Clinical Characteristics and Independent Risk Factors of Primary Gastrointestinal Lymphoma Complicated with Gastrointestinal Complications Authors: Chun-Lin Ying, Tian-Qi Zhao, Qin-Xue Pu, Lin-Yan Wan, Wei Liu Published: 2026 ABSTRACT BackgroundThis study aimed to investigate the clinical characteristics and risk factors of primary gastrointestinal lymphoma (PGIL) in patients with gastrointestinal complications (GICs).MethodsA retrospective analysis was conducted on the clinical data of 148 patients with PGIL admitted to Yichang Central People’s Hospital from January 2019 to August 2025. Univariable and multivariable logistic regression models were utilized to identify independent risk factors for GICs.ResultsPatients with PGIL exhibited a high incidence (58.1%) of GICs, primarily bleeding, obstruction, and perforation. Bleeding predominantly involves gastric ulcerative lesions, whereas obstruction and perforation frequently affect the small intestine, with obstruction typically presenting endoscopically as protruding lesions. Compared to the non-GIC cohort, patients with GICs showed a significantly higher proportion of diffuse large B-cell lymphoma (DLBCL), poorer performance status (ECOG ≥ 2), and advanced tumor invasiveness. Furthermore, the GICs group demonstrated marked metabolic depletion, elevated systemic inflammation, significant bowel wall thickening, and lymphadenopathy. Crucially, multivariable analysis identified four independent risk factors for developing GICs: history of hypertension (odds ratio [OR] =3.724, CI: 1.149–12.074), primary intestinal site (OR=3.551, 95% CI: 1.240–10.170), deep lesion infiltration into the muscularis or serosa (OR=3.615, 95% CI: 1.132–11.546), and anemia (OR=4.288, 1.457-12.615).ConclusionThe high incidence of GICs in patients with PGIL is multifactorial. Early risk stratification and tailored interventions are crucial for patients presenting with hypertension, anemia, intestinal involvement, or deep tumor infiltration at the initial diagnosis to optimize therapeutic outcomes. Introduction Primary gastrointestinal lymphoma (PGIL) represents the most prevalent manifestation of extranodal lymphoma (EL), accounting for 10–15% of all non-Hodgkin lymphomas (NHL) and 1–4% of all gastrointestinal malignancies, with its epidemiological incidence demonstrating an upward trend in recent years.1–5 The clinical manifestations in the early stage of the disease often present as insidious and non-specific.6 However, with deep infiltration of tumor tissue into the digestive tract wall, some patients often experience sudden severe gastrointestinal complications (GICs), which mainly include three types: gastrointestinal bleeding (GIB), perforation (GIP), and obstruction (GIO). The occurrence of GICs not only easily induces severe abdominal infections and multiple organ failure but also leads to the forced delay of antitumor chemotherapy cycles and poor postoperative healing, which greatly increases the short-term mortality rate of patients.7–10 Currently, the diagnosis of GICs mainly relies on urgent laboratory tests, imaging evaluations, endoscopic exploration, or confirmation by laparotomy.11 Based on their biological behavior, lymphomas are primarily classified into aggressive and indolent subtypes. Aggressive PGIL is characterized by significant tumor bulk and a highly invasive nature, rendering patients highly susceptible to developing GICs.12,13 In contrast, indolent lymphomas exhibit a relatively gradual disease progression.13,14 To date, the optimal therapeutic strategy for PGIL remains to be definitively established. Common treatment modalities encompass chemotherapy, radiotherapy, surgery, targeted therapy, anti-Helicobacter pylori (H. pylori) eradication, and combination therapies; clinicians typically formulate individualized regimens based on a comprehensive evaluation of the specific lymphoma subtype.15,16 Nevertheless, once critical GICs arise, emergent intervention is mandated, wherein surgery remains the principal salvage modality.13,17 Currently, research on the clinical characteristics and risk factors of patients with PGIL and concurrent GICs is relatively scarce, mostly consisting of case reports,18–23 and there is still a lack of standardized Multidisciplinary Team (MDT) intervention guidelines in clinical practice. Therefore, this study systematically reviewed and analyzed the differences in baseline characteristics, endoscopic phenotypes, imaging, and pathophysiology between patients with PGIL with and without concurrent GICs, aiming to identify independent risk factors for inducing GICs, thereby providing solid evidence-based medical evidence for early warning of high-risk diseases and formulating individualized comprehensive treatment plans. Patients and Methods Patients This study retrospectively analyzed 148 patients with PGIL admitted to the Yichang Central People’s Hospital between January 2019 and August 2025. The research protocol was reviewed and approved by the Ethics Committee of Yichang Central People’s Hospital (approval number: 2025–436-01), and the study was conducted in accordance with the Declaration of Helsinki. In view of the anonymized and retrospective nature of the study, the Ethics Committee formally waived the requirement for informed consent. Diagnostic and Inclusion Criteria All enrolled patients were diagnosed with gastrointestinal lymphoma through pathological examination of specimens obtained by digestive endoscopic forceps and/or surgical resection and strictly met the diagnostic criteria proposed by Dawson et al24 Tumor staging was performed using the Lugano staging criteria for PGIL.25 GIB was defined as the presence of overt bleeding (hematemesis, melena, or hematochezia) or occult bleeding (positive fecal occult blood test).26 GIO includes complete intestinal obstruction and partial obstruction,27 with the most common symptoms including abdominal pain, abdominal distension, vomiting, and/or cessation of flatus and defecation. GIP often presents as the presence of free air under the diaphragm or perforation observed during exploratory laparotomy.7 This study excluded patients with secondary gastrointestinal lymphoma, a