Anatomic distribution, clinical features, and survival data of 87 cases primary gastrointestinal lymphoma

Background The purpose of this study is to analyze the anatomic distribution, clinical features, therapeutic methods, and prognosis factors of primary gastrointestinal lymphoma (PGIL). Methods Clinical data of 87 cases PGIL in the First Affiliated Hospital of Dalian Medical University from January 1999 to December 2010 were collected. Follow-ups were made according to the clinical feature, pathological pattern, clinical stage, and therapeutic method. Kapan Meier method was used for the survival analysis. Log-rank test was used to perform univariate survival analysis. COX multivariate analysis was carried out to analyze factors of P < 0.05 in univariate survival analysis. Results The incidence of PGIL significantly increased in patients more than 40 years old (87.4 %). Clinical symptoms of PGIL were indistinguishable from other digestive system diseases, which included abdominal pain or discomfort (72.4 %), lack of appetite (16.3 %), gastrointestinal hemorrhage (14.9 %), and diarrhea (12.8 %). Some patients presented with systemic symptoms or complications, such as weight loss (35.6 %) and digestive tract obstruction (13.8 %). Primary gastric lymphoma (PGL) was the most common, followed by primary intestine lymphoma (PIL). The majority of PGIL were single lesion, which included 40 cases (87 %) PGL and 35 cases (94.5 %) PIL. The most frequent site of PGL was antrum of the stomach (43.5 %), as to PIL, the small intestine (90.2 %) was the most frequent site, especially within 100 cm far away from ileocecal valve. Most of PGIL were derived from B cell (93.1 %). The most common pathological type was mucosa-associated lymphoid tissue (MALT) (67.4 %) in the PGL group and diffuse large B cell lymphoma (DLBCL) (46.3 %) in the PIL group. Surgical treatment had been performed in most of PGIL, which included 32 cases in the PGL group and 38 cases in the PIL group. The 1-year overall survival (OS) and the 3-year OS were 82 and 77 %, respectively. Analysis of single factor affecting prognosis showed that lesion location, sources of cells, and clinical stage were associated with OS. PGL group had better OS than that of PIL group (1-year 89 vs 62 %, 3-year 84 vs 50 %, P = 0.03). B cell-originated group had better OS than that of T cell-originated group (1-year 89 vs 36 %, 3-year 85 vs 0 %, P = 0.008). Stage I + II group had better OS than that of stage III + IV group (1-year 89 vs 38 %, 3-year 87 vs 0 %, P = 0.007). Multivariate analysis showed that clinical stage and sources of cells were the significant independent prognostic factors. Conclusions It was more common to find location of PGIL in the stomach than that in the intestine. The most common pathological type was MALT in the PGL and DLBCL in the PIL. The treatment of PGL was focused on chemotherapy. It was noting that since PIL was not only difficult to make confirmed diagnosis but also likely to develop with complications, so it was usually needed surgical excision. Clinical stage and pathological pattern were related to prognosis of PGIL.


Background
Deriving from lymphoid tissue beneath mucosa of gastrointestinal wall, primary gastrointestinal lymphoma (PGIL) was extranodal non-Hodgkin's lymphomas (NHL) that account for 24-49 % of NHL [1]. It used to appear in the stomach and small intestines, accounting for 1-4 % of gastroenteric tumor [2]. PGIL was difficult to identify because clinical features of PGIL was indistinguishable from other gastroenteric diseases, especially from gastrointestinal tumor, which attributes to a high misdiagnosis rate in clinical practice. Besides, the treatment methods were quite different from each other. A retrospective analysis was made on 87 PGIL patients from January 1999 to December 2010 in the First Affiliated Hospital of Dalian Medical University; we summarized their clinical features, pathological patterns, treatment methods, and prognosis factors.  Table 1, the majority of PGL were located at the antrum of the stomach (n = 20, 43.5 %), followed by the body of the stomach (n = 16, 34.8 %). The most frequent site of PIL was the small intestine (n = 37, 90.2 %), which mostly were located at the ileum (n = 26, 70.3 %), especially within 100 cm far away from ileocecal valve (n = 25). Table 2): all patients were non-Hodgkin's lymphoma, and most of cases were derived from B cell (n = 64). In the PGL group, the most common type was MALT (n = 31), followed by DLBCL (n = 14). In the PIL group, the most common type was DLBCL (n = 19), followed by MALT (n = 10), and fewer were EATL (n = 6). The patients belonged to stage I E 46 cases, stage II E 31 cases, stage III 4 cases, and stage IV 6 cases. 4. Treatment (Table 3):

