Morphine-Modified Hepatobiliary Scanning Protocol for the Diagnosis of Acute Cholecystitis

Nuclear Medicine and Molecular Imaging Original Research Solomon et al. Imaging Diagnosis of Acute Cholecystitis Nuclear Medicine and Molecular Imaging Original Research Robert W. Solomon 1,2 Abraham Albert Harari 3,4 Robert Dragotti 3,5 Matthew D Alessandro 3,6 Nageswara Mandava 3 Solomon RW, Harari AA, Dragotti R, D Alessandro M, Mandava N Keywords: acute cholecystitis, hepatobiliary imaging, hepatoiminodiacetic acid (HIDA), HIDA scanning, mebrofenin, morphine sulfate DOI:10.2214/AJR.15.15742 Received October 6, 2015; accepted after revision March 13, 2016. Based on a presentation at the ARRS 2015 Annual Meeting, Toronto, ON, Canada. 1 Department of Radiology, Flushing Hospital Medical Center, 4500 Parsons Blvd, Flushing, NY 11355. Address correspondence to R. W. Solomon (rsolomon@jhmc.org). 2 Department of Radiology, Jamaica Hospital Medical Center, Jamaica, NY. 3 Department of Surgery, Flushing Hospital Medical Center, Flushing, NY. 4 Present address: Department of Radiology, Academic Division, University of Colorado, Aurora, CO. 5 Present address: Anderson Sobel Cosmetic Surgery, Bellevue, WA. 6 Present address: Department of Surgery, Staten Island University Hospital, Staten Island, NY. AJR 2016; 207:865 870 0361 803X/16/2074 865 American Roentgen Ray Society Morphine-Modified Hepatobiliary Scanning Protocol for the Diagnosis of Acute Cholecystitis OBJECTIVE. We report a morphine-modified hepatoiminodiacetic acid (HIDA) scanning protocol that uses 2 mg of morphine IV push at the bedside as a pretreatment. We compared this protocol with the original HIDA scanning protocol, which included delayed imaging for up to 4 hours without the use of morphine. Moreover, we contrast our results with the results of studies in the literature. MATERIALS AND METHODS. We retrospectively reviewed the charts of inpatients who underwent HIDA scanning for the diagnosis of acute cholecystitis between 2003 and 2013. The study group consisted of 374 HIDA studies of 365 patients who received 2 mg of morphine IV push at bedside and then underwent dynamic imaging for 1 hour using 222 MBq of 99m Tcmebrofenin. No delayed images were obtained. The control group consisted of 232 studies of 227 patients who underwent conventional HIDA scanning using our standard protocol with delayed imaging and without morphine. Either strict pathologic criteria or the results of a percutaneous gallbladder drainage procedure were used for the confirmation of acute cholecystitis. RESULTS. The true-negative rate in the study group was 77% and in the control group, 72%. The positive predictive value in the study group was 81% and in the control group, 45%. The negative predictive value in the study group was 98% and in the control group, 99%. The accuracy in the study group was 95% and in the control group, 84%. The sensitivity in the study group was 93% and in the control group, 93%. The specificity in the study group was 95% and in the control group, 83%. The differences in the true-negative rate, accuracy, specificity, and positive predictive value of the morphine-modified protocol used for the study group and the original protocol used for the control group were statistically significant (p < 0.0005). CONCLUSION. Pretreatment using 2 mg of IV morphine at bedside before radionuclide imaging is superior to routine HIDA scanning with only delayed images for the diagnosis of acute cholecystitis. The results of our pretreatment morphine-modified protocol are comparable to those reported in the literature for posttreatment morphine-augmented protocols. A cute cholecystitis is a common entity in patients presenting to the emergency department with acute abdominal pain, having a prevalence of 5% [1]. Occasionally, acute cholecystitis is also identified in patients hospitalized for unrelated conditions. Body imaging with ultrasound (US) of the abdomen or CT of the abdomen and pelvis is the recommended initial imaging study [2, 3] depending on whether the pain is localized to the right upper quadrant of the abdomen. When the pain is localized to the right upper quadrant and acute cholecystitis is a diagnostic possibility, hepatoiminodiacetic acid (HIDA) scanning shows