Non-perforated peptic ulcer disease: multidetector CT findings, complications, and differential diagnosis

Peptic ulcer disease (PUD) refers to development of gastric or duodenal mucosal defects which penetrate through the muscularis mucosa. Until the end of the twentieth century, PUD represented a major cause of morbidity and mortality worldwide. Since then, the availability of effective therapies, including H2-receptor antagonists, proton pump inhibitors and Helicobacter pylori (HP) eradication treatment, resulted in decreasing incidence of PUD. However, PUD continues to be a major global health issue that affects roughly 10% of the world population, due to the combined effect of persistent HP epidemics in low-income countries and of widespread use of nonsteroidal anti-inflammatory drugs (NSAIDs) [1, 2].

Patients affected by PUD may suffer from chronic symptoms such as dyspepsia, epigastric discomfort, nausea and early satiety, and often experience acute bouts of abdominal pain. Upper digestive endoscopy is definitely the mainstay diagnostic technique for PUD, but is invasive and often unfeasible in urgent conditions unless gastrointestinal bleeding is suspected. As a result, many patients with unknown PUD present to emergency departments with unexplained acute abdomen and commonly undergo multidetector CT. Since perforated gastroduodenal ulcers represent the commonest cause of spontaneous pneumoperitoneum, suggesting the underlying presence of PUD is relatively straightforward when radiologists are faced with upper abdominal free air or peritonitis without CT signs of colonic diverticulitis and digestive tract masses [3, 4].

Conversely, due to the widespread belief that CT has no role in detecting uncomplicated PUD, most non-perforated ulcers are missed at primary CT interpretation. Unfortunately, unless clinical suspicion of PUD exists, endoscopy may not be performed for days or weeks after CT, thus allowing time for complications to develop. Albeit superficial ulcers are generally inconspicuous, careful multiplanar CT interpretation and attention to subtle mural and extraluminal signs may allow diagnosing non-perforated PUD prospectively. Suggesting the possibility of unexpected PUD may allow planning early endoscopy for diagnostic confirmation and timely start of appropriate treatment with nasogastric tube aspiration, intravenous fluids, and HP “triple therapy” eradication [5‐7].

Therefore, this pictorial essay aims to provide radiologists with an increased familiarity with CT appearances of non-perforated PUD. After an overview of the current epidemiology of PUD, we explain the appropriate multidetector CT acquisition and interpretation techniques. Then, the imaging findings of uncomplicated peptic disease, PUD-related digestive haemorrhage, gastric outlet obstruction and biliopancreatic fistulisation are presented with examples and a focus on their respective differential diagnoses.

Despite progressive eradication, over half of the world’s population still harbours chronic HP infection, particularly in developing countries. Conversely the prevalence of PUD is lower (approximately 1.5% to 5%) in Europe and North America, where an increasing proportion (up to 46% of patients) of PUD cases is secondary to the widespread use of NSAIDs including low-dose aspirin. Since NSAIDs and HP result in additive risk, in Western countries PUD is increasingly encountered in elderly people (with a 1.5 male predominance), who usually suffer from lower degrees of abdominal pain compared to their younger counterparts. Other factors such as cigarette smoking and alcohol intake may contribute to the risk of developing PUD. In 10% of cases the disease is idiopathic and unrelated to either HP or NSAID medications. Furthermore, PUD probably plays a role also in the pathogenesis of atrophic gastritis, gastric adenocarcinoma and mucosa-associated lymphatic tissue (MALT) lymphoma. Nowadays, albeit the more significant decrease has been in the duodenum, duodenal ulcers remain more common than their gastric counterparts [1, 2].

Currently, digestive bleeding is by far the commonest complication, followed by perforation, gastric outlet obstruction and fistulisation, in descending order of frequency. Risk factors associated with development of PUD complications include male sex, advanced age, comorbidities, alcohol, smoking, NSAID use, anticoagulation, corticosteroids and immunosuppressant medications. Mortality increases with age and comorbidities, and is higher in patients without previous history of PUD [1, 2, 8].

