Liver rheumatoid nodules imitating liver malignancy: a rare occurrence

Background

Extra-articular manifestations are present in around 40% of patients with rheumatoid arthritis (RA).1 These are commonly in the form of rheumatoid nodules found typically along the extensor surfaces of the forearms, elbows and rarely in the visceral organs.

We report a case of an incidental finding of a rheumatoid nodule in the liver of a 61-year-old man who presented with a single-reported episode of haemoptysis. Because the favoured radiological diagnosis was of a malignant lesion but subsequently confirmed to be non-neoplastic, in particular, a rare extra-articular manifestation of RA warrants sharing this case with colleagues.

Case presentation

A 61-year-old man with a background history of RA, ischaemic heart disease and previous alcohol excess presented to the emergency department reporting a single episode of fresh haemoptysis on the background of a recent upper respiratory tract infection. The patient had been diagnosed with RA 29 years previously and has been on regular methotrexate since as well as on etanercept for the past 3 years. Previous RA treatment included leflunomide, sulfasalazine and prednisolone. The patient was an ex-smoker with a 50 pack-year history and social drinker of less than two standard drinks a day for the previous 2 years, having previously experienced alcohol dependence drinking on average five standard drinks a day for 20 years with impairment on quality of life, continued use in spite of perceived adverse consequences and withdrawal symptoms with each attempt to cut down. Furthermore, he routinely underwent monitoring colonoscopies for benign colonic polyps. On examination, the patient was afebrile with an unremarkable respiratory examination and no evidence of haemoptysis at the emergency department. He had scars from a previous extensor pollicis longus tendon rupture and repair on his right wrist, a hitchhiker’s thumb and bilateral elbow nodules noted. Given the patient’s age and smoking history, a CT scan of the chest performed in the emergency department did not identify any cause of the haemoptysis, otherwise showing extensive background centrilobular emphysematous changes and bronchial wall thickening. However, an incidental 32×20×35 mm mass in segments VII and VIII of the liver, abutting the diaphragmatic surface, showing a small area of central calcification was identified. Given that there was no evidence of haemoptysis or concerning features on clinical assessment at the emergency department, the most likely cause of his haemoptysis was attributed to an exacerbation of his underlying chronic obstructive pulmonary disease, the patient was discharged with safety netting and a letter to his general practitioner to follow-up and investigate the newly identified liver mass.

Investigations

A follow-up ultrasound (US) scan 2 weeks after discharge at a private bulk billing radiology centre confirmed a 32×19×26 mm hypoechoic lesion with no appreciable vascularity consistent with the CT scan performed earlier. A biopsy was deemed too high risk due to the close proximity of the mass to the diaphragm. Subsequently, 2 weeks later, a multiphase multidetector CT (MDCT) scan of the patient’s abdomen was conducted in order to help rule out malignancy (figure 1). MDCT reconfirmed the aforementioned lesion with central calcification, no progression in size and no enhancement throughout the course of the examination. Its features were consistent with a sclerosed hepatic haemangioma, though this was absent from a prior CT conducted in 2006. Therefore, an MRI scan to further characterised the lesion was suggested. As the patient experienced great financial difficulty, an outpatient referral for the MRI scan was made to a public tertiary centre to eliminate an out-of-pocket expense. The transition from private to public health system resulted in a 2-month delay. Ultimately, the MRI scan (figure 2) of the mass revealed that the liver lesion had increased in size (from 35 mm to 40 mm) and showed lipid content and calcification.

A full battery of tests in an attempt to identify and characterise the liver mass were performed. Liver function tests revealed no indication of hepatitis or biliary disease. C reactive protein was within normal limits. Furthermore, renal function tests, chromogranin-A, carcinoembryonic antigen, alpha-fetal protein, carbohydrate antigen 19-9 revealed no concerns. Interferon gamma release assay for mycobacterial infections was negative. The patient had no hepatitis C antibody, HbsAg or HBcAb detected but HbsAb was detected consistent with his prior immunisation to hepatitis B. An upper gastrointestinal tract endoscopy was performed which revealed gastritis and no evidence of oesophageal varices. The patient had also routine follow-up colonoscopy less than 6 months prior to presentation which revealed benign colonic polyps.

An ¹⁸F-fluorodeoxyglucose (FDG) positron emission tomography (PET) scan (figure 3) was performed 2 months after the MRI which revealed that the segment associated with the mass was moderately FDG-avid (SUVmax 3.7). A second focus of poorly defined focus in segment V was identified, raising suspicion of an additional lesion. Furthermore, there were non-specific FDG-avid lymph nodes above and below the diaphragm, none of which appear particularly enlarged. The most active nodal uptake was seen in the portacaval location (SUVmax 5.9). These lesions were also non-specific in nature and differential included granulomatous disease, inflammatory processes and immunotherapies but their significance depended on the pathology of the liver lesions, but malignancy could not be excluded.

Differential diagnosis

Given the non-specific nature of the blood and imaging findings, and proximity of the lesions to the diaphragm precluding a fine needle aspiration biopsy, the differential diagnosis remained broad. Differential diagnosis included both neoplastic and non-neoplastic entities. The latter would include infectious causes for example, tuberculosis, particularly given the patient’s history of RA and immunosuppression.

Treatment

Following up on discussions between the rheumatology, hepatobiliary, oncology and radiology multidisciplinary teams, a laparotomy with segmental liver resection was deemed to be the treatment of choice given its increasing size and concern for malignancy. An intraoperative US confirmed the liver lesion (segment VII/VIII) adherent to the diaphragm (figure 4). No intravenous extension was noted and there was no peritoneal metastasis, evidence of colonic or small bowel malignancy or other liver lesions visible or palpable. The focus of poorly defined increased activity in segment V mentioned in the FDG-PET earlier was not identified despite specific interrogation of this area, also noting no abnormality was evident in this region in the preceding MRI. Two trucut core biopsies of the lesion were taken for frozen section and reported as showing inflammatory/fibrotic changes with no evidence of malignancy. A segment VII liver resection was performed.

Outcome and follow-up

The patient’s etanercept and methotrexate were ceased postoperatively and was commenced on regular prednisolone as per advice of the respiratory team until mycobacterium was excluded. At 2 weeks post-liver segmentectomy, the patient was able to return to his regular daily activities.

Macroscopic examination of the specimen showed a well-circumscribed lobulated pale lesion 45×32×20 mm abutting the capsular surface which had adherent fatty tissue (figure 5). The margins appeared well defined. Histological analysis showed serpiginous necrosis surrounded by a palisading rim of granulomatous inflammation (figure 6). The necrosis had a necrobiotic appearance and the granulomatous inflammation comprised of epithelioid histiocytes and scattered multinucleated giant cells (figure 7).

The differential diagnosis of granulomatous inflammation in the liver is wide and includes infection, in particular mycobacterial infection, and sarcoidosis among other possibilities.2 The morphology of the lesion was not typical for either of those two possibilities. Because of the necrobiotic appearance of the necrosis with a regular thin rim of palisading granulomatous inflammation in the setting of a patient with RA, the possibility of a rheumatoid nodule was raised.

Special stains for acid-fast bacilli and fungi were negative. Furthermore, microscopy, culture and sensitivity, and PCR for mycobacterium were negative. Further blood tests, including antimitochondrial antibodies, antineutrophil cytoplasmic antibodies and proteinase 3, were performed which were negative, reducing the likelihood that lesion was due to a vasculitic pathology.

At an outpatient review 6 weeks postoperation, the patient reported experiencing multiple episodes of exacerbating symptoms relating to his RA. As a result, etanercept was recommenced under the rheumatology team’s guidance.