Comparison of clinicopathologic characteristics of gastric follicular lymphomas and duodenal follicular lymphomas

    Comparison of clinicopathologic characteristics of gastric follicular lymphomas and duodenal follicular lymphomas Akiko Miyagi Maeshima MD, Hirokazu Taniguchi MD, Tomotaka Suzuki MD, Sayako Yuda MD, Kosuke Toyoda MD, Nobuhiko Yamauchi MD, Shinichi Makita MD, Suguru Fukuhara MD, Wataru Munakata MD, Dai Maruyama MD, Yukio Kobayashi MD, Yutaka Saito MD, Kensei Tobinai MD PII: DOI: Reference:

S0046-8177(17)30155-7 doi: 10.1016/j.humpath.2017.04.025 YHUPA 4217

To appear in:

Human Pathology

Received date: Revised date: Accepted date:

1 March 2017 19 April 2017 21 April 2017

Please cite this article as: Maeshima Akiko Miyagi, Taniguchi Hirokazu, Suzuki Tomotaka, Yuda Sayako, Toyoda Kosuke, Yamauchi Nobuhiko, Makita Shinichi, Fukuhara Suguru, Munakata Wataru, Maruyama Dai, Kobayashi Yukio, Saito Yutaka, Tobinai Kensei, Comparison of clinicopathologic characteristics of gastric follicular lymphomas and duodenal follicular lymphomas, Human Pathology (2017), doi: 10.1016/j.humpath.2017.04.025

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1 / 29

Comparison of clinicopathologic characteristics of gastric follicular

RI P

T

lymphomas and duodenal follicular lymphomas

SC

Authors: Akiko Miyagi Maeshima, MD,1 Hirokazu Taniguchi, MD,1 Tomotaka Suzuki, MD,2 Sayako Yuda, MD,2 Kosuke Toyoda, MD,2 Nobuhiko Yamauchi, MD,2 Shinichi Makita, MD,2

NU

Suguru Fukuhara, MD,2 Wataru Munakata, MD,2 Dai Maruyama, MD,2 Yukio Kobayashi,

MA

MD,2 Yutaka Saito, MD,3 and Kensei Tobinai, MD 2

ED

Affiliations: 1Pathology Division, 2Department of Hematology, and 3Endoscopy Division,

PT

National Cancer Center Hospital, 5-1-1 Tsukiji, Chuo-ku, Tokyo 104-0045, Japan

AC CE

Correspondence: Akiko Miyagi Maeshima, Pathology Division, National Cancer Center Hospital, 5-1-1 Tsukiji, Chuo-ku, Tokyo 104-0045, Japan Tel: +81-3-3542-2

Fax: +81-3-5565-7029

E-mail: [email protected]

Running head: Gastric follicular lymphoma

Conflict of interest: No conflicts of interest to declare.

ACCEPTED MANUSCRIPT 2 / 29

Summary

T

We compared the incidence, esophagogastroduodenoscopy (EGD) findings, and

RI P

histopathologic characteristics of gastric and duodenal follicular lymphomas (FL). Of 626 FL

SC

cases, primary gastric FL and secondary gastric involvement of FL were observed in 1% and 5% of the cases, respectively, which were lower incidences than duodenal FL (10% and 9%,

NU

respectively). Gastric FL usually appeared as submucosal tumors (primary, 71%: secondary,

MA

79%), whereas duodenal FL as granular lesions (primary, 92%: secondary, 87%). In the granular duodenal lesions, the neoplastic follicles were located sparsely on the muscularis mucosa and

ED

could be found between villi, whereas in the stomach, similar lesions were hidden within the

PT

lamina propria, and only larger lesions such as submucosal tumors could be detected on the

AC CE

mucosal surface. The differences in the incidences and EGD findings were considered to be associated with structural differences of the lamina propria. Typical FL features: grades 1–2 histology, follicularity, and, CD10+ and/or BCL6+ and BCL2+ were usually observed in all primary and secondary gastric and duodenal FL. Gastroduodenal and bone marrow involvement were found in 12% and 33% of the cases, respectively, and there was no significant correlation between them (P = 0.095). Twenty-nine (5%) cases were up-staged by gastroduodenal-positive results. In conclusion, the histopathology of gastric FL was similar to that of duodenal and nodal FL, the differences in the incidence and EGD findings between gastric and duodenal FL were considered to be associated with structural difference of the lamina propria, and EGD was useful

ACCEPTED MANUSCRIPT 3 / 29

T

as a staging procedure.

