Gastrointestinal (GI) Involvement with Non-Hodgkin's Lymphoma (NHL) Impacts Outcome; A Retrospective Single Tertiary Care Center Analysis of Clinico-Pathologic Characteristics Provides an Alternate Insight to Presentation and Outcome

Introduction

Historically, the GI tract is involved in 10-15% of patients with newly diagnosed NHL. Despite being the predominant extranodal site for lymphoma involvement little is known about its pathophysiologic presentation given its overall rarity. Nonetheless its evaluation, diagnosis, prognosis and management is distinct from lymphomas at other sites hence their recognition irrespective of whether being the primary site is prudent.

Methods

We performed a single center retrospective study of patients to describe clinical characteristics and outcome of newly diagnosed NHL presenting with GI involvement defined by tissue biopsy between January 2003 and April 2019. Descriptive statistics were calculated by Fisher's exact test and Chi2 test for categorical variables and rank-sum analyses for continuous variables.

Results

We identified 91 patients with one patient presenting with 2 different GI NHL involvement. The median age at presentation was 70 years (range 33-92) and 61 were male. 41 (45%) patients had diffuse large B cell lymphoma (DLBCL) and 5 had concomitant low grade component, 28 (30.8%) mucosal associated lymphoid tissue (MALT) lymphoma, 10 (11%) mantle cell lymphoma (MCL), 4 (4.4%) enteropathy associated T cell lymphoma (EATL), 3 (3.3%) follicular lymphoma, 2 (2.2%) Burkitt's lymphoma and 2 (2.2%) EBV driven lymphoproliferative disorder (LPD). There were 1 case each of plasmablastic lymphoma and extranodal NK/T-cell lymphoma, nasal type. 37 with DLBCL and 2 with MCL had primary GI involvement. Gastric involvement was the commonest site, 43 (55.8%) cases followed by the large intestine, 22 (28.6%). PET-CT scan was done in 42 cases with high grade NHL; GI involvement by FDG avidity was noted in 23 cases and negative in 19 (45.2%). Serum albumin at presentation was available in all patient. Patients with presenting serum albumin < 25 g/L were more likely to die in the first 30-days post diagnosis; p<0.0001. A total of 20 patients died as a result of lymphoma; presenting serum albumin was significantly lower in that group (mean 29 g/L versus 37 g/L; p=0.0077). H pylori status was assessed by either antigen testing in stool and/or histology. This was known in 50 patients and was available by the 2 assessment methods in 31 cases but discordant in 17. A positive result was more likely noted with the antigen test. None of the T-cell LPD were H pylori positive. 12 of the 26 DLBCL and 9 (53%) of the 17 MALT lymphoma patients were H pylori status was known were positive. Nonetheless all H pylori negative MALT lymphoma cases were localized to the stomach and responded to triple eradication therapy and Rituximab monotherapy +/- surgical resection. In the DLBCL cohort, the cell of origin was known in 22 of 39 cases; 17 were non-germinal center phenotype. In the MCL cohort, all patients whom underwent surveillance endoscopies irrespective of symptoms and/or radiologic finding had histologic evidence of GI involvement with lymphoma with only 3 (18%) patients being symptomatic +/- had radiologic finding in keeping with GI involvement.

Conclusions

Despite the relatively small sample size this real-world data of GI lymphomas outlines some of its recognized features; advanced age at diagnosis, male prevalence and gastric involvement as the commonest site of involvement. Presenting serum albumin <25 g/L was associated with higher 30-day mortality highlighting its relevance and need for its inclusion as a biomarker in risk stratification of NHL where GI tract is involved. There is however several other significant findings not previously recognized; the incidence of GI involvement in MCL is much higher than reported; 100% in all patients whom have undergone endoscopy in our cohort suggesting that GI involvement is a feature of MCL. With only 53% of MALT lymphoma being H pylori positive which is much lower than previously described this renders the prototypical association of H pylori infection in MALT lymphoma pathogenesis void and suggests an alternate etiology. Nonetheless their response to antibiotic therapy with only minimal additional systemic therapy in some indicates that H pylori status does not necessary predict its efficacy and should be considered when treating MALT lymphomas irrespective of H pylori status. CT scans but also PET-CT scans are insensitive in detecting GI involvement even in high grade NHL. Endoscopy with random mapping biopsies even with normal macroscopic GI findings remains the gold standard.