Pediatric Imaging Original Research Herliczek et al. MRI of Suspected ppendicitis fter Inconclusive Ultrasound Pediatric Imaging Original Research FOCUS ON: Thaddeus W. Herliczek 1 David W. Swenson William W. Mayo-Smith Herliczek TW, Swenson DW, Mayo-Smith WW Keywords: appendicitis, MRI, pediatric DOI:10.2214/JR.12.10078 Received October 1, 2012; accepted after revision December 4, 2012. 1 ll authors: Department of Diagnostic Imaging, Warren lpert Medical School of Brown University, 593 Eddy St, Providence, RI 02904. ddress correspondence to T. W. Herliczek (thaddeus_herliczek@brown.edu). JR 2013; 200:969 973 0361 803X/13/2005 969 merican Roentgen Ray Society Utility of MRI fter Inconclusive Ultrasound in Pediatric Patients With Suspected ppendicitis: Retrospective Review of 60 Consecutive Patients OBJECTIVE. The purpose of this study is to examine the utility of appendix MRI in evaluation of pediatric patients with right lower quadrant pain and inconclusive appendix sonography findings. MTERILS ND METHODS. search of the radiology electronic database was performed for all appendix MRI examinations performed of pediatric patients within 24 hours after inconclusive appendix sonography from December 1, 2009, through pril 26, 2012. Sixty patients underwent appendix MRI within 24 hours of inconclusive sonography and represented the study cohort. MRI examinations were reviewed independently by two radiologists blinded to the diagnosis and were graded as positive, negative, or indeterminate for acute appendicitis. The final diagnosis was established by review of the surgical and pathology reports and patients electronic medical records. RESULTS. Ten of 60 patients (17%) had acute appendicitis. Both readers graded the same 12 examinations as positive and the same 48 examinations as negative for acute appendicitis, with a kappa value of 1.00 (expected agreement, 0.695). No MRI examination was interpreted as indeterminate. The sensitivity and specificity of MRI for acute appendicitis in children with inconclusive appendix ultrasound findings were 100% (95% CI, 0.72 1.00) and 96% (95% CI, 0.87 0.98), respectively. The positive predictive value for the examination was 83%, the negative predictive value was 100%, and overall test accuracy was 97%. CONCLUSION. Our study shows that MRI has a sensitivity of 100% and specificity of 96% for appendicitis in pediatric patients after inconclusive appendix sonography. We think that MRI may supplant CT as the secondary modality to follow inconclusive appendix sonography. cute appendicitis is the leading cause of emergency surgery in the pediatric population [1]. Clinical diagnosis of pediatric appendicitis can be challenging because children often present without classic signs and symptoms, making imaging an important tool to determine the cause of abdominal pain in pediatric patients [2]. The optimal imaging algorithm for pediatric right lower quadrant (RLQ) pain remains controversial [3 12]. The arguments have centered on the use of sonography or CT as the initial imaging modality [3, 4]. Graded-compression ultrasound is reported to be 88% sensitive and 94% specific for acute appendicitis [6]. Ultrasound does not use ionizing radiation and provides a relatively low-cost modality to assess pediatric appendicitis and alternative causes of RLQ pain in children [4]. CT is reported to be 94% sensitive and 95% specific for acute appendicitis [6]. However, CT exposes the pediatric patient to ionizing radiation and often uses IV or oral contrast media. Historically, our pediatric practice performed CT of pediatric patients with inconclusive ultrasound in accordance with recommendations by Garcia-Peña et al. [9] and the merican College of Radiology [13]. Given advances in body MRI software, growing literature to support the use of MRI to assess appendicitis [14 24], increased availability of MRI at our institution, the reported 97% sensitivity and 97% specificity of MRI for acute appendicitis in adults [25], and growing concerns about radiation exposure in the medical literature and lay press [26, 27], we sought to replace CT with MRI in our imaging algorithm for pediatric RLQ pain and an inconclusive ultrasound, according to the LR (as low as reasonably achievable) principle. We began imaging the pediatric appendix with MRI in December 2009. JR:200, May 2013 969