RI P

Key words: follicular lymphoma, stomach, duodenum, lamina propria, histopathology,

AC CE

PT

ED

MA

NU

SC

immunohistochemistry

ACCEPTED MANUSCRIPT 4 / 29

1. Introduction

T

Follicular lymphoma (FL) is the most common low-grade B-cell lymphoma. Primary

RI P

intestinal FL was described as a variant of FL in the World Health Organization Classification of

SC

Tumours of Haematopoietic and Lymphoid Tissues published in 2008 [1], and its name was changed to duodenal-type FL in the 2016 revision of the World Health Organization

NU

classification of lymphoid neoplasms [2]. The majority of cases of primary FL in the

MA

gastrointestinal tract occur in the small intestine, particularly in the second portion of the duodenum, presenting as white granular lesions (multiple small polyps), often as an incidental

ED

finding on esophagogastroduodenoscopy (EGD) [3–6]. The morphology, immunophenotypes,

PT

and genetic features are similar to those of nodal FLs; specifically, duodenal-type FL usually

AC CE

shows grades 1–2 histology, CD10+ and BCL2+ immunophenotypes, and IGH/BCL2 fusion [5, 7]. Duodenal-type FL has many oncogenic alterations that overlap with in-situ FL [8] and extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma) [6]; the patient outcomes appear to be excellent, including some patients managed with a watch-and-wait strategy [5, 7]. In contrast, FL in the stomach has been rarely reported, and the clinicopathological features of patients with gastric FL have not been analyzed sufficiently. The ratio of gastric FL to duodenal FL was reported as 2:111 [9] and 1:19 [10], and the endoscopic feature of primary gastric FL was non-multiple nodules [9]. Iwamuro et al. reported that the ratio of the incidence of

ACCEPTED MANUSCRIPT 5 / 29

secondary involvement of FL in the stomach to duodenum was 1:34 [11]. In our institute, EGD

T

and bone marrow aspiration and/or biopsy are performed as routine staging procedures for FL

RI P

patients; however, the association between bone marrow involvement and gastroduodenal

SC

involvement has not been analyzed previously.

Histopathogically, differential diagnosis between gastric FL and MALT lymphoma is

NU

often challenging, and gastric FL can be misdiagnosed as MALT lymphoma. FL resembling

MA

MALT lymphoma morphologically and immunophenotypically (CD10− and/or BCL6−), and FL with lymphoepithelial lesions (LEL) have been reported [12]. LEL in gastric FL lesions were

ED

observed in 42% (5/12) of cases [12].

PT

The aim of the present study was to analyze the clinicopathological characteristics of

AC CE

primary gastric FLs and secondary gastric involvement of FLs in comparison with duodenal FLs, and to discuss the reason why the incidence and endoscopic findings differ between gastric and duodenal FLs.

2. Materials and Methods 2.1. Patients The present study included 626 consecutive patients with FL who were diagnosed as having FL and underwent EGD for analyzing the primary gastroduodenal lymphoma or for a staging procedure at the National Cancer Center Hospital, Tokyo, Japan, between 2000 and

ACCEPTED MANUSCRIPT 6 / 29

2015. Clinical information, including site of the disease, EGD findings, age, sex, stage, level of

T

lactate dehydrogenase, symptoms, treatment, response, and outcomes, were obtained from

RI P

medical records. The Lugano staging system for gastrointestinal lymphoma [13] was used for

SC

staging of primary gastroduodenal FL, and the Ann Arbor staging system was used for staging of FL whose primary sites were not gastroduodenal. The staging procedures included bone

NU

marrow aspiration and/or biopsy, EGD, whole-body computed tomography (CT), and optional

MA

positron emission tomography/CT. Response to treatment was defined according to the revised response criteria for malignant lymphoma [14]. The study was approved by the institutional

AC CE

2.2. EGD findings

PT

ED

review board of the National Cancer Center.

Information regarding endoscopic findings and the number of gastric and/or duodenal

lesions was collected. Endoscopic findings were classified as granular, fold swelling, submucosal tumor, or ulcerated. Granular lesions were defined as aggregation of multiple small polyps of a few millimeters or less in diameter. Submucosal tumor was defined as large submucosal tumor >5 mm in diameter.

