Diffuse large B-cell lymphoma of the rectum in a patient with Crohn’s disease

Abstract

A 62-year-old man presented to our institute with diarrhoea and dysuria on a background of subtotal colectomy and end ileostomy and biological therapy for Crohn’s disease. He was diagnosed with urinary tract infection and acute kidney injury (AKI). Renal ultrasound suggested left hydronephrosis, with renal protocol computed tomography (CT) showing a large pelvic mass. Magnetic resonance imaging (MRI) of the pelvis demonstrated a rectal tumour invading the bladder and compressing both ureters. He underwent cystoscopy, flexible sigmoidoscopy and positron emission tomography–CT and was diagnosed with stage IV non-Hodgkin’s diffuse large B-cell lymphoma. He was treated primarily with rituximab, cyclophosphamide, hydroxydaunomycin, oncovin and prednisolone chemotherapy regimen. He had ongoing urosepsis before admission for pelvic exenteration. He underwent cystoprostatectomy, excision of rectal stump and formation of ileal conduit. Histology showed no signs of residual malignancy. One year later, the patient was admitted to the intensive care unit with aspiration pneumonia, urosepsis and AKI. Despite maximal therapy, he developed multiorgan failure and passed away.

Background

Biological therapy has revolutionised treatment of inflammatory bowel disease (IBD) and other autoimmune diseases. Not only are biologics used to treat acute ‘flares’ of IBD but also they are used commonly to induce and maintain remission of disease. There has been much debate about the safety of prescribing long-term biological therapy due to a possible risk of malignancy due to a suppressed immune system, for example, lymphoma. Recent population studies and meta-analyses remain inconclusive regarding this increased risk.

The most likely diagnosis for colorectal mass in IBD is adenocarcinoma. Our patient had primary gastrointestinal lymphoma, which required a treatment approach drastically different from the one expected for adenocarcinoma.

Due to the ever-increasing use of biologics, their perceived risk of lymphoma and the relative paucity of information on such cases, we believe our case to be useful in aiding the investigation and treatment of future similar cases.

Case presentation

A 62-year-old man was admitted through the emergency department with diarrhoea, abdominal pain, dysuria and foul-smelling dark urine. On examination, he was haemodynamically stable and tender in the left flank and suprapubic region. No masses were appreciated on abdominal or digital rectal examinations. He had an ileostomy sited in the right iliac fossa, which was functioning appropriately.

His background included Crohn’s disease (CD) (status postsubtotal colectomy and end ileostomy), type 2 diabetes mellitus and hypertension. His CD can be classified as A3, L3 and B3 via Montreal classification, with age at diagnosis being 58 years, and the presence of penetrating disease (see further) affecting the colon and ileum. He was on multiple medications, including adalimumab, deltacortril, methotrexate and folic acid.

Although the patient's prior gastroenterology follow-up was performed at an outside hospital, we know that he was diagnosed with colitis in 2011 and underwent a subtotal colectomy with end ileostomy for fulminant colitis in June 2012. Two years prior to this admission, he underwent a flexible sigmoidoscopy that showed marked active proctitis with ulceration. One year prior, he underwent an ileoscopy that showed active ileitis and a repeat flexible sigmoidoscopy that showed acute and chronic inflammation in the rectum. His most recent flexible sigmoidoscopy, 3 months prior to admission, demonstrated chronic inflammation of the rectum, with no other findings visually or on biopsy.

Three years after his subtotal colectomy and 6 months prior to this admission, he was commenced on adalimumab 40 mg every 2 weeks. Prior to commencing adalimumab, he underwent standard preimmunosuppression testing (see further). Two months later, methotrexate 17.5 mg weekly was added, as well as Deltacortril 5 mg once per day. This treatment combination is unusual, and although it was commenced by an outside institution, we believe it was due to progressive worsening of his symptoms.