the best sensitivity and specificity of these imaging modalities [1]. The main purpose of this study was to compare our morphine sulfate (MS) modified HIDA scanning protocol with the original HIDA scanning protocol. The original protocol included delayed imaging for up to 4 hours after injection without the use of MS. The primary modification we made was to administer MS before HIDA scanning as a pretreatment; for our morphinemodified protocol, 2 mg of MS was injected IV push at bedside. We practice in two community-based hospitals, each of which has approximately 300 beds, without radiology residents. Our standard HIDA scanning protocol did not include the administration of MS at any time during the image acquisition because the nuclear medicine divisions at the two hospitals AJR:207, October 2016 865

Solomon et al. TABLE 1: Results of Hepatoiminodiacetic Acid (HIDA) Scanning HIDA Scanning Results Study Group (n = 374 HIDA Studies) are unable to satisfy the regulatory requirements for storing and dispensing narcotics. We sought to make HIDA scanning simpler to perform and quicker to interpret while remaining compliant with U.S. narcotic regulations, so we changed the HIDA scanning protocol at one of our hospitals to administer 2 mg of MS IV push at bedside to inpatients being evaluated for acute cholecystitis. Patients were then transported to the nuclear medicine laboratory for imaging to begin as soon as was feasible. Materials and Methods Patients This retrospective chart review study includes the charts of all inpatients who underwent HIDA scanning in our two hospitals between 2003 and 2013 for the diagnosis of acute cholecystitis. Inpatients do not stay in the emergency department at our institutions. A HIDA scanning study was excluded if the chart could not be retrieved or if the study did not show any biliary excretion. The final study group consisted of 374 studies of 365 patients, and the control group consisted of 232 studies of 227 patients. This study was approved by our institutional review boards and was HIPAA compliant. Early Imaging Protocol: Both Groups Approximately 222 MBq (6 mci) of 99m Tc-mebrofenin was administered IV, and then anterior dynamic imaging at 1 frame per minute was performed for 1 hour. Subsequent static planar images were acquired in right anterior oblique, left anterior oblique, and right lateral projections. Study Group Protocol Starting in 2006, when the pretreatment morphine-modified protocol was begun, all inpatients at one hospital for whom HIDA scanning was ordered for acute cholecystitis were injected with 2 mg of MS IV push at bedside and were then transported to the nuclear medicine division for HIDA scanning. When patients arrived at our No. (%) of HIDA Studies Control Group (n = 232 HIDA Studies) True-positive 65 (17.4) 28 (12.1) < 0.0005 False-positive 15 (4.0) 34 (14.7) < 0.0005 True-negative 289 (77.3) 168 (72.4) < 0.0005 False-negative 5 (1.3) 2 (0.9) NS Note Percentages do not add up to 100% because of rounding. NS = not statistically significant. a Chi-square test. department, HIDA scanning was performed using the early imaging protocol described in the previous section. Outpatients were not given MS for regulatory reasons, and patients in whom a gallbladder ejection fraction was to be computed were not given MS. Neither a 4-hour fast nor delayed imaging was required for the morphinemodified protocol. Control Group Protocol All inpatients in the control group (conventional HIDA scanning protocol) underwent HIDA scanning after a 4-hour fast using the early imaging protocol described earlier, and delayed imaging was continued for up to 4 hours as long as there was enough activity within the liver. No patient received a second injection of 99m Tc-mebrofenin. This group was composed of all inpatients at the second hospital and of some inpatients at the first hospital, most of whom underwent HIDA scanning before the routine use of MS. Diagnosis of Acute Cholecystitis Acute cholecystitis was confirmed by pathologic criteria, cholangiograms acquired during the insertion of a percutaneous cholecystostomy tube, or positive results from gallbladder cultures. Our pathology department requires the presence of leukocytes in either the mucosa or the gallbladder to