SEARCH

SEARCH BY CITATION

Keywords:

  • inflammatory bowel disease;
  • Crohn's disease;
  • ulcerative colitis;
  • surgery lymphoma;
  • infliximab;
  • biologic therapy;
  • gastrointestinal carcinogenesis

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Background:

Primary intestinal lymphoma in the setting of inflammatory bowel disease (IBD) is uncommon and may be associated with immune suppressive therapy. We report clinical features and outcomes in patients with both conditions prior to use of biologic therapy.

Methods:

All patients with primary intestinal lymphoma and IBD at our institution from 1960–2000 were retrospectively identified. Data reported are frequency (proportion) or median (interquartile range). Kaplan–Meier analysis was performed.

Results:

Fifteen patients were identified: 14 (93%) were male, 10 (66%) had Crohn's disease. Median age at diagnosis of IBD and lymphoma was 30 (22–51) and 47 (28–68) years, respectively, with bloody diarrhea the most common presenting symptom for each diagnosis. Lymphoma location was colorectal in nine (60%), small bowel in four (27%), and one (6.25%) each: stomach, duodenum, and ileal pouch. Treatments were surgery plus chemotherapy (n = 6), surgery alone (n = 3), chemotherapy alone (n = 2), chemotherapy and radiation (n = 1), surgery and radiation (n = 1); two patients died before treatment. Most patients (n = 11, 73%) were Ann Arbor stages I or II. Large cell B-type histology was most common (n = 9, 60%). Three patients died within 30 days of lymphoma diagnosis. Survival free of death from lymphoma at 1- and 5-years was 78% and 63%, respectively, and was associated with advanced lymphoma stage (P = 0.004).

Conclusions:

Diagnosis and treatment of primary intestinal lymphoma in patients with IBD can be challenging and requires a high index of suspicion. Optimal survival requires multimodality therapy.(Inflamm Bowel Dis 2010;)

The influence of the chronic inflammatory process on gastrointestinal carcinogenesis in inflammatory bowel disease (IBD) has long been recognized.1 The risk of developing colorectal adenocarcinoma is estimated to be 1% annually, and may parallel the severity and chronicity of inflammation.2 The underlying etiology of IBD is unknown, but it is believed to be in part due to dysregulation of the intestinal immune system resulting in an inflammatory process. Potent immunomodulator agents such as thiopurines (azathioprine [AZA], 6-mercaptopurine [6-MP]), and more recently biological antagonists of tumor necrosis factor-alpha (TNF-α) such as infliximab (IFX), have demonstrated clinical effectiveness in reducing disease severity and preventing recurrence.3, 4

In the solid organ transplant literature, chronic immunosuppressive therapy is associated with the development of posttransplant lymphoproliferative disorders (PTLD).5 Emerging data suggest that the use of such agents in IBD may increase the risk of lymphoma.6 Whether IBD itself is associated with an increased risk of lymphoma remains unclear.7 Most studies of IBD patients have focused on the occurrence of lymphoma of any site (i.e., extraintestinal), not primary intestinal lymphoma (PIL); thus, the natural history of PIL in the IBD population remains relatively undefined. Reports of PIL in IBD have been case reports or small series, and the natural history and the prognosis of this coexisting disease state is not well established.8, 9 Therefore, we aimed to describe the clinicopathological features and long-term survival of IBD-associated PIL at our institution over the last four decades prior to the use of biologic therapies.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

After Institutional Review Board approval, we conducted a retrospective medical record review. All patients with both the diagnosis of intestinal lymphoma and IBD between January 1, 1960 and December 1, 2000 at our institution were identified from a centralized, prospective diagnostic index. This index was searched for potential cases. Inclusion criteria were that both the diagnoses of IBD and PIL had to have been made and/or confirmed at our institution. Patients without confirmed IBD or with nonprimary intestinal lymphoma were not included. All clinical, operative, pathological, and treatment data were abstracted from patients' medical records, including correspondence with referring physicians and cancer registry follow-up questionnaires. The Ann Arbor Staging system was used for pathologic tumor staging.

