Objective
The purpose of this study was to determine the diagnostic performance of magnetic resonance imaging (MRI) in the diagnosis of acute appendicitis during pregnancy in a multiinstitutional study.
Study Design
In this multicenter retrospective study, the cases of pregnant women who underwent MRI evaluation of abdominal or pelvic pain and who had clinical suspicion of acute appendicitis between June 1, 2009, and July 31, 2014, were reviewed. All MRI examinations with positive findings for acute appendicitis were confirmed with surgical pathologic information. Sensitivity, specificity, negative predictive values, and positive predictive values were calculated. Receiver operating characteristic curves were generated, and area under the curve analysis was performed for each participating institution.
Results
Of the cases that were evaluated, 9.3% (66/709) had MRI findings of acute appendicitis. Sensitivity, specificity, accuracy, positive predictive value, and negative predictive values were 96.8%, 99.2%, 99.0%, 92.4%, and 99.7%, respectively. There was no statistically significant difference between centers that were included in the study (pair-wise probability values ranged from 0.12–0.99).
Conclusion
MRI is useful and reproducible in the diagnosis of suspected acute appendicitis during pregnancy.
Acute appendicitis is the most common nonobstetric indication for emergency surgery in pregnant women, with a prevalence of 1 in 500-2000 pregnancies. Unfortunately, the clinical diagnosis of acute appendicitis remains challenging, given the wide range in differential diagnoses, both obstetric and nonobstetric and the anatomic changes during pregnancy. Classic physical examination findings of right lower quadrant tenderness, rebound, and guarding may not be present in mid to late pregnancy as the gravid uterus overlies and superiorly displaces the appendix. Physiologic leukocytosis of pregnancy also reduces the specificity of the white blood cell count.
Undiagnosed acute appendicitis has been shown to result in significant risk of fetal loss that ranges from 1.5% for nonperforated appendicitis to 35-55% for perforated appendicitis. This, coupled with the increased risk of appendiceal perforation in pregnancy, historically has resulted in a lower threshold for laparoscopic exploration. However, there is a higher rate of fetal loss in patients who undergo a negative appendectomy.
The use of preoperative imaging has decreased the negative appendectomy rate in the general population. Computed tomography has a high sensitivity and specificity for appendicitis; although the fetal dose of ionizing radiation of 1 abdominal computed tomography is not prohibitive, there is associated risk. Ultrasound scanning is safe and inexpensive; however, it is operator-dependent and potentially challenging with a gravid uterus, given the superior displacement of the appendix and poor penetration of the area of interest because of the presence of adjacent bowel loops. Frequently, the appendix is not visualized, and further imaging is required.
With its excellent soft tissue contrast, magnetic resonance imaging (MRI) offers a safe, nonionizing alternative imaging approach for the diagnosis of acute appendicitis. The use of MRI in the setting of suspected acute appendicitis during pregnancy recently has been shown to decrease the length of hospital stay without increasing hospital cost, which leads some to recommend MRI as the first-line imaging test. Unfortunately, many studies are single institution studies with relatively small patient populations.
This multicenter study was conducted to evaluate the diagnostic performance of MRI in the investigation of acute appendicitis over a 5-year experience.
Methods
Patients and data accrual
This study was approved by the institutional review board of each participating university hospital (University of North Carolina at Chapel Hill, Duke University Medical Center, Hospital of the University of Pennsylvania, University of Southern California, University of California at San Diego, Northwestern University, University of Chicago, University of Nebraska, McGill University Health Center, and Georgetown University) and was Health Insurance Portability and Accountability Act compliant. Given the retrospective review of data, individual patient consent was waived at each participating institution.
We conducted a retrospective review of all clinical MRI examinations of pregnant women who were at least 16 years old with abdominal pain and suspected appendicitis that were performed between June 1, 2009, and July 31, 2014. All MRI examinations were reviewed individually by an abdominal radiologist for the visualization of the appendix and imaging findings of appendicitis (appendiceal dilation, appendicolith, free fluid, and fat-stranding). MRI criteria for the diagnosis of acute appendicitis included a dilated appendix that measured ≥7 mm with signs of periappendiceal inflammation.
Clinical records were also reviewed for patient age, gestational age at the time of the MRI examination, and pathologic confirmation of appendicitis, when applicable. In centers where ultrasound scanning was performed as the first line of imaging, ultrasound results were also recorded (positive, negative, or nondiagnostic/nonvisualization of the appendix). Individual investigators at each university performed all data accrual. Patients with MRI findings of acute appendicitis without surgical follow up were excluded from review. Patients with MRI findings that were equivocal for acute appendicitis were considered to be positive for acute appendicitis.
MRI
All MRIs were performed on 1.5T magnets without the use of intravenous gadolinium-based contrast agents. Imaging protocols varied by institution but generally consisted of fat-suppressed HASTE imaging (half-Fourier acquisition single-shot turbo spin echo) and fat-suppressed T1-weighted imaging in all 3 planes (coronal, sagittal, and axial). Board-certified, fellowship-trained abdominal radiologists interpreted all MRI examinations.
Data analysis
Measures of central tendency were calculated with Bonferroni correction for maternal and gestational age. Kruskal-Wallis test was used to compare patient populations (maternal age and gestational age) between institutions. Contingency tables were generated for all patients (pooled data) and for individual centers. Diagnostic performance of MRI for the diagnosis of appendicitis in pregnancy was evaluated by computation of the following parameters: sensitivity, specificity, accuracy, positive predictive value, and negative predictive value. Receiver operating characteristic curves were generated, and area under the curve (AUC) analysis was performed for each participating institution. Comparison among individual institutions was performed with the DeLong’s method to test for statistically significant differences in performance (AUC) between institutions.
