Abdominal Imaging

ª Springer Science+Business Media, Inc. 2006 Published online: 7 February 2006

Abdom Imaging (2006) 31:549–554 DOI: 10.1007/s00261-005-0251-6

Omental infarct: CT imaging features A. K. Singh, D. A. Gervais, P. Lee, S. Westra, P. F. Hahn, R. A. Novelline, P. R. Mueller Division of Abdominal Imaging and Interventional Emergency Radiology, Massachusetts General Hospital, 55 Fruit Street, Boston, Massachusetts 02114, USA

Abstract Background: The aim of this study is to describe contrast-enhanced computed tomographic (CT) features of acute omental infarction and to study the evolutionary changes on follow-up CT imaging. Methods: Fifteen cases of omental infarction were evaluated for their initial CT imaging features. The imaging features evaluated included size of the fatty lesion, location, peripheral rim, and relation to colon. CT findings were correlated with etiology, clinical presentation, and leukocytosis. Follow-up CT images were available in eight patients and the imaging features were studied. Results: Eight omental infarcts were of unknown etiology and seven were secondary to abdominal surgery. In 53% of patients (eight of 15), the location of the omental infarct was in the right lower, mid, or upper quadrants. These eight right-side infarcts occurred in six patients with primary omental infarcts. In 13 of 14 patients who underwent CT within 15 days of onset of omental infarct, the margin of the lesion was ill defined. Primary omental (n = 8) infarcts were seen in younger patients (p = 0.02) and were larger on CT (p = 0.02) compared with secondary omental infarcts. CT findings evolved from an illdefined, heterogeneous fat-density lesion to a well-defined, heterogeneous fat-density lesion with a peripheral hyperdense rim in all six secondary omental infarctions for which acute stage and follow-up CT images were available for interpretation. Conclusion: There is a significant difference in the age distribution and CT findings in terms of size of the omental infarction between primary and secondary etiologies. On follow-up CT, secondary omental infarcts progressively shrank and developed a well-defined, hyperdense rim around a fatty core. Key words: Omental infarction—Omentum—Computed tomography—Acute abdomen

Correspondence to: A. K. Singh; email: [email protected]

Omental infarction is a rare cause of acute abdominal pain in children and adults. It often presents with right upper or lower quadrant pain and is often clinically indistinguishable from acute appendicitis and cholecystitis. The diagnosis of omental infarction is primarily based on cross-sectional imaging, most commonly computed tomography (CT). Although there are reports of management of omental infarction with surgical resection, the primary goal of CT is to make a confident diagnosis and allow conservative management [1–3]. There have been about 300 cases of omental infarction reported since the first case reported in 1896 by Bush [4]. Most of these are anecdotal case reports and there is a paucity of larger case series in the literature describing the imaging appearance of omental infarcts. Moreover, there are no case series describing the evolutionary changes seen on serial CT images in patients with omental infarction. We describe acute and evolutionary changes of primary and secondary omental infarctions on serial CT scans.

Materials and methods In this single-center retrospective study, we searched the radiology information system for all patients whose primary contrast-enhanced CT diagnosis was acute omental infarction. Patients who had an acute presentation in the emergency room with acute abdominal pain and a corresponding CT diagnosis of acute omental infarction (n = 8) or a new CT diagnosis of omental infarction after surgery (n = 7) as determined by two radiologists were included in this study. To exclude conditions that can mimic omental infarction, we relied on one of the following inclusion criteria: (a) patients who presented in the emergency room with acute abdomen and CT features of a heterogeneous lesion centered in the omentum and containing areas of fat density; or (b) patients who developed a heterogenous omental lesion in the postoperative period that contained areas of fat density and persisted indefinitely on follow-up CT.

