|Home | About | Journals | Submit | Contact Us | Français|
Correspondence to: Masaya Iwamuro, MD, Department of Gastroenterology and Hepatology, Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences, 2-5-1 Shikata-cho, Kita-ku, Okayama 700-8558, Japan. pj.oc.oohay@ayasamorumawi
Telephone: +81-86-2357219 Fax: +81-86-2255991
AIM: To evaluate the endoscopic manifestations and prognoses of gastrointestinal (GI) mantle cell lymphoma (MCL).
METHODS: A database search at the Department of Pathology of Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences revealed 57 MCL patients with GI involvement. Clinical records were available for 35 of the 57 patients from 21 institutions, and those 35 patients were enrolled in this study. We summarized the gross types of endoscopic features, event-free survival (EFS), and overall survival (OS) of those patients.
RESULTS: Of the 35 patients, GI involvement in the esophagus, stomach, and duodenum was found in 2 (5.7%), 26 (74.3%), and 12 (34.3%) patients, respectively. Twenty-one of the 35 patients underwent colonoscopy; among them, GI involvement in the ileum, cecum, colon, and rectum was found in 10 (47.6%), 3 (14.3%), 12 (57.1%), and 10 (47.6%), respectively. Various lesions, such as superficial, protruded, fold thickening, or ulcerative, were found in the stomach, whereas multiple lymphomatous polyposis (MLP) was dominant from the duodenum to the rectum. Twelve patients were treated with a hyper-CVAD/MA regimen, and they had better OS (3-year rate, 88.3% vs 46.4%, P < 0.01) and better EFS (3-year rate, 66.7% vs 33.8%, P < 0.05) than the remaining 23 patients who were not treated with this regimen.
CONCLUSION: MLP was a representative form of intestinal involvement, whereas a variety of lesions were found in the stomach. The hyper-CVAD/MA regimen may improve survival in these patients.
Mantle cell lymphoma (MCL) is a small B-cell neoplasm that may affect the gastrointestinal (GI) tract. In 1961, Cornes reviewed 22 case reports of multiple lymphomatous polyposis (MLP), and introduced that term as a unique form of malignant GI lymphoma. Two decades later, studies revealed that the neoplastic cells of MLP originate from the mantle zone of the lymphoid follicle in most cases, hence these days MLP is a representative intestinal manifestation of MCL. On the other hand, the gastric lesions of MCL vary from pale folds[2,4] to nodules and inflammation[5,6]. Previously, the frequency of GI involvement was reported to be up to 30%[7,8]. However, recent reports based on endoscopic examinations have revealed that 46%-49% of MCL patients had esophagogastroduodenal involvement, and that 38%-62% had colorectal involvement[5,6]. As these previous reports were conducted by hematologists, the endoscopic findings and data on frequency, particularly of the gastric lesions, have not yet been investigated thoroughly.
Importantly, MCL is clinically more aggressive and has a shorter median survival, only 3-5 years, compared with other types of small B-cell neoplasms, such as extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma) and follicular lymphoma. However, due to the rarity of MCL (it represents only about 3%-10% of non-Hodgkin lymphomas), the prognoses of cases with GI MCL have not been discussed sufficiently.
In this report, we identified 35 MCL patients with GI involvement from 21 institutions and summarized their endoscopic manifestations, chemotherapy regimens, and prognoses.
A database search at the Department of Pathology of Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences found 57 patients who were histologically diagnosed as MCL with GI involvement from August 1999 to July 2009. The diagnosis of MCL was made according to the World Health Organization classification. Briefly, histological diagnosis was made by morphologic and immunophenotypic analyses on surgically resected specimens or on endoscopically biopsied specimens. All cases were blindly reviewed and the diagnosis was confirmed by a single pathologist (Morito T). The typical features of MCL were monomorphic proliferation of small to medium-sized lymphoid cells with a vaguely nodular, diffuse, or mantle zone growth pattern accompanying architectural destruction and the expression of cyclin D1 on immunohistochemical study (Figure (Figure1).1). CD5 staining was also performed, and its positivity supported the diagnosis of MCL. All samples were also positive for BCL2 protein. Clinical records were available for 35 of the 57 patients from 21 institutions, and those 35 patients were enrolled in this study. Nine of the 35 patients were examined as subjects of our previous study. For all patients, the results of endoscopic, radiological, and biological examinations, as well as clinical information including treatment regimens and prognoses, were retrospectively reviewed from clinical records.