Descriptive univariate statistics were performed using JMP v. 7 for Windows (SAS Institute, Cary, NC), and results are reported as frequency (proportion) or median (interquartile range [IQR]). Patient survival, specifically death from lymphoma at 1 and 5 years, was estimated using the Kaplan and Meier survival method. Survival estimates were reported along with a 95% confidence interval (95% CI). Variable associations with survival free from death from lymphoma were assessed using the log-rank test. The alpha-level was set at P < 0.05 for statistical significance.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Over a 40-year period from 1960 to 2000, we estimated that ≈25,789 patients with IBD and ≈2,332 patients with primary intestinal lymphoma were seen at our institution. Of these, we identified a total of 15 patients who had been diagnosed with both IBD and PIL. Of these cases of PIL in the setting of IBD, the majority (11 [69%]) were diagnosed between 1990 and 1999, three were diagnosed prior to 1990, and one in the year 2000. The main clinicopathological features of the 15 patients are presented in Table 1.

Table 1. Baseline Characteristics and Clinicopathologic Features of 15 Patients with Inflammatory Bowel Disease and Primary Intestinal Lymphoma
Patient #IBD TypeGenderAge at Diagnosis of IBD, YearsAge at Diagnosis of PIL, YearsInterval Between Diagnoses, Years5-ASA UseSteroids Use6-MPSite of IBDSite of PILSame Segment? / Chronic InflammationPIL StagePIL Pathology
  1. IBD, inflammatory bowel disease; IQR, interquartile range; PIL, primary intestinal lymphoma; 5-ASA, 5-aminosalicylic acid; 6-MP, 6-mercaptopurine, LBCL, large B-cell lymphoma, MALT, mucosa-associated lymphomatic tissue.

1CUCMale275124nananaLeft colonRectosigmoidYes / yes2Diffuse LBCL (Hodgkin's lymphosarcoma)
2CDMale5959<1nananaTerminal ileumTerminal ileumYes / yes2Mixed type
3CDMale40455NoYesNoPancolitisRectosigmoidYes / yes4Mantle cell lymphoma
4CDMale234724NoYesNoTerminal ileumTerminal ileumYes / yes4Lymphoplasmacytic lymphoma
5CDMale51609NoYesNoTerminal ileumTerminal ileumYes / yes4Diffuse LBCL (liver involvement)
6CDMale1919<1NoYesNoTerminal ileum, right colonCecumYes / yes1Diffuse LBCL
7CUCMale276841YesNoNoPancolitisIPAANo / yes (pouchitis)1Diffuse LBCL
8CDMale67692NoNoNoTerminal ileumDuodenumNo / no1Marginal zone B-cell lymphoma, MALT type
9CDFemale14151NoYesNoIleocolitisCecumYes / yes2Diffuse LBCL
10CDMale708414YesYesNoCecum, traverse and des colonTransverseYes / yes1Diffuse LBCL
11CDMale4545<1NoNoYesIleum, duodenumTerminal ileumYes / yes1Diffuse LBCL
12CDMale307848YesNoNoTerminal ileum, Right colonStomachNo / no1MALT vs. solitary gastric plasmacytoma
13CUCMale16248YesYesYesPancolitisRectumYes / yes1Diffuse LBCL
14CUCMale22286NoYesYesCecalCecumYes / no1Diffuse LBCL (Dx at autopsy)
15CUCMale34351YesNoNoRectumRectumYes / no1Diffuse LBCL

Demographics

Overall, a total of 14 (93%) were male, 10 (66%) had Crohn's disease (CD) and five (33%) had chronic ulcerative colitis (UC). Median age at IBD diagnosis was 30 years (IQR, 22–51 years), and at diagnosis of lymphoma 47 years (IQR, 28–68 years). The median interval between diagnoses was 6 years (IQR, 1–24 years). In terms of IBD medication usage, a total of five (33%) of patients had ever received 5-aminosalicylic acid (5-ASA) derivatives, eight (62%) had ever received corticosteroids, and three (20%) had ever received 6-MP; the duration of 6-MP therapy was 1, 3, and 6 years. No patients in this series had ever received anti-TNF-α antibody or other biologic-based therapy.