history of malignant tumors, previous chemotherapy, or incomplete important medical records. All cases were rechecked and confirmed by multidisciplinary experts from the Departments of Pathology, Gastroenterology, Hematology, and Gastrointestinal Surgery at our hospital. Clinical Data Collection and Radiological Assessment Upon admission, all 148 patients in this group underwent routine blood, liver function, and blood biochemistry tests. Among them, 129 patients completed lactate dehydrogenase (LDH) determination and 77 patients completed β2-microglobulin (β2-MG) determination. In addition, inflammatory indicator tests were conducted: 60 patients completed erythrocyte sedimentation rate (ESR) determination, 120 patients completed C-reactive protein (CRP) determination, 64 patients completed procalcitonin (PCT) determination, 40 patients completed ferritin (FER) determination, 30 patients completed Interleukin-2 (IL-2) determination, 31 patients completed Interleukin-4 (IL-4) and Interleukin-10 (IL-10) determinations, and 34 patients completed Interleukin-6 (IL-6) determination. 132 patients completed abdominal or pelvic computed tomography (CT) scans, 117 patients completed ordinary endoscopy, and 15 patients completed endoscopic ultrasonography (EUS). Additionally, 67 patients (45.3%) underwent positron emission tomography-computed tomography (PET-CT) examinations. All clinical data in this study were extracted from the electronic medical record system of the Yichang Central People’s Hospital. Baseline laboratory data were defined as the latest test results obtained after the patient’s initial admission and prior to the initiation of any targeted anti-tumor treatments such as chemotherapy and radiotherapy. To minimize subjective bias, the baseline clinical characteristics were retrospectively evaluated by the MDT. The evaluators were blinded to the patient groupings and final clinical outcomes (ie, the occurrence of GICs). Any scoring discrepancies were resolved by the MDT consensus. Additionally, to avoid potential interpretation bias, two experienced radiologists independently reviewed abdominal, pelvic CT and other imaging images. During the evaluation, both radiologists were strictly blinded to patients’ clinical grouping, endoscopic or surgical findings, and final clinical outcomes. The two physicians assessed imaging features independently, and inconsistent conclusions were reconciled through mutual discussion or arbitration by a third senior radiologist. Definition of Clinical Variables To ensure the reproducibility of our findings and standardize the evaluation of clinical variables, explicit operational definitions were established for key risk factors. (1) Hypertension was defined as having a systolic blood pressure ≥ 140 mmHg and/or a diastolic blood pressure ≥ 90 mmHg on repeated measurement.28 (2) In 1968, the World Health Organization (WHO)29 established the diagnostic criteria for anemia in populations under 65 years of age, defining it as a hemoglobin level of < 130 g/L for men and < 120 g/L for women. Although hemoglobin levels naturally decline with advancing age, the vast majority of current studies focusing on elderly cohorts continue to utilize these classic WHO criteria.30 (3) Primary intestinal involvement was anatomically defined as lesions located in the duodenum, small intestine, colon, or rectum, as confirmed by endoscopic examination or surgical exploration. (4) Deep lesion infiltration was defined as tumor invasion extending into or beyond the muscularis propria (including the serosa). Statistics Statistical analysis was performed using SPSS 27.0. The Shapiro–Wilk test was used to test the normality of the measured data. Data that conformed to a normal distribution are expressed as the mean ± standard deviation (\documentclass[12pt]{minimal} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \DeclareFontFamily{T1}{linotext}{} \DeclareFontShape{T1}{linotext}{m}{n} {linotext }{} \DeclareSymbolFont{linotext}{T1}{linotext}{m}{n} \DeclareSymbolFontAlphabet{\mathLINOTEXT}{linotext} \begin{document}$\bar x \pm s$\end{document}), and an independent sample t-test was used for comparisons between groups. Data that did not conform to a normal distribution are expressed as median (interquartile range) [M(IQR)], and the Mann–Whitney U-test was used for comparison between groups. Count data are expressed as the number of cases and percentage (n, %), and the Pearson chi-square test or Fisher’s exact test was used for comparison between groups (depending on the sample size and expected frequency). Variables with a P-value < 0.05 in the univariable analysis were subsequently incorporated into a multivariable binary logistic regression model (using the backward stepwise method) to identify independent risk factors for the occurrence of GICs in patients with primary gastrointestinal lymphoma. All statistical analyses were two-tailed, and a P-value < 0.05 was considered statistically significant. Results Comparison of Baseline Characteristics A total of 148 PGIL patients were included in this study, among which 86 patients (58.1%) developed GICs, and 62 patients (41.9%) did not develop GICs. The median age of all patients was 62 years (IQR 55–72 years), with more males (83 cases, 56.1%) than females (65 cases, 43.9%). There were 35 cases (23.6%) with a history of hypertension, 12 cases (8.1%) with a history of diabetes, and 17 cases (11.5%) with a history of heart disease. The comparison of baseline data between the two groups is shown in Table 1. Compared with the non-GICs group, the proportions of patients in the GICs group who were aged ≥ 60 years, had a history of hypertension, had an Eastern Cooperative Oncology Group (ECOG) Performance Status (PS) score ≥2, had a high-risk International Prognostic Index (IPI) score, were in Lugano stage IV, had lesions invading the muscular layer/serosal layer, had lesions located in the intestine, and were initially diagnosed by surgical resection were significantly higher, with statistically significant differences (P < 0.05); there were no statistically significant differences in gender, history of diabetes, and history of heart disease