Primary Gastrointestinal Lymphoma: A Retrospective Cohort Study on Clinical Presentation, Treatment Outcomes, and Survival Trends With a Focus on Emergency Versus Elective Management

ABSTRACT

Purpose

Primary gastrointestinal lymphoma (PGIL) is a rare cancer, with diffuse large B‐cell lymphoma (DLBCL) as the most common subtype. PGIL can be acute, requiring emergency surgery, or non‐acute, allowing elective management. This article evaluates the clinical presentation, treatments, and survival outcomes of PGIL by comparing emergency and elective cases.

Methods

A retrospective cohort study from January 2013 to December 2019 included patients with histologically confirmed PGIL, excluding secondary GI involvement. Survival distributions were performed using SPSS v.20 and Kaplan–Meier analysis.

Results

Among 33 patients, 54.5% were male, with a mean age of 69. The most common site of lymphoma was the small bowel (54.5%), with DLBCL being the predominant subtype (66.7%). Emergency cases comprised 57.6% of the cohort. Poorer survival was noted in emergency cases, with significant differences in survival by age (p = 0.036) and lymphoma site (p = 0.038). Surgical excision was the main diagnostic method in emergency cases (54.5%), while endoscopic biopsy was more common in elective cases (39.4%) (p < 0.001). Chemotherapy was given to 69.7%, with R‐CHOP (Rituximab, Cyclophosphamide, Doxorubicin, Vincristine, Prednisone) (65.2%) being the most used regimen. Median survival was 17 months (interquartile range, IQR: 10–44.5). Survival differences were significant by age (p = 0.036) and lymphoma site (p = 0.038).

Conclusion

Emergency surgery is often needed for PGIL, especially with small bowel lymphomas, which have poorer outcomes. Early diagnosis and elective care may improve prognosis. Further research should explore prognostic markers and standardize treatment.

1. Introduction

Gastrointestinal (GI) lymphomas, although rare, can present with acute or non‐acute symptoms to general surgery [1]. Between 5% and 20% of extra‐nodal lymphomas occur in the GI tract, while primary GI lymphomas account for only 1%–4% of all GI cancers [2]. The incidence of primary gastrointestinal lymphomas (PGIL) lymphomas, or hollow viscus lymphomas, has been increasing globally in recent decades, with potential risk factors including celiac disease, Helicobacter pylori infection, immunosuppressive agents, EBV or HIV infection, and inflammatory bowel disease [3, 4, 5]. Primary GI lymphomas can involve any part of the GI tract, with the stomach being the commonest site (60%–70%), followed by the small intestine, ileocecum, and colorectum [6, 7]. Most primary GI lymphomas are non‐Hodgkin's lymphomas (NHL) [8], with diffuse large B‐cell lymphoma (DLBCL) and mucosa‐associated lymphoid tissue (MALT) lymphoma being the main histological types [9, 10]. The distribution of this varies geographically, with B‐cell lymphomas more common in Western countries and T‐cell lymphomas more prevalent in Eastern countries [11, 12].

Primary gastrointestinal lymphomas often lack specific clinical indicators, making them difficult to differentiate from other gastrointestinal tumors. This can unfortunately lead to missed diagnoses or misdiagnoses [13]. However, advancements in endoscopic and radiological imaging techniques have enabled earlier diagnoses in recent years. Despite this progress, the optimal treatment for PGIL remains unclear, particularly in advanced cases, due to the heterogeneity of the disease [14]. Management strategies are diverse, ranging from antibiotic therapy for H. pylori ‐associated MALT lymphoma to aggressive multi‐agent chemotherapy for DLBCL. Factors such as tumor location, immunophenotype, histological type, and tumor staging play a crucial role in determining the best treatment approach.

In this retrospective longitudinal study, we aim to evaluate the demographic characteristics, treatment outcomes, follow‐up, and survival of patients presenting with a hollow viscus lymphoma from data collected within 6 years.

2. Methods

2.1. Study Design and Study Setting

This was a retrospective cohort analysis conducted in York and Scarborough Teaching Hospitals over a seven‐year period starting from January 1, 2013, to December 31, 2019. The patients diagnosed with hollow viscus lymphoma involving the GI tract constituted the focus of the study. The study aimed to evaluate the demographics and clinical characteristics along with treatment outcomes, survival, and follow‐up data of these patients.