550

We did not include cases that were thought to represent postsurgical omental contusion, which were related to mechanisms ranging from retraction injury to omental manipulation. Cases of omental metastases were also excluded from the study group. Exclusion criteria used in this study were the following: 1. Omental contusions and retraction injuries were ruled out by noting complete resolution of CT findings on follow-up studies. 2. All patients with multifocal omental lesions or suspected or proved omental metastases were excluded from the study. Clinical follow-up longer than 6 months was available in all 15 patients to exclude metastases. 3. A previously published case of omental infarction and all cases of acute epiploic appendagitis published in a case series [5, 6] were excluded. 4. Patients with CT or laboratory evidence of acute pancreatitis.

CT evaluation Two board-certified radiologists with more than 8 years of experience evaluated all CT scans and reached a consensus. When there was no consensus between the two radiologists, a third radiologist with more than 10 years of experience was involved in the final CT interpretation. Thirty-seven initial and follow-up CT scans were available in 15 patients categorized as omental infarction. Follow-up CT scans were available in eight of the 15 patients, including all seven omental infarcts secondary to surgery and one primary omental infarct. The period between the first and follow-up CT scans was 1 day to 3 years.

CT technique Contrast-enhanced CT examinations were performed with a multidetector CT scanner (General Electric Medical Systems, Milwaukee, WI, USA; Siemens Somatom, Erlanger, Germany). All 15 patients had at least one contrast-enhanced CT evaluation in portal venous phase after injection of nonionic contrast (300 mgI/mL). Twelve patients underwent CT after injection of 140 mL of nonionic contrast at a rate of 2 to 3 mL/s. The three pediatric patients received intravenous nonionic contrast at a dose of 1 mL/lb of body weight. CT slice thickness for the abdominal and pelvic scans was 5 mm and the pitch was 1.5. Images were available for interpretation on a PACS workstation (Impax DS3000 AGFA SP4SU2).

CT findings Evaluated CT imaging features were size of the heterogeneous lesion (<5, 5–7.5, and >7.5 cm), location,

A. K. Singh et al.: Omental infarct: CT imaging features

peripheral rim (presence/absence and regular/irregular), relation to colon and liver, bowel wall thickening, and changes in density pattern, size, and margin over time. CT findings were correlated with etiology, clinical presentation, and leukocytosis. We also correlated the results of the CT with the results of abdominal ultrasound examination, which was performed within 48 hours of CT in seven patients. Data in this study were evaluated with chi-square test, where p < 0.05 was always statistically significant.

Results Demographics Fifteen patients (eight male and seven female, mean age 43 years, age range 6–81 years) diagnosed with omental infarct on CT imaging between January 1999 and November 2004 were evaluated. Patients with idiopathic omental infarcts (n = 8) were 6 to 62 years old (mean age 30 years) and those with secondary omental infarcts (n = 7) were 35 to 81 years old (mean age 56.7 years). The difference was statistically significant (p = 0.02). There were three male and five female subjects in the primary omental infarct group and five male and two female subjects in the secondary omental infarct group. Among the pediatric patients, two were males and one was female.

Leukocytosis White blood cell count performed within 24 hours of acute presentation or first positive CT scan was available for 14 of the 15 patients and was high in nine patients (64%). In six of these nine patients, the white blood cell count was lower than 15,000/mm3. These included four of the eight patients with primary omental infarcts and five of the six patients with secondary omental infarcts. The difference was not statistically significant (p = 0.3).

Etiology The 15 cases of omental infarct included seven secondary to abdominal surgery and eight with an idiopathic etiology. CT scans in six of the seven patients who had surgery before onset of the omental infarct were obtained 1 to 24 days (mean ± standard deviation 12.1 ± 8.1 days) from the day of surgery. In one patient CT was performed 2.5 years from the day of surgery and preoperative CT showed no omental infarct. Abdominal surgeries included pancreatectomy (n = 3), gastrectomy (n = 1), esophagectomy with gastric pull-through (n = 1), aortic bypass graft surgery (n = 1), and colectomy (n = 1). In the group with primary omental infarct, the only risk factor that could be recognized was obesity in two pediatric patients.