diagnose acute cholecystitis, which is a strict pathologic criterion for acute cholecystitis [4]. TABLE 2: Analysis of Computed Performance Values Performance Value [Formula to Calculate Performance Value a ] p a Statistical Analysis Data were analyzed using the chi-square test for discrete variables and ANOVA testing for continuous variables; p values < 0.05 were considered statistically significant. Results The study group contained 374 studies of 365 patients, and the control group contained 232 studies of 227 patients. The medical records department could not retrieve the charts for 100 patients in the study group and for 67 patients in the control group. We excluded five studies in which no excretion into the common bile duct (CBD) occurred: Four studies were of patients in the study group and one study was of a patient in the control group. Table 1 contains the true-positive (TP), false-positive (FP), true-negative (TN), falsenegative (FN) data, and the p values as computed by the chi-square test. Note that the TN rate is significantly better in the study group than the control group. Table 2 contains the positive predictive value (PPV), negative predictive value (NPV), accuracy, sensitivity, specificity, and p value calculations as computed by ANOVA. Table 3 shows demographic characteristics and laboratory results for the study and control groups. The delay between the study group patients receiving MS and arriving in the department to undergo HIDA scanning varied. This variation was a protocol design choice to simplify the process in a small department without radiology residents. The mean delay between the injection of MS and the injection of 99m Tc-mebrofenin dose was 47.22 minutes (range = 6 165 minutes). Prior Body Imaging Table 4 shows the number of HIDA scanning studies that were preceded by CT, US, or MRI or a combination of these studies. Only prior body imaging studies from the same hospital admission as the HIDA scan- Study Group Control Group PPV (%) [TP / total no. of studies read as positive] 81 45 < 0.0005 NPV (%) [TN / total no. of studies read as negative] 98 99 NS Accuracy (%) [(TP + TN) / n] 95 84 < 0.0005 Sensitivity (%) [TP / (TP + FN)] 93 93 NS Specificity (%) [TN / (TN + FP)] 95 83 < 0.0005 Note PPV = positive predictive value, NPV = negative predictive value, NS = not statistically significant. a Where TP is the number of true-positive studies, TN is the number of true-negative studies, n is the total number of studies obtained, FN is the number of false-negative studies, and FP is the number of false-positive studies. b ANOVA. p b 866 AJR:207, October 2016

Imaging Diagnosis of Acute Cholecystitis TABLE 3: Demographic Characteristics and Laboratory Results for the Study Group and Control Group Sex Demographic Characteristics and Laboratory Results Study Group Control Group Female 59 65 Male 41 35 Patient age (y) Mean (range) 63.9 (6 106) 59.9 (7 106) Total bilirubin (mg/dl) Mean (range) 1.22 (0 21.3) 1.41 (0.2 10.8) Alkaline phosphatase (U/L) Mean (range) 180.4 (1.7 1749) 127.7 (22 743) Time between MS injection and 99m Tc-mebrofenin injection (min) Mean (range) 47.22 (6 165) Time needed to perform control scanning (min) Mean (range) 143 (60 1440) Percentage of patients with acute cholecystitis a 19 13 Note MS = morphine sulfate. a [(TP + TN) / n], where TP is the number of true-positive studies, TN is the number of true-negative studies, and n is the total number of studies obtained. ning are tallied in Table 4. Six studies in the study group and 15 studies in the control group were of patients who underwent body imaging as part of a workup as outpatients before admission for HIDA scanning and cholecystectomy; those body imaging studies were excluded from the data presented in Table 4 because the previous body imaging studies were not available to us. For all the HIDA scanning studies of control group patients who had recently undergone previous body imaging, all body imaging studies were available at the time the HIDA study was acquired. False-Positive Studies As shown in Table 5, there were 15 FP HIDA scanning studies in the study group. In two patients, gallbladder cholangiocarcinoma was found at surgery. Three patients had choledocholithiasis on other imaging, two had a