Routine blood work was taken, and a bedside urine dipstick was performed. These revealed a urinary tract infection (UTI) with concomitant acute kidney injury (AKI). Urine culture later revealed a mixed growth with predominance of Escherichia coli. He was treated with appropriate intravenous antibiotics and intravenous fluids, and had an ultrasound (US) of his urinary tract was performed. This revealed bilateral moderate hydronephrosis, and so he was booked for urgent CT of the kidneys, ureters and bladder (CT-KUB), which showed an ill-defined pelvic mass compressing the left ureter at the vesicoureteric junction, with resultant hydronephrosis and hydroureter (figure 1).

Figure 1.

Cross-sectional imaging of the pelvis demonstrating a large, infiltrating heterogenous mass involving the rectum and bladder.

At this juncture, his diagnosis was unclear. His working diagnosis was that of rectal adenocarcinoma invading the bladder. Less likely differentials included squamous cell carcinoma, gastrointestinal stromal tumour, and primary or secondary colorectal lymphoma. He underwent MRI of the pelvis, which clarified that this was a 14.0×8.0×8.5 cm mass interposed between the posterior wall of the bladder and the rectal stump.

The mass was compressing both ureters, and while it appeared to be arising from the rectum, it was difficult to delineate tissues. A CT of the thorax, abdomen and pelvis (CT-TAP) showed an enlarged right inguinal lymph node but no signs of metastatic disease. He therefore underwent dual endoscopic investigations with cystoscopy and flexible sigmoidoscopy (figure 2). Placement of ureteric stents was attempted at this time but failed due to the presence of a large, cobbled mass and sloughy necrotic tissue obscuring the ureteric orifices. Multiple biopsies were taken from both rectum and bladder. He then underwent bilateral percutaneous nephrostomies to treat hydronephrosis and had a three-way catheter placed to aid in drainage of the bladder and decompression of the ureters and kidneys.

Figure 2.

Flexible sigmoidoscopy images demonstrating extensive mural necrosis and sloughing.

Endoscopic biopsies returned a diagnosis of non-Hodgkin’s diffuse large B-cell lymphoma (DLBCL). The patient’s care was transferred to the haematology team at this time. Initial positron emission tomography (PET)–CT gave a diagnosis of stage IV DLBCL, with uptake in a necrotic pelvic mass, right inguinal lymph node, liver lesion, stomach lesion and right testicle. He was discussed at the haematology multidisciplinary team (MDT) meeting and was commenced on six cycles of rituximab, cyclophosphamide, hydroxydaunomycin, oncovin and prednisolone (R-CHOP) regimen. Intrathecal methotrexate was also used to supplement treatment, in order to decrease the probability of brain involvement in the future.

His inpatient stay was complicated by recurrent urosepsis, neutropaenia, hydronephrosis, Vancomycin Resistant Enterococcus and Extended Spectrum Beta-Lactamase producing organisms. These were treated with cystoscopy and washout and with multiple courses of intravenous antibiotics. He was transferred to the high dependency unit for one instance of septic shock that required pressor support from which he recovered quickly. Repeat bilateral percutaneous nephrostomies were inserted with anterograde ureteric stents.

Following R-CHOP, a repeat PET–CT scan showed reduced uptake in the rectal stump with no distant disease. However, following MDT discussion, it was recommended that the patient undergo surgical excision of the rectal stump and bladder. This was primarily due to multiple admissions with urosepsis (including intensive care unit (ICU) admission) secondary to a persistent rectovesical fistula. Indeed, preoperative optimisation was difficult as he required multiple admissions via the haematology day ward with episodes of urosepsis. He was later admitted electively for anterior pelvic exenteration, in which he underwent cystoprostatectomy, excision of rectal stump and formation of ileal conduit. He was discharged 1 week later. Histology showed necrotic tissue with fistulation between the anterior rectal wall, base of prostate and posterior bladder wall. It showed no signs of dysplasia or malignancy.