between the two groups (P > 0.05).Table 1Comparison of Baseline Characteristics Between the GICs and Non-GICs CohortsItemNon-GICs(n=62, %)GICs(n=86, %)Total(148, %)P valueAge ≥60 years old29(46.8)58(67.4)87(58.8)0.012Sex0.939 Male35(56.5)48(55.8)83(56.1) Female27(43.5)38(44.2)65(43.9)Hypertension history9(14.5)26(30.2)35(23.6)0.026Diabetes history3(4.8)9(10.5)12(8.1)0.216Heart disease history7(11.3)10(11.6)17(11.5)0.949ECOG PS score0.003 <258(93.5)64(74.4)122(82.4) ≥24(6.5)22(25.6)26(17.6)IPI grouping<0.001 Low-risk46(83.6)37(48.1)83(62.9) Low-intermediate-risk8(14.5)22(28.6)30(22.7) Intermediate-high-risk1(1.8)5(6.5)6(4.5) High-risk0(0.0)13(16.9)13(9.8)Lugano stage0.027 I–IIE53(85.5)60(69.8)113(76.4) IV9(14.5)26(30.2)35(23.6)Depth of lesion<0.001 Mucosa/submucos52(83.9)41(47.7)93(62.8) Muscular/serosal layer10(16.1)45(52.3)55(37.2)Location<0.001 Stomach45(72.6)32(37.2)77(52.0) Intestine17(27.4)54(62.8)71(48.0)Methods of diagnosis<0.001 Endoscope52(83.9)38(44.2)90(60.8) Surgery10(16.1)48(55.8)58(39.2)Notes: Data are presented as n (%). Values highlighted in bold italics indicate statistically significant results.Abbreviations: GICs, gastrointestinal complications; ECOG PS, Eastern Cooperative Oncology Group Performance Status; IPI, International Prognostic Index; Lugano stage, Staging system for primary gastrointestinal lymphoma. Table 1Comparison of Baseline Characteristics Between the GICs and Non-GICs CohortsItemNon-GICs(n=62, %)GICs(n=86, %)Total(148, %)P valueAge ≥60 years old29(46.8)58(67.4)87(58.8)0.012Sex0.939 Male35(56.5)48(55.8)83(56.1) Female27(43.5)38(44.2)65(43.9)Hypertension history9(14.5)26(30.2)35(23.6)0.026Diabetes history3(4.8)9(10.5)12(8.1)0.216Heart disease history7(11.3)10(11.6)17(11.5)0.949ECOG PS score0.003 <258(93.5)64(74.4)122(82.4) ≥24(6.5)22(25.6)26(17.6)IPI grouping<0.001 Low-risk46(83.6)37(48.1)83(62.9) Low-intermediate-risk8(14.5)22(28.6)30(22.7) Intermediate-high-risk1(1.8)5(6.5)6(4.5) High-risk0(0.0)13(16.9)13(9.8)Lugano stage0.027 I–IIE53(85.5)60(69.8)113(76.4) IV9(14.5)26(30.2)35(23.6)Depth of lesion<0.001 Mucosa/submucos52(83.9)41(47.7)93(62.8) Muscular/serosal layer10(16.1)45(52.3)55(37.2)Location<0.001 Stomach45(72.6)32(37.2)77(52.0) Intestine17(27.4)54(62.8)71(48.0)Methods of diagnosis<0.001 Endoscope52(83.9)38(44.2)90(60.8) Surgery10(16.1)48(55.8)58(39.2)Notes: Data are presented as n (%). Values highlighted in bold italics indicate statistically significant results.Abbreviations: GICs, gastrointestinal complications; ECOG PS, Eastern Cooperative Oncology Group Performance Status; IPI, International Prognostic Index; Lugano stage, Staging system for primary gastrointestinal lymphoma. Comparison of Baseline Characteristics Between the GICs and Non-GICs Cohorts Notes: Data are presented as n (%). Values highlighted in bold italics indicate statistically significant results. Abbreviations: GICs, gastrointestinal complications; ECOG PS, Eastern Cooperative Oncology Group Performance Status; IPI, International Prognostic Index; Lugano stage, Staging system for primary gastrointestinal lymphoma. The clinical manifestations of patients with PGIL lack specificity. The most common clinical manifestations were abdominal pain (98 cases, 66.2%) and abdominal distension (48 cases, 32.4%), followed by nausea and vomiting (28 cases, 18.9%), anorexia and fatigue (25 cases, 16.9%), hematemesis or melena (24 cases, 16.2%), altered bowel habits (24 cases, 16.2%), epigastric discomfort (13 cases, 8.8%), acid regurgitation (13 cases, 8.8%), eructation (12 cases, 8.1%), singultus (5 cases, 3.4%), borborygmi (4 cases, 2.7%), palpable abdominal mass (4 cases, 2.7%), hematochezia (4 cases, 2.7%), dysphagia (2 cases, 1.4%), and visible abdominal protuberance (1 case, 0.7%). Notably, the incidences of nausea and vomiting, as well as hematemesis or melena, were significantly higher in the GICs cohort compared to the non-GICs cohort (P < 0.005). Comparison of Laboratory Findings A comparison of the laboratory test results between the GICs and non-GIC groups is presented in Table 2. The analysis showed that the occurrence of GICs was accompanied by significant bodily nutritional depletion and metabolic disorders. Compared with the non-GICs group, the levels of hemoglobin (Hb) [108.00 g/L (IQR 82.50–118.00) vs. 123.00 g/L (IQR 106.00–141.00), P < 0.001], serum albumin (Alb) [34.70 g/L (IQR 29.64–38.78) vs. 40.18 g/L (IQR 35.98–43.42), P < 0.001], and serum calcium (sCa) [2.13 mmol/L (IQR 2.03–2.25) vs. 2.19 mmol/L (IQR 2.10–2.32), P=0.043] in the GICs group all exhibited statistically significant decreases. In addition, in the assessment of systemic inflammatory response and tumor burden, multiple relevant markers in the GICs group showed significant differences compared with the non-GICs group. Specifically, LDH levels in the GICs group were significantly elevated [226.50 IU/mL (IQR 176.25–312.50) vs. 197.00 IU/mL (IQR 174.50–233.50), P = 0.049]. Inflammatory indicators in this group, including ESR [48.00 mm/h (IQR 17.00–89.00) vs. 12.00 mm/h (IQR 8.50–34.00), P = 0.003], CRP [18.67 mg/L (IQR 4.57–50.19) vs. 4.60 mg/L (IQR 1.14–24.50), P < 0.001], and PCT [0.24 ng/mL (IQR 0.05–2.70) vs. 0.04 ng/mL (IQR 0.04–0.40), P = 0.019], all showed significant increases. Detection of the cytokine microenvironment showed that in the GICs group, not only the pro-inflammatory factor IL-6 was significantly upregulated [22.78 pg/mL (IQR 14.97–61.82) vs. 11.14 pg/mL (IQR 7.13–33.45), P = 0.044], but also the anti-inflammatory factor IL-10 was significantly highly expressed [8.79 pg/mL (IQR 6.73–22.33) vs. 5.13 pg/mL (IQR 3.75–6.72), P = 0.017]. In addition, it is worth noting that two composite indicators were explored in 120 PGIL patients. The C-reactive protein/albumin ratio (CAR) in the GICs group was significantly higher than that in the non-GICs group [0.49 (IQR 0.13–1.42) vs. 0.12 (IQR 0.03–0.60), P < 0.001], while the CRP-albumin-lymphocyte (CALLY) index was significantly lower [0.19 (IQR 0.05–0.85) vs. 1.31 (IQR 0.20–5.25), P < 0.001]. However, no statistically significant differences were observed between the two groups in the expression levels of β2-MG, ferritin (FER), and cytokines IL-2 and