severely contracted stone-filled gallbladder, six had chronic cholecystitis, and one had fatal acute necrotizing pancreatitis. We could not find an explanation for the FP HIDA study in the remaining patient. False-Negative Studies Five HIDA scanning studies of study group patients and two of the control group patients were classified as FN by either surgical pathology or percutaneous cholangiography. Discussion The basic standard protocol for HIDA scanning was established [5, 6] nearly 4 decades ago during an era when HIDA scanning was the primary imaging study performed in patients with abdominal pain. One part of this original protocol required bowel to be seen to exclude the possibility of CBD obstruction. We believe that modern crosssectional body imaging makes this requirement obsolete and unnecessary. HIDA scanning has not been the primary abdominal imaging study in this clinical setting [2, 3] for a long time because of the superior anatomic detail and broader range of diagnoses that can be made by body imaging with CT, US, or MRI. Table 4 shows how many HIDA scanning studies were preceded by recent body imaging studies and which modality or modalities were used. Two HIDA scanning studies in the study group were of patients who had not undergone recent body imaging before HIDA scanning. For the first case, abdominal US was performed 2 days after HIDA scanning. This patient was admitted 2 months later and underwent CT of the abdomen and pelvis, repeat abdominal US, and another HIDA scanning study in that order. For the second case, abdominal US that was performed 2.5 months before HIDA scanning was performed during a previous hospitalization, so it was not counted in Table 4. A repeat abdominal US study was performed 4 days after HIDA scanning during the same hospital admission. Neither of these patients had acute cholecystitis. The first use of MS to improve the accuracy of HIDA scanning was reported by Choy et al. [7] using 0.04 mg/kg IV infused over 3 minutes. Since then, multiple authors have investigated the use of MS including Kim et al. [8]. Kim et al. found MS augmentation to be superior to delayed imaging for the diagnosis of acute cholecystitis. That group also administered 0.04 mg/kg of MS as an IV infusion over 3 minutes, 1 hour after the injection of the first dose of diisopropyl iminodiacetic acid ( DISIDA) and reinjected a TABLE 4: Prior Body Imaging Performed During the Same Admission as Hepatoiminodiacetic Acid (HIDA) Scanning Prior Body Imaging Studies Performed During Same Admission as HIDA Scanning Study Group (n = 368 Studies a ) No. (%) of HIDA Studies Control Group (n = 217 Studies a ) None 2 (0.5) 0 CT only 28 (7.6) 20 (9.2) US only 132 (35.9) 99 (45.6) MRI only 0 0 CT and US 196 (53.3) 95 (43.8) CT and MRI 2 (0.5) 1 (0.5) US and MRI 0 1 (0.5) CT, US, and MRI 8 (2.2) 1 (0.5) Note Percentages do not add up to 100% because of rounding. US = ultrasound. a Six studies in the study group and 15 studies in the control group were of patients who underwent body imaging as part of a workup as outpatients before admission for HIDA scanning and cholecystectomy; those body imaging studies were excluded from the data presented because the previous body imaging studies were not available to us. AJR:207, October 2016 867

Solomon et al. smaller dose of DISIDA immediately after they gave MS. The standard use of MS is to inject MS after the bowel is visualized. Interestingly, Tulchinsky et al. [9] observed the following: Some suggest that morphine infusion should be started only after activity is seen in the bowel because it can compound CBD obstruction. However, there is no evidence to support this concern, and the experience of Louridas et al. [10] supports the safety of blind morphine administration. TABLE 5: False-Positive Morphine-Modified Studies Patient No. Age (y) Sex The results of our study are consistent with their statement. Our study group patients received MS injected IV push at the bedside before the administration of the 99m Tc-iminodiacetic acid derived radiopharmaceutical, and we did not identify any adverse effects. The protocol investigated by Louridas et al. [10] pretreated all patients with cholecystokinin (CCK), followed 10 minutes later by concurrent IV administration of