Pathologic features and stages (
As for 46 patients in PGL group, (1) 32 patients accepted surgical treatment, including radical operation (n = 27) and palliative surgery (n = 5). Radical operation included total gastrectomy (n = 6), radical proximal gastrectomy (n = 4,) and radical distal gastrectomy (n = 17). Eighteen patients accepted surgery alone, and 10 patients were treated with CHOP after surgery, while 4 patients were treated with R-CHOP (rituximab + phosphoric acid amide + doxorubicin + vincristine + prednisone).  (Fig. 1). B cell-originated group had better OS than that of T cell-originated group (1-year 89 vs 36 %, 3-year 85 vs 0 %, P = 0.008) (Fig. 2). There was no significant difference of the 1-year and 3-year OS between stage I and stage II, stage III and stage IV patients (data not shown). But stage I + II group had better OS than that of    (Fig. 3). Multivariate analysis showed that clinical stage and sources of cells were the significant independent prognostic factors ( Table 4).

Discussion
Although PGIL pathogenesis remained unknown, certain factors had been considered to be related with its incidence for a long time, including virus infection, autoimmune deficiency, and environment pollution. [3,4]. The incidence of PGIL had increased in Asia, North America, and Europe [5][6][7][8]. Many studies showed that the stomach was the most commonly involved site followed by the intestine, while in Pacific Ocean, small intestines were mostly seen followed by the stomach and colon [9,10]. In our study, we found that 53.9 % patients were PGL and 46.1 % were PIL. Furthermore, the majority of PGL were located at the antrum of the stomach, followed by the body of the stomach, and the most frequent site of PIL was the small intestine, especially within 100 cm far away from ileocecal valve. Besides, it should be noticed in clinical work that there might be multiple lesion locations in PGIL. Clinical symptoms of PGIL were indistinguishable from other digestive system diseases. The main symptom included abdominal pain or discomfort, together with weight loss and nausea and other intestinal symptoms [11,12]. Imageological examination might show wall thickened and intestinal masses; it was usually difficult to identify from other gastrointestinal cancer. Endoscopy and biopsy were the most reliable methods for confirming diagnosis [13,14].
The surgical treatment was traditionally considered as the main treatment methods of PGIL. Most of patients accepted the radical resection. Palliative resection might due to huge size of tumor or extensive transfer of lymph node. However, as lymphoma was highly sensitive to chemotherapy, the main treatment of PGIL was non-surgery now. A prospective study showed that surgery treatment could not improve the 10-year survival rate of PGIL by comparing of surgery plus chemotherapy with chemotherapy alone [6]. Recently, there was a study showed that it had equivalent efficacy whether patients accepted operation or not [11]. Moreover, more and more studies demonstrated that non-surgery strategies had better OS [15,16]. In our study, 50 patients accepted non-surgery methods, such as CHOP or R-CHOP, which account for 54.5 % of total patients. Rituximab is a chimeric monoclonal antibody against the protein CD20, which is primarily found on the surface of immune system B cells. Rituximab destroys both normal and malignant B cells that have CD20 on their surfaces. The addition of rituximab has improved the overall survival of lymphoma.
Many studies have showed that rituximab can improve the efficacy of chemotherapy after relapse [17,18]. In our retrospective study, there were not all patients whose CD20-positive approved rituximab therapy due to economic reasons. Nowadays, surgery had gradually been replaced by non-surgery treatment. However, many studies showed that surgery was benefit to patients who present with hemorrhage, perforation, or ileus [19,20], especially to PIL patients. PIL was not only difficult to make confirmed diagnosis but also likely to develop with complications, so it was usually needed surgical excision and then diagnosed by pathologic analysis. In our study, there were 41 PIL patients, which 38 patients accepted surgical treatment. We believed that surgery was the main treatment method of undetermined diagnosis of PGIL patients, with its irreplaceable advantages as follows: (1) surgery was an important means to gain the pathological diagnosis and determine diagnosis; (2) postoperative specimens could be graded and staged correctly in order to judge prognosis more exactly; (3) surgery might alleviate tumor load, relieve clinical symptoms, and strengthen effect of other treatment, such as postoperative chemotherapy; and (4) patients who were insensitive to other treatment or appeared life-threatening complications should choose surgery as soon as possible.
In summary, the reasonable clinical treatment method of PGIL should be made according to the location, clinical stage, pathologic pattern, and with complications or not.