Introduction:
Millions of patients are seen in emergency departments (EDs) for abdominal pain. Appendicitis is a frequent cause of such visits, leading to 250,000 appendectomies performed annually in the United States.[
1]
Diagnosing appendicitis based solely on clinical symptoms is often incorrect in approximately 30% of cases, potentially resulting in unnecessary surgical interventions [
2]. Conversely, a missed diagnosis of appendicitis carries significant morbidity. And yet, appendicitis is diagnosed clinically with investigations that include blood tests and imaging studies, but no test exists that can reliably identify it with 100% accuracy.
The potential complications of a ruptured appendix are far too dangerous and carry a significant mortality and morbidity risk. To effectively manage patients suffering from appendicitis, once a diagnosis of appendicitis is made, the traditional treatment is surgical excision of the appendix (appendectomy) via open or laparoscopic approaches to the abdomen [
3]
An incorrect diagnosis of appendicitis may lead to unnecessary surgery if the underlying etiology is self-limiting or requires medical treatment. Surgery will result in a negative appendectomy (NA), where the appendix is excised but tissue analysis reveals no inflammation (a negative appendectomy).
The risks of a missed diagnosis, as well as one associated with unnecessary surgery due to an incorrect diagnosis, both underline the necessity of an effective diagnostic algorithm based on imaging that offers accuracy as well as convenience. Diagnostic imaging with ultrasonography or computed tomography (CT) has been proven accurate to detect acute appendicitis [
4]. However, these imaging modalities do not seem to be as effective in differentiating simple from perforated appendicitis. For ultrasonography, the reported sensitivities vary from 29% to 84% [
5]. More recent studies evaluating the accuracy of CT in detecting perforated appendicitis have documented sensitivities between 28% and 62% and specificities between 81% and 91% [
6] [
7] [
8]. While these numbers appear to be encouraging, they do not offer 100% accuracy for diagnosing appendicitis, which has been established to be a condition that would require prompt, accurate diagnosis followed by efficient interventions to be managed safely.
Magnetic resonance imaging (MRI) is a promising modality in the evaluation of suspected acute appendicitis owing to its high diagnostic accuracy [
9] [
10] and avoidance of ionizing radiation and intravenous contrast medium. In the past, MRI availability and affordability, especially in emergency department settings, have posed challenges. Nevertheless, the cost of MRI has gradually approached that of CT over time, and its accessibility has significantly improved in recent years. If MRI is confirmed to be an accurate, radiation-free imaging test, then it could be a valid alternative or even a first-line imaging modality for appendicitis. This would be particularly true in children and pregnant women, to whom avoidance of radiation is especially desirable.
The purpose of this meta-analysis was to ascertain the accuracy of MRI as a diagnostic modality in the context of diagnosing appendicitis, to pave the way for effective, efficient, and accurate management of the same, thereby addressing concerns about diagnostic modalities and their adverse effects, as well as the necessity and success of relevant interventions.
A comprehensive literature search was conducted using PubMed, Google Scholar, and the Cochrane Library databases to identify relevant articles. Only full-text articles in English were included in this analysis. Medical subject headings (MeSH) and keywords like 'CT scan,' 'MRI Scan,' 'MRI vs CT scan for Acute Appendicitis,' 'Diagnosis imaging for Acute Appendicitis,' and 'Acute Appendicitis' were employed. Additionally, we reviewed references, reviews, and meta-analyses for potential additional articles.
Initial screening of titles and abstracts was performed to eliminate duplicate entries and citations. References of relevant papers were also reviewed to identify possible additional articles. Selection criteria included papers with detailed patient information and statistically supported results.
This study's primary objective was to assess the diagnostic accuracy of MRI in detecting appendicitis in diverse patient populations.
We included studies comparing MRI and CT scan outcomes for suspected appendicitis in the general population, encompassing children, pregnant patients, and adults. Hence, this study aimed to conduct a systematic review and meta-analysis of MRI's diagnostic efficacy for appendicitis in the broader population, not limited to specific subgroups like pregnant patients or children. Our primary outcomes of interest were MRI's sensitivity and specificity for this indication.
Inclusion criteria were: (1) studies reporting accurate diagnosis with MRI and CT scan, (2) studies published in English, and (3) studies comparing MRI with CT for Acute Appendicitis. Exclusion criteria included: (1) non-full-text articles, (2) unpublished articles, and (3) articles in languages other than English.
Each eligible paper was independently evaluated by two reviewers. They examined the number of patients, age, modality of procedures, and incidence of predetermined complications. Any conflicts were resolved through discussion with the author or a third party. We assessed the research quality using a modified Jadad score. According to the PRISMA guidelines, we selected a total of 35 randomized controlled trials (RCTs) involving 5206 patients.
The quality of each included study was independently assessed by two reviewers using the QualSyst tool, consisting of ten questions with scores ranging from 0 to 2. The maximum possible total score is 20. Two authors evaluated each article based on the mentioned criteria. Interobserver agreement for study selection was determined using the weighted Cohen's kappa (K) coefficient. To assess the risk of bias in RCTs, we also employed the Cochrane tool. No assumptions were made about missing or unclear information, and no external funding was received for data collection or review.
Statistical analyses were conducted using software packages, including RevMan (Review Manager, version 5.3), SPSS (Statistical Package for the Social Sciences, version 20), and Stata 14 in Excel. Data were collected and entered into analytical software. We estimated sensitivity, specificity, positive predictive value (PPV), diagnostic odds ratios (DOR), and relative risk (RR) using fixed- or random-effects models, with 95 percent confidence intervals to evaluate critical clinical outcomes. We computed diagnosis accuracy and the Youden index for each outcome. Individual study sensitivity and specificity were depicted in Forest plots and receiver operating characteristic (ROC) curves. Additionally, we described the prior odds ratio, positive and negative likelihood ratios, and positive and negative post-test ratios in Fegan's analysis.
To assess the risk of bias, QUADAS-2 analysis was employed. This tool includes 4 domains Patient selection, Index test, Reference standard, Flow of the patients, and Timing of the Index tests.
Table 1.
Table of the description of papers.
Table 1.
Table of the description of papers.
First Author |
Year |
Location |
Study type |
Study dates |
Index test |
Reference standard |
Total Patients |
Females |
Mean Age |
Aggawala [12] |
2018 |
Australia |
Prospective |
2017-2018 |
MRI |
Histology or Follow-Up |
52 |
37 |
Unclear (range 10-39 years) |
Aguilera [13] |
2018 |
U.S.A |
Retrospective |
2014-2016 |
MRI |
Histology or Follow-Up |
52 |
52 |
Median age 25 |
Amitai [14] |
2016 |
Isreal |
Retrospective |
2007-2013 |
MRI |
Histology or Follow-Up |
49 |
49 |
Unclear |
Aspelund [15] |
2014 |
U.S.A |
Retrospective |
2008-2012 |
MRI |
Histology or Follow-Up |
397 |
Unclear |
Unclear |
Avcu [16] |
2013 |
Turkey |
Prospective |
2009-2010 |
MRI |
Histology or Follow-Up |
55 |
26 |
35.6 |
Batool [17] |
2016 |
Canada |
Prospective |
Unclear |
MRI |
Histology or Follow-Up |
100 |
56 |
Unclear |
Bayraktutan [18] |
2014 |
Turkey |
Prospective |
Unclear |
MRI |
Histology or Follow-Up |
45 |
19 |
7 |
Burke [19] |
2015 |
U.S.A,multicentre |
Retrospective |
2009-2014 |
MRI |
Histology or Follow-Up |
709 |
709 |
27.5 |
Burns [20] |
2018 |
Canada |
Retrospective |
2006-2012 |
MRI |
Histology or Follow-Up |
63 |
63 |
31 |
Corkum [21] |
2018 |
U.S.A |
Retrospective |
2015-2016 |
MRI |
Histology or Follow-Up |
135 |
Unclear |
11.5 |
des Plantes [22] |
2016 |
Netherlands |
Prospective |
Unclear |
MRI |
Histology or Follow-Up |
112 |
112 |
22 |
Dibble [23] |
2017 |
U.S.A |
Retrospective |
2011-2012 |