DISIDA with 0.04 mg/kg of MS as an IV infusion over 3 minutes. Our protocol is simpler because it does not use CCK, allows the bedside administration of a fixed dose of MS (2 mg IV push), and waits a variable amount of time before scanning is started (Table 3) while the patient is being transported to the nuclear medicine section. We found this morphine-modified protocol to be safe and effective and to yield si gnificantly improved TN rate, accuracy, specificity, and PPV compared with the conventional HIDA scanning protocol used for the control group. The low PPV of the conventional HIDA scanning protocol for our control group has been reported in several studies [1, 8, 11, 12], especially in studies in which strict pathologic criteria for diagnosing acute cholecystitis were used, as in this study. It is worth noting that acute cholecystitis was diagnosed in only 19% of our study group and 13% of our control group (Table 3). We discovered that oftentimes elderly patients with comorbidities do not undergo urgent cholecystectomy surgery, so we needed to use another standard for determining acute cholecystitis in these nonoperative patients. Because these patients have a percutaneous cholecystostomy tube inserted, the cholangiograms obtained as part of this drainage procedure are able to show whether the cystic duct and CBD are patent. Additionally, the bile cultures obtained directly from the gallbladder are included as a means by which to confirm the diagnosis of acute cholecystitis. Time Between MS Injection and Scanning (min) Pathology or Cholangiography Results Relevant Body Imaging Findings by Modality 1 85 F 60 Percutaneous transhepatic cholangiograms showed patent cystic duct and one stone in CBD 2 75 F 40 Percutaneous transhepatic cholangiograms showed patent cystic duct and one stone in CBD 3 65 F 60 No clinical indications for percutaneous drainage or surgery 4 28 F 30 Chronic cholecystitis without gallstones on surgical pathology 5 71 M 65 Chronic cholecystitis and cholelithiasis on surgical pathology 6 50 F 30 Chronic cholecystitis and cholelithiasis on surgical pathology US: no gallstones, irregular thickened GB walls, no biliary dilatation; CT: GB wall enhancement Portable US: GB distended, GB shows mild wall thickening, no gallstones, no biliary dilatation US: contracted stone-filled GB with wall-echoshadow sign US: no GB stones, no pericholecystic fluid, no GB wall thickening US: 2-cm stone in GB neck with mild wall thickening and edema, no pericholecystic fluid; MRCP: at least three small stones in CBD US: gallstones with mild GB wall thickening and edema, no pericholecystic fluid, or no Murphy sign, CBD mildly dilated 7 81 M 45 Infiltrating adenocarcinoma of GB US and MRI: no mass was seen; mass could have been obscured by what was read as multiple stones 8 71 M 35 Cholecystostomy tube inserted at outside hospital; tube fell out and was not reinserted US: severely contracted stone-filled GB 9 83 F 45 Marked chronic fibrosing cholecystitis and cholelithiasis CT: GB wall enhancement; US: gallstones, wall thickening, and pericholecystic fluid 10 76 F 60 Adenocarcinoma of GB (T2N1MX) and cholelithiasis US, CT, and MRI: no mass was seen; mass could have been obscured by multiple stones 11 60 M 30 Final diagnosis: fatal necrotizing pancreatitis 12 89 M 75 No clinical indication for surgery or percutaneous GB drainage 13 78 F 30 Percutaneous transhepatic cholangiograms showed normal cystic duct and CBD 14 79 F 40 GB pathology showed mild chronic cholecystitis and cholelithiasis 15 59 F 45 GB pathology showed chronic cholecystitis with cholelithiasis Note MS = morphine sulfate, CBD = common bile duct, US = ultrasound, GB = gallbladder. US: GB with mobile small stones, sludge, and thickened walls US: 1-cm GB stone in neck, no GB wall thickening, small amount of pericholecystic fluid US: multiple 1-cm GB stones, mild GB wall thickening, no pericholecystic fluid US: 1.6-cm gallstone in the neck, no GB wall thickening, no pericholecystic fluid 868 AJR:207, October 2016