551

A. K. Singh et al.: Omental infarct: CT imaging features

Indication for CT The primary clinical indications for the initial CT scan in the primary omental infarct group was suspected cholecystitis in four patients, appendicitis in three, and diverticulitis in one. In the secondary omental infarct group, the indication for the initial postoperative CT was to rule out postoperative abscess in four patients, bowel perforation in one, and pancreatic tumor in two. CT scans revealed one abscess, one pancreatic pseudocyst, and no bowel perforations.

Location Primary omental infarcts (n = 8) were located along the right omental margin in six of the eight patients. Omental infarctions were located in the left lower quadrant in one patient and epigastrium in one patient. In four patients, the primary omental infarct showed extension of inflammation into the pericholecystic fat and left lobe of the liver. Secondary omental infarcts (n = 7) were located on the left in four patients, epigastrium in one, and right in two. Omental infarctions were located in or near the surgical bed in all seven patients. Differences in location between primary and secondary omental infarcts were not statistically significant (p = 0.2). The infarcted omentum was not in anatomic contact with the colonic wall and was centered in the omentum rather than in the bowel wall in eight of the 15 patients. In six patients, the omental infarction abutted the colonic wall on CT. The remaining patient previously underwent colectomy.

Size The size of acute omental infarct could be determined in 14 of 15 patients on CT. In one patient in the secondary omental infarct group, there was no CT performed acutely, so this patientÕs CT scan was not included in the size comparison of the acute omental infarction. In the primary omental infarct group, the inflammatory lesion was 5 to 7.5 cm in four patients, larger than 7.5 cm in two, and smaller than 5 cm in two. The lesion was larger than 7.5 cm in all six patients with secondary omental infarct who underwent CT during the acute stage (within 15 days). The difference in size was statistically significant (p = 0.02).

Morphology CT findings consisted of an ill-defined, heterogeneous fat-density lesion (Fig. 1) and no definable continuous peripheral rim in the greater omentum in 13 of 14 patients who underwent CT within 15 days of surgery or

onset of a clinically acute abdomen. In one patient, there was a hyperdense rim around the omental infarction on CT that was obtained during acute presentation.

Colonic wall The colonic wall was normal in thickness in all 15 patients.

Evolutionary changes CT findings progressed from an ill-defined, heterogeneous fat density lesion in the early stage (<15 days) to a well-defined, smaller lesion with a hyperdense rim in six of seven patients with secondary omental infarct. These findings were seen on multiple CT scans that were performed over more than 30 days (Figs. 2, 3). In one patient, the initial ill-defined, heterogeneous fat-density lesion progressed to a well-defined abscess in the omentum, which was successfully drained by imageguided catheter drainage after an unsuccessful surgical drainage. On CT, the liquefied abscess contained nondependent floating fat, believed to represent necrotic fatty tissue. In one patient in the secondary omental infarct group, CT scans performed approximately 2.5 years from the date of gastrectomy showed a small, 3-cm diameter, well-defined fat-density lesion with a hyperdense peripheral rim, which decreased on a 3-year follow-up study. The location of this lesion on the left side, at the expected location of the descending colon, favored the diagnosis of secondary over primary omental infarction. In the one patient in the primary omental infarct group (n = 8) who had an available follow-up CT, the findings were an ill-defined, heterogeneous fat-density lesion adjacent to the left liver lobe on acute presentation. The lesion shrank and continued to have illdefined margins on a follow-up CT performed 1 year later.