Invaded GI organs were evaluated by esophagogastroduodenoscopy and colonoscopy. Neither capsule endoscopy nor double-balloon/single-balloon enteroscopy was performed to evaluate the jejunum and ileum. The GI tract was subdivided into three parts: the esophagus, the stomach, and the intestines (including the duodenum, ileum, cecum, colon, and rectum). GI lesions in each part were classified into the following six subtypes by gross findings: (1) the protruded type (solitary or fewer than 10 elevated lesions forming tumorous nodules; these lesions often resemble submucosal tumors and sometimes accompany ulcers on their tops); (2) the fold thickening type (thickened mucosal folds like large cerebroid folds; typically seen only in the stomach); (3) the MLP type (multiple micropolyps with or without some large polyps; the number of polyps is 10 or more); (4) the ulcerative type (solitary or multiple lowered lesions due to ulcers); (5) the superficial type (changes in the mucosal color and/or changes in mucosal morphology); and (6) the mixed type (combinations of these five subtypes). All cases were reviewed and their subtypes were classified by at least two board certified endoscopists (Iwamuro M and Okada H).
The Lugano staging system for the classification of GI tract lymphoma[11,12] was used to determine the patients’ clinical stages. Patients were diagnosed with primary GI MCL based on a set of criteria established by Dawson et al. The response evaluation included a physical examination, a complete blood count, serum biochemistry profile, endoscopic examination, bone marrow aspirate and biopsy, chest X-ray, abdominal ultrasound, and computed tomography scans of the neck, chest, and abdomen. A complete response (CR), partial remission, stable disease, and progressive disease were defined according to the International Lymphoma Workshop response criteria. Overall survival (OS) was measured from diagnosis until death from any cause, and event-free survival (EFS) was measured from diagnosis until documented progression/relapse, death from any cause, or off-protocol treatment for any reason.
Therapeutic regimens, international prognostic index (IPI), a recently introduced prognostic index for advanced-stage MCL (MCL IPI score: MIPI score), and mitotic rates in the histological specimens were investigated as predictive prognostic factors[17,18]. The MIPI score was calculated as follows: [0.03535 × age (years)] + 0.6978 (if the Eastern Cooperative Oncology Group performance status > 1) + [1.367 × log10 (LDH/upper limits of normal LDH)] + [0.9393 × log10 (white blood cell count per 10-6 L)].
Patients were classified as low risk (score < 5.7), intermediate risk (5.7 ≤ score < 6.2), or high risk (score ≥ 6.2) according to their MIPI scores, as previously reported. To estimate the mitotic rate, Ki-67 staining was performed. Positivity of 0%-10% was classified as a low mitotic rate, while more than 10% positivity was classified as high. Cox proportional hazards regression analysis was used to analyze the prognostic factors. Factors exhibiting significant values in the univariate analysis were further analyzed by multivariate analysis. Kaplan-Meier curves were generated for OS and EFS. We compared the curves for the two groups with the log-rank test. Statistical analyses were performed by JMP 8.0.1 software (SAS Institute, Cary, NC, USA), and P < 0.05 was considered significant.
Thirty-five patients (32 male, 3 female) were enrolled; their characteristics are summarized in Table Table1.1. The median age at diagnosis was 67 years (range: 47-86 years). Twenty-nine of 35 patients (82.9%) were Lugano stage IV, whereas only one patient was stage I with gastric involvement. Only this case fulfilled Dawson’s criteria for primary GI MCL. Essentially all the patients in this study had GI lesions. Other extranodal sites involved were bone marrow (n = 12), spleen (n = 5), liver (n = 3), kidney (n = 3), Waldeyer’s ring (n = 3), peripheral blood (n = 3), skin (n = 2), tongue (n = 1), and ureter (n = 1).