Imaging Diagnosis of Acute Cholecystitis Fig. 1 Morphine response in 71-year-old woman in study group who underwent pretreatment morphinemodified hepatoiminodiacetic acid (HIDA) scanning protocol. HIDA image shows liver, common bile duct, and gallbladder without showing bowel. Visualization of bladder is incidental. We and others [2] have observed that since the advent of interventional radiology (IR) there has been a shift in clinical practice away from urgent surgery. We believe this shift to be an important factor contributing to the low incidence of acute cholecystitis that we observed in both of our HIDA scanning groups. Patients with classic symptoms of acute cholecystitis and with a confirmatory body imaging study will often be referred to IR for insertion of a percutaneous cholecystostomy tube without being referred to nuclear medicine for HIDA scanning. The surgeons at both of our hospitals commonly request this procedure from IR for either nonoperative patients or patients in whom other causes of sepsis have been excluded because a percutaneous cholecystostomy tube procedure can be both diagnostic and therapeutic, particularly for ICU patients [2]. HIDA scanning is used more often in atypical cases such as possible acute acalculous cholecystitis or when patients are stable enough to wait for nonemergent surgery. Corroborating nuclear HIDA scanning is requested by the clinical service before some stable patients are scheduled for gallbladder surgery. Many patients with classic acute cholecystitis are no longer referred for HIDA scanning; consequently, the effective pretest probability for acute cholecystitis when a patient is referred for HIDA scanning is reduced. This lowered pretest probability of acute cholecystitis at our institutions may have contributed to the low PPVs in the control group. Additionally, the current American College of Radiology (ACR) appropriateness criteria for right upper quadrant pain [2] recommends HIDA scanning only after US has been performed, especially if the result of the US is equivocal. False-Negative Studies Five HIDA scanning studies in our study group and two in our control group were classified as FN. The chi-square method states that any subcategory that contains five items or fewer is too few to be used for the analysis. Therefore, the difference between the study and control FN subcategories is not significant in this study. Morphine Response We call the pattern in which the liver, bile ducts, and gallbladder are seen but without the visualization of the duodenum or intestines a morphine response (Fig. 1). The presence of this morphine effect makes these studies easier to interpret because the lack of activity in the duodenal bulb avoids confusion between activity in the gallbladder versus activity in the duodenal bulb. In 62% (183/294) of the studies read as negative in which MS pretreatment was given, the morphine response was shown (i.e., CBD and gallbladder were seen but bowel was not seen). These studies were interpreted as showing no radionuclide evidence of acute cholecystitis. In 12% (20/170) of the studies read as negative in the control group, the CBD and gallbladder were seen but bowel was not. Because of logistic factors, we were not able to determine which medications were administered to these 16 patients shortly before HIDA scanning was performed. Therefore, we cannot confidently explain the presence of the morphine effect in these control subjects because we did not ask that these patients be given MS. Prolonged Response of Morphine We found that even when the delay between the injection of the MS dose and beginning of the dynamic imaging approached 3 hours (i.e., 159 and 165 minutes), a morphine response (Fig. 1) was still shown. This effect persisted on the final set of static images. When including the 1 hour or so it takes to acquire HIDA scans, we have been able to show that the spasmodic effect of 2 mg of MS IV push on the sphincter of Oddi can persist for at least 4 hours. Lack of Consequences of the Morphine Response We did not identify any adverse consequences to the lack of visualization of small bowel due to the morphine effect. We have observed that HIDA scanning is not the primary modality used in our departments to establish a diagnosis of CBD obstruction; our clinicians widely use serum chemistries and anatomic body imaging to make that diagnosis. Because radiologic body imaging has higher spatial resolution than nuclear imaging, we do not believe it to be