Diagnosis of IBD and PIL

Bloody diarrhea was the most common presenting symptom at the time of the diagnosis of IBD, and at the time of diagnosis of PIL, occurring in 11 (73%) at time of diagnosis of IBD and eight (53%) at time of diagnosis of PIL. Other symptoms included abdominal pain, present at the time of diagnosis of IBD and PIL in 11 (73%) and eight (53%) of patients, and weight loss in two (13%) and five (33%), respectively. Less common symptoms were related to mass effect (n = 3), fistulas (n = 2), nausea/vomiting (n = 2), gastrointestinal hemorrhage (n = 1), and drainage from enterocutaneous fistula (n = 1). In this series no case of IBD presented as peritonitis, but two cases (13%) of PIL presented with acute intestinal perforation and required emergency surgery.

Lymphoma was diagnosed intraoperatively in eight (53%), by endoscopic biopsy in six (40%), and by autopsy in one patient (6%). Endoscopic descriptions of the appearance of the PIL lesions included: rubbery nodules, polyploidy masses, submucosal masses, and ulcerations. Ulcerations were described as large, deep, and associated with edema and mucosal friability. One area of lymphomatous involvement was found in an area of inflammation and narrow in the transverse colon which was not otherwise endoscopically suspicious for malignancy.

Of the eight patients diagnosed intraoperatively, indications for surgery included perforation (n = 3), abscess (n = 1), refractory chronic UC (CUC) (n = 1), ileal pouch–anal anastomosis (IPAA) dysfunction (n = 1), ileal stricture (n = 1), and right colonic mass (n = 1). Thus, not including the patient who presented with ileal stricture and the patient presenting with right colonic mass, PIL was unsuspected and incidental in six of eight (75%) of patients who required surgery. A gross pathologic specimen of a PIL in the setting of Crohn's colitis is shown in Figure 1.

thumbnail image

Figure 1. Gross pathologic specimen from a patient with Crohn's colitis and primary intestinal lymphoma. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

The extent of disease of IBD and the anatomic location of PIL are summarized in Table 2. Most commonly, patients had CD with terminal ileitis which was also the location of the PIL (n = 4, 26%), while pancolitis (two CUC, one CD) and development of rectosigmoid PIL (n = 3, [20%]) was the next most common anatomic combination. Terminal ileitis was also associated with PIL occurrence in the gastroduodenum (n = 2), small bowel (n = 4), and cecum (n = 2). One case of PIL developed in an IPAA and has been previously reported.10 Thus, the site of PIL was congruent with the location of IBD involvement in 12 patients (80%), 10 of which had pathologic evidence of chronic inflammation.

Table 2. Relationship Between Location and Types of Inflammatory Bowel Disease and Site of Primary Intestinal Lymphoma
Frequency (%) of PatientsType of IBDSite of IBDSite of PILSame Segments?
  1. IBD, inflammatory bowel disease; PIL, primary intestinal lymphoma; CUC, chronic ulcerative colitis; CD, Crohn's disease.

4 (26%)CUC (2), CD (2)PancolitisRectosigmoid (2), Transverse colon (2)Yes
4 (26%)CDTerminal ileumTerminal ileumYes
2 (13%)CDTerminal ileumStomach (1), Duodenum (1)No
2 (13%)CDTerminal ileumCecumYes
1 (6%)CUCPancolitisCecumYes
1 (6%)CUCProctitisRectumYes
1 (6%)CUCPancolitisIleal pouchNo

Treatment of IBD and PIL

Patients who ultimately received treatment for PIL fell into three treatment categories: 1) surgery plus adjuvant chemotherapy (n = 6); 2) surgery only (n = 4); and 3) chemotherapy alone without any surgical intervention (n = 3). Two (13%) patients also received adjuvant radiation therapy (both stage 4, one had right hemicolectomy followed by chemoradiation, and one had primary chemoradiation). Two patients died before specific therapy for PIL (one from myocardial infarction prior to planned chemotherapy, another from fulminant hepatitis with diagnosis of PIL at autopsy).