He had ongoing stable hydroureter bilaterally, consistent with expected reflux disease. One year later, the patient was transferred to the ICU from a peripheral hospital, with a combination of aspiration pneumonia, urosepsis and AKI. He deteriorated, requiring dialysis and pressor support. Unfortunately, he developed multiorgan failure and passed away in the ICU.

Investigations

Initial investigations on admission, revealing UTI and AKI

• Haemoglobin: 119g/L.

• Mean corpuscular volume: 83.4 fL.

• White cell count: 10.9×10⁹/L.

• Neutrophils: 8.0×10⁹/L.

• Platelets: 484×10⁹/L.

• Sodium: 132 mmol/L.

• Potassium: 5.0 mmol/L.

• Urea: 10.7 mmol/L.

• Creatinine: 265 mmol/L.

• Urine dipstick: 3+blood, 3+protein, 3+leukocytes.

• Mid-stream specimen of urine: WCC>1000, colony count>100 000 colony-forming units/mL; mixed growth, predominantly E. coli.

Preimmunosuppression investigations in outside institution, prior to commencement of adalimumab

• HIV – negative.

• Hepatitis A antibodies: positive.

• Hepatitis B surface antigen: negative.

• Antibody to Hepatitis B core antigen: negative.

• Hepatitis C antibodies: negative.

• Varicella Zoster Virus Immunoglobulin G >100 mIU/mL: protected against primary inf.

• Cytomegalovirus Immunoglobulin M and Immunoglobulin G: negative.

• Toxoplasma Immunoglobulin G: positive.

• Epstein-Barr virus (EBV) Immunoglobulin M: negative.

• EBV viral capsid antigen Immunoglobulin G : positive.

• Treponema pallidum: negative.

• Borrelia burgdorferi IgG: negative.

• Legionella urine Ab: negative.

Radiological imaging, in chronological order

In turn, these revealed signs of compression of the urinary system, a pelvic mass, invasion of this mass from rectum to bladder and no distant metastasis.

• Renal US: moderate bilateral hydronephrosis.

• CT-KUB: left hydronephrosis and hydroureter, ill-defined pelvic mass compressing the left ureter at the vesicoureteric junction.

• MRI of the pelvis: 14.0×8.0×8.5 cm rectal mass invadingthe posterior wall of the bladder and compressing the ureters bilaterally.

• CT-TAP: 13 cm pelvic mass, centred on the rectal stump with suspected involvement of the adjacent urinary bladder; moderate hydroureter and hydronephrosis; enlarged right inguinal lymph node; no evidence of metastatic disease.

Endoscopic procedures

Cystoscopy and flexible sigmoidoscopy were planned for visualisation of the pelvic mass, biopsy for diagnosis and ureteric stenting for treatment of hydroureter and hydronephrosis.

• Cystoscopy and examination under anaesthesia: unable to place stents bilaterally due to the presence of a large, cobbled mass. The posterior wall and trigone were covered in sloughy necrotic tissue.

• Flexible sigmoidoscopy to 20 cm: an area from 7 to 10 cm was suspicious for neoplasia, but this was difficult to determine due to dense sloughy exudate. Again, biopsy was sent for cellular pathology.

Cellular pathology of bladder and rectum specimens

• Non-Hodgkin’s DLBCL, diffuse infiltration of lymphoid cells, CD20 positive, CD10 negative, Multiple Myeloma 1 (MUM1) positive, non-germinal centre B subtype and Technetium-99M sestamibi (MIBI) 50%–60%.

Differential diagnosis

We initially felt on sigmoidoscopy that this necrotic mass may have been a large adenocarcinoma due to his history of IBD. Less likely differentials that we also considered were a large gastrointestinal stromal tumour, or primary or secondary lymphoma. Biopsies quickly demonstrated that the appearances were that of a lymphoma rather than carcinoma or stromal tumour. Further immunohistochemical testing with CD20 and CD10 and MUM1 confirmed primary DLBCL.

