy is required immedi- ately for most patients with acute appendicitis.’ How- ever, the accurate diagnosis of acute appendicitis often remains difficult, because clinical features are atypical and physical findings are not specific in some patients.” In clinical practice, some patients without acute appendicitis are removed the normal appendix as a re- sult of erroneous diagnosis, whereas some patients with acute appendicitis diagnosis lose the opportunity to receive early appendectomy as a result of missed diagnosis.” The negative appendectomy is associated with some complications postoperatively, and the missed appendicitis will result in perforation followed by peri- tonitis making the treatment more complicated-* ° Therefore, methods to diagnose acute appendicitis quickly and accurately before surgery ate popular pur- suits in modem surgery. Address corespondence and reprint requests to Feng Chea, MD. Department of Raskology, the Fst Aiiated Hospital of Medical School of Zhejiang Universiy, 79 Qingchun Road, Hangzhou 310003, Zhejiang Province, China, E-mail: fengchenxb@ 163 com, Carrently, the diagnosis of appendicitis is established by a combination of physical findings, laboratory results, and imaging tests such as ultrasonic and radiographic features. Accumulating evidence demonstrates that non- contrast CT scanning of abdomen has great advantage in the diagnosis of appendicitis.*~ It is simple to perform, casy to interpret, and quick to diagnose. However, the efficiency of noncontrast CT on the diagnosis of acute appendicitis still remains controversial This study was (0 provide the current high-level evidence to evaluate the diagnostic accuracy of non- contrast CT in detecting acute appendicitis. This meta- analysis was performed in accordance with the PRISMA statement on reporting preferred items for systematic reviews and meta-analyses.” Methods Search Strategy ‘The PubMed, EMBASE, and Cochrane Library databases were electronically searched for identified articles on evaluating the diagnostic accuracy of non- contrast CT in detecting acute appendicitis. Searched terms used were appendicitis, appendectomy, and CT. 626No. 6 ‘The references of included articles, relevant reviews, relevant systematic reviews, and meta-analyses were searched by hand as well for any other potential arti- cles. No language restriction was applied. Study Selection ‘We identified all relevant studies that evaluated the diagnostic accuracy of noncontrast CT in detecting acute appendicitis. Studies were included if they met the following two criterion: 1) the study design was prospective and 2) surgery (operative findings or his- tological findings after surgery) was regarded as the reference standard for diagnosing acute appendicitis; ‘a clinical follow-up for two weeks or more could also be used as a reference standard? ‘When duplicate publications that reported on the same cor parts of the same study population were identified, only the largest or the most detailed and informative article was included. All studies that were retrospective studies, review articles, case reports, or editorials were excluded, Studies that could not provide available data for extraction were also excluded. If any doubt of suit- ability existed after the abstract was reviewed, the full manuscript was obtained. Data Extraction Two review authors independently extracted data from eligible studies independently, without considera- tion ofthe results, Extracted data were then cross-checked between the two authors to resolve any discrepancy. ‘The data of test accuracy (true positive, false negative, true negative, false positive) were extracted to pool sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic odds ratio (DOR). The data of study characteristics (first author, year of publication, country, number of patients, reference standard) were also extracted. Quality Assessment ‘Two review authors critically assessed the method- ological quality of eligible studies independently. The quality was assessed by the Quality Assessment of Diagnostic Accuracy Studies tool.’ The tool consisted of 14 items and each item was graded as yes, no, or unclear. Any discrepancy was resolved by consensus discussions with the authors Statistical Analysis Al statistical analysis was carried out by Meta-DiSe Version 1.4 software (Unit of Clinical Biostatistics team of the Ramén y Cajal Hospital, Madrid, Spain). Pooled sensitivity, specificity, PLR, NLR, and DOR of non- contrast CT in detecting acute appendicitis were obtained DIAGNOSTIC ACCURACY OF NONCONTRAST CT Xiong et al 627 using a random-effects model. A summary receiver- operating characteristic (SROC) curve was conducted and the area under the curve (AUC) was calculated, assessing an overall diagnostic accuracy of the non- contrast CT. The precision for the characteristics of noncontrast CT test was presented with 95 per cent confidence intervals (95% CD). Threshold effect was assessed by calculating spearman correlation co- efficient between the logit of sensitivity and logit of (L-speciticity). Statistical heterogeneity was assessed by the x? test and F° > SO per cent was considered substantial heterogeneity.