IL-4 (P > 0.05).Table 2Comparison of Laboratory Findings Between the GICs and Non-GICs CohortsItemNon-GICs(n=62)GICs(n=86)P valueHb (g/L)123.00(106.00,141.00)108.00(82.50,118.00)<0.001ALB (g/L)40.18(35.98,43.42)34.70(29.64,38.78)<0.001sCa (mmol/L)2.19(2.10,2.32)2.13(2.03,2.25)0.043LDH (IU/L)197.00(174.50,233.50)226.50(176.25,312.50)0.049β2-MG (ug/L)2.96(2.24,3.98)3.05(2.28,4.19)0.596ESR (mm/h)12.00(8.50,34.00)48.00(17.00,89.00)0.003CRP (mg/L)4.60(1.14,24.50)18.67(4.57,50.19)<0.001PCT (ng/mL)0.04(0.04,0.40)0.24(0.05,2.70)0.019FER (ng/mL)246.00(124.55,310.45)190.00(86.30,665.00)0.784IL-2 (pg/mL)3.53±2.053.02±1.240.416IL-4 (pg/mL)3.35±1.633.55±1.810.780IL-6 (pg/mL)11.14(7.13,33.45)22.78(14.97,61.82)0.044IL-10 (pg/mL)5.13(3.75,6.72)8.79(6.73,22.33)0.017CAR0.12(0.03,0.60)0.49(0.13,1.42)<0.001CALLY1.31(0.20,5.25)0.19(0.05,0.85)<0.001Notes: Continuous variables: Data are expressed as median (interquartile range) [M(IQR)] or mean ± standard deviation (\documentclass[12pt]{minimal} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \DeclareFontFamily{T1}{linotext}{} \DeclareFontShape{T1}{linotext}{m}{n} {linotext }{} \DeclareSymbolFont{linotext}{T1}{linotext}{m}{n} \DeclareSymbolFontAlphabet{\mathLINOTEXT}{linotext} \begin{document}$\bar x \pm s$\end{document}). Values highlighted in bold italics indicate statistically significant results.Abbreviations: GICs, gastrointestinal complications; Hb, hemoglobin; Alb, albumin; sCa, serum calcium; LDH, lactate dehydrogenase; β2-MG, β2-microglobulin; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; PCT, procalcitonin; FER, ferritin; IL, interleukin; CAR, C-reactive protein/albumin ratio; CALLY, CRP-albumin-lymphocyte. Table 2Comparison of Laboratory Findings Between the GICs and Non-GICs CohortsItemNon-GICs(n=62)GICs(n=86)P valueHb (g/L)123.00(106.00,141.00)108.00(82.50,118.00)<0.001ALB (g/L)40.18(35.98,43.42)34.70(29.64,38.78)<0.001sCa (mmol/L)2.19(2.10,2.32)2.13(2.03,2.25)0.043LDH (IU/L)197.00(174.50,233.50)226.50(176.25,312.50)0.049β2-MG (ug/L)2.96(2.24,3.98)3.05(2.28,4.19)0.596ESR (mm/h)12.00(8.50,34.00)48.00(17.00,89.00)0.003CRP (mg/L)4.60(1.14,24.50)18.67(4.57,50.19)<0.001PCT (ng/mL)0.04(0.04,0.40)0.24(0.05,2.70)0.019FER (ng/mL)246.00(124.55,310.45)190.00(86.30,665.00)0.784IL-2 (pg/mL)3.53±2.053.02±1.240.416IL-4 (pg/mL)3.35±1.633.55±1.810.780IL-6 (pg/mL)11.14(7.13,33.45)22.78(14.97,61.82)0.044IL-10 (pg/mL)5.13(3.75,6.72)8.79(6.73,22.33)0.017CAR0.12(0.03,0.60)0.49(0.13,1.42)<0.001CALLY1.31(0.20,5.25)0.19(0.05,0.85)<0.001Notes: Continuous variables: Data are expressed as median (interquartile range) [M(IQR)] or mean ± standard deviation (\documentclass[12pt]{minimal} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \DeclareFontFamily{T1}{linotext}{} \DeclareFontShape{T1}{linotext}{m}{n} {linotext }{} \DeclareSymbolFont{linotext}{T1}{linotext}{m}{n} \DeclareSymbolFontAlphabet{\mathLINOTEXT}{linotext} \begin{document}$\bar x \pm s$\end{document}). Values highlighted in bold italics indicate statistically significant results.Abbreviations: GICs, gastrointestinal complications; Hb, hemoglobin; Alb, albumin; sCa, serum calcium; LDH, lactate dehydrogenase; β2-MG, β2-microglobulin; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; PCT, procalcitonin; FER, ferritin; IL, interleukin; CAR, C-reactive protein/albumin ratio; CALLY, CRP-albumin-lymphocyte. Comparison of Laboratory Findings Between the GICs and Non-GICs Cohorts Notes: Continuous variables: Data are expressed as median (interquartile range) [M(IQR)] or mean ± standard deviation (\documentclass[12pt]{minimal} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \DeclareFontFamily{T1}{linotext}{} \DeclareFontShape{T1}{linotext}{m}{n} {linotext }{} \DeclareSymbolFont{linotext}{T1}{linotext}{m}{n} \DeclareSymbolFontAlphabet{\mathLINOTEXT}{linotext} \begin{document}$\bar x \pm s$\end{document}). Values highlighted in bold italics indicate statistically significant results. Abbreviations: GICs, gastrointestinal complications; Hb, hemoglobin; Alb, albumin; sCa, serum calcium; LDH, lactate dehydrogenase; β2-MG, β2-microglobulin; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; PCT, procalcitonin; FER, ferritin; IL, interleukin; CAR, C-reactive protein/albumin ratio; CALLY, CRP-albumin-lymphocyte. Anatomical Distribution and Endoscopic Findings of GICs Table 3 details the characteristics of the 86 patients with PGIL and concurrent GICs. The incidence of GIB was the highest (48 cases, 55.8%), followed by that of GIO (20 cases, 23.3%) and GIP (6 cases, 7.0%). In addition, 12 patients (14.0%) had two of the above complications simultaneously: 4 patients (4.7%) had both GIB and GIP, 7 patients (8.1%) had both GIO and GIB, and 1 patient (1.2%) had both GIO and GIP. The distribution of the different GICs varied by location. Gastric lesions were highly prone to bleeding (29 cases, 60.4%), while the small intestine was a common site for perforation (3 cases, 50.0%), obstruction (9 cases, 45.0%), and multiple complications (9 cases, 75.0%). Meanwhile, colonic involvement accounted for a certain proportion of bleeding (10 cases, 20.8%) and obstruction (8 cases, 40.0%), but involvement of the duodenum, rectum, and multiple sites was less common.Table 3Anatomical Sites and Endoscopic Findings of Various GICsGIB(n=48, %)GIP(n=6, %)GIO(n=20, %)Multiple Complications (n=12, %)N (%)LocationStomach29(60.4)1(16.7)0(0.0)2(16.7)32(37.2)Duodenum4(8.3)0(0.0)1(5.0)0(0.0)5(5.8)Small Intestine3(6.3)3(50.0)9(45.0)9(75.0)24(27.9)Colon10(20.8)1(16.7)8(40.0)1(8.3)20(23.3)Rectum2(4.2)0(0.0)0(0.0)0(0.0)2(2.3)Multiple sites0(0.0)1(16.7)2(10.0)0(0.0)3(3.5)Endoscopic featuresPolypoid16(35.6)1(50.0)6(60.0)1(50.0)24(40.7)Ulcerative19(42.2)0(0.0)1(10.0)1(50.0)21(35.6)Diffuse infiltrating8(17.8)1(50.0)1(10.0)0(0.0)10(16.9)Lymphofollicular2(4.4)0(0.0)0(0.0)0(0.0)2(3.4)Notes: Multiple complications: Refers to patients experiencing two or more complications simultaneously. Endoscopic features: Based on the evaluation of 59 patients who underwent endoscopic examination.Abbreviations: GIB, gastrointestinal bleeding; GIP, gastrointestinal perforation; GIO, gastrointestinal obstruction. Table 3Anatomical Sites and Endoscopic Findings of Various