MRI |
Histology or Follow-Up |
77 |
Unclear |
11.5 |
Didier [24] |
2017 |
U.S.A |
Retrospective |
2013-2015 |
MRI |
Histology or Follow-Up |
98 |
60 |
11 |
Dillman [25] |
2016 |
U.S.A |
Retrospective |
2013-2014 |
MRI |
Histology or Follow-Up |
103 |
56 |
11.5 |
Donlon [26] |
2015 |
Ireland |
Retrospective |
2008-2014 |
MRI |
Histology or Follow-Up |
29 |
29 |
Unclear |
Fonseca [27] |
2014 |
U.S.A |
Retrospective |
2000-2011 |
MRI |
Histology or Follow-Up |
31 |
31 |
Unclear |
Kearl [28] |
2016 |
U.S.A |
Retrospective |
2010-2013 |
MRI |
Histology or Follow-Up |
192 |
Unclear |
14.8 |
Kennedy [29] |
2018 |
U.S.A |
Retrospective |
2014-2017 |
MRI |
Histology or Follow-Up |
612 |
353 |
11.7 |
Khalil [30] |
2018 |
U.S.A |
Retrospective |
2014-2017 |
MRI |
Histology or Follow-Up |
568 |
Unclear |
Unclear |
Kinner [31] |
2017 |
U.S.A |
Prospective |
2012-2014 |
MRI |
Histology or Follow-Up |
230 |
28 |
17.1 |
Koning [32] |
2014 |
U.S.A |
Retrospective |
2012-2013 |
MRI |
Histology or Follow-Up |
364 |
223 |
11.3 |
Konrad [33] |
2015 |
U.S.A |
Retrospective |
2009-2011 |
MRI |
Histology or Follow-Up |
114 |
114 |
Unclear |
Kulaylat [34] |
2015 |
U.S.A |
Retrospective |
2011-2013 |
MRI |
Histology or Follow-Up |
510 |
23 |
11.3 |
Shin [35] |
2017 |
Korea |
Retrospective |
2008-2015 |
MRI |
Histology or Follow-Up |
125 |
125 |
30.6 |
Theilen [36] |
2015 |
U.S.A |
Retrospective |
2007-2012 |
MRI |
Histology or Follow-Up |
171 |
171 |
Unclear |
Thieme [37] |
2014 |
Netherlands |
Prospective |
2009 |
MRI |
Histology or Follow-Up |
104 |
57 |
12 |
Tsai [38] |
2017 |
U.S.A |
Retrospective |
2003-2015 |
MRI |
Histology or Follow-Up |
233 |
233 |
28.4 |
Nitta [39] |
2005 |
Japan |
Retrospective |
Unclear |
MRI |
Histology or Follow-Up |
37 |
19 |
37.1 |
Cobben [40] |
2009 |
Netherlands |
Prospective |
2005-2006 |
MRI |
Histology or Follow-Up |
138 |
80 |
29 |
|
|
|
|
|
|
|
|
|
|
Singh [41] |
2009 |
U.S.A |
Retrospective |
2001-2007 |
MRI |
Histology or Follow-Up |
40 |
unknown |
34 |
Inci E [42] |
2011 |
Turkey |
Prospective |
Unclear |
MRI |
Histology or Follow-Up |
119 |
36 |
27 |
Chabanova [43] |
2011 |
Denmark |
Prospective |
Unclear |
MRI |
Histology or Follow-Up |
48 |
29 |
37.1 |
Heverhagen JT [44] |
2012 |
Germany |
Prospective |
2008 |
MRI |
Histology or Follow-Up |
52 |
21 |
44.7 |
Zhu [45] |
2012 |
China |
Prospective |
2009-2011 |
MRI |
Histology |
41 |
23 |
41.5 |
Leeuwenburgh [46] |
2014 |
Netherlands |
Prospective |
2010 |
MRI |
Histology or Follow-Up |
223 |
138 |
38 |
- MRI
vs CT scan for Appendicitis
A total of 35 RCTs with 5206 patients were selected for the study(figure 2). Out of these tests, 19 tests showed a sensitivity of over 95%, and 22 tests provided a specificity of over 95%. And 12 tests showed both specificity and sensitivity over 95%. The value of True positive was 1481, True Negative was 3440, False negative was 74, and False Positive was 211. With a confidence interval of 95%, Sensitivity, specificity, and Positive Predictive values were calculated. A summary of these findings is depicted in Figure 2. The Sensitivity of the test is 0.952 with a CI of 95% in a range of (0.904 to 1) the mean being (0.048). The Specificity of the test is 0.942 with a CI of 95% in a range of ( 0.874 to 1.01) the mean being (0.068). The PPV is 0.875 with a CI of 95% in a range of (0.82 to 0.93) the mean being (0.055).
The summary of the ROC curve (Figure 3) shows that the area under the ROC (AUC) was 0.9477. The overall diagnostic odds ratio (DOR) was 326.287. Diagnostic Accuracy is 0.945 and The Younden Index is 0.895.