In the emergency setting, CT is largely used to rapidly assess patients with acute abdominal complaints, aiming to identify complications requiring hospitalization and surgery. In our experience, clinical suspicion of PUD is exceptional in urgent CT requests: as a result the disease most usually represents an unexpected finding in CT studies performed using routine abdominal protocols including a preliminary precontrast, an optional pancreatic phase (using a 35 s delay) and a mandatory portal-venous phase enhanced acquisition obtained 70–75 s after automated power injection of 110 to 130 ml of non-ionic iodinated contrast medium at a 2.5–3 ml/s flow rate, followed by saline flushing. When gastrointestinal bleeding is suspected, an arterial-dominant acquisition is performed using faster contrast injection (3.5–5 ml/s), bolus tracking with region-of-interest placed in the upper abdominal aorta and maximum-intensity projection (MIP) reconstructions; this technique is beneficial for identifying active contrast extravasation, which mandates endoscopic or interventional treatment. On our 64-slice CT scanner, the usual acquisition parameters used for abdominal studies are 120 KV, 300 mAs, 0.891 pitch, 0.75 s rotation time, and 64 × 0.625 mm collimation.

In most routine contrast-enhanced CT studies, the decompressed stomach and duodenum with peristalsis or retained enteric content are inherently difficult to assess. The inadequately distended stomach shows prominence of the gastric folds and apparent mural thickening; in absence of disease, the gastric wall thickness may reach up to 12 mm in the antrum. In patients with clinical suspicion or endoscopic detection of gastro-duodenal disease, the attending radiologist may opt for a dedicated CT study with gastric distension to assess the true mural thickness and structure. After fasting for at least 6 h, the patient is instructed to drink 500 to 750 ml of tap water as negative oral contrast during 15–20 min before examination and is then placed on the scanner table in either supine or prone (when antropyloric abnormalities are suspected) position. After hypotonisation using 20 mg of N-butylscopolamine intravenously, contrast-enhanced CT is obtained in pancreatic-dominant and portal venous phases [5‐7, 9].

Multidetector CT studies are routinely reconstructed along axial and coronal planes. Furthermore, as seen in the following imaging examples, thorough search for gastric and duodenal abnormalities benefits from additional sagittal and oblique study review, particularly focused on elucidating anatomy and abnormalities of the antropyloric tract and proximal duodenum.

The CT diagnosis of PUD relies on a combination of direct and indirect signs. Among the latter, the hallmark of gastro-duodenitis is represented by mural thickening of the affected portion of the stomach and of the proximal duodenum, which typically shows a stratified appearance corresponding to hypoattenuating oedematous submucosa and enhancing hyperaemic mucosa. This appearance may be either diffuse or focal surrounding the ulceration site, and is best perceptible in views perpendicular to the long axis, and in arterial-dominant or pancreatic phases rather than in portal-venous acquisitions. In the setting of HP- or NSAID-related PUD, gastro-duodenitis is generally segmental and limited to the distal gastric body and antrum (Figs. 1, 2, 3 and 4). The other useful indirect sign which alerts the keen radiologist to the possible presence of an active inflammation is represented by oedematous “stranding” of the perigastric or periduodenal fat (Figs. 1, 3 and 4). Sometimes, adjacent adenopathies may be present (Fig. 1) [5‐7].

The two direct signs of PUD include focal discontinuity of the mucosal hyperenhancement, which reflects disease reaching the muscularis mucosa, and identification of luminal outpouching: the latter corresponds to the ulcer crater which extends through and beyond the gastroduodenal wall (Figs. 2, 3 and 4). Literature from the multidetector CT era reported sensitivities of CT for ulcers of 29.6% to 54%, with an average size of 25 mm for visible ulcers compared to 16 mm of missed ones. Whereas even large ulcers may not be visible on CT with poor luminal distension, deep or penetrating ulcers may be identified with careful multiplanar scrutiny and paying attention to secondary findings [5‐7].

Bleeding is nowadays the most common PUD complication, which affects up to 15% of patients (incidence 19–57 cases/100.000 individuals/year) with NSAID use as key risk factor. Manifestations range from asymptomatic blood loss with positive faecal blood test, to melaena or haematemesis, to hypovolemic shock. Despite effective treatment, PUD-related haemorrhage tends to recur and up to 31% of patients will experience rebleeding within a month [1, 15, 16].

Albeit endoscopy remains the standard of care, as it allows both diagnosis and haemostasis, multidetector CT may prove useful in patients with suspected gastroduodenal bleeding. On preliminary unenhanced scans, dependent hyper attenuating debris from recent blood may be detected in the gastric fundus or duodenal lumen (Fig. 8). Blood products measure 30 to 45 Hounsfield units (HU), and the highest attenuation clot (45–70 HU) is generally found nearby the bleeding site. Caution should be paid to not misinterpret ingested material, medications, foreign bodies or residual contrast medium as blood, leading to falsely positive diagnosis.