2.3. Histopathologic and immunohistochemical analyses Materials obtained by biopsy or surgery were fixed in 10% neutral-buffered formalin,

ACCEPTED MANUSCRIPT 7 / 29

embedded in paraffin, cut into 4 μm-thick sections, and stained with hematoxylin and eosin for

T

histological evaluation. All specimens were diagnosed as FL by two hematopathologists (AMM

RI P

and HT) according to the World Health Organization Classification of Tumours of

SC

Haematopoietic and Lymphoid Tissues published in 2008 [1]. The presence of lymphoepithelial lesions and histologic transformation were also observed.

NU

Immunohistochemical analysis of formalin-fixed, paraffin-embedded tissues was

MA

performed using a panel of monoclonal antibodies. Sections 4-μm thick were cut from each paraffin block, deparaffinized, and incubated at 121 °C in citrate buffer (pH 6.0) for 10 min for

ED

antigen retrieval. Antibodies against the following antigens were used to diagnose FL: CD3

PT

(PS1, 1:25; Novocastra, Newcastle upon Tyne, UK), CD5 (4C7, 1:100; Novocastra), CD10

AC CE

(56C6, 1:100; Novocastra), CD20 (L26, 1:200; Dako, Glostrup, Denmark), BCL2 (124, 1:100; Dako), BCL6 (PG-B6p, 1:50; Dako), and cyclin D1 (SP4, ready for use; Nichirei, Tokyo, Japan).

2.4. Interphase fluorescence in situ hybridization (FISH) Four-μm-thick sections were cut from each paraffin block and used for FISH analysis. An LSI IGH Spectrum Green/LSI BCL2 Spectrum Orange Dual Fusion Translocation Probe (Vysis, Downers Grove, IL) was used to detect the t(14;18) translocation that gives rise to IGH/BCL2 fusion. The fusion gene was identified as previously described [15]. In brief, 50–200

ACCEPTED MANUSCRIPT 8 / 29

nuclei were scored per case, and if ≥2% of the tumor cells had 2 fusion signals on IGH/BCL2

RI P

T

examination, the sample was defined as being positive for the fusion.

SC

2.5. Statistical analysis

Patient outcomes after diagnosis of FL were estimated using the Kaplan–Meier

NU

method, and the P value was calculated by log rank test. The correlation between two

MA

parameters was estimated using the Spearman-coefficient test or the χ2 test. Differences were considered significant when the P value was ≤0.05. The progression-free survival (PFS) duration

ED

was calculated from the date of initial diagnosis to the date of disease progression, relapse, or

PT

death from any cause. The overall survival (OS) duration was calculated from the date of initial

AC CE

diagnosis to the date of death from any cause.

3. Results

3.1. Characteristics of patients with primary or secondary FLs in the stomach The characteristics of seven patients with primary gastric FL compared to 63 patients with primary duodenal FL are summarized in Table 1. The seven gastric FL cases corresponded to 1% of the 626 total FLs. In contrast, 63 primary duodenal FL cases (10%), including 2 primary duodenal FL cases with gastric involvement, were observed. The primary gastric FLs were detected by medical check-up or by symptoms such as abdominal discomfort or

ACCEPTED MANUSCRIPT 9 / 29

melena/anemia. The treatment strategies included watchful waiting (2 patients), rituximab,

T

cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) ± radiotherapy (RT) (2

RI P

patients), eradication of Helicobacter pylori (2 patients), and resection (1 patient). Two patients,

SC

who underwent eradication of H. pylori, were initially diagnosed with low-grade B-cell lymphoma, either MALT lymphoma or FL, and neither patient responded to the therapy. The

NU

five-year PFS and OS rates were 43% and 75%, respectively. One patient died of histologic

MA

transformation of FL. The PFS rate of patients with gastric FL was significantly shorter than that of patients with duodenal FL (P = 0.042).

ED

The characteristics of FL patients with secondary involvement of the stomach and/or

PT

duodenum are summarized in Table 2. Of 556 patients with FL whose primary sites were not

AC CE

stomach or duodenum, 29 patients (5%) had involvement of FL in the stomach and 48 patients (9%) in the duodenum. Both gastric and duodenal involvement was observed in six patients. A diffuse large B-cell lymphoma component was found in more cases in the stomach (5 cases, 17%) than in the duodenum (1 case, 2%). Among six diffuse large B-cell lymphoma components, only one in the stomach was tested and proven to have IGH/BCL2 fusion by FISH.