In Figure 4, a summary of Fagan’s analysis can be observed, in conclusion, the prior probability of the test was 30. The positive likelihood ratio was 16 and the post-test ratio was 87. The negative likelihood ratio was 0.05 and the post-test ratio was 2.
Table 2. Risk of bias and applicability concern
The summary of publication bias is shown in (Table 2 and Figure 5). For the publication bias, Inpatient selection was low in 14 studies out of 35, unclear in 18, and high in 3. In the index test, it was low in 14 studies, unclear in 20, and high in 1. While the reference standard was low in 8, high in 13, and unclear in 14. The flow and timing were high in 23, low in 9, and unclear in 3. The applicability concerns in patient selection were low in 30 high in 4 and unclear in 1. The Reference standard was low in 32, and high in 3. The index test was high in 7 and low in 26, unclear in 2.
Figure 6.
Funnel plot for publication bias.
Figure 6.
Funnel plot for publication bias.
Discussion:
Acute appendicitis is one of the most commonly encountered emergency diseases that need urgent surgical intervention [
47]. This demonstrates the need for urgent and precise diagnostic modalities. Commonly, ultrasound (US) and computed tomography (CT) are recognized as valuable imaging techniques for diagnosing acute appendicitis. The reported sensitivity of US in diagnosing acute appendicitis ranges from 75% to 90%, and the specificity and accuracy are greater than 90% [
48] [
49] [
50] [
51]. However, there are several disadvantages to using the US to diagnose acute appendicitis. Its diagnostic accuracy depends on the examiner's skill and experience. The graded compression technique, although effective, demands a high degree of technical skill and experience. The retro-cecal appendix is hard to visualize in US. Approximately 30% of cases missed on US examination are attributed to the appendix being in the retro-cecal position. In obese patients, visualizing the appendix using US can be challenging, and US may not accurately depict widespread inflammation. In addition, gaseous distension of the bowel loop in paralytic or obstructive conditions hinders the visualization of the appendix. These findings further establish the need to seek a more effective diagnostic tool.
CT is considered a highly accurate diagnostic modality for identifying acute appendicitis. According to previous reports, even when patients present with atypical symptoms, CT demonstrates sensitivity and specificity exceeding 80%. Reported accuracy typically ranges from 93% to 94%, with sensitivity values of 87% to 98% and specificity values of 83% to 97% [
52] [
53] [
54] [
55]. CT offers several advantages over ultrasound (US) in diagnosing appendicitis. It excels in correctly identifying periappendiceal inflammation by detecting increased fatty density, linear strands, or fluid collection.
Complicated appendicitis such as abscess formation or severe phlegmonous change around the cecal region is easily demonstrated, as is an extrapelvic extension of the inflammation. CT can also easily assess the remainder of the abdomen and pelvis, allowing non-appendiceal diseases with the same symptoms including cholecystitis, pancreatitis, gynecological disease, and urological disease to be diagnosed [
56] [
57].
In line with the findings of our study, there has been increasing awareness of the potential harms associated with the use of ionizing radiation from CT, despite its very high accuracy for imaging acute pathology in the abdomen. Certain reports have shown a strong trend of increasing use of CT for the evaluation of patients presenting to the ED with abdominal pain, without a corresponding increase in the number of cases of surgical emergencies identified [
58]. This potentially led to the exploration of MRI as a safer and more accurate diagnostic modality for appendicitis.
MRI is an excellent technique for the diagnosis of acute appendicitis and the exclusion of diseases requiring surgical/ interventional treatment. By using MR imaging, the unnecessary appendectomy rate was decreased. MR imaging has reported a sensitivity of 95.2%, a specificity of 94.2% for the diagnosis of acute appendicitis, along with a diagnostic accuracy of 94.5% and Younden Index of 0.895. Cobben et al., showed the value of a shorter and simpler MRI protocol in patients with clinically suspected appendicitis. According to this study sensitivity and specificity were 100% and 98%, respectively [
59]. It is also prudent to note the effect of an MRI scan of the appendix on the use of hospital resources, concluding that an abdominal MRI in the evaluation of patients suspected of having appendicitis is a reliable, safe, and potentially cost-effective technique, in line with the findings of Cobben et al.
However, a noteworthy drawback of MRI would be that when compared with CT, MRI is said to not be easily accessible for emergency studies. [
60]