® Publication bias was eval- uated by Decks’ funnel plot using Stata/SE software (StataCorp, College Station, TX). A two-tailed P < 0.05 was determined statistically significant Results Search Results The initial search identified 1076 studies, of which 305 were excluded by duplicate checking, leaving 771 studies for preliminary screening. Based on the cligi- bility criteria, 749 studies were excluded for reviews, meta-analyses, systematic reviews, retrospective stud- ies, and other ineligible studies by title and abstract reading, leaving 22 studies for more detailed screening. Subsequently, 15 studies were excluded by full-text reading, including four retrospective studies, three studies without detailed data, three studies without ap- propriate reference standard, three studies including contrast CT, and two studies duplicately reporting on similar patient data. Finally, seven prospective studies were taken into account in this meta-analysis.®: 10-15 Characteristies of the Included Studies ‘The studies included in the meta-analysis were pub- lished between 1998 and 2010 and reported on a total of 345 patients, with a range of 49 to 296 patients per study. There were clear definitions of standard refer- ences for diagnosing acute appendicitis, including sur- gery or clinical follow up for two weeks or more. Methodological Quality of Included Studies ‘The scores of five out of seven included studies were over 10 rated with a ‘yes’, indicating that most of the included studies were of moderate quality. The weak- ness of most studies was the results of differential verification and the reference standard lacking blind- ing from the noncontrast CT test. ‘Meta-analysis Results The forest plot of sensitivity and specificity for noncontrast CT in detecting active appendicitis were628 tie) rman Fic. 1. Forrest plot demonstrating the sensitivity of noncontrast (CT in detecting active appendicitis, + Ries 58 bran ai00 So rtm =} ioe oe 1 (100 * | Sagat ages ta tatty een Fro. 2._ Forrest plot demonstrating the specificity of non- contrast CT in the diagnosis of active appendicitis, shown in Fig. 1 and 2. The sensitivity ranged from 0.76 to 0.97 (mean 0.90; 95% CI: 0.86-0.92), and specificity from 0.75 to 0.90 (mean 0.94; 95% CI: 0.92-0.97). The pooled data also showed that PLR was 12.90 (95% CI 4.80-34.67), NLR was 0.09 (95% CI: 0.04-020), and DOR was 162.76 (95% CI: 31.05-853.26). The x test showed that 2° values of sensitivity, specificity, PLR, NLR, and DOR were 84.0 per cent (P < 0.05), 75.0 per cent (P < 0.05), 73.5 per cent (P < 0,05), 86.0 per cent (P< 0.08), and 82.7 per cent (P < 0.05), respectively, all, suggesting that there was a significant heterogeneity among the studies. However, the SROC curve was sym- metrical and the AUC was 0.97 (95% CI: 0.95-0.99), suggesting that diagnostic accuracy of noncontrast CT was high in detecting active appendicitis. In addition, according to the SROC curve analysis,* there was no significant correlation between sensitivity and L-specificity (P = 0.18), indicating that there was no significant threshold effect in this meta-analysis. Publication Bias The funnel plot showed no statistically significant asymmetry (P > 0.05), suggesting that there was no ‘major publication bias in this meta-analysis. Discussion ‘The results of this meta-analysis suggest that non- contrast CT is reliable in detecting acute appendicitis with high sensitivity and good specificity. ‘THE AMERICAN SURGEON June 2015 Vol. 81 Acute appendicitis is 2 common surgical emer- gency in clinical practice. A quick and accuracy di- agnosis is important for the clinical decision making. When the appendicitis is diagnosed unclearly, it is, difficult to make the subsequent treatment’ 5 Therefore, selecting a simple, quick, and low-risk diagnostic tool appears to be very important in the emergency department. Noncontrast CT is a usual diagnostic technique in detecting various diseases. Previous studies showed noncontrast CT has a high value in diagnosing acute appendicitis.