3.2. EGD and histopathologic findings The EGD and histopathologic finding are summarized in Table 3. All seven primary gastric FL cases had a single submucosal or ulcerated tumor. In contrast, the majority of primary

ACCEPTED MANUSCRIPT 10 / 29

duodenal FL cases had multiple (97%) granular (92%) lesions; however, some cases had both

T

granular and submucosal and/or ulcerated tumor. The majority of secondary involvement of FL

RI P

in the stomach were submucosal tumors (79%, Figures 1A and 1B), and granular lesions (4%)

SC

were rarely seen. The majority of secondary involvement of FL in the duodenum were granular lesions (87%, Figures 1C and 1D). In the stomach, both primary FL and secondary

NU

involvement of FL had similar EGD findings, with the majority being submucosal (71% and

MA

79%, respectively) or ulcerated tumors (29% and 17%, respectively). In the duodenum, both primary FL and secondary involvement of FL showed similar EGD findings, with the majority

ED

being granular lesions (92% and 87%, respectively).

PT

In the granular lesions of the duodenum, the neoplastic follicles were located sparsely

AC CE

on the muscularis mucosa, and they could be found between villi (Figure 1D). One case with secondary involvement of FL in the stomach was found incidentally in the endoscopic submucosal dissection material for early gastric adenocarcinoma (Figures 1E and 1F). The neoplastic follicles were located sparsely on the muscularis mucosa but were hidden within the lamina propria and could not be detected by observation of the mucosal surface. Among primary FL and secondary involvement of FL in the stomach and duodenum, the histologic grade was usually grade 1–2, follicularity was observed, and CD10 and/or BCL6 and BCL2 were positive. Lymphoepithelial lesions, which are typical features of MALT lymphoma, were seen in 5 cases (4%, Figure 2E). In gastric FL cases, a diffuse pattern without

ACCEPTED MANUSCRIPT 11 / 29

CD21+ follicular dendritic cell meshwork but CD10+ and/or BCL6+ phenotype (primary, 28%;

T

secondary, 50%) (Figures 2A-2D) was more frequently observed than in duodenal FL cases

RI P

(primary, 6%; secondary, 4%). A total of 4 cases (3%) with diffuse pattern, CD10−, and BCL6−

SC

but IGH/BCL2 fusion positivity were also observed.

NU

3.3. Comparison of EGD findings of six patients with FL showing both gastric and

MA

duodenal involvement

The EGD findings of six patients with FL with secondary involvement of both the

ED

stomach and duodenum are summarized in Table 4. As with primary and secondary

PT

involvement of FL in the stomach or duodenum, the most frequent EGD finding in the stomach

AC CE

was submucosal tumor (100%) and in the duodenum was granular lesion (83%).

3.4. Correlation between EGD findings and histologic grade of FL The correlation between EGD findings and histologic grade of FL in the stomach and

duodenum is summarized in Table 5. The majority were grades 1–2 (135 cases, 97%), and a few cases were grade 3a (3 cases, 2%) or grade 3b (2 cases, 1%). There was no significant correlation between EGD findings and histologic grade (P = 0.667). However, two grade 3b cases had ulcerated lesions.

ACCEPTED MANUSCRIPT 12 / 29

3.5. Correlation between secondary gastroduodenal and bone marrow involvement

T

Of 556 FL cases whose primary sites were not gastroduodenal, gastroduodenal and

RI P

bone marrow involvement were found in 66 (12%) and 185 cases (33%), respectively. There

SC

was no significant correlation between gastroduodenal and bone marrow involvement of the 556 FL cases (P = 0.095, Table 6). Of 556 cases, 29 (5%) were up-staged by

MA

NU

gastroduodenal-positive results.