The indications for surgery were perforation (n = 3), mass (n = 2), and one of each the following: hemorrhage, refractory CUC, IPAA dysfunction, ileal ulcer, and stricture. Surgery performed included: ileocolic resections (four right hemicolectomies, one extended right hemicolectomy, and one re-do ileocolic resection); and other single procedures (total proctocolectomy with Brooke ileostomy, left hemicolectomy with Hartmann's pouch, IPAA excision, and small bowel resection). One patient had concurrent splenectomy, one had concurrent small bowel resection, and one patient underwent resection of the PIL by extensive (complex) endoscopic submucosal dissection—this patient was included in the surgical subgroup. Three (30%) of these 10 operations were emergent (right hemicolectomy, Hartmann's, small bowel resection).

Histopathologic examination of the specimens revealed that four (26%) of the lymphomas were not associated with chronic gastrointestinal inflammation. In terms of the lymphoma pathology, histological grading demonstrated that 10 patients (66%) had diffuse large B-cell lymphoma, two patients had mucosa-associated lymphoid tissue (“MALT-oma”), and one patient each had mantle cell, plasmacytoma, and lymphoplasmacytic subtypes. By the Ann Arbor classification, tumors were stage IE in nine (60%), stage IIE in four (25%), and stage IV in two (12.5%).

Epstein–Barr Virus (EBV) positivity was reported in four patients, including all three who had received 6-MP.

Short-term Survival

Three patients died within 30 days of the diagnosis of PIL. Two patients who presented with perforation and underwent surgery died within 30 days (one of pulmonary complications, the other from diffuse intravascular coagulation), while the third patient died of fulminant hepatitis prior to initiation of any specific therapy for PIL. There was no significant association between the need for surgery and death within 30 days of diagnosis of PIL (20% in each group, P = 1.0, Fisher's exact test).

Long-term Survival

Overall, lymphoma was the cause of death in four patients: three after surgery without subsequent chemotherapy (including the two short-term surgical deaths mentioned above, and one from progressive PIL not further specified), and one after primary chemotherapy without surgery (specific cause unknown but related to PIL). Other causes of death included congestive heart failure (after chemotherapy), myocardial infarction, fulminant hepatitis, sepsis related to endstage renal disease, and small cell carcinoma of the lung (n = 1 each). Overall 1- and 5-year survivals (any cause) were 67% (95% confidence interval [CI], 47%–95%) and 40% (20%–80%), respectively.

Survival free of death from lymphoma at 1- and 5-years was 78% (95% CI, 59%–100%) and 63% (37%–100%), respectively. Variables associated with mortality from lymphoma are shown in Table 3 and included only stage (P = 0.004). Survival free of death from lymphoma is depicted in Figure 2.

thumbnail image

Figure 2. Long-term survival free of death from lymphoma in IBD patients with PIL.

Download figure to PowerPoint

Table 3. Univariate Assessment of Variables with Survival Free of Death from Lymphoma
VariableNOne-Year Survival (95% C.I.)Five-Year Survival (95% C.I.)P-value*
  1. *Log-rank test.

Overall survival1578% (59 to 100)63% (37 to 100)
Age at diagnosis of PIL   0.83
 <50 years886% (63 to 100)57% (24 to 100) 
 ≥50 years769% (40 to 100)69% (34 to 100) 
Gender   0.61
 Female1100% 
 Male1477% (56 to 100)61% (36 to 100) 
IBD type   0.67
 Crohn's disease1080% (58 to 100)53% (23 to 100) 
 Ulcerative colitis575% (43 to 100)75% (39 to 100) 
Steroid use, ever   0.14
 No5100%100% 
 Yes886% (62 to 100)43% (10 to 100) 
6-MP use, ever   0.50
 No1090% (72 to 100)68% (37 to 100) 
 Yes3100%100% 
Abdominal pain at diagnosis of PIL   0.06
 No8100%80% (52 to 100) 
 Yes757% (30 to 100) 
Surgery   0.58
 Yes5100%67% (30 to 100) 
 No1070% (47 to 100)70% (36 to100) 
Stage   0.004
 19100%100% 
 >1650% (23 to 100) 

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

In our high-volume IBD practice we found PIL in the presence of IBD to be a rare entity. In patients with both PIL and IBD, symptoms of each were similar. Most lymphomas were seen in males, and areas afflicted by PIL found in the colorectum and the small intestine corresponded to a location of chronic disease. Additionally, most lymphomas were Stage I or II at diagnosis. Overall long-term survival free of death from lymphoma was reasonable at both 1 and 5 years (78% and 63%).