All patients underwent esophagogastroduodenoscopy, and 21 of the 35 patients underwent colonoscopy. Of the 35 patients, GI involvement in the esophagus, stomach, and duodenum was found in 2 (5.7%), 26 (74.3%), and 12 (34.3%) patients, respectively. Among the 21 patients who also received colonoscopy, 10 (47.6%), 3 (14.3%), 12 (57.1%), and 10 (47.6%) patients showed GI involvement in the ileum, cecum, colon, and rectum, respectively (Table (Table1).1). Twenty-two patients had at least one site of involvement in the intestines (from duodenum to rectum). Endoscopic features are summarized in Table Table2.2. Gross findings of the esophageal lesions were the protruded type in one patient and the superficial type in another. The former lesion was a solitary nodule in the esophagogastric junction, and measured about 7 mm in diameter (Figure (Figure2A).2A). The latter lesion had a unique form showing slightly elevated multiple white plaques resembling glycogenic acanthosis (Figure (Figure2B).2B). The number of these plaques increased 13 mo after the initial endoscopy, and a biopsied specimen revealed infiltration by MCL cells. Gastric lesions varied morphologically: the superficial type was found in 7 cases (26.9%), the protruded type in 6 (23.1%), the fold thickening type in 6 (23.1%), the ulcerative type in 6 (23.1%), and the combined (protruded and ulcerative) type in 1 (3.8%) (Figure (Figure3).3). In stark contrast to the morphological variety of the gastric lesions, MLP was dominant in the intestines; it was identified in 17 of the 22 cases (77.3%). The remaining patients showed protruded type lesions in 4 (18.2%) and the superficial type in 1 (4.5%) (Figure (Figure44).
Various regimens were employed for treatment of the 35 patients, because they were treated at 21 different institutions. CHOP-like regimens, which included CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) and THP-COP (pirarubicin, cyclophosphamide, vincristine, and prednisone), were used for 16 patients, and rituximab was also administered to 10 of those 16 patients. The hyper-CVAD/MA regimen (fractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone, alternating with high doses of methotrexate and cytarabine) was used for 12 patients, and rituximab was added to this regimen in 7 of the 12 patients. In the hyper-CVAD/MA regimen, cycles 1, 3, 5, 7 (Hyper-CVAD) consisted of cyclophosphamide (300 mg/m2, iv every 12 h for 6 doses) on Days 1-3, doxorubicin (16.6 mg/m2 per day, iv continuous infusion for 72 h) on Days 4-6, vincristine (1.4 mg/m2 iv, maximum 2 mg) on Days 4 and 11, and dexamethasone (40 mg/d, iv or orally) on Days 1-4 and Days 11-14. Cycles 2, 4, 6, 8 (MA) consisted of methotrexate (200 mg/m2, iv) on Day 1, methotrexate (800 mg/m2, iv continuous infusion for 22 h) on Day 1, and cytarabine (3000 mg/m2, reduced to only 1000 mg/m2 if age > 60 years or creatinine > 1.5 mg/dL, iv every 12 h for 4 doses) on Days 2-3. Cycles 1 and 2 were alternated every 21 d. After the hyper-CVAD/MA regimen, high-dose chemotherapy with autologous peripheral blood stem cell transplantation (PBSCT) was performed in 8 patients. Allogeneic PBSCT was carried out in one patient for relapse after the hyper-CVAD/MA regimen.
For OS, only the therapeutic regimen (hyper CVAD/MA regimen vs other treatment) had a significant impact in the univariate Cox regression analysis (Table (Table3).3). For EFS, neither sex, Lugano stage (I and II vs III and IV), LDH levels, white blood cell count, bone marrow involvement, Ki-67 positivity, IPI (high risk vs other), nor MIPI score showed prognostic relevance in the univariate analysis. In contrast, age and therapeutic regimen (hyper CVAD/MA regimen vs other treatment) had a significant impact on EFS (Table (Table4),4), whereas in the multiple Cox regression, neither age nor the therapeutic regimen was of prognostic relevance (Table (Table55).
Survival curves are shown in Figure Figure5.5. Median OS had not been reached after a median follow-up of 33.8 mo, and the 3-year OS rate was 61.1%. Median EFS was 15.2 mo, and the 3-year EFS rate was 44.4%. As shown in Figure Figure5,5, patients treated with the hyper-CVAD/MA regimen had both markedly longer OS and longer EFS than the group without hyper-CVAD/MA; the 3-year OS rate and 3-year EFS rate of the patients treated with hyper-CVAD/MA were 83.3% and 66.7%, respectively, whereas those of the patients without hyper-CVAD/MA were 46.4% and 33.8%. The differences between the two groups were statistically significant.
In our patients, the gastric lesions varied in form; superficial, protruded, fold thickening, and ulcerative lesions appeared equally. On the other hand, intestinal involvement showed a clear predominance of MLP. Cornes described gastric lesions as pale folds, like the convolutions of the brain, with some larger lobules standing out like solitary tumors. Ruskoné-Fourmestraux et al also characterized gastric lesions as having large cerebroid folds; this is equal to the fold thickening type in our study. Romaguera et al reported that the most frequent abnormal findings in the upper GI tract in MCL cases were nodules and inflammation. A recent report by Salar et al described 3 cases of gastric lesions, with antral gastritis in 2 and pangastritis in 1. Thus, the gastric involvement of MCL shows diverse forms in endoscopic examinations.