2.2. Inclusion and Exclusion Criteria

The inclusion criteria included:

• –Age was not a consideration, offered with a confirmed histopathological diagnosis of primary gastrointestinal lymphoma affecting any part of the GI tract.

The exclusion criteria were:

• –The exclusion was for patients with secondary involvement of the GI tract by systemic lymphoma.
• –The other exclusion included patients who presented with insufficient clinical information or could not be located after follow‐up.

2.3. Case Definitions

An acute case was defined as one requiring immediate surgery for either bowel obstruction, perforation, or acute GI bleeding causing hemodynamic instability.

A non‐acute case was defined as one not requiring immediate surgery due to symptoms such as paraneoplastic syndrome, weight loss, chronic abdominal pain, anemia, or any other non‐specific GI symptoms.

2.4. Data Collection and Variables

Data were obtained retrospectively through electronic medical records review from the hospital's main patient database. Included in variable analyses were demographics, age, gender, and comorbidities, clinical presentation, acute versus non‐acute cases, presenting symptoms, and initial findings, as well as work‐up, imaging studies including CT, MRI, and PET‐CT scans, endoscopic findings from gastroscopy, colonoscopy, and capsule endoscopy, and histopathology to confirm the lymphoma subtypes and staging. Details of treatment, including surgical interventions (laparotomy, bowel resection, anastomosis, and stoma formation), chemotherapy regimens, radiotherapy, and additional therapeutic measures taken, were collected. Finally, patient outcomes were evaluated via the assessment of postoperative complications (anastomotic leak, ileus, wound infections, and bleeding), with due completion of chemotherapy, survival status and mortality, and follow‐up months. The follow‐up period for each patient was calculated from the date of diagnosis to the date of death or the study's end date, December 31, 2019.

2.5. Statistical Analysis

Data were collected from the study using version 20.0 of the SPSS Package. Descriptive statistics, focusing on demographic information and patient clinical characteristics with the frequency and percentage for categorical variables and median interquartile ranges (IQR) for continuous variables, were used. Comparison between categorical data was performed using the chi‐square test or Fisher's exact test where appropriate, particularly when cell counts were less than five, while that for continuous data was done first by checking for normality using the Shapiro–Wilk test. Data on continuous variables were analyzed using either the independent samples t‐test (when the variables are normally distributed) or the Mann–Whitney U test (when the variables are not normally distributed). The survival analysis was performed using the Kaplan–Meier method, and differences between groups were compared with the log‐rank test. Survival was defined as overall survival (OS), calculated from the date of diagnosis to the date of death from any cause or the last follow‐up date. Univariate factors for survival analysis were performed using a Cox proportional hazards regression analysis (backward stepwise method), and hazards were expressed as hazard ratios (H.R. and 95% CI). Statistical significance was defined as p < 0.05 for all the data analyses.

3. Results

A total of 33 patients were recruited for the study, the majority of them being males (n = 18) and emergency patients (n = 19). The mean age at diagnosis was just over 69 years. Abdominal pain was the most common presenting symptom, being experienced by approximately half of the patients. In terms of the location of lymphoma, the small bowel was the most common site, while diffuse large B cell lymphoma was the most common subtype (2/3rd patients). A detailed breakdown of the baseline demographic variables has been presented in Table 1.

TABLE 1.

Baseline demographic variables.

Variable	Value
Gender (n, %)
Male	18 (54.5)
Female	15 (45.5)
Age at diagnosis, mean SD	69.2 ± 14.3
Source (n, %)
Elective	14 (42.4)
Emergency	19 (57.6)
Location (n, %)
Scarborough	10 (30.3)
York	23 (69.7)
Frailty status (n, %)
1	7 (21.2)
2	7 (21.2)
3	6 (18.2)
4	9 (27.3)
5	3 (9.1)
6	1 (3.0)
Dementia (n, %)	1 (3.0)
Previous history of lymphoma (n, %)	4 (12.1)
Symptoms on presentation (n, %)
Abdominal pain	16 (48.5)
Bowel obstruction	6 (18.2)
Paraneoplastic syndrome	11 (33.3)
Location of lymphoma (n, %)
Gastric	7 (21.2)
Small bowel	18 (54.5)
Colon	8 (24.2)
Metastasis (n, %)	1 (3.0)
Subtype of lymphoma (n, %)	22 (66.7)
Diffuse large B cell
Hodgkin	1 (3.0)
Follicular	3 (9.1)
Mantle cell	1 (3.0)
Marginal zone (MALT)	3 (9.1)
Burkitt's	2 (6.1)
T cell	1 (3.0)
Mode of diagnosis (n, %)
Endoscopic biopsy	13 (39.4)
Lymph node biopsy	1 (3.0)
Percutaneous biopsy	1 (3.0)
Surgical excision	18 (54.5)