Discussion Patients with acute omental infarct are most commonly adults who present with acute abdominal pain that is localized to the right lower quadrant. Other presenting symptoms include nausea, vomiting, anorexia, diarrhea, and fever. Because the incidence is low and the clinical features nonspecific, omental infarct is not included in the clinical differential diagnosis in most of patients on initial evaluation. Further, based on clinical presentation, most patients are suspected to present with acute appendicitis or cholecystitis [1, 2]. Primary omental torsion has been postulated to be due to congenital variations such as bifid omentum, congenital attachment abnormalities, or vascular variations that predispose to venous thrombosis. Secondary

552

A. K. Singh et al.: Omental infarct: CT imaging features

Fig. 1. Acute omental infarct. A Contrast-enhanced CT scan in a 36year-old female shows an oval mixeddensity lesion (arrow) that is well separated from the colon by small bowel loops in the right lower quadrant. B Contrast-enhanced CT performed after esophagectomy in a 54-year-old male shows a heterogeneous density, fatty lesion (arrow) in the left upper quadrant. C Contrast-enhanced CT in a 64-year-old female shows streaky densities (arrow) in the greater omental fat anterior to the left lobe of the liver.

Fig. 2. Evolutionary changes in omental infarct in a 78-year-old female after colectomy. A Postcolectomy contrast-enhanced CT shows an illdefined area of increased density (arrows) from an early acute omental infarct. B Contrast-enhanced CT performed 25 days later shows development of a hyperdense rim (arrow) surrounding the omental infarct. C, D Follow-up CT images at 7 months and 3 years show a residual, smaller, fat-density lesion with a welldefined hyperdense rim (arrow).

torsion is often due to acquired attachment of the omentum to pathologic foci such as cysts, masses, or scars. The right-side predilection is thought to be due to the omentum being longer and more mobile on that side. Omental infarction without torsion may be associated with predisposing conditions such as obesity, trauma, overeating, overexertion, laxative use, surgery, sudden posture change, and congestive heart failure. Superior mesenteric artery occlusion and remote abdominal surgery as a cause of omental infarction have also been described in the literature [5, 7, 8].

The diagnosis of omental infarction is based primarily on CT findings of a triangular or oval heterogeneous fatty mass that is located between the anterior abdominal wall and the transverse or ascending colon fat density [9]. Omental torsion as a cause of omental infarction is suggested at CT when a whirled pattern of concentric linear strands is seen. Distinction from diverticulitis can be made by lack of an adjacent diverticulum, abscess, and bowel wall thickening. However, rarely there may be bowel wall thickening adjacent to an area of omental infarction due to reactive bowel wall changes [10].

A. K. Singh et al.: Omental infarct: CT imaging features

553

Fig. 3. Evolutionary changes in omental infarct in a 61-year-old female after partial pancreatic resection. A Contrast-enhanced CT shows an illdefined area of streaky densities (arrow) in the left upper quadrant. A small pancreatic pseudocyst seen at the pancreatic resection margin. B, C Contrast-enhanced CT images obtained 4 and 7 months later shows a smaller infarct and development of a hyperdense rim surrounding the omental infarct (arrow).

The fatty lesion was larger than 5 cm in most patients (64%) in our study. This is unlike acute epiploic appendagitis, where the lesion is most commonly smaller than 5 cm in diameter [6]. Unlike acute epiploic appendagitis, where the fat-density lesion is surrounded by a hyperdense rim abutting the colonic wall, cases of acute omental infarction in this study did not abut the colonic wall in nine of the 15 patients. Moreover, a continuous hyperdense rim around the fatty lesion in cases of acute omental infarction was absent in 13 of 14 patients whose CT scan was obtained in the acute stage. A hyperdense rim was seen in only one 6-year-old pediatric patient whose diagnosis of omental infarction was made based primarily on the large (6.7 cm) fatty lesion and young age of the patient. We also did not see central hyperdense foci in the omental infarction, which is seen in more than 50% of acute epiploic appendagitis cases [6]. The importance of making the correct CT diagnosis of acute omental infarct is especially important to avoid unnecessary surgical intervention. Clinical management of acute omental infarct includes conservative treatment with pain medication. Omental infarcts have also been managed by laparoscopic resection of the inflamed omentum, which is sometimes indicated by worrisome peritoneal signs and suspected complications [11, 12]. This report describes the imaging features of acute omental infarction and their evolution changes. The commonest CT appearance of acute omental infarction seen in our study was an ill-defined, heterogeneous fatdensity lesion with surrounding inflammatory changes centered in the greater omentum. Unlike acute appendagitis, we did not see a central high attenuation