GICsGIB(n=48, %)GIP(n=6, %)GIO(n=20, %)Multiple Complications (n=12, %)N (%)LocationStomach29(60.4)1(16.7)0(0.0)2(16.7)32(37.2)Duodenum4(8.3)0(0.0)1(5.0)0(0.0)5(5.8)Small Intestine3(6.3)3(50.0)9(45.0)9(75.0)24(27.9)Colon10(20.8)1(16.7)8(40.0)1(8.3)20(23.3)Rectum2(4.2)0(0.0)0(0.0)0(0.0)2(2.3)Multiple sites0(0.0)1(16.7)2(10.0)0(0.0)3(3.5)Endoscopic featuresPolypoid16(35.6)1(50.0)6(60.0)1(50.0)24(40.7)Ulcerative19(42.2)0(0.0)1(10.0)1(50.0)21(35.6)Diffuse infiltrating8(17.8)1(50.0)1(10.0)0(0.0)10(16.9)Lymphofollicular2(4.4)0(0.0)0(0.0)0(0.0)2(3.4)Notes: Multiple complications: Refers to patients experiencing two or more complications simultaneously. Endoscopic features: Based on the evaluation of 59 patients who underwent endoscopic examination.Abbreviations: GIB, gastrointestinal bleeding; GIP, gastrointestinal perforation; GIO, gastrointestinal obstruction. Anatomical Sites and Endoscopic Findings of Various GICs Notes: Multiple complications: Refers to patients experiencing two or more complications simultaneously. Endoscopic features: Based on the evaluation of 59 patients who underwent endoscopic examination. Abbreviations: GIB, gastrointestinal bleeding; GIP, gastrointestinal perforation; GIO, gastrointestinal obstruction. In terms of initial diagnostic methods, the GICs group often relied on surgical resection for confirmation (48 cases, 55.8%), while endoscopic forceps biopsy was relatively less used (38 cases, 44.2%). Due to the lack of specific endoscopic manifestations for PGIL, a universally standardized macroscopic classification system is currently unavailable. Therefore, based on conventional clinical practice and dominant macroscopic features, lesions were descriptively categorized into polypoid (protruding masses), ulcerative (mucosal excavations), diffuse infiltrating (thickened and rigid folds), and lymphofollicular (multiple granular elevations) types. Among the 59 patients (68.6%) in this group who completed endoscopic evaluation (Table 3), the endoscopic findings were mainly of the polypoid type (24 cases, 40.7%) and ulcerative type (21 cases, 35.6%), with the diffuse infiltrative type (10 cases, 16.9%) and lymph follicular type (2 cases, 3.4%) being relatively rare; no giant fold type was observed. Patients with GIB mostly presented with the ulcerative type endoscopically (19 cases, 42.2%), while obstruction patients were mainly of the intraluminally growing polypoid type (6 cases, 60.0%). In addition, 1 patient (1.7%) showed no obvious abnormalities, and 1 patient (1.7%) could not tolerate endoscopy, so no lesions were observed. 5 patients underwent endoscopic ultrasound examination, which showed wall thickening (3 cases, 60.0%) and unclear wall layers (1 case, 20.0%), with 1 patient (20.0%) showing both manifestations. Comparison of Radiological Findings Based on the imaging data of 132 patients with PGIL who completed abdominal or pelvic CT scans (Table 4). The most common imaging finding was gastrointestinal bowel wall thickening (78 cases, 59.1%), followed by lymphadenopathy (57 cases, 43.2%), stranding of the surrounding fat (39 cases, 29.5%), pelvic effusion (30 cases, 22.7%), intraperitoneal fluid collection (20 cases, 15.2%), bowel dilatation with gas and fluid accumulation (20 cases, 15.2%), space-occupying lesions (18 cases, 13.6%), peritoneal thickening (9 cases, 6.8%), luminal narrowing (4 cases, 3.0%), pneumoperitoneum (3 cases, 2.3%), and mesenteric thickening (2 cases, 1.5%). Intergroup comparison showed that the incidences of wall thickening, lymphadenopathy, stranding of surrounding fat, and Bowel dilatation with gas and fluid accumulation in the GICs group were significantly higher than those in the non-GICs group (P < 0.05). Additionally, 67 patients (45.3%) underwent PET-CT examinations, which assisted in identifying hypermetabolic regional lymph nodes and occult distant involvements.Table 4Comparison of Radiological Findings Between the GICs and Non-GICs CohortsNon-GICs(n=62, %)GICs(n=86, %)Total(N, %)P valueWall thickening24(46.2)54(67.5)78(59.1)0.015Regional Lymphadenopathy15(28.8)42(52.5)57(43.2)0.007Stranding of surrounding fat9(17.3)30(37.5)39(29.5)0.013Pelvic effusion9(17.3)21(26.3)30(22.7)0.231Intraperitoneal fluid collection4(7.7)16(20.0)20(15.2)0.054Bowel dilatation with gas and fluid accumulation3(5.8)17(21.3)20(15.2)0.015Space-occupying lesion6(11.5)12(15.0)18(13.6)0.571Peritoneal thickening3(5.8)6(7.5)9(6.8)0.974Luminal narrowing1(1.9)3(3.8)4(3.0)0.937Pneumoperitoneum1(1.9)2(2.5)3(2.3)1.000Mesenteric thickening0(0.0)2(2.5)2(1.5)0.519Notes: Data are presented as n (%) based on 132 patients who completed abdominal or pelvic CT. Values highlighted in bold italics indicate statistically significant results. Table 4Comparison of Radiological Findings Between the GICs and Non-GICs CohortsNon-GICs(n=62, %)GICs(n=86, %)Total(N, %)P valueWall thickening24(46.2)54(67.5)78(59.1)0.015Regional Lymphadenopathy15(28.8)42(52.5)57(43.2)0.007Stranding of surrounding fat9(17.3)30(37.5)39(29.5)0.013Pelvic effusion9(17.3)21(26.3)30(22.7)0.231Intraperitoneal fluid collection4(7.7)16(20.0)20(15.2)0.054Bowel dilatation with gas and fluid accumulation3(5.8)17(21.3)20(15.2)0.015Space-occupying lesion6(11.5)12(15.0)18(13.6)0.571Peritoneal thickening3(5.8)6(7.5)9(6.8)0.974Luminal narrowing1(1.9)3(3.8)4(3.0)0.937Pneumoperitoneum1(1.9)2(2.5)3(2.3)1.000Mesenteric thickening0(0.0)2(2.5)2(1.5)0.519Notes: Data are presented as n (%) based on 132 patients who completed abdominal or pelvic CT. Values highlighted in bold italics indicate statistically significant results. Comparison of Radiological Findings Between the GICs and Non-GICs Cohorts Notes: Data are presented as n (%) based on 132 patients who completed abdominal or pelvic CT. Values highlighted in bold italics indicate statistically significant results. Comparison of Pathological Subtypes The PGIL patients included in the study had the following different pathological types. The most prevalent subtypes were Diffuse Large B-Cell Lymphoma (DLBCL, 69 cases, 46.6%) and Mucosa-Associated Lymphoid Tissue (MALT) lymphoma (49 cases, 33.1%), followed by Follicular Lymphoma (FL) in 8 cases (5.4%), Burkitt Lymphoma (BL), Enteropathy-Associated T-Cell Lymphoma (EATL), and Natural Killer/T-cell (NK/T-cell) lymphoma in 4 cases each (2.7%), Peripheral T-Cell Lymphoma (PTCL) in 3 cases (2.0%), and Mantle Cell Lymphoma (MCL) and plasmablastic Lymphoma (PBL) in 2 cases each (1.4%). In addition, 3 cases (2.0%) of patients diagnosed with PGIL could not be determined for their specific types. Comparative analysis showed that the proportion of DLBCL in the GICs group was significantly increased, while the proportion of indolent MALT lymphoma was significantly decreased. The difference in the composition ratio of pathological subtypes between the two groups was statistically significant (P = 0.028). Comparison of Treatment Modalities There were significant differences in the treatment management of patients with PGIL (Table 5). In this study, 15 patients (10.1%) received only anti-H. pylori treatment, 21 patients (14.2%) received only surgical treatment, 40 patients (27.0%) received only chemotherapy, 6 patients (4.1%) received only radiotherapy, 5 patients (3.4%) received targeted therapy, 35 patients (23.6%) received surgery combined with chemotherapy, and 10 patients (6.8%) received two or more of the above treatments. In addition, 13 patients (8.8%) did not receive special treatment, 1 patient was transferred to a higher-level hospital after surgery, 1 patient underwent endoscopic mucosal resection (EMR), and 1 patient was transferred to a higher-level hospital after conservative treatment. Comparing the distribution of treatment methods between the two groups, the proportions of patients in the GICs group who received only surgical treatment and surgery combined with chemotherapy were significantly higher than those in the non-GICs group, while the proportions of patients in the non-GICs group who received only anti- H. pylori treatment and only radiotherapy were significantly higher than those in the GICs group (P < 0.001).Table 5Comparison of Treatment Modalities Between the GICs and Non-GICs CohortsNon-GICs(n=62, %)GICs(n=86, %)Total(n=148)P valueAnti-H. pylori12(19.4)3(3.5)15(10.1)<0.001Surgery3(4.8)18(20.9)21(14.2)Chemotherapy19(30.6)21(24.4)40(27.0)Radiotherapy5(8.1)1(1.2)6(4.1)Targeted therapy0(0.0)5(5.8)5(3.4)Surgery + chemotherapy7(11.3)28(32.6)35(23.6)Other combination therapies7(11.3)3(3.5)10(6.8)No specific treatment8(12.9)5(5.8)13(8.8)Note: Values highlighted in bold italics indicate statistically significant results.Abbreviation: GICs, gastrointestinal complications. Table 5Comparison of Treatment Modalities Between the GICs and Non-GICs CohortsNon-GICs(n=62, %)GICs(n=86, %)Total(n=148)P valueAnti-H. pylori12(19.4)3(3.5)15(10.1)<0.001Surgery3(4.8)18(20.9)21(14.2)Chemotherapy19(30.6)21(24.4)40(27.0)Radiotherapy5(8.1)1(1.2)6(4.1)Targeted therapy0(0.0)5(5.8)5(3.4)Surgery + chemotherapy7(11.3)28(32.6)35(23.6)Other combination therapies7(11.3)3(3.5)10(6.8)No specific treatment8(12.9)5(5.8)13(8.8)Note: Values highlighted in bold italics indicate statistically significant results.Abbreviation: GICs, gastrointestinal complications. Comparison of Treatment Modalities Between the GICs and Non-GICs Cohorts Note: Values highlighted in bold italics indicate statistically significant results. Abbreviation: GICs, gastrointestinal complications. Risk Factors for Gastrointestinal Complications This study included 148 patients with PGIL. The results of univariable logistic regression analysis (Table 6) showed that advanced age, history of hypertension, intestinal lymphoma, invasion of the muscularis propria or serosa, ECOG PS score ≥ 2, Ki67 ≥ 40%, anemia, hypoproteinemia, hypocalcemia, and elevated LDH were all risk factors for GICs in patients with PGIL (P < 0.05). To exclude potential confounding biases, the above positive variables were further included in a multivariable logistic regression model. It was found that a history of hypertension (OR = 3.724, 95% CI: 1.149–12.074, P = 0.028), primary lesion located in the intestine (OR = 3.551, 95% CI: 1.240–10.170, P = 0.018), deep invasion of the lesion into the muscularis propria or serosa (OR = 3.615, 95% CI: 1.132–11.546, P = 0.030), and complicated with anemia (OR = 4.288, 95% CI: 1.457–12.615, P = 0.008) were the four independent risk factors for GICs in PGIL patients. It is worth noting that anemia not only reflects potential gastrointestinal bleeding but also is the independent warning factor with the highest predictive power.Table 6Univariable and Multivariable Logistic Regression Analyses of Risk Factors for GICs in PGIL PatientsRisk FactorsUnivariable AnalysisMultivariable AnalysisOR95% CIP valueOR95% CIP valueAge≥ 602.3571.203–4.6180.0121.3650.512–3.6360.534Male0.9740.505–1.8820.939–Hypertension history2.5521.098–5.9300.0293.7241.149–12.0740.028Diabetes history2.2990.596–8.8660.227–Intestinal4.4672.199–9.076<0.0013.5511.240–10.1700.018T-cell lymphoma1.9410.364–10.3590.438–Invasion of muscular/serosal layer5.7072.569–12.679<0.0013.6151.132–11.5460.030ECOG PS score≥24.9841.621–15.3230.0052.1850.464–10.2810.322Ki-67≥40%2.5701.304–5.0670.0060.6720.236–1.9170.457Anemia5.3862.550–11.377<0.0014.2881.457–12.6150.008Hypoproteinemia4.5732.141–9.768<0.0012.1550.484–9.6010.314Hypocalcemia2.2761.119–4.6310.0231.1150.337–3.6830.859Elevated LDH3.3681.439–7.8840.0050.8420.234–3.0280.793Elevated β2-MG1.2240.473–3.1710.677–Elevated FER3.2350.593–17.6580.175–Notes: Values highlighted in bold italics indicate statistically significant results; Multivariable analysis: Includes variables with P < 0.05 in the univariable multivariable analysis to identify independent risk factors. Reference categories: For categorical variables, the comparison group is the first category listed.Abbreviation: OR, odds ratio; CI, confidence interval; ECOG PS, Eastern Cooperative Oncology Group Performance Status; LDH, lactate dehydrogenase; β2-MG, β2-microglobulin; FER, ferritin. Table 6Univariable and Multivariable Logistic Regression Analyses of Risk Factors for GICs in PGIL PatientsRisk FactorsUnivariable AnalysisMultivariable AnalysisOR95% CIP valueOR95% CIP valueAge≥ 602.3571.203–4.6180.0121.3650.512–3.6360.534Male0.9740.505–1.8820.939–Hypertension history2.5521.098–5.9300.0293.7241.149–12.0740.028Diabetes