Table 2 shows the breakdown of type of therapies and patient outcomes. Chemotherapy was utilized in approximately 70% of the patients, with more than half of these patients being subjected to 6 chemotherapy cycles. In terms of type of chemotherapy, RCHOP was the most common, seen in 15/23 patients. Majority of the procedures were emergency procedures (n = 19) as compared to elective (n = 1). Small bowel resection was the most common type of bowel resection (n = 12), followed by right hemicolectomy (n = 5). Majority of the patients did not suffer from post‐operative complications, and for those who did, ileus, wound infection/dehiscence, and anastomotic leak was seen in 2 patients each. Overall, mortality was seen in 12 patients, and 75% of these patients had a reason of death related to lymphoma/chemotherapy/surgery (Table 2).

TABLE 2.

Type of therapies and patient outcomes.

Variable	Value
Chemotherapy (n, %)	23 (69.7)
Cycles of chemotherapy (n, %) a
1	3 (13.6)
2	3 (13.6)
3	3 (13.6)
4	2 (9.1)
5	12 (54.5)
Type of chemotherapy (n, %) a
– RCHOP—Rituximab + Cyclophosphamide, Hydroxydaunorubicin (Doxorubicin), Oncovin (Vincristine), Prednisone	15 (65.2)
– R DHAP—Rituximab + Dexamethasone, High‐dose Ara‐C (Cytarabine), Platinum (Cisplatin)	1 (4.3)
– RCVP—Rituximab + Cyclophosphamide, Vincristine, Prednisone	3 (13.0)
– ChIVVP—Chlorambucil, Vinblastine, Procarbazine, Prednisone	1 (4.3)
– Rituximab	1 (4.3)
– R‐CODOC‐M Rituximab + Cyclophosphamide, Vincristine (Oncovin), Doxorubicin (Adriamycin), high‐dose Methotrexate OR R‐IVAC—Rituximab + Ifosfamide, Etoposide (VP‐16), high‐dose Ara‐C (Cytarabine)	1 (4.3)
– R Bendamustine—Rituximab + Bendamustine	1 (4.3)
–Radiotherapy (n, %)	1 (3.0)
Emergency vs. elective (n, %)
Emergency	19 (57.6)
Elective	1 (3.0)
Type of bowel resection (n, %)
Small bowel resection	12 (36.4)
Hartmann's procedure	1 (3.0)
Right hemicolectomy	5 (15.2)
Duodenal resection	1 (3.0)
Segmental colectomy	1 (3.0)
Laparoscopic (n, %)	3 (9.1)
Stoma (n, %)	3 (9.1)
Post‐operative complications (n, %)
None	13 (39.4)
Ileus	2 (6.1)
Upper GI bleed	1 (3.0)
Wound infection/dehiscence	2 (6.1)
Anastomosis leak	2 (6.1)
Intraabdominal collection	1 (3.0)
Temporary vision loss	1 (3.0)
Post‐operative PE (n, %)	1 (3.0)
Length of stay, median (IQR)	11.0 (6.3–20.5)
Mortality (n, %)	12 (36.4)
Reason of death related to lymphoma/chemotherapy/surgery (n, %) a	9 (75)
Survival in months, median (IQR)	17.0 (10.0–44.5)

^a Percentage derived from total applicable patients.

In terms of the relationship between source and study variables, a few statistically significant differences were seen. For location, Scarborough had more elective patients while York had more emergency patients (p = 0.002). In terms of presenting symptoms, majority of those with elective procedures had paraneoplastic syndrome on presentation while abdominal pain was more common in emergency patients (p = 0.013). All patients with gastric lymphomas were treated via elective procedure, while majority of those with small bowel lymphoma were treated via emergency procedure (p = 0.001). For elective procedures, the most common mode of diagnosis was endoscopic biopsy, while surgical excision was the most common modality for emergency procedures (p < 0.001) (Table 3).