focus within the fat or apposition of the fatty lesion to the colonic wall. Unlike acute appendagitis, in which CT changes resolve completely, 6 months after the acute presentation omental infarcts in our series had shrunk but did not resolve in any patient. In general, the heterogeneous fatty lesion in the omentum shrank and developed a well-defined hyperdense rim on follow-up CT scan. One might consider that the lack of pathologic confirmation a weakness of this study, but pathologic evaluation would not be expected in the current management of this condition. We manage omental infarction at our institution by conservative means, and the CT imaging features provide confident diagnosis for such management to be carried out. However, because of the availability of follow-up CT imaging in patients with secondary omental infarction and the persistence of omental findings on follow-up CT, we were able to exclude cases of omental injury that might be considered an omental contusion, retraction injury, or postsurgical changes. Potential complications of omental infarct are bowel obstruction from adhesions, omental abscess, and a reported fatality [13–15]. In this study, we encountered one omental abscess, which was managed by CT-guided percutaneous catheter drainage. No cases of bowel obstruction were found in this study. As the number of CT scans for the workup of acute abdominal pain increases and the incidence of obesity increases, we are more likely to see changes of acute omental infarction in emergency room patients who present with an acute abdomen. Awareness of acute and follow-up imaging features of acute omental infarction is

554

important in making the correct diagnosis and directing proper management. References 1. Puylaert JB (1992) Right-sided segmental infarction of the omentum: clinical, US, and CT findings. Radiology 185:169–172 2. Miguel Perello J, Aguayo Albasini JL, Soria Aledo V, et al. (2002) Omental torsion: imaging techniques can prevent unnecessary surgical interventions. Gastroenterol Hepatol 25:493–496 3. van Breda Vriesman AC, Lohle PN, Coerkamp EG, Puylaert JB (1999) Infarction of omentum and epiploic appendage: diagnosis, epidemiology and natural history. Eur Radiol 9:1886–1892 4. Bush P (1896) A case of hemorrhage into the greater omentum. Lancet 1:286 5. Singh AK, Gervais DA, Hahn PF, et al. (2004) CT appearance of acute appendagitis. AJR 183:1303–1307 6. Singh AK, Alhilali LM, Gervais DA, Mueller PR (2004) Omental infarct: an unusual CT appearance after superior mesenteric artery occlusion. Emerg Radiol 10:276–278 7. Phillips RW, Peterson CM (1988) Infarction of the omentum after cesarean section. A case report. J Reprod Med 33:382–384

A. K. Singh et al.: Omental infarct: CT imaging features

8. Schwartzman GJ, Jacobs JE, Birnbaum BA (2001) Omental infarction as a delayed complication of abdominal surgery. Clin Imaging 25:341–343 9. Hollerweger A, Rettenbacher T, Macheiner P, Gritzmann N (1996) Spontaneous fatty tissue necrosis of the omentum majus and epiploic appendices: clinical, ultrasonic and CT findings. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb 165:529–534 10. Garant M, Taourel P, Fried GM, Bret PM (1995) Thickening of the transverse colon associated with torsion of the greater omentum. AJR 165:1309 11. Helmrath MA, Dorfman SR, Minifee PK, et al. (2001) Right lower quadrant pain in children caused by omental infarction. Am J Surg 182:729–732 12. Kimber CP, Westmore P, Hutson JM, Kelly JH (1996) Primary omental torsion in children. J Paediatr Child Health 32:22–24 13. Vertuno LL, Dan JR, Wood W (1980) Segmental infarction of the omentum: a cause of the semi-acute abdomen. Am J Gastroenterol 74:443–446 14. Balthazar EJ, Lefkowitz RA (1993) Left-sided omental infarction with associated omental abscess: CT diagnosis. J Comput Assist Tomogr 17:279–381 15. Patchell RD (1977) A fatal case of idiopathic omental infarction. W Va Med J 73:29–30