history2.2990.596–8.8660.227–Intestinal4.4672.199–9.076<0.0013.5511.240–10.1700.018T-cell lymphoma1.9410.364–10.3590.438–Invasion of muscular/serosal layer5.7072.569–12.679<0.0013.6151.132–11.5460.030ECOG PS score≥24.9841.621–15.3230.0052.1850.464–10.2810.322Ki-67≥40%2.5701.304–5.0670.0060.6720.236–1.9170.457Anemia5.3862.550–11.377<0.0014.2881.457–12.6150.008Hypoproteinemia4.5732.141–9.768<0.0012.1550.484–9.6010.314Hypocalcemia2.2761.119–4.6310.0231.1150.337–3.6830.859Elevated LDH3.3681.439–7.8840.0050.8420.234–3.0280.793Elevated β2-MG1.2240.473–3.1710.677–Elevated FER3.2350.593–17.6580.175–Notes: Values highlighted in bold italics indicate statistically significant results; Multivariable analysis: Includes variables with P < 0.05 in the univariable multivariable analysis to identify independent risk factors. Reference categories: For categorical variables, the comparison group is the first category listed.Abbreviation: OR, odds ratio; CI, confidence interval; ECOG PS, Eastern Cooperative Oncology Group Performance Status; LDH, lactate dehydrogenase; β2-MG, β2-microglobulin; FER, ferritin. Univariable and Multivariable Logistic Regression Analyses of Risk Factors for GICs in PGIL Patients Notes: Values highlighted in bold italics indicate statistically significant results; Multivariable analysis: Includes variables with P < 0.05 in the univariable multivariable analysis to identify independent risk factors. Reference categories: For categorical variables, the comparison group is the first category listed. Abbreviation: OR, odds ratio; CI, confidence interval; ECOG PS, Eastern Cooperative Oncology Group Performance Status; LDH, lactate dehydrogenase; β2-MG, β2-microglobulin; FER, ferritin. Discussion One of the core challenges in the clinical management of PGIL lies in the sudden onset and severity of acute complications. This study systematically reviewed the clinical characteristics of 148 patients with PGIL and found that the overall incidence of GICs was as high as 58.1%, with GIB being the most common. Consistent with previous literature,5,31 our cohort confirmed that PGIL is more prevalent in elderly male patients overall; however, gender does not significantly affect the risk of developing GICs (P = 0.939). The disease typically presents with non-specific clinical manifestations, with abdominal pain being the most common symptom. Furthermore, it revealed that impaired performance status (ECOG PS score ≥ 2), high-risk IPI stratification, and a history of hypertension were closely associated with the occurrence of GICs. This suggests that the high invasive burden of the tumor and the deterioration of the host’s underlying microvascular conditions together form the basis for the outbreak of complications. The most common pathological types of PGIL are DLBC and MALT lymphoma.32 The incidence of GICs in DLBCL is significantly higher than in MALT lymphoma, and this difference may be attributed to the rapid proliferation of highly aggressive B cells causing full-thickness destruction of the vessel wall. A further analysis of the relationship between lesion locations and complications found that the stomach is the most prone to complications, followed by the small intestine, colon, duodenum, and rectum. The stomach is the most common site of bleeding, while the small intestine is the main site prone to perforation and obstruction. This distribution difference may depend on the anatomical and physiological characteristics of the gastrointestinal tract. This study conducted a comparative analysis of laboratory test results between the GICs group and the non-GICs group, and found that patients with GICs have significant inflammation and nutritional imbalance. Other studies have demonstrated that tumor-associated inflammation is a hallmark of its development and progression,33 LDH can reflect the tumor burden.34 In the GICs group of this study, the significant increase in acute-phase proteins such as ESR, CRP, and PCT suggests that the tumor-related inflammatory microenvironment may further damage the intestinal mucosal barrier function and accelerate tissue necrosis. Integrating these findings with the baseline data analysis reveals that patients in the GICs cohort generally presented with more advanced tumor stages and deeper tumor infiltration. Consequently, the elevation of these indices in this group stems not only from the complication-induced acute inflammation but also indicates a heavier tumor burden and greater tumor invasiveness. These results demonstrate that advanced tumor stage, high tumor burden, and systemic inflammation act synergistically to jointly induce and drive the onset and progression of GICs. Concurrently, the high tumor burden in this cohort is closely intertwined with nutritional imbalance. In addition, this study found that CAR and CALLY indices showed significant differences between the two groups of patients, suggesting that these two composite inflammatory and nutritional indices are expected to become potential markers for risk stratification of GICs in PGIL patients, but multi-center studies are still needed to further verify their clinical value. It is very important to evaluate the disease through a combination of multiple auxiliary examinations. This study found that protuberant lesions often cause GIO, while ulcerative lesions often cause GIB. Although ulcers or protuberant lesions under direct endoscopic vision can provide core clues, there are still a few patients with no obvious abnormalities found under endoscopy, which may be related to deep tumor infiltration and limited endoscopic field of view. This reminds clinicians that for PGIL patients suspected of having GICs but with negative endoscopic findings, further examinations such as endoscopic ultrasonography, abdominal CT, and PET-CT should be performed to avoid missed diagnosis. Thickening of the gastrointestinal wall is the most common imaging finding in PGIL. Patients with GIBs are more likely to have wall thickening, lymphadenopathy, stranding of surrounding fat and intestinal dilation with gas or fluid accumulation. In addition, PET-CT has irreplaceable value in accurately distinguishing high metabolic