Herliczek et al. Previous literature regarding MRI of the pediatric appendix has described the utility of MRI in preselected pediatric patients with acute appendicitis that was visualized conclusively with ultrasound [19], the use of MRI as an initial imaging modality in children with RLQ pain [20, 23], the ability of MRI to reveal the normal appendix in asymptomatic children [21], and the speed and utility of 3-T MRI in children with appendicitis found on ultrasound, CT, or both [24]. In 2002, Hörmann et al. [21] suggested that MRI should be considered as an alternative to CT in pediatric patients with inconclusive appendix ultrasound. Yet, to the best of our knowledge, no previous work has analyzed the utility of MRI performed after inconclusive ultrasound in pediatric patients with RLQ pain and clinically suspected appendicitis. The purpose of this study is to evaluate the utility of appendix MRI in evaluation of pediatric patients with RLQ pain and inconclusive appendix sonography findings. Materials and Methods This retrospective study was approved by our institutional review board, and the need for informed consent waived. It was conducted in accordance with the HIP. Patient Population search of the radiology electronic database was performed for all appendix MRI examinations performed for pediatric patients from December 1, 2009, through pril 26, 2012. This search yielded 67 patients. Next, the electronic medical record (EMR) was searched for all of these patients to identify those for whom appendix MRI was performed within 24 hours of an inconclusive ultrasound examination. This yielded 60 patients (28 boys and 32 girls) with an average age of 13.4 years (range, 7 17 years) who represented the study cohort. Ultrasound was considered inconclusive if the appendix was not identified and if either of the following two scenarios was present: there were secondary ultrasound signs of appendicitis (RLQ inflammation, intraperitoneal collection, or fluid), or there was a high clinical concern for appendicitis. Clinical criteria used to decide whether to perform appendix MRI were left to the discretion of the referring pediatrician, pediatric emergency medicine physician, or pediatric surgeon. The average time between the inconclusive ultrasound and appendix MRI was 5.1 hours (range, 1 24 hours), with 88% (53/60) of MRI examinations performed within 6 hours of the inconclusive ultrasound. Proof of Diagnosis Twelve of the 60 patients underwent surgery, and 48 patients were followed clinically. Ten of the 60 patients (17%) in our cohort had acute appendicitis on review of operative and pathology reports. The average age of the 10 patients with acute appendicitis was 13.1 years (range, 9 16 years), with nine of 10 being male. Fifty of the 60 patients (83%) in our study did not have appendicitis. Two of these 50 patients underwent surgery, but a normal appendix was found at surgery and on pathologic examination. The first of these two patients was a 17-year-old boy who underwent appendectomy because of clinical suspicion for acute appendicitis. His appendix was normal according to both the operative and pathology reports. He was discharged on postoperative day 2 with a diagnosis of mesenteric adenitis. The second patient with a normal appendix at surgery and on pathologic examination was a 14-year-old girl. She was discharged on postoperative day 4 with a diagnosis of abdominal pain, not otherwise specified. Forty-eight of the 50 patients without appendicitis improved clinically and did not undergo surgery at our institution. Twentynine of the 48 patients who did not undergo appendectomy after MRI were observed for an average of 1.4 days (range, 1 4 days). Nineteen of the 48 patients who did not undergo appendectomy after MRI were discharged from the emergency department without observation after MRI. Most (36/48) patients who did not undergo appendectomy after MRI were discharged with a diagnosis of gastroenteritis or abdominal pain, not otherwise specified. One of these patients was readmitted 2 weeks later and was diagnosed with Crohn disease at colonoscopy. The other 12 patients were discharged with the following diagnoses: four with ruptured or hemorrhagic ovarian cysts, two with mesenteric adenitis, two with omental infarcts, two with pelvic inflammatory disease, and one each with cholelithiasis and nephrolithiasis. MRI Technique Forty-nine MRI examinations were performed on 1.5-T systems (Magnetom Espree or Symphony, Siemens Healthcare), and 11 examinations were performed on a 3-T system (Verio, Siemens Healthcare). The 3-T system was installed at our institution in December 2010. Patients were imaged on the next available MRI system without regard to field strength. ll appendix MRI sequences were acquired from the inferior poles of the kidneys through the urinary bladder without sedation or contrast media. The MRI protocol varied over time as our experience with pediatric appendix MRI grew. However, all examinations included a single-plane STIR sequence