With help from focused MIP reformations, multidetector CT may allow direct identification of active haemorrhage as intraluminal “jet-like” or “blush” contrast extravasation at the bleeding site, isoattenuating with enhanced vessels (Figs. 9, and 10). In PUD, arterial bleeding may originate from the gastroduodenal artery, such as in duodenal ulcers, or from the left gastric artery in gastric ulcers along the lesser curvature, and in our experience is generally focal and subtle. Alternatively, CT may suggest different causes of digestive haemorrhage, such as tumour or varices [5‐7].

When endoscopic haemostasis cannot be achieved, transcatheter arterial embolisation (TAE) (Fig. 10) may be helpful, particularly in frail patients. Compared to surgery, TAE has a slightly lower rate of haemostasis (91% versus 100%) and more frequent rebleeding (40% versus 15%), but is associated with fewer complications and better overall outcome [17].

Gastric outlet obstruction (GOO) refers to blocked passage of gastric contents into the duodenum from a partial or complete obstacle located at the distal stomach, pylorus, or duodenal bulb. Manifestations generally include vomiting and inability to eat. Currently, obstruction represents an uncommon PUD complication with a 2% incidence, in 80% of cases secondary to scarring from prolonged inflammation in untreated or long-standing bulbar ulcers [1, 15].

Plain abdominal radiographs generally show gastric overdistension with scarce or no air distally (Fig. 11a). Before endoscopy and biopsy, CT imaging allows confirming mechanical nature, level (transition point) and probable cause of GOO. The stagnant gastric content provides luminal distension without administering oral contrast, which is poorly tolerated and may potentially cause inhalation. GOO secondary to PUD is suggested by more or less symmetric, oedematous pyloric-duodenal mural thickening with interrupted mucosal enhancement or ulcer outpouching (Fig. 11). Alternatively, in our experience it may present as a tight stricture measuring 2–3 cm in length, without abnormal mural thickening, solid tissue or extrinsic mass (Fig. 12). Endoscopic balloon dilatation of benign GOO is feasible but rarely allows long-term cure [5, 7, 19].

Contrary to intraperitoneal perforation, which causes free peritonitis in the abdominal cavity, posterior duodenal ulcers may occasionally penetrate into the retroperitoneum or fistulise to adjacent organs such as the pancreas (Fig. 16), common bile duct (Fig. 17), gallbladder (Fig. 18) or liver. Clinical manifestations of retroperitoneal infections may be either hyperacute (epigastric pain and tenderness, fever) or insidious (abdominal or back pain, malaise weight loss). Leukocytosis and elevation of acute phase reactants, serum amylase, lipase and hepatic transaminases are generally observed [23‐27].

CT consistently shows peripherally enhancing pancreatic abscesses (Fig. 16), which most commonly result from superinfection of post-necrotic collections following severe acute pancreatitis. Conversely, the presence of air-attenuation components and signs of severe gastro-duodenitis suggest the correct diagnosis of fistulising PUD, which is crucial to prevent failed laparoscopy. Open surgical treatment including gastric resection or duodenectomy is generally required [23, 24].

Alternatively, PUD may fistulise into the gallbladder or biliary tract, causing superimposition of jaundice and ascending cholangitis on chronic PUD complaints. Spontaneous pneumobilia without previous surgery or instrumentation is the CT hallmark of bilio-digestive fistulas. Other suggestive findings include close adhesion and possible communication between posterior duodenum and either ventral aspect of pancreatic head (Fig. 17) or gallbladder (Fig. 18). The gold standard technique for biliary fistulas, endoscopic retrograde cholangiopancreatography (ERCP) is often unfeasible because of duodenal deformity and narrowing. Surgical management is indicated in cases refractory to medical therapy. Coledochoduodenal fistulas from PUD are very uncommon compared to cholelithiasis-related perforation into a normal duodenum: the latter are suggested by contracted gallbladder with mural thickening, and may give rise to further complications, such as gallstone ileus and Bouveret syndrome [7, 22, 25, 27, 28].

Albeit not the diagnostic technique of choice for suspected PUD, multidetector CT may allow aware radiologists to diagnose both uncomplicated PUD and complications such as haemorrhage, GOO, and biliary or pancreatic fistulisation, thus allowing for timely appropriate treatment [5, 7, 19].