ED

4. Discussion

PT

We analyzed the incidence and the histopathologic and immunophenotypic

AC CE

characteristics of gastric FL compared with those of duodenal FL. First, we discuss why duodenal FL appears as granular lesions and gastric FL appears as submucosal or ulcerated tumors. The surface of the lamina propria of the stomach is smooth, but the surface of the small intestine is villous. The thickness of the lamina propria of the stomach was reported as 1.08 ± 0.07 mm [16], and the length of the villi in the small intestine was reported as 1 ± 0.5 mm [17], depending on the degree of distention of the intestinal wall and the state of smooth muscle fiber contraction within these organs. The ratio of villous length to crypt length in the small intestine varies from approximately 3:1 to 5:1 [18]; therefore, the surface of the crypt is 0.23 ± 0.15 mm from the muscularis mucosa. The diameter of a single neoplastic follicle of FL has been reported

ACCEPTED MANUSCRIPT 13 / 29

as 0.4 ± 0.08 mm [19]. Because neoplastic follicles are located on the muscularis mucosa,

T

neoplastic follicles 1.08 ± 0.07 mm or less in diameter cannot be detected by endoscopic

RI P

examination of the stomach. However, in the small intestine, neoplastic follicles more than 0.23

SC

± 0.15 mm in diameter can be found as small granules by endoscopic examination. Therefore, in the duodenum, granular lesions, which reflect early lesions of FL with individual neoplastic

NU

follicles sparsely distributed on the muscularis mucosa, were frequently found, and the incidence

MA

of primary and secondary involvement of FL was high, at 10% and 9%, respectively, in the present study. However, in the stomach, granular lesions were rarely detected, and gastric FL

ED

was found in progressed forms such as submucosal or ulcerated tumors. The incidence of

PT

primary FL found by EGD was lower in the stomach (1%) than that in the duodenum, and

AC CE

secondary involvement of FL was lower in the stomach (5%) than that in the duodenum. The PFS of patients with gastric FL was significantly shorter than that of patients with duodenal FL, suggesting that the gastric FL had progressed much further than the duodenal FL. It was suggested that involvement of early FL lesions is easily found in the small intestine because of the specific surface structure of the lamina propria; the surface is villous, and the crypt length is only 0.23 ± 0.15 mm. In contrast, in the stomach, the surface is flat, and the crypt length is 1.08 ± 0.07 mm; therefore, granular lesions 1.08 ± 0.07 mm or less in diameter cannot be detected, and only large and progressed tumors more than 1.08 ± 0.07 mm in diameter can be detected as submucosal or ulcerated tumors. We believe the real incidence of gastric FL is higher but they

ACCEPTED MANUSCRIPT 14 / 29

were missed by EGD.

T

Whether they are gastric or duodenal, primary or secondary involvement, typical

RI P

histopathologic features of FLs: the presence of follicularity, CD10+ and/or BCL6+, and BCL2+

SC

were observed in the majority of the cases. However, in the stomach, FLs with a diffuse pattern but CD10+ and/or BCL6+ were observed in 28% of the primary cases and 50% of the secondary

NU

involvement cases. It is possible that only the surface of the tumor was biopsied, and

MA

follicularity could not be seen, or FL showed diffusion more frequently in the stomach. Rare FL cases had a diffuse pattern and CD10− and BCL6− immunophenotypes, but were positive for

ED

IGH/BCL2 fusion or had LELs. These are challenging cases for differentiating between MALT

PT

lymphoma and FL. Because the EGD findings of gastric MALT lymphoma are varied but

AC CE

usually involve minimal-to-mild surface erosion [20], whereas gastric FL involved submucosal or ulcerated tumors in the present study, EGD findings might be helpful to differentiate MALT lymphoma and FL in the stomach. The EGD findings of FL were considered to progress from granular, fold swellings, to submucosal or ulcerated tumors, and the EGD findings were associated with tumor volume but not histologic grade in the present study. However, because two ulcerated tumors corresponded to grade 3b cases, grade 3b or more should be considered if ulcerated tumors are found. EGD performed as the staging procedure could detect 5% (29 cases) of gastric stage IV lesions and 9% (48 cases) of duodenal stage IV lesions in the present study. Of them, 29

ACCEPTED MANUSCRIPT 15 / 29

(5%) were up-staged by gastroduodenal-positive results. Because bone marrow involvement

T

and gastroduodenal involvement were not significantly correlated in the present study, EGD is

RI P

recommended as a staging procedure in FL cases.

SC

In conclusion, primary FL and secondary involvement of FL in the stomach was observed in 1% and 5% of the cases, respectively, which were lower incidences than in the

NU

duodenum. Both primary FL and secondary involvement of FL in the stomach usually appeared

MA

as submucosal or ulcerated tumors, in contrast to granular lesions in the duodenum. The differences in incidence and EGD findings were considered to be associated with structural

AC CE

Acknowledgements:

PT

ED

differences of the lamina propria.

This work was supported in part by the National Cancer Centre Research and Development Fund (26-A-4 and 26-A-24), and JSPS Grants-in-Aid for Scientific Research (C-16K09865) in Japan.

ACCEPTED MANUSCRIPT 16 / 29

References

T

1. Harris NL, Sweldrow SH, Jaffe ES, et al. Follicular lymphoma. In: Swerdlow SH, Campo E,

RI P

Harris NL, et al. eds. WHO Classification of Tumours of Haematopoietic and Lymphoid

SC

Tissues, 4th edn. Lyon: International Agency for Research on Cancer; 2008:158–166. 2. Sweldrow SH, Camo E, Pileri SA, et al. The 2016 revision of the World Health Organization

NU

classification of lymphoid neoplasms. Blood. 2016;127:2375–2390.

MA

3. Yoshino T, Miyake K, Ichimura K, et al. Increased incidence of follicular lymphoma in the duodenum. Am J Surg Pathol. 2000;24:688–693.

ED

4. Shia J, Teruya-Feldstein J, Pan D, et al. Primary follicular lymphoma of the gastrointestinal

PT

tract: a clinical and pathologic study of 26 cases. Am J Surg Pathol. 2002;26:216–224.

AC CE

5. Mori M, Kobayashi Y, Maeshima AM, et al. The indolent course and high incidence of t(14;18) in primary duodenal follicular lymphoma. Ann Oncol. 2010;21:1500-1505. 6. Takata K, Sato Y, Nakamura N, et al. Duodenal follicular lymphoma lacks AID but expresses BACH2 and has memory B-cell characteristics. Mod Pathol. 2013;26:22-31. 7. Schmatz AI, Streubel B, Kretschmer-Chott E, et al. Primary follicular lymphoma of the duodenum is a distinct mucosal/submucosal variant of follicular lymphoma: a retrospective study of 63 cases. J Clin Oncol. 2011;29:1445–1451. 8. Mamessier E, Song JY, Eberle FC, et al. Early lesions of follicular lymphoma: a genetic perspective. Hematologica. 2014;99:481–488.

ACCEPTED MANUSCRIPT 17 / 29

9. Takata K, Okada H, Ohmiya N, et al. Primary gastrointestinal follicular lymphoma involving

T

the duodenal second portion is a distinct entity: a multicenter, retrospective analysis in Japan.

RI P

Cancer Sci. 2011;102:1532–1536.

SC

10. Misdraji J, Harris NL, Hasserjian RP, et al. Primary follicular lymphoma of the gastrointestinal tract. Am J Surg Pathol. 2011;35:1255–1263.

NU

11. Iwamuro M, Okada H, Takata K, et al. Diagnostic role of 18F-fluorodeoxyglucose positron

MA

emission tomography for follicular lymphoma with gastrointestinal involvement. World J Gastroenterol. 2012;18:6427–6436.

ED

12. Ohnishi N, Takata K, Miyata-Takata T, et al. CD10 down expression in follicular lymphoma

PT

correlates with gastrointestinal lesion involving the stomach and large intestine. Cancer Sci.

AC CE

2016;107:1687-1695.

13. Rohatiner A, d’Amore F, Coiffier B, et al. Report on a workshop convened to discuss the pathological and staging classifications of gastrointestinal tract lymphoma. Ann Oncol. 1994;5:397–400.

14. Cheson BD, Pfistner B, Juweid ME, et al. International Harmonization Project on Lymphoma. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25:579– 586. 15. Sekiguchi N, Kobayashi Y, Yokota Y, et al. Follicular lymphoma subgrouping by fluorescence in situ hybridization analysis. Cancer Sci. 2005;96:77–82.

ACCEPTED MANUSCRIPT 18 / 29

16. Lee EY, Wang TC, Clouse RE, et al. Mucosal thickening adjacent to gastric malignancy:

T

association with epidermal growth factor. Mod Pathol. 1989;2:397–402.

RI P

17. Bloom W, Fawcett DW. Textbook of Histology, 11th edn. Philadelphia, PA: W. B. Saunders;

SC

1986:641–678.

18. Gramlich TL, Petras RE. Small intestine. In: Mills SE eds. Histology for Pathologists, 3rd

NU

edn. Philadelphia, PA: Lippincott Williams & Wilkins; 2007:603–626.

MA

19. Crocker J, Jones EL, Curran RC. A quantitative study of the size of benign and malignant lymphoid follicles. J Clin Pathol. 1983;36:1055–1061.