Whether the risk of lymphoma overall is higher among IBD patients compared to the general population continues to be debated.7 Several studies from hospitals or referral centers suggest a lymphoma risk in the setting of IBD, but these studies are subject to selection bias because they are nonpopulation-based studies. Some11–14 but not all15–18 population-based studies suggest that the risk is no higher than that of the general population. The largest and most recent study from several Swedish cohorts suggested that patients with CD carried a risk 30% higher than the general population, while CUC patients harbored no increased risk.15

The biological mechanism of this possible increased risk is not well understood. In general, chronic immunosuppressive therapy with AZA and 6-MP may play a role in the development of lymphoma. This phenomenon is observed in solid organ transplant recipients, who possess a higher risk of developing posttransplant lymphoproliferative disorder due to chronic immunosuppressive therapy.5 EBV infection, or carrier status in IBD patients, may also be a risk factor for PIL, as demonstrated by Dayharsh et al,19 who found that IBD patients who were treated with thiopurines and subsequently developed lymphoma were more likely to have tumors positive for EBV. The exact role of chronic intestinal inflammation in the pathogenesis of these tumors remains unclear, but the fact that most of these tumors occurred in areas chronically affected by IBD suggests that inflammation is a risk factor. Due to the rarity of PIL, the association and the etiology of PIL in IBD may be difficult to establish.

Nonetheless, because IBD and PIL can occasionally coexist, it is important for physicians caring for IBD patients to be aware of this relatively rare coincidence. As demonstrated in the present study, the clinical features of IBD and PIL can be similar. Bloody diarrhea and abdominal pain are the most common presenting symptoms for both conditions. Most of the patients in this series were diagnosed by endoscopic biopsy. Endoscopy with biopsies should be performed in patients with a change in clinical condition and as part of the evaluation of medically refractory disease. A histologic diagnosis of lymphoma is dependent on several factors and this should be conveyed to an experienced histopathologist.

Not surprisingly, decreased overall long-term survival is associated with advanced stage and the presence of abdominal pain at the time of diagnosis of PIL. This latter finding supports the notion that new abdominal pain in IBD patients must not automatically be ascribed to IBD. Also as expected, survival free of death from lymphoma appears mostly dependent on the stage of PIL. Interestingly, the mode of management of PIL (i.e., surgical or nonsurgical) did not affect long-term outcomes, but this is likely related to selection and coordinated multidisciplinary treatment. In the modern era, the majority of patients who present with PIL, especially in the setting of IBD, will likely be offered surgical resection (as indicated) followed by chemotherapy in order to consolidate the chance of a cure.

This descriptive study has limitations, including a small number of patients and the long study period, which limited our ability to accurately assess disease-free survival; thus, survival free of death from lymphoma was used as a surrogate outcome. In addition, this group of patients had a heterogeneous group of tumors. The treatment of both lymphoma and IBD has evolved considerably over this span of time, making it difficult to extrapolate to survival in a contemporary cohort. Current chemotherapeutic regimens such as C/EBP homologous protein (CHOP) have proven to be quite effective in treating lymphoma. Surgery in the management of lymphoma has evolved during this period as well, with staging laparotomy of historic interest only. Recently, Tondini et al20 reviewed their experience with a combined treatment approach that included both surgical resection and systemic chemotherapy. They demonstrated 99% complete tumor remission for patients with limited disease with an overall survival of 86%.8 Other reports have shown similar results proposing that adjuvant chemotherapy in completely resected localized PIL can decrease local and systemic relapse resulting in long-term disease free survival and overall survival compared to operation alone.9–11 Despite these limitations, the results of this study may offer healthcare providers some insight into the expected clinical features and outcomes in the rare patient with both conditions, and also may provide ongoing research into the relationship between various therapies, including biologic therapies, and lymphoma in IBD patients.

In conclusion, the diagnosis and treatment of primary intestinal lymphoma in patients with IBD can be challenging and requires a high index of suspicion. When discovered, a multidisciplinary approach with chemotherapy and surgical intervention may optimize survival.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES