X-ray for Appendicitis

X-rays were a common historical starting point for evaluation of appendicitis, but they have little diagnostic value and should generally be avoided. The classic x-ray findings of appendicitis include a radiodensity in the right lower quadrant, called an appendicolith. Yet this is found on x-ray in fewer than 20% of cases of appendicitis.¹³⁰ Multiple studies have shown a low diagnostic yield for x-ray. Other findings sometimes seen on x-ray include obscuration of the normal right psoas muscle and properitoneal fat interface, presumably because of periappendiceal inflammation, but these are seen with equal frequency in patients with and without appendicitis and therefore have no discriminatory value.¹³¹ Rao et al.¹³² reviewed 821 consecutive patients hospitalized for suspected appendicitis, 78% of whom underwent x-ray. Radiographic abnormalities were seen in 51% of patients with appendicitis and 47% of those without appendicitis, with no radiographic finding being sensitive or specific. Even in the 10% of cases in which the x-ray suggested a specific condition, these failed to match the final diagnosis in 57%. Although the cost of x-ray was low ($67), the utility was so low that the cost per specific diagnosis was $1593, exceeding the cost for CT ($270).

One important reason for the lack of utility of x-ray is that treatment of appendicitis has evolved from open appendectomy to laparoscopic appendectomy for uncomplicated cases and frequent percutaneous drainage with antibiotic therapy in cases with complications such as loculated abscess. Even in the minority of cases in which x-ray findings of appendicolith support the diagnosis of appendicitis, x-ray does not provide additional information relevant to modern management strategies.

Barium Enema for Appendicitis

Before the advent of CT, barium enema was used periodically in assessment of appendicitis. Although a mass effect on the cecum and nonfilling of the appendix were described as diagnostic of appendicitis, these findings can be seen in nonsurgical disease processes such as small-bowel obstruction, enterocolitis, pelvic adhesions, and pelvic inflammatory disease.¹³³ Barium enema is generally no longer indicated for diagnosis of appendicitis. The ACR rates contrast enema as two or three (meaning “usually not appropriate”) out of nine on its appropriateness scale, depending on the clinical scenario.¹³⁴

Computed Tomography for Appendicitis

Despite intense research into CT for appendicitis, many controversies remain, which we address in the text that follows:

Ultrasound for Appendicitis

Ultrasound is a common alternative diagnostic modality to CT scan for appendicitis. Advantages of ultrasound include portability, the absence of any exposure to ionizing radiation (a major concern with CT, especially in children, young adults, and pregnant women), and the absence of need for oral or injected contrast agents. Ultrasound disadvantages include operator dependence (a term used to describe the variable diagnostic performance of ultrasound in the hands of ultrasonographers of varying skill and experience) and limited ability to visualize the appendix in patients with significant fat stores or overlying bowel gas. Ultrasound is generally found to be quite specific (~93%-94% in multiple studies) but relatively insensitive (83%-88%) (Table 9-9).^135,139-141 In patients without appendicitis, the normal appendix can be difficult to identify, with as few as 2.4% of normal appendixes found in one study.¹⁴²

Table 9-9 Sensitivity and Specificity of Ultrasound for Appendicitis, Compared With Computed Tomography

Van Randen et al.¹⁴¹ performed a metaanalysis of studies that compared CT and ultrasound in the same population and found that CT performed more accurately, though the performance of both tests was quite dependent on the prevalence of disease. In patients with low pretest probability and a positive ultrasound or CT, the possibility of a false-positive test should be considered. Overall, CT had a positive likelihood ratio of 9.29, indicating a high probability of appendicitis with a positive CT. Ultrasound had a positive likelihood ratio of only 4.5, which may not be adequate to rule in disease in cases of low pretest probability (recall that a positive likelihood ratio greater than 10 is generally felt to increase the probability of disease to a clinically actionable level). Negative likelihood ratios for CT and ultrasound were 0.1 and 0.27, respectively. Recall that a negative likelihood ratio less than 0.1 is felt to exclude disease with a high degree of certainty, except in cases of very high pretest probability. Thus a negative CT rules out appendicitis in most patients, whereas a negative ultrasound does not.

Multiple authors have proposed the following use of ultrasound. In patients in whom appendicitis is clinically suspected but additional confirmation is desired, ultrasound should be the first diagnostic imaging test performed. If ultrasound confirms appendicitis, no further imaging may be required because of high specificity. If ultrasound does not confirm appendicitis, CT may be required, as a result of the poor sensitivity of ultrasound. Target populations in whom ultrasound use as the first diagnostic test should be encouraged are young patients and pregnant patients, arising from concerns about radiation exposures from CT. Thin patients of any age may be reasonable candidates for ultrasound. Conversely, ultrasound is likely to be nondiagnostic in obese patients, and CT may more reasonably be the first test. In older adults and the elderly, in whom radiation exposures from CT are of little concern and in whom multiple other serious abdominal conditions may be present, CT may be the more appropriate initial test.

Ultrasound for appendicitis is a focused examination. The examiner inspects the right lower quadrant, often starting with the location of tenderness identified by the patient. An abnormal appendix exceeds 6 mm in diameter and is incompressible. Sometimes a hypoechoic or heterogeneously echoic fluid collection representing abscess may also be identified. A normal appendix may be difficult to identify because of its small size. Figure 9-57 demonstrates ultrasound findings of appendicitis.

Magnetic Resonance Imaging for Appendicitis

MRI represents a third option for imaging of suspected appendicitis. MRI is generally not a first-line test because of expense and limited availability but is useful in patients with equivocal results of other modalities or in patients in whom radiation exposures are a particular concern, such as children and pregnant females. The normal appendix on MRI is less than or equal to 6 mm in diameter and may be filled with air and oral contrast material. MRI findings of appendicitis include an appendiceal diameter greater than 7 mm and the presence of periappendiceal inflammatory changes.¹⁴³ MRI has high reported sensitivity and specificity. Pedrosa et al.¹⁴³ reported MRI to be 100% sensitive and 93% specific in a consecutive series of 51 pregnant patients—though only 4 patients in this series had appendicitis. In another study, Pedrosa et al.¹⁴⁴ reviewed 148 consecutive pregnant patients evaluated for acute appendicitis, 140 of whom had both ultrasound and MRI. MRI was 100% sensitive (14 of 14) and 93% specific, although the small number of patients in this study results in wide CIs.

Israel et al.¹⁴⁵ found MRI to be 100% sensitive and specific for appendicitis when the appendix was visualized, though only four patients with appendicitis were diagnosed by MRI. MRI was unable to identify the appendix in a fifth patient with appendicitis.

Cobben et al.¹⁴⁶ found MRI to be 100% sensitive and specific for appendicitis in pregnancy—but in a series of only 12 patients, 3 with appendicitis.

Chabanova et al.¹⁴⁷ examined a fast, unenhanced MRI technique in 48 patients and found sensitivity of 83% to 93% and specificity of 50%-83%. This study was not restricted to pregnant patients.

Basaran and Basaran¹⁴⁸ performed a systematic review and compared the pooled sensitivity and specificity of CT and MRI, following a nondiagnostic ultrasound in a pregnant patient. CT was 87.5% sensitive and 97.4% specific, whereas MRI was 80% sensitive and 99% specific. Although the studies on MRI in pregnancy are limited, the ACR¹³⁴ rates MRI as seven on its nine-point appropriateness scale, just below ultrasound (with a score of eight). In children, the ACR gives MRI a rating of five of nine, below both ultrasound (eight of nine) and CT (seven of nine).

Appendicitis Decision Rules

Imaging in suspected appendicitis carries economic costs, radiation exposures (in the case of CT), and potential harms, including adverse effects of IV contrast administration for CT. In patients with acute appendicitis, therapeutic delays while awaiting diagnostic imaging could lead to perforation, with increased morbidity or even mortality. Therefore a clinical decision rule to stratify appendicitis risk could improve imaging utilization with cost, radiation, time, and morbidity benefits. Patients at low risk for appendicitis might require no imaging, only clinical observation for signs of worsening. Patients at high risk might require no imaging but, rather, empiric surgical management. Patients with intermediate risk might represent a group in whom imaging could provide benefit by preventing unnecessary appendectomy, reducing perforation rate, and identifying alternative diagnoses.

Several scoring systems have been proposed, though none are widely accepted at this time. The Alvarado score incorporates elements from the classic history, physical examination, and laboratory abnormalities associated with appendicitis (Table 9-10).¹⁴⁹ The mnemonic MANTRELS is sometimes used to indicate the criteria (see Table 9-10). In a retrospective validation study of 150 patients aged 7 or older, an Alvarado score of 7 or higher predicted appendicitis with 100% specificity, whereas a score of 3 or lower ruled out appendicitis without use of imaging (sensitivity = 96.2%). Patients with intermediate Alvarado scores (4-6) required further testing for appendicitis. Further study is required to narrow CIs.¹⁵⁰

Table 9-10 Alvarado Score for Appendicitis, Also Known by the MANTRELS Mnemonic^149-150

Alvarado Scores of 3 or lower rule out appendicitis (sensitivity = 96.2%). Scores of 7 or higher rule in appendicitis (specificity = 100%). Scores of 4 to 6 require imaging. This score requires further validation.

Adapted from Alvarado A: A practical score for the early diagnosis of acute appendicitis. Ann Emerg Med 1986;15:557-64; and McKay R, Shepherd J. The use of the clinical scoring system by Alvarado in the decision to perform computed tomography for acute appendicitis in the ED. Am J Emerg Med 25:489-493, 2007.

Other investigators have suggested that the Alvarado score is too nonspecific for clinical use without CT. Kim et al.¹⁵¹ prospectively scored 157 patients using the Alvarado score and a separate clinical impression by an emergency physician, in which patients were judged to have “clinically evident appendicitis” or not. Of 71 patients with clinically evident appendicitis, 19 did not have appendicitis (specificity = 71.6%). Of 52 patients with Alvarado scores greater than or equal to 8 (predicted high likelihood of appendicitis), 14 did not have appendicitis (specificity = 79%). CT was 95.5% specific, leading the authors to suggest that CT should be used even in patients with apparently high pretest probability to prevent negative appendectomy.

The Pediatric Appendicitis Score similarly incorporates classic features of appendicitis to develop a risk assessment (Table 9-11).¹⁵² In a validation study of 246 children with a 34% rate of appendicitis, the score performed well in reducing imaging, limiting the negative appendectomy rate, and avoiding missed appendicitis. A score of five or greater detected appendicitis with a sensitivity of 97.6%, meaning that children with scores of four or less need no imaging. A score greater than or equal to eight gave a specificity of 95.1%, indicating a need for appendectomy without imaging. Patients with scores of five to seven required imaging to exclude or confirm appendicitis. Using this rule would have eliminated 41% of imaging, achieved a negative appendectomy rate of 8.8%, and resulted in a missed appendicitis rate of 2.4%.¹⁵² Larger studies are required to validate these findings.

Table 9-11 Pediatric Appendicitis Score¹⁵²

A score of 4 or less excludes appendicitis (sensitivity = 97.6%, negative predictive value = 97.7%). A score of 8 or more confirms appendicitis (specificity = 95.1%, positive predictive value = 85.2%).

Scores of 5-7 require imaging. This scoring system requires further validation.

Adapted from Bhatt M, Joseph L, Ducharme FM, et al: Prospective validation of the pediatric appendicitis score in a Canadian pediatric emergency department. Acad Emerg Med 16:591-596, 2009.

Birkhahn et al.¹⁵³ prospectively derived an appendicitis likelihood model in 439 patients. Patients with a white blood cell count of less than 9500 x 10⁹/L and either no right lower quadrant tenderness or a neutrophil count of less than 54% were classified as low likelihood. Patients were classified as high likelihood if they had a white blood cell count greater than 13,000 x 10⁹/L and either rebound tenderness or voluntary guarding with a neutrophil count greater than 82%. The authors suggest that patients with neither high nor low likelihood undergo imaging, whereas patients with high and low likelihood undergo laparotomy and observation, respectively. Compared with actual clinical practice in the study population, use of this model would have reduced the number of missed cases of appendicitis from 10% to 1% and negative laparotomy from 14% to 5.6% while reducing imaging from 71% to 62%. This study was underpowered to prove statistical significance of these findings.

Clinical scoring systems have not yet achieved wide acceptance, partly because of disparate results among studies and lack of broad external validation of studies with promising results.

Numerous studies have documented the historical rate of negative appendectomy and perforation, correlated these with rising rates of CT utilization, and attempted to draw conclusions about the effect of CT on these important outcomes. Overall, most of these reports suggest a correlation between CT use and reduced rates of negative appendectomy in women of childbearing age, with lesser or no benefit in men or older women. Some studies suggest benefits in older adults. In children, some studies suggest lower rates of negative appendectomy with CT, whereas others suggest no benefit from CT. A handful of studies suggest lower perforation rates with CT use, though others show no difference. Most of the studies we review here (Table 9-12) suffer from the same methodologic flaw: they are not randomized, so patients undergoing CT may differ in important ways from patients not undergoing CT. As a consequence, the differences in outcomes observed between the two groups may have more to do with those initial differences than with the choice of diagnostic strategy. Consider this example before we examine the study findings: Patients with a classic presentation of appendicitis may be relatively unlikely to be subjected to CT by the treating physician, whereas patients with an atypical presentation may be more likely to be subjected to CT or other diagnostic tests. At the same time, classic-presentation patients may be more likely to have appendicitis (and therefore have a low rate of negative appendectomy) compared with patients with atypical presentations (who would therefore have a higher rate of negative appendectomy). If CT prevents unnecessary appendectomy in some patients with atypical presentations, the negative appendectomy rate in this group would decline but might remain equal to or higher than the rate in patients with classical presentations. Using a study design that compares the rates of negative appendectomy in patients with and without CT, it might appear that CT had no effect or benefit—when in truth it might have substantially decreased appendectomy in patients with equivocal presentations. With this in mind, consider some studies on CT and surgical outcomes.

Table 9-12 Studies of the Association Between CT Scan Use and Clinical Outcomes in Appendicitis

Coursey et al.¹⁵⁵ retrospectively reviewed records from 925 patients undergoing appendectomy from 1998 to 2007 and found a dramatic increase in CT use from 18.5% to 93.2%. At the same time, the overall negative appendectomy rate remained unchanged, whereas the negative appendectomy rate fell in women 45 years and younger from 42.9% to 7.1%.

Balthazar et al.¹⁵⁶ retrospectively compared negative appendectomy and perforation rates in patients undergoing preoperative CT with historical rates from previous publications. The negative appendectomy rate was 4%, compared with historical rates of 15% to 20%. The perforation rate was 22%, similar to previous publications. In women ages 15 to 50 years, the negative appendectomy rate was 8.3%, with a 19% perforation rate.

Rao et al.¹⁵⁷ compared rates of negative appendectomy and perforation from 1992 to 1995 (before introduction of appendiceal CT) and from 1997, when 59% of patients underwent preoperative CT. Before appendiceal CT was introduced, the negative appendectomy rate was 20%, falling to 7% during a period of intensive use of preoperative CT. Perforation rates fell from 22% to 14% during the same period. Negative appendectomy rates fell from 11% to 5% in men, from 35% to 11% in women, from 10% to 5% in boys, and from 18% to 12% in girls. The authors advocated preoperative CT in most female and many male patients.

Karakas et al.¹⁵⁸ retrospectively reviewed the charts of 633 children and adolescents with suspected appendicitis, noted an increased risk for perforation in patients undergoing CT, and suggested a causal relationship. However, this retrospective study cannot prove causation, and it may be that patients with perforation are more difficult to diagnose and therefore are more likely to undergo CT than patients with unperforated appendicitis.

Applegate et al.¹⁵⁹ retrospectively examined rates of perforation and negative appendectomy in children. Perforation rates were similar with preoperative CT, preoperative ultrasound, and no preoperative imaging. Negative appendectomy rates were lower in patients undergoing CT alone (2%) than in patients with CT with or without prior ultrasound (7%), ultrasound alone (17%), and no imaging (14%).

McDonald et al.¹⁶⁰ observed no statistically significant change in negative appendectomy or perforation rates with increasing rates of CT utilization. In a retrospective study, Bendeck et al.¹⁶¹ found the negative appendectomy rate to be significantly decreased, from 28% to 7%, in adult women undergoing CT compared with women undergoing no preoperative imaging, leading the authors to suggest that women undergo preoperative CT when appendicitis is suspected. Negative appendectomy rates were not significantly different in men or male and female children, with or without CT imaging.

In a retrospective review, Fuchs et al.¹⁶² noted a negative appendectomy rate of 11.9% in patients undergoing no CT versus 6.3% in patients undergoing preoperative CT. In women, the negative appendectomy rate was 23.5% without CT and 5.3% with preoperative CT.

In a retrospective study in 616 children, Partrick et al.¹⁶³ noted an increase in preoperative CT utilization from 1.3% in 1997 to 58% in 2001, with a nonsignificant decrease in negative appendectomy from 8% to 7%.

Torbati and Guss¹⁶⁴ prospectively observed 310 emergency department adult patients with suspected appendicitis. Patients with “clinically evident appendicitis” did not undergo CT, whereas those with equivocal presentations underwent CT. Comparison with historical controls was performed. In males, the perforation rate fell from 38% in historical controls to 25% with selective use of CT. In females, perforation rate fell from 23% historically to 6% with selective CT. This study also showed a decrease in time to operative intervention of nearly 5 hours in females with selective CT compared with historical controls.

Antevil et al.¹⁶⁵ retrospectively reviewed 633 adult patients undergoing appendectomy from 2000 to 2002, subdividing by age and gender. They found an association between CT use and lower rates of negative appendectomy in female patients of all ages and in patients of both genders older than 30 years (23% without CT vs. 8% with CT).

Vadeboncoeur et al.¹⁶⁶ compared the rate of negative appendectomy with selective CT use to a historical period before introduction of CT and found no significant difference (13.5% vs. 14.8%).

Chooi et al.¹⁶⁷ reviewed 380 patients undergoing appendectomy between 2000 and 2004 and found lower rates of negative appendectomy in both male and female patients undergoing (CT or ultrasound) imaging (11.4%) than in those without imaging (22.2%). The authors also reported a lower perforation rate in patients with preoperative imaging.

Riesenman et al.¹⁶⁸ compared emergency department length of stay and perforation rate in patients with and without CT for suspected appendicitis. Statistically, length of stay was significantly longer in patients with CT (606 vs. 321 minutes). Perforation rate was not significantly different between groups.

Randomized Controlled Trials on the Clinical Effect of Abdominal Computed Tomography for the Diagnosis of Appendicitis

A better research methodology for determining the effect of diagnostic imaging on patient outcomes is a randomized controlled trial. In trials of this type, randomization attempts to create two equivalent patient populations before the study intervention, the use or not of diagnostic imaging for appendicitis. Several trials of this type have been conducted, using slightly different comparisons. All of these studies suffer from a lack of statistical power, so their results may not accurately depict the likely effect of imaging on negative appendectomy rates, rupture, or other patient outcomes. Although existing studies do not lay to rest questions about the clinical benefit of CT, we examine the evidence because it may shed light on future studies to resolve this debate.

Hong et al.¹⁶⁹ randomized 182 patients with suspected appendicitis to clinical assessment alone or clinical assessment with mandatory abdominal CT. The authors found clinical assessment to be more sensitive (100%) than CT (91%) but less specific (73% for clinical assessment vs. 93% for CT). Perforation rates were low in both groups and not statistically different (6% for clinical assessment vs. 9% for CT). Hospital length of stay and costs were equal in the two groups. Patients assessed without CT were operated upon more quickly. The authors concluded that CT was not routinely warranted.

Although this study begins to apply strong methodology to this important clinical question, it may fail to reveal a benefit of CT because of the inclusion of all patients with suspected appendicitis, regardless of pretest probability. It stands to reason that CT would be unlikely to benefit patients with classic clinical presentations of appendicitis, because these patients likely have the disease process. CT in these patients would be expected to delay operative care, raise costs, and potentially increase perforation rates by delaying operative care. Because most patients in this group would be expected to have appendicitis, CT would be unlikely to reduce the negative appendectomy rate, which would be near zero. CT would also be expected to have little benefit in patients with very low clinical probability of appendicitis. In these patients, appendicitis would be expected to be rare, and few would undergo appendectomy on clinical grounds if CT were not used. CT might identify patients in this group with unexpected appendicitis, but this would be expected to be a rare event. Consequently, CT use would be expected to raise costs while having little effect on important outcomes such as perforation and negative appendectomy rates. CT would be most likely to have clinical benefit in patients with intermediate pretest probability of appendicitis. In this group, a relatively high rate of negative appendectomy might be expected without CT; a relatively high rate of perforation might also be expected if appendectomy were delayed until clinical signs progressed. Hong et al.¹⁶⁹ might have found clinical benefit to CT if the patient population had been selected in this way based on pretest probability.

Lee et al.¹⁷⁰ randomized 152 adult patients with right lower quadrant pain to selective CT or mandatory CT use. CT was performed in 68% of patients in the selective group and 97% of patients in the mandatory group. Mandatory CT imaging showed a trend toward decreased negative appendectomy (11% reduction) and perforation rates (8% reduction), not reaching statistical significance. A larger study is required to determine whether the observed trends represent a true benefit of mandatory CT.

Lopez et al.¹⁷¹ randomized 98 women of childbearing age with possible appendicitis to clinical observation alone or observation with CT scan. Charges and length of stay were similar in the two groups. Clinical observation alone was 100% sensitive and 87.5% specific, whereas CT was 89.5% sensitive and 95.6% specific. The study was underpowered to determine superiority of one method over the other. While acknowledging this limitation, the authors concluded that CT reliably identified patients requiring operation and appeared equivalent to observation. In addition to limitations because of poor statistical power, this study again suffers from inclusion of patients with very low and very high pretest probability of appendicitis, extremes at which CT would be predicted to have little benefit. The authors included patients with Alvarado scores of 2 to 8 (see Table 9-10). Alvarado scores of 7 or greater strongly predict appendicitis, whereas scores of 3 or less strongly exclude the diagnosis.¹⁵⁰ Inclusion of patients with these scores dilutes the possible benefit of CT in patients in whom true clinical uncertainty exists.

Appendicitis Imaging in the Pediatric Patient

Imaging for appendicitis in the pediatric patient warrants several special considerations. First, the differential diagnosis for acute abdominal pain is usually far more restricted than in adults. Often a binary question can be asked: Is the cause of the patient’s pain appendicitis? No further diagnostic testing is required when the answer is no. In contrast, a multitude of important causes in adults may simulate appendicitis, including ovarian pathology, obstructing renal stones, and vascular pathology. Although a broad differential diagnosis should be considered in pediatric patients, the requirement for an imaging test to fully evaluate all possible causes of abdominal pain is diminished. As a consequence, ultrasound, with its more limited ability to assess right lower quadrant abdominal pain, becomes a reasonable alternative to CT (see Figure 9-57).

Second, radiation exposures are a growing concern in pediatric patients, with risks varying depending on age at exposure. Brenner and Hall¹⁸⁵ estimate lifetime attributable mortality from a single abdominal CT to be approximately 1 in 700 if performed at birth, 1 in 1000 if performed at 5 years, and 1 in 2000 if performed at age 25. In contrast, the attributable risk is thought to be less than 1 in 5000 if CT is performed at 35 years, with a continued decrease with advancing age. Whereas these risks are small for the individual patient compared with baseline risks for cancer death in the United States (~22%, according to Centers for Disease Control and Prevention data from 2005),¹⁸⁶ on a population level these risks are estimated to translate into as many as 500 additional pediatric deaths annually in the United States.¹⁸⁷ Again, a diagnostic imaging test with no ionizing radiation exposure becomes more desirable; ultrasound and MRI meet this criterion.

Third, the CT diagnosis of appendicitis depends partly on the presence of intraperitoneal fat, the substrate for inflammatory fat stranding. Although CT has been shown to be relatively sensitive (84%) and highly specific (99%) for appendicitis in children,¹⁵⁸ some thin pediatric patients have little intraperitoneal fat, which may make diagnosis more difficult.¹⁷⁹ At the same time, a low body mass index typically improves the diagnostic performance of ultrasound.

Therefore a growing trend has emerged toward ultrasound as the first diagnostic imaging test for pediatric patients. As we discussed earlier, ultrasound is quite specific but relatively insensitive. Consequently, the ACR recommends that ultrasound be used as the first diagnostic test in children younger than 14 years, followed by CT for equivocal ultrasound results (e.g., nonvisualization of the appendix by ultrasound).¹³⁴ MRI is another option, although its expense and limited emergency availability make it a less viable strategy in many centers.¹³⁴

Imaging the Pregnant Patient With Possible Appendicitis

Pregnant patients with possible appendicitis pose another clinical diagnostic dilemma. Appendicitis is one of the most common surgical emergencies in pregnancy. Changes in the location of the appendix have been posited as causing atypical presentations in pregnant patients, although there is little supporting evidence in the medical literature. Leukocytosis is a normal finding in pregnancy, also complicating the diagnosis. Missed appendicitis can lead to fetal loss, so accurate diagnosis through the use of imaging is particularly desirable. As with the pediatric patient, minimizing radiation exposure in the pregnant patient is a priority. Consequently, alternatives to CT imaging using ultrasound or MRI are often sought.

Ultrasound is the first-line imaging modality in pregnant patients with suspected appendicitis. In pregnant patients, ultrasound has limited sensitivity (~50%) but is highly specific (100%) for appendicitis when the appendix is seen.¹⁴⁵ Ultrasound may fail to visualize the appendix in a large minority of pregnant patients, resulting in a nondiagnostic study. Shetty et al.¹⁸⁸ found that CT was 100% sensitive for appendicitis in pregnancy, compared with only 46% for ultrasound. Given the low sensitivity of ultrasound for appendicitis, further imaging is prudent in patients with negative or equivocal ultrasound.

Ironically, physicians generally overestimate the radiation risk from CT in pregnancy. The radiation dose from a single CT of the abdomen and pelvis is approximately 8 mSv, below the cumulative radiation exposure generally held to result in a measurable increase in risk for fetal anomaly or loss.^189-190 Despite this, physician knowledge lags behind, perhaps resulting in unnecessary fear of CT by physicians and patients. Ratnapalan et al.¹⁹⁰ surveyed family physicians and obstetricians on the risk for fetal anomaly following a single CT scan in pregnancy. In the survey, 61% of family physicians and 34% of obstetricians rated the likelihood of anomaly as greater than 5%—although the consensus among experts in this field is that no appreciable risk results from these exposures. This is not to say that CT should be used with abandon in pregnancy. Some evidence suggests that in utero radiation exposures might result in an increased risk for childhood leukemias, so if other imaging or diagnostic options are available, they should be considered. Ultrasound and MRI are reasonable options, with CT used when these are unavailable. Currently the ACR recommends right lower quadrant ultrasound, followed by MRI and CT when ultrasound is nondiagnostic for suspected appendicitis in pregnancy (Table 9-14).¹³⁴,¹⁹¹ Neither ultrasound nor MRI exposes the patient to radiation. Both ultrasound and MRI can evaluate for appendicitis and other causes of right lower quadrant or pelvic pain. Gadolinium safety in pregnancy is unknown; avoid gadolinium use when possible.

Table 9-14 American College of Radiology Appropriateness Criteria: Right Lower Quadrant Pain —Suspected Appendicitis, Fever, Leukocytosis, Pregnant Woman

This table is abbreviated from the complete ACR appropriateness criteria. Only the primary imaging recommendations (i.e., procedures rated as “usually appropriate”) and alternatives (i.e., procedure rated as “may be appropriate”) are included in this truncated list.Reprinted with permission of the American College of Radiology, Reston, VA. No other representation of this material is authorized without expressed, written permission from the American College of Radiology. Refer to the ACR website at www.acr.org/ac for the most current and complete version of the ACR Appropriateness Criteria®.”

From American College of Radiology: ACR appropriateness criteria: Right Lower Quadrant Pain — Suspected Appendicitis. 2010. Available at: http://www.acr.org/SecondaryMainMenuCategories/quality_safety/app_criteria/pdf/ExpertPanelonGastrointestinalImaging/RightLowerQuadrantPainDoc12.aspx.

Diverticulosis

Diverticulosis (see Figures 9-60 through 9-62) is the presence of diverticuli, small outpouchings of the bowel, most common in the colon. Risk factors include low-fiber diets, common in the United States and rare in the developing world. Diverticuli occur with increasing frequency in older patients but are present in patients in the United States as early as age 30. Diverticulosis occurs in 10% of U.S. adults younger than 40 and 50%-70% of adults 80 years or older.²⁰⁷ Thus diverticulosis is not a condition restricted to the elderly in the United States. Colonic diverticuli more commonly occur in the distal colon, where stool is more solid in consistency; the sigmoid and descending colon are the most frequent sites, though ascending and transverse colon diverticuli occur.²⁰⁷ Pseudodiverticuli, composed only of mucosa and submucosa with no muscular layer, can occur at entry sites of perforating arteries, which represent weak areas in the muscle wall.

Diverticuli are not seen on x-ray. Historically, contrast barium enema was used to diagnosis diverticuli, but today the diagnosis is more commonly made using CT in U.S. emergency departments. On contrast enema, the appearance is of small outpouchings, often with a narrow neck. A similar appearance is seen with CT scan, although no enteral contrast is needed to identify diverticuli. Imaging of diverticular bleeding is discussed in the later section on GI hemorrhage.