A language and environment for statistical computing (R Core Team, Vienna, Austria) was used for statistical computing. Descriptive data values were reported as mean ± standard deviation. All reported probability values were 2-sided, and statistical significance was defined as a probability value of < .05.
Results
A total of 714 pregnant women (16-46 years old; mean age, 27.5 ± 6.2 years) were reviewed. Five of these patients with MRI findings of acute appendicitis were excluded from data analysis because of the absence of pathologic confirmation. Four of these patients were read as “early appendicitis” on MRI and successfully treated with antibiotics. The remaining patient was read as acute appendicitis, but she declined surgical intervention. She was also treated successfully with antibiotics. All 5 patients went on to normal deliveries.
A total of 709 pregnant women (16-46 years old; mean, 27.8 ± 6.2 years) were included for review. Gestational age ranged from 1–39 weeks, with a mean of 17 ± 8.5 weeks. Patient demographics by site are further described in Table 1 . There was a significant difference in maternal age and gestational age across the institutions ( P = .0006 and .0007 respectively).
| Institution | Patients, n | Maternal age, y (mean ± SD) | Gestational age, wk (mean ± SD) | Sensitivity, % | Specificity, % | Positive predictive value, % | Negative predictive value, % | Area under the curve |
|---|---|---|---|---|---|---|---|---|
| A | 50 | 28.3 ± 5.6 | 21.9 ± 8.9 | 100 | 95.7 | 66.7 | 100 | 0.98 (95% CI, 0.95–1.0) |
| B | 15 | 30.4 ± 6.6 | 21.1 ± 8.2 | 100 | 100 | 100 | 100 | 1.0 (95% CI, 1.0–1.0) |
| C | 70 | 26.6 ± 6.0 | 15.7 ± 7.5 | 100 | 100 | 100 | 100 | 1.0 (95% CI, 1.0–1.0) |
| D | 138 | 27.5 ± 6.2 | 16.0 ± 7.7 | 100 | 99.2 | 92.3 | 100 | 0.99 (95% CI, 0.98–1.0) |
| E | 58 | 25.8 ± 5.8 | 16.7 ± 8.7 | 100 | 100 | 100 | 100 | 1.0 (95% CI, 1.0–1.0) |
| F | 61 | 29.3 ± 5.8 | 17.2 ± 8.0 | 100 | 100 | 100 | 100 | 1.0 (95% CI, 1.0–1.0) |
| G | 45 | 26.8 ± 6.2 | 16.8 ± 8.8 | 100 | 97.7 | 66.7 | 100 | 0.99 (95% CI, 0.97–1.0) |
| H | 107 | 29.1 ± 6.6 | 15.9 ± 8.6 | 100 | 100 | 100 | 100 | 1.0 (95% CI, 1.0–1.0) |
| I | 130 | 27.2 ± 5.8 | 17.5 ± 9.0 | 83.33 | 99.2 | 90.9 | 98.3 | 0.91 (95% CI, 0.80–1.0) |
| J | 35 | 30.8 ± 6.7 | 19.4 ± 7.6 | 100 | 100 | 100 | 100 | 1.0 (95% CI, 1.0–1.0) |
Sixty-six of 709 patients (9.3%) had MRI findings compatible with acute appendicitis (appendiceal dilation, appendicolith, free fluid, and/or fat-stranding; Figure 1 ). Of these 66 patients, 61 patients were confirmed to have acute appendicitis by pathologic evaluation. There were 5 false positives; pathologic evaluation revealed a torsed right ovary (n = 1), appendicolith with mild lymphoid hyperplasia (n = 1), fibrous obliteration of the appendiceal lumen without changes of acute appendicitis (n = 1), and normal appendices (n = 2). Pooled performance data across all institutions are provided in Table 2 with 95% confidence intervals. Sensitivity and specificity for the detection of acute appendicitis were high, 96.8% and 99.2%, respectively. Accuracy was 99.0%. Positive and negative predictive values were 92.4% and 99.7%, respectively. The pooled AUC was 0.98 (95% confidence interval, 0.96–1.0). There were no statistically significant differences of the AUC values between institutions; pairwise comparisons of AUC ranged between 0.83 and 1 ( P values: .12–.99; Figure 2 ).
| Magnetic resonance image | Clinical/pathologic outcome a | |
|---|---|---|
| Acute appendicitis | No appendicitis | |
| Acute appendicitis | 61 | 5 |
| No appendicitis | 2 | 641 |
a Positive predictive value, 92.4% (range, 83.2–97.5%); Negative predictive value, 99.7% (range, 98.9–99.9%); Sensitivity, 96.8% (range, 89–99.6%); Specificity, 99.2% (range, 98.2–99.8%); Accuracy, 99.0% (range, 98.0–99.6%).
Of the patients without evidence of acute appendicitis by MRI, 72 patients (72/643; 11.2%) had alternative findings that may have accounted for the patient’s acute abdominal pain. A list of these alternative diagnoses is listed in Table 3 .
| Alternative diagnosis | Patients, n |
|---|---|
| Pyelonephritis | 18 |
| Cholelithiasis | 10 |
| Obstructing renal stone | 7 |
| Ovarian torsion | 6 |
| Degenerating fibroid tumor | 6 |
| Dermoid | 5 |
| Acute inflammatory bowel disease | 4 |
| Ectopic pregnancy | 3 |
| Endometriosis | 3 |
| Pancreatitis | 2 |
| Cystitis | 2 |
| Subchorionic hemorrhage | 2 |
| Uterine rupture | 1 |
| Omental infarct | 1 |
| Colitis | 1 |
| Closed loop obstruction, small bowel | 1 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