ED

20. Jain D. Lymphoproliferative disorders of the gastrointestinal tract. In: Riddell R, Jain D, eds.

PT

Gastrointestinal Pathology and its Clinical Implication, 2nd edn. Philadelphia, PA: Wolters

AC CE

Kluwer; 2014:123–189.

ACCEPTED MANUSCRIPT 19 / 29

Figure legends Figure 1. (A) Endoscopic and (B) semi-macroscopic findings of a gastric FL grade 3a case with

T

a submucosal tumor. CD10-positive lymphoid cells had proliferated in the follicles and

RI P

interfollicular area. (C) Endoscopic finding of a duodenal FL grade 1 case. Granular lesions (multiple small polyps) were noted. (D) Corresponding histopathologic findings of duodenal FL.

SC

The villous length of the lamina propria was approximately 1 mm, and the crypt length was

NU

approximately 0.2 mm. Neoplastic follicles approximately 0.5 mm in diameter in the lamina propria were observed. (E) A case of incidentally detected secondary involvement of FL grade 1

MA

in the stomach in the endoscopic submucosal dissection material for adenocarcinoma. Neoplastic follicles approximately 0.5 mm in diameter with (F) CD10+ and BCL2+

ED

immunophenotypes located on the muscularis mucosa were hidden within the lamina propria of

PT

the stomach.

AC CE

Figure 2. (A) Histopathologic findings of a gastric FL case with a diffuse pattern, (B) CD20+, (C) CD10+, and low (D) Ki67 labeling index. Rare gastric FL cases showed lymphoepithelial lesions (E, HE and AE1/3).

ACCEPTED MANUSCRIPT

AC CE

PT

ED

MA

NU

SC

RI P

T

20 / 29

ACCEPTED MANUSCRIPT

AC CE

PT

ED

MA

NU

SC

RI P

T

21 / 29

ACCEPTED MANUSCRIPT 22 / 29

Table 1: Characteristics of patients with primary FLs in the stomach and duodenum

Total (n = 626)

Stomach

Duodenum1

7 (1%)

63 (10%)1

Age 58y (52-79)

61y (42-85)

6 (86%) / 1 (14%)

37 (59%) / 26

Sex Male/Female

SC

Median (range)

(41%)

NU

FLIPI Low/Intermediate/High

5/1/1

52/10/1

2

4/0/0/3

LDH 0

Symptom None (Medical check-up)

Melena/Anemia Not available

AC CE

Treatment

Watchful waiting

Resection

8 (13%)

50 (82%)

2 (29%)

8 (13%)

1 (14%)

3 (5%) 2

2 (29%)

27 (44%)

R monotherapy

R-CHOP ± RT Eradication of H. pylori3 RT

46/0/8/9

4 (57%)

PT

Abdominal discomfort

ED

High

MA

Stage

I/II1/II2/IV

12 (20%) 2 (29%)

9 (15%)

2 (29%)

8 (13%) 5 (8%)

1 (14%)

Not available

2

Response CR

3 (60%)

25 (73%)

PR NC/PD

6 (18%) 2 (40%)

3 (9%)

Not available Not applicable

T

Primary site

RI P

Parameters

2 2

27

Outcome 10-year PFS rate

43%

10-year OS rate

75%

4

77% No event

ACCEPTED MANUSCRIPT 23 / 29

1)

including 2 primary duodenal FL cases with gastric involvement using staging system for gastrointestinal lymphomas 3) Response for eradication was NC in 5 cases, PR in 4, and CR in 1. 4) P value by log-rank test: P = 0.042 CHOP, cyclophosphamide, doxorubicin, vincristine, and prednisone; CR, complete response; FL, follicular lymphoma; FLIPI, Follicular Lymphoma International Prognostic Index; LDH, lactate dehydrogenase (normal range: 119-229 U/L); NC, no change; OS, overall survival; PD, progressive disease; PFS, progression-free survival; PR, partial response; R, rituximab; RT, radiotherapy

AC CE

PT

ED

MA

NU

SC

RI P

T

2)

ACCEPTED MANUSCRIPT 24 / 29

Stomach

Duodenum

Total (n = 556)

29 (5%)1

48 (9%)1

Age Median (range)

56 (35-82)

54 (25-81)

10 (34%) / 19

18 (37%) / 30 (63%)