TABLE 3.

Relationship between source and study variables.

Variable	Category	Elective	Emergency	p
Location	Scarborough	8	2	0.002*
	York	6	17
Gender	Female	9	6	0.131
	Male	5	13
Frailty status	1	2	5	0.723
	2	3	4
	3	2	4
	4	5	4
	5	1	2
	6	1	0
Dementia	No	13	19	0.424
	Yes	1	0
Previous history of lymphoma	No	13	16	0.620
	Yes	1	3
Age at diagnosis	—	74.9 ± 6.9	65.0 ± 16.9
Mortality	No	10	11	0.324
	Yes	4	8
Reason of death related to lymphoma/chemotherapy/surgery	No	12	12	0.141
	Yes	2	7
Survival in months	—	22.5 (15.3–45.8)	11.0 (9.0–45.0)	0.054
Symptoms on presentation	Abdominal pain	6	10	0.013*
	Bowel obstruction	0	6
	Paraneoplastic syndrome	8	3
Location of lymphoma	Gastric	7	0	0.001*
	Small bowel	4	15
	Colon	3	4
Subtype of lymphoma	Diffuse large B cell	7	15	0.367
	Hodgkins	1	0
	Follicular	2	1
	Mantle cell (MALT)	1	0
	Marginal zone	2	1
	Burkitt's	1	1
	T cell	0	1
Chemotherapy	No	6	4	0.118
	Yes	8	15
Cycles of chemotherapy	1	1	2	0.622
	3	1	2
	4	1	2
	5	0	2
	6	5	7
Type of chemotherapy	R CHOP	3	12	0.148
	R DHAP	0	1
	RCVP	1	2
	ChIVVP	1	0
	Rituximab	1	0
	R‐CODOC‐M/R‐IVAC	1	0
	R Bendamustine	1	0
Radiotherapy	No	13	19	0.424
	Yes	1	0
Mode of diagnosis	Endoscopic biopsy	12	1	< 0.001*
	Lymph node biopsy	0	1
	Percutaneous biopsy	1	0
	Surgical excision	1	17
Emergency bowel resection	No	13	0	< 0.001*
	Yes	1	19
Type of bowel resection	Small bowel resection	0	12	0.332
	Hartmann's procedure	0	1
	Right hemicolectomy	1	4
	Duodenal resection	0	1
	Segmental colectomy	0	1
Laparoscopic	No	1	16	0.412
	Yes	0	3
Stoma	No	1	16	0.510
	Yes	0	3
Post‐operative length of stay	—	7.0 (7.0–7.0)	12.0 (6.0–21.0)	0.243

^*Statistically significant.

Only one patient received external beam radiotherapy to the abdomen with a total dose of 30 Gy delivered in 15 fractions over 3 weeks. Treatment was planned using CT‐based simulation, with 3D conformal fields encompassing the primary gastric lesion and regional lymphatics. No acute Grade ≥ 2 toxicities were documented during or immediately after radiotherapy.

A log rank test was run to determine if there were differences in the survival distribution for the different age groups: < 60 years old and 60 years old and above. The survival distributions for the two groups were statistically significantly different, χ² (1) = 4.40, p = 0.036. Figure 1 shows the Kaplan–Meier curve for survival in months based on these age groups. Overall, cumulative survival was noted to be better in patients aged < 60 years old from 0 to 20 months, and then 70 months onwards. For those aged more than 60 years, the cumulative survival was noted to be slightly better than those < 60 years from 20 to 70 months. The overall difference in cumulative survival rates was statistically significant.

FIGURE 1.

Kaplan Meier curve for survival distribution based on age groups.

A log rank test was also run to determine if there were differences in the survival distribution for the different sites of lymphoma involvement: gastric, small bowel, and colon. The survival distributions for the three sites were statistically significantly different, χ ² (2) = 6.34, p = 0.042. As seen in Figure 2, those with colon lymphomas had the highest cumulative survival rate up until approximately 20 months, after which the rate declined steeply and was the worst survival rate of the 3 groups from 20 months onward. Overall cumulative survival was noted to be better for those with gastric lymphomas compared to small bowel lymphomas.

FIGURE 2.

Kaplan Meier curve for survival distribution based on sites of lymphoma involvement.

4. Discussion

PGIL is a rare neoplastic disease forming a small percentage of all gastrointestinal tumors, with the commonest histological form of DLBCL [15]. Diagnosing and managing this disease is quite a challenge, from mild gastrointestinal discomfort and non‐specific symptoms to an acute surgical disaster that mandates intervention [15, 16]. Where chemotherapy continues to be the mainstay of treatment, the role of surgery is still a matter of debate in terms of obstruction, perforation, or other instances of hemorrhage [17]. Advances in immunochemotherapy, particularly the addition of rituximab to standard regimens like R‐CHOP, have significantly improved survival outcomes in patients with PGIL [18]. Still, it is of concern how emergency or elective presentation might influence the determination of prognosis. In this study, we undertook an analysis of the clinical presentation and treatment of PGIL that would assess the difference between elective and emergency presentations.

Among the cohort of 33 patients, there was a slight male preponderance (54.5%); the mean age at diagnosis was 69.2 years. Small bowel lymphoma (54.5%) was the most involved site, followed by the colon (24.2%) and gastric lymphoma (21.2%). DLBCL was the most common histological subtype (66.7%), confirming previous reports. Abdominal pain (48.5%) was the most reported symptom, while paraneoplastic syndrome (33.3%) and bowel obstruction (18.2%) were also notable.

A striking finding from our study is that a majority of patients (57.6%) presented as emergencies requiring immediate surgical intervention. This reflects the often vague and non‐specific initial symptoms of PGIL, such as abdominal pain, which can lead to delayed diagnosis until a life‐threatening complication like bowel obstruction or perforation occurs. This high rate of emergency surgery, particularly for small bowel cases, is a central theme of our findings and directly impacts prognosis.

Diagnosis occurred via many modes in elective and emergency cases, with surgical excision being the primary diagnostic modality for emergencies (54.5%). Endoscopic biopsy was more common in elective cases (39.4%); p < 0.001. Chemotherapy was given to 69.7% of patients, R‐CHOP (65.2%) being the most commonly used regimen. This rate is explained by the clinical realities of this cohort, which included a significant number of emergency presentations and patients with notable frailty. Consequently, some patients were too unwell to tolerate chemotherapy, died before treatment could be initiated, or declined therapy. The overall mortality was 36.4%, with a sobering median survival of 17 months. This poor survival outcome is influenced by the high proportion of acute surgical cases and is also linked to the fact that nearly a third of patients did not receive chemotherapy, often due to postoperative morbidity that precluded or delayed subsequent oncological treatment.

Overall mortality was 36.4%, of which 75% was due to lymphoma, chemotherapy, or surgical complications. Statistically significant differences in survival were observed based on lymphoma site (p = 0.021) and age (p = 0.038), with poorer survival trends noted in emergency cases.

Our findings support those from previous studies, particularly those from the group of Owattanapanich et al. [19], who conducted a 10‐year cohort study on 175 PGIL cases in Thailand. They too found DLBCL as the most common subtype, with the stomach (38.9%) and small intestine (23.4%) being the most common sites. However, unlike their cohort, which was mainly from elective presentations, in our study, there was a higher proportion of emergency cases—57.6%, which may have translated to shorter median survival (17 months) compared to their 5‐year overall survival rate of 64.4%.

Furthermore, our findings can be contrasted against those of Virijevic et al. [20] in terms of prognosis for MALT lymphoma, where the significance of hypoglycemia and high lactate dehydrogenase (LDH) in the gastrointestinal lymphoma was stressed. While these were not major points of analysis in our study, they could serve as potential markers for future research. Of note is the finding in their study of a 5‐year overall survival rate of 64%, which is high compared with our observed survival rates; this is in all likelihood due to the less aggressive nature of MALT lymphoma compared to DLBCL.

A key finding of our study is the predominance of small bowel lymphoma (54.5%) and its strong association with emergency presentation (p < 0.001). This contrasts with some literature where gastric lymphoma is more common. Intestinal lymphomas may present acutely with obstruction or perforation more frequently than their gastric counterparts due to the narrower lumen of the bowel, leading to a higher rate of emergency surgery. This underscores the need for heightened clinical suspicion for lymphoma in patients presenting with acute abdominal symptoms, particularly small bowel obstruction, to facilitate earlier diagnosis and potentially avoid emergent, high‐risk surgery.