and a coronal T2-weighted 3D turbo spin-echo sequence (SPCE) with multiplanar reconstructions and/or T2-weighted single-shot turbo spin-echo (i.e., HSTE) without fat saturation in multiple planes. Most examinations (51/60) included an axial true fast imaging with steadystate precession (TrueFISP) sequence. Seventeen of 60 examinations included in-phase and out-ofphase T1-weighted imaging. ll sequences except SPCE were obtained with breath-holding. SPCE was acquired during shallow free breathing. Imaging parameters are noted in Table 1. Image acquisition time varied from 10 to 66 minutes, with an average of 30.5 minutes. ll patients undergoing MRI after inconclusive ultrasound tolerated the examination successfully. Image Interpretation The 60 MRI examinations were reviewed independently by one board-certified fellowshiptrained pediatric radiologist with 3 years of experience and one third-year radiology resident physician. Readers were aware of the indication for the examination, but were otherwise blinded to surgical-pathologic findings and clinical follow-up. Reader 1 provided the initial interpretation for 28 of 60 examinations and performed review of these examinations at least 8 weeks (range, 8 60 weeks) removed from the time of initial interpretation. Examinations were graded by each reader independently as positive, negative, or indeterminate regarding the presence of acute appendicitis. Readers assessed the following factors: appendix visualization, appendiceal diameter (> 7 mm was considered pathologic), appendiceal mural thickening of more than 3 mm, appendiceal mural edema (i.e., increased mural signal intensity on T2-weighted imaging), the presence of appendiceal intraluminal fluid (i.e., increased intraluminal signal intensity on T2-weighted imaging), the presence of periappendiceal inflammation (i.e., increased periappendiceal signal intensity on T2-weighted imaging), the presence of RLQ inflammation (i.e., increased signal intensity on STIR), and the presence of an appendicolith (i.e., focal round well-circumscribed intraluminal hypointensity on all sequences). If an abnormal appendix and secondary findings were identified, the reader assigned a grade of positive (Figs. 1 and 2). If a normal appendix was identified (Fig. 3), the reader assigned a grade of negative. If the appendix was not identified, the reader assigned a grade of negative if no inflammatory changes were identified in the RLQ on the STIR sequence. If the appendix was not identified, but RLQ inflammatory changes were present on the STIR image and no alternate cause of RLQ pain was identified, the reader was to assign a grade of indeterminate. lternate causes of RLQ pain were recorded when present. 970 JR:200, May 2013
MRI of Suspected ppendicitis fter Inconclusive Ultrasound TBLE 1: ppendix MRI Parameters, by Sequence and Magnet Strength Sequence, Magnet Strength TR/TE Slice Thickness (mm) Gap (mm) Coronal T2-weighted 3D turbo spin-echo 1.5 T 1500/203 205 0.9 1 0 3 T 2000 2180/125 135 1 0 HSTE 1.5 T 900 1420/100 102 4 5 4 5.5 3 T 1000 1300/83 101 4 6 4 6 STIR 1.5 T a 2400 9770/90 97 4 4 3 T b 2300 4030/89 100 4 4 True fast imaging with steady-state precession 1.5 T 4.3 9/2.15 2.4 4 5 4 6 3 T 3.5 4.4/1.5 2.5 4 4 T1 weighted 1.5 T In-phase 147/4.77 8 10 Out-of-phase 147/2.33 8 10 3 T In-phase 4.76/2.45 3 10 Out-of-phase 4.36/1.33 3 1 Note The FOV varied with patient size. Inversion time was 150 160 ms. Inversion time was 220 ms. Patients with acute appendicitis documented in operative notes and pathology reports were considered true-positives. Patients with normal appendix documented in either operative or pathology reports and patients who did not have an appendectomy at our institution were considered true-negatives. The EMR of the patients discharged without appendectomy were reviewed for length of admission or observation, discharge diagnosis, and subsequent appendectomy at our institution. The EMR was reviewed, on average, 14.5 months (range, 4 30.5 months) after the MRI to obtain as complete a picture as possible of the patients final diagnoses and follow-up. Statistical nalysis The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of appendix MRI for acute appendicitis in pediatric patients after an inconclusive ultrasound were calculated using Excel (Microsoft). The kappa value for interobserver reliability of the two readers was also calculated using Excel. kappa value greater than 0.8 is defined as very good agreement beyond chance, a value between 0.40 and 0.8 indicates moderate-togood agreement beyond chance, and a value of less than 0.40 indicates fair-to-poor agreement [28]. The expected rate of agreement (Pe) was