Treatment

This lymphoma was initially treated with six cycles of of R-CHOP chemotherapy and intrathecal methotrexate. Once follow-up PET–CT revealed no presence of residual active lymphoma, he was planned for surgical resection. He underwent pelvic exenteration, consisting of a cystoprostatectomy, excision of the rectal stump and formation of an ileal conduit.

Outcome and follow-up

He was discharged 1 week postpelvic exenteration. Histology showed necrotic tissue, with fistulation between the anterior rectal wall, the base of the prostate and the posterior bladder wall. There were no signs of dysplasia or malignancy.

He had ongoing stable hydroureter bilaterally, consistent with expected reflux disease. One year later, the patient was transferred to the ICU from a peripheral hospital, with a combination of aspiration pneumonia, urosepsis and AKI. He deteriorated, requiring dialysis and pressor support. Unfortunately, he developed multiorgan failure and passed away in the ICU.

Discussion

IBD, which can be broken down into CD and ulcerative colitis, refers to a chronic disease with characteristic changes to the gastrointestinal tract due to infiltration of inflammatory cells.¹ These diseases both have phases of relapse and remission, and the goal of therapy is to achieve and maintain remission.²

A large repertoire of immunosuppressive agents exists for treating CD, and they can be used in myriad combinations. Typical agents for treating CD include prednisolone, cyclosporine, methotrexate, 6MP, azathioprine and tumour necrosis factor alpha (TNFα) inhibitors, such as infliximab and adalimumab.³ Immunosuppressive agents are most often chosen for patients with moderate to severe disease, or for those who do not respond to steroid therapy.¹ Important considerations for treatment include disease severity, presence of perianal or extraintestinal manifestations, and when to start/stop immunosuppressive agents.²

Thiopurines (6MP and azathioprine) are best recommended for long-term maintenance and are often used with corticosteroids for induction immunosuppression as they take too long to be effective for monotherapy. Azathioprine is one of the more widely used immunosuppressive agents and is effective for the treatment of active CD or for maintaining remission.¹ Methotrexate is effective for active CD and has an extensive list of potential adverse effects, but these can be minimised by coprescribing folic acid.¹

TNFα inhibitors, such as infliximab and adalimumab, have been used to treat IBD for over 25 years.⁴ They target TNFα, a well-known proinflammatory cytokine. While there is concern that high-dose and long-term use of TNFα inhibitors may increase the risk of opportunistic infection and malignancy, they have been proven effective at treating IBD that is refractory to other medical therapies.²

Infliximab is recommended for patients who have failed other medical therapies, are intolerant of other therapies or have fistulating CD.⁵ The Clinical Assessment of Adalimumab Safety and Efficacy Studied as Induction Therapy (CLASSIC-I), Gauging Adalimumab Efficacy in Infliximab Non-Responders (GAIN) and Crohn’s Trial of the Fully Human Antibody Adalimumab for Remission Maintenance (CHARM) trials advocated for the use of adalimumab in induction of remission, maintenance of remission and in non-responders to infliximab, respectively.⁶

As IBD and immunosuppressive agents are both linked with an increased risk of malignancy, it is crucial to attempt to tease out specific risks. In our case, it is important to assess the risk of lymphoproliferative disorders (LPDs), in particular, as our patient developed non-Hodgkin’s DLBCL. Although CD is a well-known risk factor for small and large bowel adenocarcinoma, there is no conclusive link between CD and LPDs to date.^{1 3} In fact, the presence of primary intestinal lymphoma in the setting of CD is actually quite uncommon.⁷ The incidence of primary colorectal lymphoma has been reported as comprising approximately 0.5% of large bowel malignancies.⁸ The most common of these is DLBCL.⁹

An explanation for the possible increased incidence of LPD in patients with CD has been attributed to immunosuppression. Rationale for this theory has borrowed heavily from similar patient cohorts, such as post-transplant patients or patient with rheumatoid arthritis (RA), as the effects of immunosuppression in these groups have been studied in more depth.³ The incidence of non-Hodgkin’s lymphoma (NHL) has increased in the last 50 years, and this coincides with an increasing number and popularity of immunosuppressive agents.^{8 9}

Another contributing factor which has been largely discussed is the presence of EBV infection.^{1 3} A link between LPD and EBV is well established in the literature. As EBV is not well studied in IBD, researchers have again referred to patients suffering from post-transplantation LPDs, or from RA as there is a significant crossover of immunosuppressive agents between these cohorts. Patients with immunosuppressed RA were found to have a higher risk of Epstein-Barr Virus-associated LPD, and this risk was proportional to dosage.

The risk of malignancy imparted by thiopurines is one of the most clearly described in the literature. Dayharsh et al cited that use of thiopurines worsened the risk of EBV-associated LPD from 17% to 50%.¹⁰ The duration of treatment increases the risk of developing lymphoma, but stopping treatment reduces the risk.³ A prospective, observational cohort study in France—the Cancers Et Surrisque Associé aux Maladies inflammatoires inestinales En France or CESAME cohort study—found that 75% of incident cases of NHL occurred in patients taking thioprine medications.¹¹ They found a multivariate-adjusted HR of 5.28 between patients taking thioprine medications and patients who had never taken them.¹²

A meta-analysis of LPD in IBD found a fourfold to sixfold increase in lymphoma risk in patients taking azathioprine or 6MP. As this risk returned to baseline on stopping treatment, the authors suggested that immunosuppression, not permanent damage to DNA, was responsible for the increased risk of lymphoma. This increased risk was seen in patients treated for longer than 1 year.¹³

With regard to TNFα antagonists, to date, there is no clear evidence of the risk of malignancy that they imbue on patients. A systematic review of 22 randomised controlled trials revealed that these therapies were not associated with an increased risk of malignancy in patients with IBD. Unfortunately, these trials were limited as none provided follow-up for greater than 1 year.¹⁴

A further overview of systematic reviews and meta-analyses of patients with various inflammatory diseases found some studies linked anti-TNFα medications and lymphoma risk.¹⁵ However, these were limited again by insufficient follow-up and by inadequate sample size. It was hypothesised that it may be effects of the more advanced disease in cases requiring TNFα medications than the effects of the drugs themselves.

There are multiple challenges within this area of research. One problematic confounder is disease severity because it is difficult to tease out which complications could be due to disease severity when assessing the effects of medications.^{3 6} It is also difficult to single out the effects of any one drug, as the vast majority of patients have been on multiple combinations of drugs at any one time, and adverse effects may be due to a different medication or even due to a combined effect.^{3 16} There is also a risk of ascertainment bias, as the higher rates of lymphoma seen in level 1 evidence may be due to patients with worse disease being sent to referral/specialist centres.^{1 13} In our patient’s case, many of the aforementioned confounders apply. Our patient had severe disease, had been exposed to multiple immunomodulatory drugs and had varying combinations of these drugs.

In conclusion, it is difficult to study the impact of either CD or one of the drugs of choice for treating CD on risk of lymphoma for multiple reasons. Confounding factors such as disease severity and polypharmacy, a lack of long-term follow-up and ascertainment bias all complicate this field of research. Theories can tentatively be put forward by studying similar groups, but these theories are largely untested to date. Further research is needed in order to ascertain (1) the actual impact of each drug on lymphoma risk, (2) any particular subgroups at increased risk (eg, due to age and EBV status) and (3) whether the risks inferred from analysis of other patient groups translate to patients with CD.

Learning points.

• In patients presenting with rectal lymphoma that is causing compression of the ureters, hydronephrosis and acute kidney injury can develop rapidly and require prompt treatment.

• Rectal lymphoma requires the coordination of multiple disciplines, including colorectal surgery and haematology/oncology, and possibly other services such as nephrology and urology as outlined in our case.

• In patients with refractory inflammatory bowel disease requiring long-term use of biological therapy, the risk of lymphoma is higher than that in the general population. While it is difficult to define this risk (as discussed previously), a higher index of suspicion should be kept for these patients.