“ but the results were still controversial In this study, seven prospective studies were included, and a meta-analysis was conducted to evaluate the di- agnostic accuracy of noncontrast CT in detecting acute appendicitis. The results showed that the pooled sensi- tivity, specificity, PLR, NLR, and DOR were 0.90, 0.94, 12.90, 0.09 and 162.76, respectively, ‘The SROC was symmetrical and the AUC was 0.97. All these data suggest that noncontrast CT has high diagnostic accu- racy in detecting acute appendicitis, which is adequate for clinical decision making, In addition, it should be noted that noncontrast CT should be combined with other diagnostic tools such as ultrasound or biochemical index in order to further increase diagnostic accuracy of acute appendicitis, when the clear diagnosis is difficult to make through noncontrast CT test! However, this meta-analysis has some limitations that require careful consideration. First, the overall method- ‘logical quality of the included studies was moderate with several sources of bias, especially of the measure- ment bias from differential verification and the reference standard lacking blinding from the noncontrast CT test. Second, a publication bias may exist, with all studies published in English, although the funnel plot showed no evidence of publication bias. Possible unpublished studies may influence our results. Third, clinical het- ‘erogencity among studies may be an issue because of different severity of acute appendicitis in all the included studies. Therefore, these limitations above may weaken the conclusions of this meta-analysis. Future studies are needed to confirm these results. In conclusion, based on the seven prospective studies, noncontrast CT has high diagnostic accuracy in detecting acute appendicitis, which is adequate for clinical decision making. However, the evidence is limited by the moderate quality. Further higher ‘quality studies should be undertaken to confirm these findings, Acknowledgment We thank Dr. Shanjun Tan of Research Institute of Gen- cral Surgery, Jinling Hospital, Medical School of Nanjing University for his help in conducting this meta-analysis.No. 6 [REFERENCES 1. Suen K, Hayes IP, Thomson BN, et al. Bifect of the in- lwoduetion of an emergency general surgery service on outcomes ftom appendivectomy. Br T Surg 2014;10L-e141-6 2. Schellekens DH, Hulsewe KW, van Acker BA, etal Bvalu- ation of the diagnostic accuracy of plasma markers for early di- ‘agnosis in patients suspected for acute appendicitis, Acad Emerg ‘Med 2013;20:703-10, 3. Hlibemuk V, Dattaro JA, Jin Z, etal. Diagnostic accuracy of | noncontrast computed tomography for appendicitis in adults: a systematic review. Ann Emerg Med 2010;55:51-9, 4, in't Hof KH, van Lankeren W, Krestin GP, et al, Surgical validation of unenhanced helical computed tomography in acute appendicitis. Br J Surg 2004:91:1641-5, 5. Christopher FL, Lane MJ, Ward JA. eal. Unenbanced helical T scanning of the abdomen and pelvis changes disposition of patients presenting to the emergency department with possible acute appendicitis. J Emerg Med 2002;23:1-7 6. Poortman P, Lohle PN, Schoemaker CM, etl, Improving the false-negative rat of CT in acute appendicitis Reassessment of CT images by body imaging radiologists: a blinded prospective study Bur J Radiol 2010;74:67-10. 7. Maher D, Liberati A, Tetzlaff J, et al. Prefered reporting items for systematic reviews and metianelyses: the PRISMA statement. Int J Surg 2010;8:336-41. 8. Whiting P Rutjes AW, Reitsma JB, etal The development of| QUADAS: a tool for the quality assessment of studies of diagnostic DIAGNOSTIC ACCURACY OF NONCONTRAST CT Xiong et al 629 accuracy included in systematic reviews. BMC Med Res Methodol 2003;3.25 9, ‘Tan $J, Zhou P, Yu WK, etal. Fast track programmes vsta- ditional care in laparoscopic colorectal surgery: a meta-analysis of randomized controlled tials. Hepatogastoenterology 201461: 78-84 10. D'Lppolito G, de Mello GG, Szejnfeld J. The value of ‘unenhanced CT in the diagnosis of acute appendicitis, Sao Paulo Med J 1998;116:1838-45. 11, Bge G, Akman H, Sahin A, et al. Diagnostic value of unenhanced helical CT in adult patients with suspected acute ap- pendicits, Br J Radiol 2002:75:72 12. GamanagattiS, VashishtS, Kapoor A, etal. Comparison of graded compression ultrasonography and unenhanced spiral com- puted tomography in the diagnosis of acute appendicitis. Singapore ‘Med J 2007:48:80-7 13. Horton MD. Counter SE Florence MG, eal. A prospective teal of computed tomography and ultrasonography for diagnosing appendicitis in the atypical patient. Am 3 Surg 2000;179:379-81 14. in't Hof KH, Krestin GP, Steijerberg EW, ct al. In terobserver variability in CT scan interpretation for suspected acute appendicitis. Emerg Med J 2009:26:92-4 15, Keyzer C, Zaleman M, De Maertelaer V, tal. Comparison of US and unenhanced mult-detector row CT in patient suspected of having acute appendicitis. Radiology 2005:236:527-34, 16. Bachar I, Peny ZH, Dukihno L, ea. Diagnostic value of lap- sxoseopy, abdominal computed tomography, and ultrasonography in acute appendicitis, FLaparoendose Adv Surg Tech A 2013:23:982-9,