Tumour or not tumour – that is the question
Presentation, investigations and management
A 73-year-old man presented with 1 week’s history of progressive epigastric pain. The pain was dull in nature and the patient reported no vomiting or radiation of pain. On presentation, the patient was found to have a low-grade fever (temperature, 37.8°C). There was no tea-coloured urine. The patient had enjoyed good past health except for a previous surgery for spinal injury.
Physical examination showed no jaundice. The patient’s upper abdomen was soft and non-tender. No mass lesion was felt. His bowel sound was normal.
Investigation showed elevated white blood cell count of 15.9 x 109/L and neutrophil count of 84.7 percent. Haemoglobin was 10.5 g/dL. Derangement of liver function was found, with an alkaline phosphatase level of 183 U/L, aspartate aminotransferase level of 41 U/L, alanine aminotransferase level of 48 U/L, and gamma-glutamyl transpeptidase level of 233 U/L. The patient’s C-reactive protein level was 278 mg/L, carcinoembryonic antigen level was 0.9 μg/L, and cancer antigen 19-9 level was 2 U/mL.
Ultrasound examination and non-contrast CT scan showed a normal gall bladder with no gallstone disease. The bile duct was not dilated. The pancreatic head and uncinate process were swollen. (Figure 1)
PET scan was performed, showing a markedly hypermetabolic lesion around the posterior aspect of the pancreatic head and the uncinate process (4.5 x 2.3 cm; SUVmax = 8.5). The lesion showed probable infiltration to the adjacent right mesenteric region. Another large, markedly hypermetabolic mass was detected around the portal region (5.8 x 4.1 cm; SUVmax = 7.9). In addition, focally hypermetabolic peripancreatic nodes were identified (3.1 x 1.4 cm; SUVmax = 4.7). (Figure 2) The common bile duct and intrahepatic bile ducts were not dilated. The gall bladder was distended without focally hypermetabolic lesions. The liver appeared normal. Radiological findings suggested primary pancreatic cancer with metastatic lymph nodes. Differential diagnoses included primary cholangiocarcinoma, inflammation and infection. Since a tissue diagnosis was needed to formulate a further treatment plan, the patient was referred for an endoscopic ultrasound (EUS) examination plus fine needle aspiration (FNA).
EUS with luminal examination showed shallow gastric ulcers at the gastric antrum. The first part of the duodenum was mildly deformed due to external compression. The major papilla was normal. Furthermore, a large heterogeneous mass measuring 5 cm was found in the caudate lobe of the liver. (Figure 3) The gall bladder was distended with sludge. The cystic duct was dilated due to sludge material and obstruction at the distal end. The cystic duct wall was thickened, mimicking a tumour mass lesion. The intrahepatic duct and common bile duct were not dilated, and there was no filling defect inside. An inflamed mass lesion measuring 3.5 cm was found at the pancreatic head. The background parenchyma of the pancreas was normal and the pancreatic duct was not dilated.
EUS-guided FNA (EUS-FNA) was performed using 25G aspiration needles for all punctures. Separate punctures were done to the caudate lobe lesion, cystic duct and pancreatic head lesion. The aspirate appeared pus-like. The EUS diagnosis was acute inflammation of the cystic duct due to obstruction by sludge material, with infection spreading to the liver caudate lobe and pancreatic head causing abscess collections, and stress-induced gastric ulcer disease.
Gastric ulcer biopsy showed benign ulcer. No Helicobactor pylori infection was identified. The fluid was sent for gram stain, culture and cytology examination. Cytology examination showed inflammatory cells with no malignancy. Culture of the aspirate showed mixed growth of Streptococcus mitis and Streptococcus intermedius, both being resistant to common antibiotics but sensitive to vancomycin and linezolid. The patient was started on vancomycin and linezolid, and the fever subsided. His liver function test improved after treatment. The patient was transferred to a university hospital for further treatment. Follow-up CT scan showed decrease in size of the liver and pancreatic lesions. The patient was scheduled to undergo elective cholecystectomy after complete resolution of the abscesses.
In daily clinical management of intra-abdominal emergency, ultrasonography and CT scan are commonly employed for diagnosis. However, when a mass lesion is detected, both examinations cannot differentiate between malignant disease, inflammation and abscess collection. PET scan can assess the metabolic activities of the lesion, but it cannot tell for sure the nature of the lesion. A tissue diagnosis is often required for planning the next step of treatment, either with surgery or palliative care if malignancy is confirmed.
Percutaneous biopsy is commonly used for tissue diagnosis. However, ultrasonography and CT-guided biopsy often cannot access deep-seated lesions such as pancreatic mass due to intervening major blood vessels and bowel structures. Although needle tract seeding is uncommon, the long needle path via the percutaneous route may theoretically increase the risk.1,2 EUS-FNA is shown to be effective in obtaining tissue diagnosis from the pancreas and other internal organs that are close to the gastrointestinal tract. In a recent meta-analysis of 41 studies including 4,766 patients, EUS-FNA demonstrated a pooled sensitivity of 86.8 percent, pooled specificity of 95.8 percent, positive likelihood ratio of 15.2, and negative likelihood ratio of 0.17 in correctly diagnosing the nature of solid pancreatic masses.3 Furthermore, the needle path is short and frequently included in the resection field, thus possibly reducing the risk of needle path seeding.4
In the current case, EUS-FNA of the liver and pancreas excluded the initial diagnosis of metastatic pancreatic cancer. It also provided an abscess sample for bacterial identification and antibiotic sensitivity testing, which had significantly altered the antibiotics used to target the infection. This echoes with findings of a previous study, which showed that in patients with acute pancreatitis and peri-pancreatic fluid collection, direct EUS-FNA fluid sampling could provide a more accurate bacterial diagnosis than blood culture.5
The infection in the current case was considered to be arising from the cystic duct obstruction by sludge disease despite normal ultrasound and CT findings. The cystic duct is a curved structure and the whole path may not be easily identified.6,7 EUS scope can be placed close to the common bile duct and cystic duct, and with the high-frequency ultrasound probe used, small stone and sludge disease can be easily identified. Furthermore, the detection rate is not affected by the size of the duct.8 Multiple studies have also confirmed the use of EUS as a triage method, allowing unnecessary endoscopic retrograde cholangiopancreatography (ERCP) to be avoided in up to two-thirds of patients with suspected bile duct stone disease.9
In conclusion, EUS is a valuable tool in the assessment of pancreaticobiliary diseases. EUS-FNA can provide cytology to differentiate between benign and malignant diseases, as well as fluid sample for bacteriological study. It could have a major impact in patient management.