important that we may miss such an obstruction when we see a morphine effect. Additionally, because MRCP is becoming more common, it is not rare for stones in the biliary tree to be directly imaged using these body imaging modalities. In fact, we were able to directly image three patients who had CBD stones, two by cholangiography and one by MRCP. Fasting The classic HIDA scanning protocol suggests a minimum fast of 2 hours [13, 14]. We do not require patients to be fasting for our morphine-modified protocol. We agree with others who have observed that patients who have an acute abdomen and may have acute cholecystitis have been fasting since the onset of abdominal pain [13], thus rendering the need for a minimum period of fasting to be moot. Interestingly, there is a paradoxical role of CCK in this scenario. We know that exogenous CCK (sincalide) administration causes gallbladder contraction and can be used to reduce FPs [14, 15] when patients have been fasting for more than 24 hours. However, studies have also shown that pretreatment CCK (sincalide) does not improve the accuracy of HIDA scanning performed with MS posttreatment [9, 12, 16, 17]. The paradox is that exogenous CCK (sincalide) reduces FPs, but endogenous CCK from a recent meal is believed to increase FPs. We do not know what would be an optimal fasting period or whether fasting is necessary when MS is used for either a pretreatment protocol or a posttreatment protocol. Again, this point is largely a moot one because patients with an acute abdomen are fasting anyway. Comparison With Performance Values in the Literature Kiewiet et al. [1] reported the results of their extensive meta-analysis of imaging for the diagnosis of acute cholecystitis. They evaluated overall diagnostic accuracy and then separated the included studies into two subgroups: The first subgroup consisted of studies published before January 1, 1990, and the second subgroup included those published after that AJR:207, October 2016 869

Solomon et al. date up to March 2011. They reported the overall sensitivity estimate for HIDA scanning to range from 78% to 100%, with a summary estimate of 96% (95% CI, 94 97%). Their reported overall specificity estimate ranged from 50% to 100%, with a summary estimate of 90% (95% CI, 86 93%). Surprisingly, their subgroup analysis showed that the sensitivity and specificity of HIDA scanning were comparable for the two subgroups: That is, they had essentially divided HIDA scanning into subgroups that approximately correspond to the era before and after the use of MS after treatment and found no statistically significant difference between these two subgroups. This result is interesting because of the study by Kim et al. [8] that showed the use of posttreatment MS to be superior to the use of delayed scanning alone. We cannot explain the difference between these published studies to yield this unexpected conclusion. When we compare the results for our pretreatment morphine-modified protocol with the overall performance values reported by Kiewiet et al. [1], our sensitivity of 93% is slightly below their CI range of 94 97%, and our specificity of 95% is slightly greater than their CI range of 86 93%. Kim et al. [8] did not report a sensitivity or specificity. However, they reported their raw numbers, so these statistics can be easily calculated. The sensitivity of their posttreatment MS protocol was 92%, and the sensitivity of our pretreatment MS protocol was 93%. Their specificity was 79%, and ours was 95%. Moreover, of 45 posttreatment MS studies, they reported that there were four (8.9%) FP and two (4.4%) FN studies; of our 374 pretreatment MS studies, there were 15 (4.0%) FP and five (1.3%) FN studies. That is, our sensitivity is not different, our specificity is much better, and we had lower rates of FPs and FNs than those reported by Kim et al. Conclusion We found that our new protocol for morphine-modified HIDA scanning in which patients are pretreated with MS at bedside and are then transported to the nuclear medicine section is quicker to perform and interpret than the HIDA scanning protocol with delayed imaging that was used for the control group. The TN rate, accuracy, specificity, and PPV are all significantly improved in our study protocol