Ge et al. [16] find gastric lymphoma to be more common than small bowel, a conclusion contrary to this study's results that small bowel was more frequently involved. They indicated that surgical treatment is a vital aspect of primary intestinal lymphoma (PIL) treatment, granting further veracity to the finding that emergency bowel resection was done significantly more often in small bowel lymphomas (p = 0.001).

Virijevic et al. [20] identified hypoproteinemia and high levels of lactate dehydrogenase (LDH) as the key prognostic factors in MALT lymphoma, variables not examined in the present study but possibly arising as promising prognostic factors in cohort studies in the future. Also, a key study is that of Juárez‐Salcedo et al. [21], which emphasized the great correlation of H. pylori infection with MALT lymphoma, which could suggest that eradication therapy might be a treatment option for a subgroup of PGIL patients. Although we did not assess H. pylori status in our study, these findings warrant the need for routine screening, should they actually present within a gastric lymphoma case.

We also noted a low rate of laparoscopic surgery (9.1%). This is largely attributable to the high proportion of emergency presentations in our cohort (57.6%). In an emergency setting, such as acute bowel obstruction or perforation, an open laparotomy is often required to gain rapid surgical control, manage widespread contamination, and safely resect compromised bowel segments. The complexity and hemodynamic instability of patients in these situations frequently preclude a minimally invasive approach.

4.1. Strengths and Limitations

We present several compelling advantages behind our study. First, we focus on the diagnostic approaches and the impact of emergency versus elective presentation; the literature has missed this comparison. We made a detailed comparison of diagnostic approaches in these cases, wherein surgical excision was the primary diagnostic method used in emergency cases; on the other hand, endoscopic biopsy was most often used in elective cases. The study included Kaplan–Meier survival analysis to evaluate trends in survival; some other studies fail to have this included.

However, several limitations must be acknowledged. The small number of cases (n = 33) and its single‐center nature hinder broader statements about this association's generalizability. Further, the retrospective nature of our study might bring in bias from missing data and treatment heterogeneity, of which we have no control. While other studies have analyzed specific prognostic markers such as LDH and albumin levels, we did not assess such biochemical markers, which would help in gaining further insights into survival outcomes. Detailed histological manifestations and immunophenotypic profiles of PGIL and specific manifestations of paraneoplastic syndromes were not uniformly documented in the patient records, which limited our ability to provide detailed subgroup analyses. Finally, H. pylori testing was not performed in this study, thereby limiting our effort in understanding any potential role of H. pylori in the prognosis of PGIL. Future prospective studies on PGIL, particularly gastric MALT lymphoma cases, should routinely include H. pylori testing to better stratify patients and explore its prognostic significance.

5. Conclusion

The present study reiterates the high incidence of emergency presentations in PGIL, especially with small bowel involvement, underscoring the dire consequences of acute presentations and survival. This finding underscores the necessity of an early diagnosis with elective management to obtain optimal results for the patients. Additional research should preferably focus on discovering other new prognostic markers, including biochemical parameters, and multi‐center prospective studies to substantiate these findings. Addressing those related to emergency presentations and modification of protocols for these patients would greatly impact the management of surgical PGIL for improving survival.

Ethics Statement

This audit was registered with the Scarborough General Hospital Clinical Audit Department under reference number AUD325. In accordance with local guidelines, ethical approval was not required for this project as it was conducted as a clinical audit.

Consent

As the audit involved the analysis of anonymized patient data, individual consent to participate was not required. The principles of confidentiality and data protection were adhered to throughout the audit process.

Conflicts of Interest

The authors declare no conflicts of interest.

Alshamari T., Chong P., Yi L. M., Pergassam D., and Gnanendran D., “Primary Gastrointestinal Lymphoma: A Retrospective Cohort Study on Clinical Presentation, Treatment Outcomes, and Survival Trends With a Focus on Emergency Versus Elective Management,” JGH Open 9, no. 10 (2025): e70283, 10.1002/jgh3.70283.

Data Availability Statement

The data that supports the findings of this study is available from the corresponding author, DG, upon reasonable request.