activity residual lesions from necrotic fibrotic tissues, comprehensively evaluating tumor burden, and guiding postoperative re-staging.35–37 There is no unified guideline for the optimal treatment strategy of PGIL.38–40 For cases without gastrointestinal complications, medical treatment is typically preferred for this disease.41 The treatment regimens for patients in the non-gastrointestinal complication group in this study were highly consistent with current clinical trends, as these patients could opt for more flexible conservative treatment modalities. Severe complications can delay and complicate treatment, impair quality of life, and elevate mortality rates.8 Once life-threatening gastrointestinal complications occur, the standard oncological treatment protocol must inevitably be modified.13 Studies have shown that surgical treatment is mainly used for the treatment of GICs.17 In this study, the primary treatment regimens for patients in the GICs group were mainly simple surgery or surgery combined with chemotherapy. This heavy reliance on surgical treatment reflects the emergency interventions adopted to save lives in critical and acute conditions. However, radical resection of malignant tumors may seriously disrupt the physiological function of the gastrointestinal tract, so further research is necessary to explore strategies that balance therapeutic effects and safety.5,42 In this study, patients in the non-GICs group had more flexible treatment choices, including anti-H. pylori therapy, radiotherapy, simple chemotherapy, etc. This suggests that the occurrence of gastrointestinal complications directly dictates the clinical therapeutic strategy, necessitating a comprehensive assessment by clinicians to formulate individualized treatment plans.43 The significant differences in treatment strategies observed between the two groups demonstrate that the occurrence of gastrointestinal complications directly dictates the clinical therapeutic approach. Furthermore, these findings suggest that clinicians should prioritize the development of individualized, multidisciplinary treatment plans based on the patient’s specific complications. The risk factor analysis of this study provides a valuable clinical reference for risk stratification of PGIL with GICs. It is important to interpret these findings prudently; rather than establishing direct causality, our analysis demonstrates that a history of hypertension, intestinal lymphoma, lesion invasion into the muscularis or serosa layer, and anemia are four factors independently associated with an increased risk of gastrointestinal complications (P<0.05). In particular, patients with anemia at the first visit (OR=4.288) not only have a direct manifestation of chronic blood loss but also a dangerous signal of body compensation exhaustion, which is not conducive to controlling tumor progression. These findings provide important clues for early identification of high-risk groups of GICs in PGIL patients and formulation of targeted prevention and intervention measures. It is worth noting that while anemia was identified as a pivotal risk factor for GICs, its clinical management was primarily symptomatic and supportive. The definitive resolution of these complications and the ultimate improvement in patient prognosis rely fundamentally on the targeted treatment of the underlying primary disease, namely lymphoma, rather than merely the symptomatic correction of hemoglobin levels. This study has certain objective limitations that must be explicitly acknowledged. First, the single-center retrospective design inherently introduces selection bias. Although our cohort is representative of the western Hubei region, this geographical confinement restricts the external validity of our findings. Additionally, the sample size was determined by available clinical archives without an a priori statistical power calculation, which may limit the statistical power to detect more subtle differences. Second, the lack of long-term follow-up data restricts our ability to evaluate overall survival, tumor recurrence, and the long-term oncological benefits of specific treatments. Third, because DLBCL and MALT lymphomas comprised nearly 80% of our cohort, the identified risk factors primarily reflect major B-cell lymphomas and may not be generalizable to rare subtypes. Fourth, therapeutic interventions were dictated by acute clinical presentations rather than randomization, precluding the establishment of matched control groups to robustly validate treatment efficacy. Future multicenter, prospective studies are warranted to verify these results. Conclusion This study found that the incidence of GICs in patients is relatively high (58.1%), among which GIB was the most common, followed by GIO and GIP. The stomach is the most frequent site of bleeding, while the small intestine is more prone to obstruction and perforation, which is closely related to the anatomical structures of different parts of the gastrointestinal tract and the tumor growth patterns. Patients with complications usually have poorer physical status, higher tumor invasiveness, as well as more significant systemic inflammatory responses and malnutrition. In addition, a history of hypertension, intestinal lymphoma, lesions invading the muscularis propria or serosa, and anemia were identified as four independent factors significantly associated with an increased risk of GICs in this specific PGIL cohort. Anemia is not only a manifestation of bleeding but also serves as the strongest independent warning indicator for evaluating the potential risk of complications (OR=4.288). For high-risk groups with the above risk factors, clinicians should strengthen the evaluation of imaging and endoscopy at the initial diagnosis and formulate individualized monitoring plans. For high-risk patients, initial chemotherapy should be cautiously titrated to prevent rapid tumor lysis-induced perforation or bleeding, complemented by multidisciplinary surgical standby. To address the inherent limitations of this single-center study, we are planning a prospective, multicenter cohort to further validate these tailored therapeutic strategies.