Diverticulitis

When food particles become entrapped within diverticuli, infection and inflammation of diverticuli may occur, termed diverticulitis (see Figures 9-62 through 9-66). Mirroring the locations of diverticuli, diverticulitis occurs in the descending and sigmoid colon in 90% of cases. CT is the preferred imaging modality for diagnosis of diverticulitis, because it can display inflammatory change, complications of diverticulitis such as perforation or abscess formation, and alternative diagnoses including ischemic or infectious colitis, cancers, ureteral stones, and aortic aneurysm. Ultrasound and MRI are alternatives, though studies are limited. X-ray plays little role in the diagnosis. The ACR appropriateness criteria for imaging of suspected diverticulitis are shown in Table 9-15.²⁰⁸

Table 9-15 American College of Radiology Appropriateness Criteria: Left Lower Quadrant Pain, Older Patient With Typical Clinical Presentation for Diverticulitis

The full ACR Appropriateness Criteria document includes other clinical variations with additional imaging recommendations.Reprinted with permission of the American College of Radiology, Reston, VA. No other representation of this material is authorized without expressed, written permission from the American College of Radiology. Refer to the ACR website at www.acr.org/ac for the most current and complete version of the ACR Appropriateness Criteria®.

From American College of Radiology: ACR appropriateness criteria: Left lower quadrant pain, 2008. (Accessed at http://www.acr.org/SecondaryMainMenuCategories/quality_safety/app_criteria/pdf/ExpertPanelonGastrointestinalImaging/LeftLowerQuadrantPainDoc8.aspx

Sensitivity of Diagnostic Modalities for Inflammatory Bowel Disease

The sensitivity and specificity of CT, MRI, ultrasound, and nuclear scintigraphy have been compared in a metaanalysis of prospective studies. Horsthuis et al.²³⁵ evaluated 33 studies and found similar diagnostic performance among the modalities on a per-patient basis, as opposed to a per-bowel-segment diagnostic basis (Table 9-18). The authors suggest that MRI and ultrasound be considered for disease monitoring, because these modalities do not result in ionizing radiation exposure.

Table 9-18 Diagnostic Performance of Modalities for Inflammatory Bowel Disease

Other authors have found better sensitivity of dedicated CT techniques such as CT enteroclysis, though the number of included patients is small. Boudiaf et al.²³⁶ reported CT enteroclysis to have a sensitivity of 100% and specificity of 95% for small-bowel diseases including small-bowel masses, Crohn’s disease, small-bowel tuberculosis, small-bowel lymphoma, and low-grade small-bowel obstruction. However, the total number of patients in this study was only 107 and represented a range of disease processes, so the sensitivity and specificity of CT for each disease process have wide CIs. In small studies, CT enterography has been shown to correlate with disease progression, may reveal clinically unsuspected bowel strictures, and may rule out clinically suspected strictures.^237-238

Computed Tomography Findings of Inflammatory Bowel Disease

Figures 9-67 through 9-70 demonstrate CT changes of inflammatory bowel disease. On CT scan, hallmarks of bowel inflammation include bowel wall thickening (often visible with no contrast agents) and surrounding fat stranding (also visible without contrast agents). Bodily et al.²³⁹ found that increased wall thickness and increased mural attenuation correlated with disease activity measured by endoscopy and histology. Lymphadenopathy may be seen. Findings such as free abdominal fluid may be present and should raise concerns about perforation. Inspection for extraluminal air should also be performed. If IV contrast is administered, the bowel wall should be inspected for increased enhancement, consistent with inflammation, infection, or ischemia, and decreased enhancement—which suggests bowel ischemia. Mural pneumatosis suggests regional bowel gangrene, which can occur from ischemia or infection. Occasionally, fistula formation may be visible without enteral contrast, but enteral contrast can assist in the diagnosis by demonstrating extraluminal contrast leak via the fistula. Abscess formation can occur with inflammatory bowel disease and appears the same as described in the earlier section on diverticulitis or in the later section on abscesses.

The “fat halo” sign is a finding of submucosal fat deposition in the bowel wall, associated with Crohn’s disease. A density of less than −10 Hounsfield units is diagnostic. The ileum and ascending colon are the most common locations. It is not a marker of acute inflammation but is thought to occur with increasing duration of disease. In one study, the fat halo sign was seen in only 3.8% of patients with disease duration of less than 1 year and in 21.9% of patients with longer duration of disease.²⁴⁰ Unfortunately, the fat halo sign may also be seen in obese patients with no apparent inflammatory bowel disease and is thus nonspecific.²⁴¹

Differentiating the cause of bowel abnormalities by CT criteria may be difficult or impossible, and clinical history or other testing such as microbiologic or endoscopic findings may be required. Some clues from CT can narrow the differential diagnosis (Table 9-19). First, diverticuli should be sought; their absence generally rules out diverticulitis. However, sometimes severe inflammation around diverticuli may mask their presence. If previous CT or other imaging from the patient is available, it may be useful to review that imaging for the presence of diverticuli. The bowel should be followed sequentially from the rectum proximally, including both the small and the large bowel, for signs of inflammation. Ulcerative colitis should demonstrate only contiguous regions of inflammation starting at the rectum and moving proximally. In contrast, Crohn’s disease is well known for discontiguous regions of inflammation (sometimes called “skip lesions”), which can involve any part of the intestine from mouth to anus. Vascular disease causing ischemic bowel can appear identical to other inflammatory or infectious abnormalities of bowel; it is discussed in detail in Chapter 11. However, clues to a vascular cause include a region of inflammation or bowel abnormality matching a common vascular distribution. Inspection of blood vessels for filling defects can allow direct detection of stenosis or thrombosis; this requires IV contrast administration. Beware the possibility of bowel ischemia; a noncontrast CT cannot fully discriminate bowel ischemia and inflammatory or infectious bowel disease. Neoplastic disease can result in bowel wall abnormalities and inflammation, sometimes mimicking infection or benign inflammatory bowel disease. Inspection of solid organs may reveal lesions consistent with metastatic disease, suggesting a malignant cause of the bowel abnormalities. Lymphadenopathy can occur with infectious, inflammatory, or malignant disease. Radiation colitis typically causes circumferential bowel wall thickening and adjacent fat stranding, but the appearance is nonspecific. Graft-versus-host disease in patients with a history of bone marrow transplant can result in a similar appearance. ²⁴²

Table 9-19 CT Findings Suggesting Inflammatory, Infectious, or Ischemic Colitis, or Colonic Neoplasm

Infectious enteritis or colitis, including C. difficile colitis, can appear similar to other forms of inflammatory or ischemic disease and should be considered.²⁴² Fishman et al.²⁴³ described the CT findings of C. difficile pseudomembranous colitis, including significant bowel wall thickening with a mean thickness of 14.7 mm, and enteric contrast trapped between broad transverse tissue bands. The entire colon or localized segments can be involved. An accordion appearance can be seen, in which the colon wall is so thickened that only a thin string of contrast is seen within the lumen between abutting walls. The appearance is not specific and must be correlated with the clinical presentation. Toxic megacolon from C. difficile or other causes can be evaluated with CT. Perforation and septic thrombosis of the portal system complicating toxic megacolon can be identified with CT.²⁴⁴

Bowel disease can be further evaluated with a variety of modalities when CT does not adequately differentiate disease. Angiography can further explore bowel ischemia. Upper GI or lower GI contrast studies, including enteroclysis and contrast enema, can assist in identifying changes of inflammatory bowel disease. Direct endoscopic visualization can also assist when imaging is nondiagnostic. MRI, ultrasound, and nuclear scintigraphy provide other options but are rarely indicated in the emergency department.

Ultrasound of Cholelithiasis and Cholecystitis

Ultrasound is commonly used for the diagnosis of cholelithiasis and cholecystitis and is increasingly performed by emergency physicians. The test is portable, requires no radiation exposure, and can rapidly provide diagnostic information. Ultrasound findings of cholecystitis must be differentiated from those of cholelithiasis—because the presence of gallstones is not sufficient for the diagnosis of cholecystitis. Although gallstones are present in most cases of cholecystitis, acalculous cholecystitis can occur in critically ill patients, particularly following burns or trauma, during sepsis, and while receiving total parenteral nutrition.²⁵⁰ Currently ultrasound is recommended by the ACR as the most appropriate imaging test in patients with right upper quadrant pain.²⁵¹

Ultrasound Findings of Cholelithiasis

With ultrasound, gallstones are visible as hyperechoic (bright white) structures within the gallbladder, casting dark acoustic shadows (see Figures 9-73 through 9-87). Gallstones are typically mobile (see Figure 9-80) unless lodged in the gallbladder neck (see Figure 9-83), cystic duct, or common bile duct—distinguishing them from gallbladder polyps, which are immobile and do not cast acoustic shadows (see Figure 9-84). Gallstone mobility can be assessed by rolling the patient during the ultrasound examination to observe any resulting motion of gallbladder contents. Smaller gallstones may be as small as a grain of sand or even finer particles—leading to terms such as “gallbladder sand” and “sludge.” Even at these small sizes, gallstones remain hyperechoic and are mobile. Gallbladder sludge forms a dependent layer within the gallbladder but may not be so dense as to cast an acoustic shadow. When the gallbladder is completely filled with stones, an appearance called the wall–echo–shadow sign may be seen. In this case, the ultrasound beam reflects off of stones immediately deep to the superficial gallbladder wall. Deeper structures are not seen. This appearance can be similar to that seen with a heavily calcified gallbladder (porcelain gallbladder; Figure 9-111), which is associated with gallbladder carcinoma. In one recent study, this association has been found to be less strong than previously described.²⁵² In addition, in gangrenous cholecystitis, air within the gallbladder wall creates a bright echo and posterior shadowing.²⁵³

Figure 9-111 Porcelain gallbladder versus air in gallbladder wall, ultrasound.

Porcelain gallbladder, a finding concerning for carcinoma of the gallbladder, can be recognized on ultrasound by a highly echogenic anterior wall of the gallbladder with posterior acoustic shadowing. This patient has findings that do not quite meet this definition. The anterior wall is highly echogenic, but the posterior wall is visible. With a true porcelain gallbladder, the densely echogenic anterior gallbladder wall would create a dense shadow, preventing visualization of the posterior wall. Other possible causes of this finding in this patient are air within the gallbladder wall (which scatters and reflects the ultrasound signal) or confluence of “comet tail artifact” from multiple adjacent cholesterol crystals. In this case, the patient had severe pulmonary disease and was not an operative candidate, so no pathology report is available. His CT suggested acute cholecystitis, and he was treated successfully with percutaneous drainage and antibiotics.

Findings of cholecystitis include gallbladder wall thickening (>3-4 mm), pericholecystic fluid, and sonographic Murphy’s sign.^253-254 The normal gallbladder wall is a pencil-lead-thin (2-3 mm) hyperechoic line (see Figures 9-73 and 9-85). With inflammatory change, the wall thickens and may take on a striated appearance (see Figure 9-86). Thickening may occur in other disease states besides cholecystitis, including cirrhosis, viral hepatitis, heart failure, hypoalbuminemia, and chronic renal failure.²⁵⁵ Pericholecystic fluid appears hypoechoic (black) on ultrasound (see Figure 9-87). The finding is nonspecific in isolation, because fluid may be present in the gallbladder fossa from other processes including diffuse ascites. A sonographic Murphy’s sign is the presence of tenderness to direct palpation over the gallbladder with the ultrasound probe. The sensitivity of this finding in isolation for acute cholecystitis is only around 86%, with 35% specificity.²⁵⁶ Some radiologists decline to comment on the sonographic Murphy’s sign in a patient who has received analgesics. An enlarged gallbladder is another nonspecific sign of possible cholecystitis. The normal size of the gallbladder is 7-10 cm in length and 2-3 cm in width. A transverse diameter greater than 5 cm suggests cholecystitis.²⁵⁴

Sensitivity and Specificity of Ultrasound and Nuclear Scintigraphy for Acute Cholecystitis

Numerous studies have evaluated the sensitivity and specificity of ultrasound and nuclear scintigraphy (described shortly) for cholelithiasis and acute cholecystitis. Most studies suggest that ultrasound is highly sensitive and specific for gallstones. It is sensitive but less specific for cholecystitis. Nuclear scintigraphy is considered more sensitive and specific than ultrasonography for acute cholecystitis—but is more costly and time-consuming and exposes the patient to ionizing radiation.²⁵¹ Shea et al.²⁵⁷ performed a metaanalysis of studies of ultrasound and nuclear scintigraphy and corrected the calculated sensitivity and specificity for a common methodologic bias: verification bias. This error occurs when a gold standard test is not performed in all patients to confirm the findings of the test whose performance is being evaluated. This can result in inaccurate calculation of sensitivity and specificity, because the true presence or absence of disease is not confirmed. The authors found that the sensitivity of ultrasound and nuclear scintigraphy were likely lower than previously reported, when corrected for verification bias (Table 9-20). They recommended that a sensitivity midway between corrected and uncorrected values be assumed.

Table 9-20 Sensitivity and Specificity of Ultrasound and Nuclear Scintigraphy for Cholelithiasis and Cholecystitis

False-positive ultrasound findings can occur in a number of settings. Gallbladder wall thickening can occur postprandially because of normal gallbladder contraction.²⁵⁴ Consequently, some centers require that the patient fast at least 4 hours before gallbladder ultrasound. As previously mentioned, pericholecystic fluid may represent fluid within the abdomen from another source such as ascites, rather than inflammatory fluid related to gallbladder pathology.

Nuclear Scintigraphy (Hepatobiliary Iminodiacetic Acid Scan) of Cholecystitis and Biliary Obstruction

Hepatobiliary iminodiacetic acid (HIDA) nuclear scintigraphy (also called cholescintigraphy, hepatobiliary scintigraphy, or a hepatobiliary scan) represents another means of assessing the gallbladder and biliary tree for pathology. In this test, HIDA labeled with a radioactive tracer (usually technetium-99m) is injected intravenously, is taken up by the liver, and is excreted into the bile. It then travels in bile from the liver through the biliary ducts to the gallbladder and on to the small intestine. A series of images is captured over time with a gamma camera, recording the sequence of passage of bile. In a normal HIDA scan, the liver becomes visible first, followed by the hepatic duct. The gallbladder then becomes visible, because it fills retrograde through the cystic duct. Filling of the common bile duct also occurs, followed by transit of bile through the Sphincter of Oddi into the small bowel. Patients are usually required to fast for 2 hours before cholescintigraphy.

In cases of obstruction of biliary ducts, portions of the biliary system distal to the obstruction are not visualized, because HIDA-labeled bile does not pass the point of obstruction. If the cystic duct or gallbladder neck becomes obstructed by a stone, as is common in cases of cholecystitis, the gallbladder is not visualized. In cases of choledocholithiasis, distal portions of the common bile duct and small bowel do not receive HIDA-labeled bile and are not visualized. An abnormal HIDA scan is not specific for biliary obstruction by a stone, because obstruction by a mass or stricture would have a similar appearance.

In some cases of cholecystitis or gallbladder dysfunction leading to chronic abdominal pain, the gallbladder fills normally but does not contract normally. When the gallbladder is visualized on cholescintigraphy but doubt remains about the function of the gallbladder, cholecystokinin can be administered intravenously and the response observed. Usually brisk emptying of the gallbladder is observed, with subsequent nuclear scintigraphy images showing an absence of the gallbladder in the image after HIDA-laden bile is expelled. In the case of an abnormal gallbladder, the gallbladder image remains on images acquired after administration of cholecystokinin, because HIDA-labeled bile remains within the gallbladder. Quantitative measurement of the gallbladder ejection fraction can be performed; an ejection fraction less than 50% after cholecystokinin is abnormal.²⁵⁸ Figures 9-88 through 9-90 demonstrate normal and abnormal HIDA studies.

HIDA–nuclear scintigraphy can be used to assess for conditions other than cholecystitis, biliary obstruction, and gallbladder dysfunction. Because the test traces the route of bile excreted from the liver, it can be used to assess for bile leaks, as may occur as a complication of hepatobiliary surgery. If radiotracer accumulates in an unexpected location, a leak can be diagnosed (see Figure 9-101). If tracer accumulates locally, a biloma may be diagnosed. If tracer is distributed diffusely in the abdomen, a bile leak into the peritoneal cavity is present. HIDA scintigraphy can also diagnose iatrogenic obstruction, such as inadvertent clipping of the common bile duct during surgery. Because the method records scintigraphic emission from bile, it provides no specific information about the cause of the obstruction (tumor, stone, stricture, or surgical clip), because the body tissues themselves are not imaged.

The high sensitivity and specificity of HIDA cholescintigraphy is achieved with imaging up to 4 hours after radiotracer administration, an important limitation to use in the emergency department. Sensitivity and specificity are decreased to only 80% to 88% when imaging is continued for only 1 hour. Administration of morphine can be performed if the gallbladder is not seen by 40 to 60 minutes after administration of radiotracer and if tracer is seen in the small bowel. Morphine is thought to cause some constriction of the sphincter of Oddi, enhancing retrograde filling of the gallbladder when the cystic duct is patent. Cholescintigraphy with morphine augmentation in this manner has been shown to achieve high sensitivity (94.6%) and specificity (99.1%) with only 1.5 hours of imaging.²⁵⁹

The sensitivity and specificity of HIDA cholescintigraphy has been studied extensively (see Table 9-20).^251,257 It remains a second-line test to ultrasound in most circumstances, because it may take 2 to 4 hours to complete, exposes the patient to ionizing radiation, and is more expensive than ultrasound. In addition, it does not directly visualize organs, unlike ultrasound and CT, and so may be less useful in evaluating for alternative pathology. Although metaanalysis data suggests high specificity, some studies have suggested very low specificity of HIDA scans, as low as 38%.²⁶⁰

Computed Tomography for Cholelithiasis and Acute Cholecystitis

CT is generally not considered the first-line test for cholecystitis and related biliary disorders, because ultrasound is well-studied and offers an imaging option free from radiation or the use of potentially nephrotoxic contrast agents. However, CT is frequently used for the imaging of nonspecific abdominal pain and sometimes reveals biliary pathology as the source of pain. CT can disclose the presence of gallstones, although many gallstones are isodense with bile by CT and are not seen. Normal bile has a density somewhat greater than water (~20 Hounsfield units), and appears dark gray on a CT soft-tissue window. Bile does not normally enhance with IV (or oral) contrast. The gallbladder wall is normally a pencil-thin line and isodense or slightly denser than bile, thus appearing as a slightly brighter line around the gallbladder. The normal gallbladder fossa does not contain fluid. Gallstones may be visible as hyperdense (brighter) rounded or irregular structures in a dependent position within the gallbladder. Figures 9-92 through 9-94 demonstrate gallstones seen with CT.

Fidler et al.²⁶¹ reported the CT findings of acute cholecystitis in 29 patients with proven disease (Table 9-21). Gallbladder wall thickening may be seen as thickness greater than 4 mm. The gallbladder may appear distended. An abnormal gallbladder wall may enhance with IV contrast, appearing bright. A gangrenous gallbladder may fail to enhance. Pericholecystic fluid may be visible within the gallbladder fossa and appears an intermediate gray, around 0 to 20 Hounsfield units. CT cannot assess for the presence of a Murphy’s sign. Complications of gangrenous cholecystitis, including gas in the gallbladder, gallbladder wall, or gallbladder fossa, may be seen. Inflammatory fat stranding may be seen surrounding the gallbladder. In some cases, an abscess in the gallbladder fossa may be seen, with rim enhancement, air, and fluid.²⁶² Figures 9-95 through 9-98 demonstrate CT findings of cholecystitis.

Table 9-21 CT Findings of Acute Cholecystitis

Additional complications such as ascending cholangitis are discussed later.

X-ray for Gallstones and Biliary Disease

Some gallstones are visible on x-ray (see Figure 9-72), but many are not, and the sensitivity and specificity of x-ray for clinically important conditions such as cholecystitis are too low for clinical utility. In some complicated cases such as ascending cholangitis, air may be seen on x-ray overlying the gallbladder fossa, or branching over the liver, suggesting biliary duct or portal venous gas. Generally, x-rays should not be obtained for assessment of suspected biliary pathology. Rothrock et al.²⁴⁷ found that major abnormalities such as pneumobilia were extremely rare in patients with biliary disease—and did not occur in their series of patients. Minor findings possibly suggesting biliary disease (e.g., right upper quadrant calcifications, ileus, and right basilar atelectasis) occurred with equal frequency in patients with and without acute biliary disease—making them useless to differentiate biliary and nonbiliary sources of symptoms.

Contrast-Biliary Imaging for Gallstones

Contrast biliary studies using fluoroscopy and contrast agents injected directly into the biliary tract represented the first imaging method to visualize the biliary system—before the advent of ultrasound, CT, or MRI. Today the modality is primarily used for interventions such as placement of percutaneous biliary drains or stents under image guidance (see Figure 9-103). The ACR does not generally recommend these studies, except in hospitalized patients in whom invasive cholangiography with percutaneous cholecystostomy can be both diagnostic and therapeutic.²⁵¹

Magnetic Resonance Cholangiopancreatography for Biliary Obstruction and Cholecystitis

Magnetic resonance cholangiopancreatography (MRCP) allows visualization of the gallbladder, biliary ducts, and pancreas. The method can detect obstruction and can often reveal the cause, including gallstone, extrinsic compression from mass, and intrinsic obstruction from neoplasm. This technique is used more often in the diagnosis of painless jaundice from presumed pancreatic mass than in the diagnosis of cholecystitis, because less expensive and more available alternatives exist.²⁶⁴

MRI has been compared with ultrasound for the diagnosis of acute cholecystitis. Hakansson et al.²⁶⁵ studied 35 patients with both MRI and ultrasound and found that MRI was more sensitive. Loud et al.²⁶⁶ also reported that MRI was useful in the preoperative evaluation of 14 patients with suspected acute cholecystitis, though no comparison with other modalities was made.

Acalculous Cholecystitis

Acalculous cholecystitis can be a difficult diagnosis, with nonspecific imaging findings. Large prospective studies do not exist because of the rarity of the condition. Mirvis et al.²⁶⁰ retrospectively reported findings in 56 patients during a 6-year period. Ultrasound and CT were sensitive (92% and 100%, respectively) and specific (96% and 100%, respectively), whereas cholescintigraphy with HIDA was only 38% specific. Findings were similar to those of gallstone-associated cholecystitis, but with an absence of stones.

Choledocholithiasis and Biliary Ductal Obstruction

Choledocholithiasis is the presence of bile stones within the biliary ducts. It can be assessed with several modalities, including ultrasound, magnetic resonance, CT, cholescintigraphy, and contrast fluoroscopy.

Ascending Cholangitis

Ascending cholangitis is a life-threatening infection of the biliary ducts, usually in the context of biliary ductal obstruction. Imaging plays both direct and indirect roles in the diagnosis. The primary role of imaging is to rule out other causes of symptoms such as abdominal pain, tenderness, jaundice, and fever. In addition, imaging may reveal biliary ductal obstruction from an intrinsic process, such as gallstone, or extrinsic obstruction from a compressing mass or abscess. In some cases, imaging may reveal more direct evidence of ascending cholangitis such as gas within the biliary ducts or portal venous system, caused by the presence of gas-forming anaerobic organisms. Direct evidence of ascending cholangitis may not be present; the emergency physician may need to infer the presence of ascending cholangitis from imaging findings of biliary obstruction, in the presence of other clinical findings such as fever, jaundice, altered mental status, or hypotension. Treatment requires decompression of obstruction, which can be performed surgically, endoscopically by endoscopic retrograde cholangiopancreatography, or by percutaneous biliary drainage under ultrasound or fluoroscopic guidance. Thus imaging plays a role in both diagnosis and treatment. Treatment need not relieve the actual site of pathologic obstruction; instead, as a temporizing measure, percutaneous drainage of the biliary system can suffice.

Computed Tomography Findings of the Normal Pancreas and Pancreatitis

The normal pancreas is an intermediate gray band of soft-tissue density (~+40 Hounsfield units) in the upper abdomen (see Figures 9-112 and 9-113). The pancreatic head is located just right of the midline, with the pancreatic body crossing to the patient’s left anterior to the abdominal aorta. The pancreatic tail terminates in the left abdomen medial to the spleen and anterior to the left kidney. The surface of the normal pancreas has a frondlike or crenellated appearance. The surrounding fat is quite black, with no inflammatory change. No free fluid is seen around the normal pancreas. When IV contrast is administered, the normal pancreas enhances uniformly. Oral contrast is not needed to assess the pancreas.

CT findings of pancreatitis include peripancreatic free fluid, peripancreatic fat stranding, and edema of the pancreas, with loss of the normal crenellated appearance (see Figures 9-114 through 9-116).²⁸⁴ The adjacent fat appears smoky or hazy, and the borders of the pancreas become indistinct. A pancreatic mass or pseudocyst may also be seen (see Figures 9-117 through 9-119). Pancreatitis can be recognized by CT without oral or IV contrast. However, if IV contrast is given, areas of necrosis will fail to enhance and remain dark, whereas less-affected areas will enhance and appear brighter.²⁸⁵ The demarcation between areas of normal enhancement and necrotic unenhancing regions is often sharp.²⁸⁴ The pancreatic head is often spared because of a redundant vascular supply.²⁸⁴ Several investigators have shown a correlation between lack of pancreatic parenchymal enhancement on CT with IV contrast and surgically proven pancreatic necrosis.^285-286 Areas of necrosis identified by CT do not recover enhancement but rather eventually resorb in survivors.²⁸⁷ Mesenteric edema surrounding blood vessels creates a low-attenuation cuff around the contrast-enhanced blood vessel. A pancreatic pseudocyst may be seen as a well-circumscribed area of hypodensity (typically close to 0 Hounsfield units, or dark gray by CT with a soft-tissue window) within or immediately adjacent to the pancreas, often with an enhancing rim. This may be difficult to differentiate from abscess, although abscesses may be more heterogeneous, sometimes containing air, as well as fluid and semi-solid debris. Patients with extrapancreatitic phlegmon and pancreatic necrosis are at high risk for complications and poor clinical outcomes.²⁸⁸ CT severity of pancreatitis has been shown to be more predictive of clinical outcomes than are clinical scoring systems such as Ranson’s criteria in some studies, although this remains an area of research (described in detail later).^289-290

Clinical and Computed Tomography Grading Systems for Pancreatitis

Ranson’s criteria for pancreatitis are a clinical grading scale that correlates with mortality—though validation of these criteria in subsequent trials shows less impressive predictive power than some proposed CT systems. Nonetheless, it is interesting to observe a qualitative relationship among Ranson’s criteria, CT findings, and the pathophysiology of pancreatitis (Table 9-22). Ranson’s criteria include several clinical parameters that may have visible CT analogues.²⁹³

Table 9-22 Ranson’s Criteria for Pancreatitis and Related CT Findings

A number of grading systems for the severity of pancreatitis have been developed, based on CT criteria. Increasing severity by CT predicts increasing mortality. A direct relationship between CT findings and therapeutic interventions has not been developed, so although CT may have prognostic significance, the clinical benefit to patients is not proved. Presumably, additional information about the severity of pancreatitis would improve clinical care by allowing treatment of complications such as abscess. We examine some of the studies on CT severity scores here.

Balthazar et al.²⁹⁴ graded 83 patients with acute pancreatitis based on CT findings and correlated CT severity with clinical outcome (Table 9-23). Length of hospitalization correlated positively with CT severity. Abscesses occurred in 21.6% of all patients, no patients with grades A and B, and 60.0% of patients with severe pancreatitis (grade E). Pleural effusions were more common in grade E patients. No patients with CT grade A or B died. The authors concluded that early CT had prognostic value for morbidity and mortality. Whether early CT results in improved care has not been studied and would require a prospective study randomizing patients to CT or no CT to determine the effect on outcome.

Table 9-23 CT Pancreatitis Severity Index, Which Combines the Balthazar Grading System With Additional Points for Degree of Pancreatic Necrosis

Subsequently, Balthazar et al.²⁹⁵ correlated the degree of pancreatitic necrosis identified with CT in 88 acute pancreatitis patients with morbidity and mortality. Based on these findings, they developed a CT severity index to predict morbidity and mortality (see Table 9-23).

Clavien et al.²⁹⁶ compared early CT findings (within 36 hours) and clinical scores to clinical course in 202 patients with acute pancreatitis. The authors classified CT findings in three categories and clinical scores as three groups, which correlated with mortality (Table 9-24). In patients with CT grade III and clinical score 2 or 3, mortality was particularly high: 32% or 58%, respectively. Complications developed in all survivors in these two groups, and all but two pancreatic abscesses developed in patients with CT grade III. The authors suggested that early CT be performed in all patients with a clinical score of 2 or 3, given the high mortality in these groups. This study did not explicitly examine any therapeutic inventions performed in response to CT findings, so it is unproven that the additional information provided by CT would alter management or improve mortality.

Table 9-24 CT Findings and Clinical Scores in Pancreatitis, Compared With Mortality

London et al.²⁹⁷ studied 126 patients with acute pancreatitis undergoing contrast-enhanced abdominal CT within 72 hours of presentation. Several CT criteria including pancreatic enhancement, loss of peripancreatic tissue planes, and a pancreatic size index (calculated by multiplying the maximum AP diameter of the pancreatic head and body) correlated with severity of pancreatitis. A pancre atic size index of at least 10 cm² was 71% sensitive and 77% specific for a severe attack—but was less predictive than the modified Glasgow criteria (similar to Ranson’s Criteria), which were 85% sensitive and 79% specific.

Basterra et al.²⁹⁸ compared Ranson’s criteria with Balthazar’s CT criteria in 100 consecutive patients admitted with pancreatitis and found CT criteria to have superior sensitivity and specificity in predicting outcomes. Pancreatic necrosis was most predictive of outcome.

Ju et al.³⁰⁰ compared clinical scores (Ranson’s criteria) with three CT scoring systems: the Balthazar ABCD score, the Balthazar CT severity index (see Table 9-23), and London’s pancreatic size index. Ranson’s criteria and the pancreatic size index performed best, with the pancreatic size index being more predictive of local complications. Earlier investigators had found little additional prognostic benefit to CT compared with clinical scores.³⁰¹

Rickes et al.³⁰² reported the sensitivity and specificity of contrast-enhanced ultrasound (using an IV sonography contrast agent) for detection of severe pancreatitis, defined by a Balthazar score of D or E or CT evidence of pancreatic necrosis. Compared with CT, ultrasound was 82% sensitive and 89% specific.