when compared with our control protocol, which consisted of only delayed imaging without the posttreatment use of MS. The sensitivity of our pretreatment morphine-modified protocol is slightly less than the summary estimate 95% CI reported by Kiewiet et al. [1], but the specificity of our protocol is slightly better than their reported summary estimate 95% CI. We also found that our sensitivity is comparable to that reported by Kim et al. [8], but our specificity is much better than theirs. Overall, we believe that our pretreatment morphine-modified protocol is comparable to the commonly used posttreatment MS protocol reported in the published literature. Our modification to the morphineaugmented protocol increases the usefulness of HIDA scanning to referring clinicians by making the scanning acquisition shorter, and this modification also makes the scans easier to interpret because of the frequent lack of activity in the duodenal bulb. Acknowledgment We thank Elena Yee for painstakingly collecting most of the data needed for this article. References 1. Kiewiet JJ, Leeuwenburgh MM, Bipat S, Bossuyt PM, Stoker J, Boermeester MA. A systematic review and meta-analysis of diagnostic performance of imaging in acute cholecystitis. Radiology 2012; 264:708 720 2. American College of Radiology website. American College of Radiology Appropriateness Criteria: right upper quadrant pain. www.acr.org/~/ media/acr/documents/appcriteria/diagnostic/ RightUpperQuadrantPain.pdf. Accessed January 15, 2016 3. American College of Radiology website. American College of Radiology Appropriateness Criteria: acute (nonlocalized) abdominal pain and fever or suspected abdominal abscess. www.acr.org/~/ media/acr/documents/appcriteria/diagnostic/ Ac ut e Ab d o m i n a l P a i n Feve r Su s p e c t e d Ab d o m i n a l Abscess.pdf. Accessed January 15, 2016 4. Freitas JE, Coleman RE, Nagle CE, Bree RL, Krewer KD, Gross MD. Influence of scan and pathologic criteria on the specificity of cholescintigraphy: concise communication. J Nucl Med 1983; 24:876 879 5. Weissmann HS, Frank MS, Bernstein LH, Freeman LM. Rapid and accurate diagnosis of acute cholecystitis with 99mTc-HIDA cholescintigraphy. AJR 1979; 132:523 528 6. Weissmann HS, Sugarman LA, Freeman LM. The clinical role of technetium-99m iminodiacetic acid cholescintigraphy. In: Freeman LM, Weissmann HS, eds. Nuclear medicine annual 1981, 1st ed. New York, NY: Raven Press, 1981:35 89 7. Choy D, Shi EC, McLean RG, Hoschl R, Murray IP, Ham JM. Cholescintigraphy in acute cholecystitis: use of intravenous morphine. Radiology 1984; 151:203 207 8. Kim CK, Kevin KM, Tse MJ, et al. Cholescintigraphy in the diagnosis of acute cholecystitis: morphine augmentation is superior to delayed imaging. J Nucl Med 1993; 34:1866 1870 9. Tulchinsky M, Colletti PM, Allen TW. Hepatobiliary scintigraphy in acute cholecystitis. Semin Nucl Med 2012; 42:84 100 10. Louridas G, Botha JR, Esser JD, et al. Role of morphine administration with 99m-technetium labeled di-isopropyl iminodiacetic acid in the diagnosis of acute cholecystitis. S Afr J Surg 1987; 25:148 150 11. Weissmann HS, Freeman LM. The biliary tract. In: Freeman LM, ed. Freeman and Johnson s clinical radionuclide imaging, 3rd ed., vol. 2. Orlando, FL: Grune & Stratton, 1984:914 916 12. Chen CC, Holder LE, Maunoury C, Drachenberg CI. Morphine augmentation increases gallbladder visualization in patients pretreated with cholecystokinin. J Nucl Med 1997; 38:644 647 13. Mettler FA, Guiberteau MJ. Essentials of nuclear medicine imaging, 6th ed. Philadelphia, PA: E lsevier-saunders, 2012:250 14. Vergara-Wentland P, Bennett P. Acute cholecystitis and biliary obstruction. In: Bennett P, Oza U, Trout A, Mintz A, eds. Diagnostic imaging: nuclear m edicine, 2nd ed. Philadelphia, PA: Elsevier, 2016:71 15. Mettler FA, Guiberteau MJ. Essentials of nuclear medicine imaging, 6th ed. Philadelphia, PA: E lsevier-saunders, 2012:253 16. Yen TC, King KL, Chang SL, et al. Morphineaugmented versus CCK-augmented cholescintigraphy in diagnosing acute cholecystitis. Nucl Med Commun 1995; 16:84 87 17. Kim CK. Pharmacologic intervention for the diagnosis of acute cholecystitis: cholecystokinin pretreatment or morphine, or both? J Nucl Med 1997; 38:647 649 870 AJR:207, October 2016