In this study, esophageal, gastric, and intestinal lesions were found by endoscopic examinations in 2 (5.7%), 26 (74.3%), and 22 (62.9%) patients, respectively. The prevalence of gastric lesions seems to be higher than the reported incidence of 46%-49%, whereas the prevalence of intestinal lesions is compatible with the reported rate of 38%-62%[5,6]. Importantly, histological evidence of MCL involvement reportedly exists in most cases even those with endoscopically intact mucosa[5,6]. Therefore, random biopsies by endoscopy at the time of diagnosis will probably reveal a higher incidence of GI involvement in MCL patients. Nevertheless, the impact of microscopic involvement in the outcome of MCL patients has not yet been elucidated.
MALT lymphomas and follicular lymphomas are classified as small B-cell neoplasms as well as MCL. It is well known among endoscopists that MALT lymphomas exhibit various lesions in the stomach including erosions, ulcers, polyps, protruded tumors, and swollen mucosal folds, but the involvement of other parts of the GI tract, such as the small intestine, colon, and rectum, is uncommon[20-23]. Follicular lymphomas often arise in the duodenum around the ampulla of Vater with multiple whitish granules, and they rarely form bulky masses or ulcers. Endoscopists should bear in mind that MALT lymphomas and follicular lymphomas can affect the entire GI tract and even form MLP[25-27]. Thus, evaluation of the entire GI tract enables endoscopists to discriminate MCL from other small B-cell neoplasms, although a differential diagnosis is sometimes difficult and requires immunostaining for CD5, CD10, cyclin D1, and BCL2.
The prognosis of patients with GI MCL has not yet been thoroughly discussed. The present study revealed that the hyper-CVAD/MA regimen plus rituximab and PBSCT is effective for MCL patients with GI involvement as well as for systemic MCL. The prognosis of MCL patients is poorest among those with B-cell lymphoma. Conventional chemotherapeutic regimens such as CHOP, with or without rituximab, obtained only short (less than 2 years) remission periods, despite a high remission rate (75%-96%)[29-32]. To overcome the unfavorable outcomes of CHOP-like regimens, a hyper-CVAD/MA regimen was established as an effective cytoreductive regimen for MCL patients[33,34]. An excellent response rate, over 90%, and a CR rate of 38%-68% were achieved by this intensified initial chemotherapy. Additionally, rituximab in combination with hyper-CVAD/MA has augmented CR rates to 87%. High-dose chemotherapy with PBSCT after hyper-CVAD/MA with or without rituximab could extend the CR period and achieve 3-year OS of 72%-92%[33,36]. In this study, the hyper-CVAD/MA regimen exhibited a superior response and higher rates of survival among MCL patients over other regimens. Therefore, this regimen is a promising therapeutic option for MCL patients with GI involvement as well as for systemic MCL. We believe the hyper-CVAD/MA regimen plus rituximab and PBSCT should be administered to MCL patients with GI involvement who are young (< 60-65 years) and fit (no relevant co-morbidity).
A small subset of patients with MCL may show indolent behavior and have extended survival even with little or no treatment. To identify this subset of patients and classify MCL patients according to their prognoses, several researchers have attempted to establish a prognostic index. Ki-67 positivity, which represents cell proliferation, has been reported as a predictor; high mitotic rates were associated with adverse prognoses[17,18]. The recently introduced MIPI scoring system successfully differentiated OS based on four independent prognostic factors: age, performance status, LDH, and leukocyte count. In this study, however, Ki-67 positivity and MIPI score failed to classify our patients’ prognoses. We speculate that any one of several factors may explain this. First, MCL patients with GI involvement might have different prognoses from all other MCL patients. In this study, all but 3 patients were at advanced stages (II-2 or IV). GI involvement of MCL might represent a more advanced disease status, but the impact of these GI findings on the outcomes of MCL patients is not known. Second, because this study was retrospective and the treatment regimens administered to patients were not uniform, the non-uniform backgrounds might have distorted analysis. Further investigation concerning the prognosis of MCL with GI involvement is required in a study using a larger patient group.
In conclusion, 35 cases of MCL with GI involvement were included in this report. Esophageal, gastric, and intestinal lesions were identified in 2 (5.7%), 26 (74.3%), and 22 (62.9%) patients, respectively. MLP was a representative form of intestinal involvement, whereas a variety of lesions were found in the stomach. Prompt diagnosis based on the above findings and administration of initial cytoreductive chemotherapy, such as the hyper-CVAD/MA regimen followed by PBSCT, may improve the survival of MCL patients.
Mantle cell lymphoma (MCL) is clinically more aggressive and has a shorter median survival, only 3-5 years, as compared with other types of small B-cell neoplasms. MCL can affect the gastrointestinal (GI) tract and its frequency is reportedly up to 60% in patients undergoing endoscopic examinations. However, the prognosis of MCL patients with GI involvement has not yet been revealed in detail.
Recently, the hyper-CVAD/MA regimen was established as an effective cytoreductive regimen for MCL patients. An excellent response rate, over 90%, and a complete remission (CR) rate of 38%-68% were achieved by this intensified initial chemotherapy. Additionally, rituximab in combination with hyper-CVAD/MA has augmented the CR rates to 87%. High-dose chemotherapy with autologous peripheral blood stem cell transplantation (PBSCT) after hyper-CVAD/MA with or without rituximab could extend the CR period and achieve 3-year overall survival (OS) of 72%-92%.
Patients treated with the hyper-CVAD/MA regimen had both markedly longer OS and longer event-free survival (EFS) than the group without hyper-CVAD/MA; the 3-year OS rate and 3-year EFS rate of the patients treated with hyper-CVAD/MA were 83.3% and 66.7%, respectively, whereas those of the patients without hyper-CVAD/MA were 46.4% and 33.8%, respectively. The differences between the two groups were statistically significant.
Hyper-CVAD/MA regimen plus rituximab and PBSCT is effective for MCL patients with GI involvement as well as for systemic MCL. The hyper-CVAD/MA regimen plus rituximab and PBSCT should be administered to MCL patients with GI involvement who are young (< 60-65 years) and fit (no relevant co-morbidity).
In the hyper-CVAD/MA regimen, cycles 1, 3, 5, 7 (Hyper-CVAD) consisted of cyclophosphamide, doxorubicin, vincristine, and dexamethasone. Cycles 2, 4, 6, 8 (MA) consisted of methotrexate and cytarabine. Cycles 1 and 2 were alternated every 21 d. Rituximab is a chimeric monoclonal antibody against the protein CD20, which is primarily found on the surface of B cells. Rituximab is used for the treatment of many lymphomas, leukemias, and some autoimmune disorders.
This is a multi-institutional retrospective review of 35 MCLs of the GI tract. Its strength is the multiple institutions that patients were treated allowing for comparison of different therapeutic regimens.
We are grateful to the following doctors for their important contributions to the search of patients’ medical records: Dr. Atsushi Imagawa, Mitoyo General Hospital; Dr. Atsushi Yao, Okayama Rosai Hospital; Dr. Hiraku Fujita, Tottori Municipal Hospital; Dr. Jun Tomoda, Fukuyama Medical Center; Dr. Junji Shiode, Okayama Saiseikai General Hospital; Dr. Junnosuke Shimamura, Kurashiki Riverside Hospital; Dr. Katsuhisa Ohashi, Juzen General Hospital; Dr. Kazutaka Sunami, Okayama Medical Center; Dr. Kyouichi Hayashi, Okayama Red Cross General Hospital Tamano Branch Hospital; Dr. Mamoru Nishimura, Okayama Citizens’ Hospital; Dr. Masashi Araki, Kagawa Rosai Hospital; Dr. Masatsugu Miyoshi, Kurashiki Central Hospital; Dr. Motowo Mizuno, Hiroshima City Hospital; Dr. Seiyu Suzuki, Sumitomo Besshi Hospital; Dr. Shigeatsu Fujiki, Tsuyama Central Hospital; Dr. Shouichi Tanaka, Iwakuni Clinical Center; Dr. Tadashi Matsumura, St. Mary’s Hospital; Dr. Takeshi Shimomura, Hiroshima-Nishi Medical Center; Dr. Tomohiko Mannami, Chugoku Central Hospital; Dr. Tomoki Inaba, Kagawa Prefectural Central Hospital.
Peer reviewers: Chakshu Gupta, MD, FCAP, Pathology and Laboratory Medicine, Heartland Regional Medical Center, 5325 Faraon Street, St. Joseph, MS 64506, United States; Shotaro Nakamura, MD, Department of Medicine and Clinical Science, Kyushu University, Maidashi 3-1-1, Higashi-ku, Fukuoka 812-8582, Japan
S- Editor Wang YR L- Editor Webster JR E- Editor Zheng XM