calculated. Results ll 10 MRI examinations of the patients with acute appendicitis were graded as positive by both readers. There were two false-positive MRI readings by both readers (the same two patients). The first was a 17-year-old boy imaged at 1.5 T (Fig. 4). He was discharged from our emergency department the same day without appendectomy after resolution of clinical symptoms. His discharge diagnosis was abdominal pain, not otherwise specified. Review of the EMR indicated that he did not return to our institution for appendectomy. The second was a 14-year-old girl who was imaged at 3 T and underwent subsequent appendectomy. Her appendix was noted to be normal in both the operative note and pathology report. She was discharged on postoperative day 4 with a diagnosis of abdominal pain, not otherwise specified. ll true-positives had an appendix diameter of at least 9 mm, whereas the two false-positive appendixes measured 8 mm in diameter at the tip only. There were 48 true-negative MRI readings by both readers (the same 48 patients). The sensitivity and specificity of MRI for acute appendicitis in children with inconclusive ultrasound findings were 100% (95% CI, 0.72 1.00) and 96% (95% CI, 0.87 0.98), respectively. The Fig. 1 12-year-old boy with acute appendicitis. xial 1.5-T STIR image shows hypointense appendicolith (arrow) centrally within lumen of distended appendix, with mural edema and periappendiceal inflammation. Fig. 2 16-year-old boy with acute appendicitis., xial 1.5-T HSTE image shows retrocecal appendix (arrow) with mural edema, intraluminal fluid, and periappendiceal edema. B, Coronal 1.5-T STIR image shows conspicuous inflamed appendix (arrow) and periappendiceal edema. B JR:200, May 2013 971
Herliczek et al. positive predictive value for the examination was 83% (95% CI, 0.55 0.95), and the negative predictive value was 100% (95% CI, 0.93 1.00), with an overall test accuracy of 97%. None of the MRI examinations was interpreted as indeterminate regarding the presence or absence of acute appendicitis by either reader. Interobserver agreement was very good, with a kappa value of 1.00 (Pe, 0.695). The normal appendix was identified in 83% of cases by reader 1 (35/41 [85%] at 1.5-T MRI and 5/7 [71%] at 3-T MRI) and in 88% of the cases by reader 2 (36/41 [88] at 1.5-T MRI and 6/7 [86%] at 3-T MRI). Readers 1 and 2 excluded acute appendicitis when the appendix was not visualized and no RLQ inflammation was present in eight patients and six patients, respectively. MRI detected an alternate cause of RLQ pain in four patients, including two patients with omental infarction (Fig. 5), one with a hemorrhagic ovarian cyst, and one with cholelithiasis. Fig. 3 16-year-old girl with abdominal pain, not otherwise specified. Normal retrocecal appendix (arrow) is seen on 3-T coronal HSTE image. Discussion Our study shows that MRI has 100% sensitivity and 96% specificity for acute appendicitis in pediatric patients with RLQ pain and inconclusive ultrasound. MRI had a negative predictive value of 100% and an overall test accuracy of 97%. ll 10 patients with proven acute appendicitis had a positive MRI interpretation. Furthermore, there were no false-negative examinations, implying that a negative MRI may be sufficient to exclude appendicitis. Two MRI examinations were false-positive in that the examination was graded as positive by both reviewers, but neither patient had acute appendicitis. In both of these patients, the abnormality found on MRI was seen only in the tip of the appendix, and the diameter was 8 mm, compared with 9 mm or larger in all true-positive MRI examinations. Our MRI readings for the presence or absence of acute appendicitis were completely concordant between both a fellowship-trained pediatric radiology attending physician and a third-year radiology resident. This finding implies that MRI readings are sensitive, specific, and reproducible despite variable levels of training and experience with MRI. Our results are similar to published sensitivity and specificity rates of CT for appendicitis, which, to date, has been considered the reference standard for pediatric appendicitis. The meta-analysis by Doria et al. [6] notes that CT has a sensitivity and specificity of 94% and 95%, respectively, for acute appendicitis. We show that MRI confirms or excludes appendicitis conclusively because we had neither falsenegative nor inconclusive examinations. Every patient with proven appendicitis had an abnormal appendix identified on MRI by both observers, with very good interobserver agreement (κ = 1.00). It has been previously found that in cases where a normal appendix is not visualized by CT, the absence of RLQ inflammatory changes accurately excludes acute appendicitis in 98% of patients 15 91 years old, with a negative predictive value of 98.7% in patients 0.1 18.6 years old [29, 30]. lthough a large prospective analysis is needed to illustrate the utility of MRI in the setting of the nonvisualized appendix, our data suggest that appendicitis may be excluded by MRI, even when the normal appendix is not clearly identified. Fig. 4 17-year-old boy with abdominal pain, not otherwise specified. xial 1.5-T spatial and chemicalshift encoded excitation image shows mural thickening, intraluminal fluid, and periappendiceal stranding at 8-mm diameter appendix tip (arrow). Image was interpreted by both readers as positive for appendicitis, but his symptoms improved without appendectomy. Previous investigations of pediatric appendix MRI differ from our work. Hörmann et al. [21] and Baldisserotto et al. [22] performed appendix MRI in asymptomatic pediatric volunteers. Baldisserotto et al. identified 48% of normal appendixes in this population. This is lower than our observers rates of 83% and 87.5%. Our rates of identification of the normal appendix are similar to the 86% rate reported by Hörmann et al. Yet, our study population differs from the patients in the studies by Baldisserotto et al. and Hörmann et al. because they imaged asymptomatic volunteers only, whereas children with RLQ pain and inconclusive ultrasound (who would have proceeded previously to CT) comprised our study group. The performance of MRI in symptomatic children B Fig. 5 11-year-old boy with omental infarction., Coronal 1.5-T spatial and chemical-shift encoded excitation image shows both omental infarction (white arrow) and normal appendix (black arrow). B, xial 1.5-T STIR image shows omental infarction (white arrow) with conspicuous edema and surrounding fluid (black arrow). 972 JR:200, May 2013
MRI of Suspected ppendicitis fter Inconclusive Ultrasound with inconclusive ultrasound is likely to be more relevant clinically than the utility of MRI in asymptomatic patients. There are several limitations to our study, including a small sample population, limited direct clinical follow-up for all of our true-negatives, and variation in MRI field strength and protocols over time as our practice evolved. lthough it is possible that our true-negative patients underwent appendectomy at another hospital after discharge from our institution, this is unlikely because ours is the only pediatric hospital in our state. lthough comparing the utility of 1.5- and 3-T systems would be of interest because the shorter acquisition time and greater signal-to-noise ratio of 3-T systems could be beneficial in smaller younger patients, we did not attempt to compare the utility of different field strengths for pediatric appendicitis because of the small sample size. We imaged patients with inconclusive ultrasound in the next available MRI system at our institution, regardless of field strength. One other possible limitation is that one of the readers had interpreted the original examinations at the time of original diagnosis, but we attempted to minimize this potential bias by the long interval between the initial read and the secondary read (a minimum of 8 weeks, up to a maximum of 60 weeks). Our results and those of earlier studies [14 24] depict the evolving roles of ultrasound, CT, and MRI in evaluating pediatric appendicitis. Moore et al. [23] postulated that an algorithm of ultrasound followed by MRI might be the best imaging algorithm for pediatric RLQ pain. Imaging algorithms for pediatric appendicitis without CT lack ionizing radiation and follow the LR principle. n algorithm wherein MRI follows inconclusive ultrasound in pediatric patients with RLQ pain is likely to be used at institutions currently using CT after inconclusive ultrasound. We think that staged MRI following inconclusive ultrasound is the best imaging algorithm for pediatric RLQ pain because of its demonstrated sensitivity and specificity for appendicitis in this population and its lack of ionizing radiation. Still, further research is needed to elucidate the effectiveness of this staged algorithm compared with other imaging algorithms for pediatric RLQ pain (ultrasound first, MRI first, and staged CT following inconclusive ultrasound) and to optimize MRI protocols. In conclusion, our study shows that MRI has high sensitivity and specificity for appendicitis in pediatric patients with inconclusive ultrasound. References 1. ddiss DG, Shaffer N, Fowler BS, Tauxe RV. The epidemiology of appendicitis and appendectomy in the United States. m J Epidemiol 1990; 132: 910 925 2. Brown JJ. cute appendicitis: the radiologist s role. Radiology 1991; 180:13 14 3. Hernanz-Schulman M. CT and US in the diagnosis of appendicitis: an argument for CT. Radiology 2010; 255:3 7 4. Strouse PJ. Pediatric appendicitis: an argument for US. Radiology 2010; 255:8 13 5. Krishnamoorthi R, Ramarajan N, Wang N, et al. Effectiveness of a staged US and CT protocol for the diagnosis of pediatric appendicitis: reducing radiation exposure in the age of LR. Radiology 2011; 259:231 239 6. Doria S, Moineddin R, Kellenberger CJ, et al. US or CT for diagnosis of appendicitis in children and adults? meta-analysis. Radiology 2006; 241:83 94 7. Kaiser S, Frenckner B, Jorulf H. Suspected appendicitis in children: US and CT a prospective randomized study. Radiology 2002; 223:633 638 8. Sivit CJ, Newman KD, Boenning D, et al. ppendicitis: usefulness of US in a pediatric population. Radiology 1992; 185:549 552 9. Garcia-Peña BM, Mandl KD, Kraus SJ, et al. Ultrasonography and limited computed tomography in the diagnosis and management of appendicitis in children. JM 1999; 282:1041 1046 10. Birnbaum B, Jeffrey RB Jr. CT and sonographic evaluation of acute right lower quadrant pain. JR 1998; 170:361 371 11. Karakas SP, Guelfuat M, Leonidas JC, Springer S, Singh SP. cute appendicitis in children: comparison of clinical diagnosis with ultrasound and CT imaging. Pediatr Radiol 2000; 30:94 98 12. Sivit CJ, pplegate KE, Stallion, et al. Imaging evaluation of suspected appendicitis in a pediatric population: effectiveness of sonography versus CT. JR 2000; 175:977 980 13. Rosen, MP, Ding,, Blake, M et al.; merican College of Radiology. CR appropriateness criteria: right lower quadrant pain suspected appendicitis. merican College of Radiology website. gm.acr. org/secondarymainmenu Categories/quality_safety/ app_criteria/pdf/expertpanelonpediatric Imaging/ Othertopics/RightLowerQuadrantPain.aspx. Published 1996. Updated 2010. ccessed January 4, 2013 14. Pedrosa I, Levine D, Eyvazzadeh D, Siewert B, Ngo L, Rofsky NM. MR imaging evaluation of acute appendicitis in pregnancy. Radiology 2006; 238:891 899 15. Birchard KR, Brown M, Hyslop WB, Firat Z, Semelka RC. MRI of acute abdominal and pelvic pain in pregnant patients. JR 2005; 184:452 458 16. Oto, Srinivasan PN, Ernstet RD, et al. Revisiting MRI for appendix localization during pregnancy. JR 2006; 186:883 887 17. Oto, Ernst RD, Shah R, et al. Right-lower-quadrant pain and suspected appendicitis in pregnant women: evaluation with MR imaging initial experience. Radiology 2005; 234:445 451 18. Pedrosa I, Lafornara M, Pandharipande PV, Goldsmith JD, Rofsky NM. Pregnant patients suspected of having acute appendicitis: effect of MR imaging on negative laparotomy rate and appendiceal perforation rate. Radiology 2009; 250:749 757 19. Hörmann M, Paya K, Eibenberger K, et al. MR imaging in children with nonperforated acute appendicitis: value of unenhanced MR imaging in sonographically selected cases. JR 1998; 171: 467 470 20. Cobben L, Groot I, Kingma L, Coerkamp E, Puylaert J, Blickman J. simple MRI protocol in patients with clinically suspected appendicitis: results in 138 patients and effect on outcome of appendectomy. Eur Radiol 2009; 19:1175 1183 21. Hörmann M, Puig S, Prokesch SR, Partik B, Helbich TH. MR imaging of the normal appendix in children. Eur Radiol 2002; 12:2313 2316 22. Baldisserotto M, Valduga S, da Cunha C. MR imaging evaluation of the normal appendix in children and adolescents. Radiology 2008; 249:278 284 23. Moore MM, Gustas CN, Choudhary K, et al. MRI for clinically suspected pediatric appendicitis: an implemented program. Pediatr Radiol 2012; 42:1056 1063 24. Johnson K, Filippi CG, ndrews T, et al. Ultrafast 3-T MRI in the evaluation of children with acute lower abdominal pain for the detection of appendicitis. JR 2012; 198:1424 1430 25. Barger RL Jr, Nandalur KR. Diagnostic performance of magnetic resonance imaging in the detection of appendicitis in adults: a meta-analysis. cad Radiol 2010; 17:1211 1216 26. Brenner DJ, Hall EJ. Computed tomography: an increasing source of radiation exposure. N Engl J Med 2007; 357:2277 2284 27. Gee. Radiation concerns rise with patients exposure. The New York Times, June 12, 2012:18 28. ltman DG. Practical statistics for medical research. London, UK: Chapman and Hall, 1991 29. Nikolaidis P, Hwang CM, Miller FH, Papanicolaou N. The nonvisualized appendix: incidence of acute appendicitis when secondary inflammatory changes are absent. JR 2004; 183:889 892 30. Garcia K, Hernanz-Schulman M, Bennett DL, Morrow SE, Yu C, Kan JH. Suspected appendicitis in children: diagnostic importance of normal abdominopelvic CT findings with nonvisualized appendix. Radiology 2009; 250:531 537 JR:200, May 2013 973