NU

(66%) FLIPI (NA5) 8/8/8

LDH (NA2) High

11 (38%)

Histology

15 (31%)

24 (83%) /5

ED

FL/DLBCL

20/14/14

MA

Low/Intermediate/High

SC

Sex Male/Female

(17%)

47 (98%) /1 (2%)

including 6 cases with both gastric and duodenal involvement

PT

1)

RI P

Parameters

T

Table 2: Characteristics of FL patients with secondary involvement of the stomach and/or duodenum

FL, follicular lymphoma; FLIPI, Follicular Lymphoma International Prognostic Index; DLBCL, diffuse large B-cell

AC CE

lymphoma; LDH, lactate dehydrogenase (normal range: 119-229 U/L); NA, not available

ACCEPTED MANUSCRIPT 25 / 29

Stomach

Duodenum

Primary

Secondary

Primary

No. of patients

7

24

Single/Multiple

MA N

No. of FL lesions (NA2) 7 (100%) / 0

12 (52%) / 11 (48%)

ED

Endoscopic finding (including the overlap)

5 (71%)

Submucosal tumor

2 (29%)

CE

Fold swelling

NA Histological grade 1-2/3a/3b

Secondary involvement

63

2 (3%) / 60 (97%)

471

5 (11%) / 42 (89%)

1 (4%)

58 (92%)

41 (87%)

8 (33%)

9 (14%)

9 (19%)

19 (79%)

12 (19%)

4 (9%)

4 (17%)

3 (5%)

3 (6%)

1 (4%) 1

AC

Ulcerated Normal (incidental)

PT

Granular

US

involvement 1

CR

Parameters

IP T

Table 3: EGD and histopathologic findings of FLs in the stomach and duodenum

4/1/2

24/0/0

62/1/0

46/1/0

5 (71%)

10 (42%)

58 (92%)

43 (91%)

1 (14%)

2 (8%)

1 (2%)

1 (2%)

Follicularity Present Lymphoepithelial lesion Present Histologic transformation

ACCEPTED MANUSCRIPT 26 / 29

Present

1 (14%)

3 (13%)

6 (86%)

22 (92%)

1 (2%)

Positive

IP T

CD10 and/or BCL6

61 (97%)

7 (15%)

36 (77%) (NA2)

CR

(NA1)

BCL2

23 (96%) (NA1)

US

6 (86%)

MA N

Positive

IGH/BCL2 fusion Positive

10 (of 11, 91%)

7 (of 8, 88%)

5 (72%)

10 (42%)

58 (92%)

43 (92%)

12 (50%)

4 (6%)

2 (4%)

1 (4%)

1 (2%)

2 (4%)

1 (4%, possible)2

0

0

2 (28%) 0

0

4. Diffuse and negative for CD10, BCL6, and IGH/BCL2 fusion

AC

but IGH/BCL2 fusion (+)

CE

BCL6(+) 3. Diffuse, CD10 (-), and BCL6 (-)

ED

2. Diffuse but CD10 (+) and/or

(NA1)

2 (of 2, 100%)

PT

BCL2(+)

37 (79%) (NA2)

3 (of 5, 60%)

Summary 1. Follicularity (+), CD10(+), and

62 (98%)

EGD, esophagogastroduodenoscopy; FL, follicular lymphoma; NA, not available 1)

6 cases had both gastric and duodenal involvement

2)

The sample from FL of the lymph node was also negative for IGH/BCL2 fusion.

ACCEPTED MANUSCRIPT 27 / 29

Stomach

Duodenum

No. of patients

6

6

RI P

Parameters

T

Table 4: EGD findings of 6 FL patients with both gastric and duodenal involvement

Endoscopic finding

1 (17%)

5 (83%)

Fold swelling

2 (33%)

1 (17%)

Submucosal tumor

6 (100%)

1 (17%)

Ulcerated

3 (50%)

NU

Granular

PT

ED

MA

EGD, esophagogastroduodenoscopy; FL, follicular lymphoma

AC CE

SC

(including the overlap)

ACCEPTED MANUSCRIPT 28 / 29

Table 5: Correlation between EGD findings and histological grade of FL in the stomach and duodenum Fold

ar only

swelling

Grade of FL 69

19

36

3a

1

1

1

3b

0

0

Total

70

20

Total

11

135 (97%)

0

3 (2%)

0

2

2 (1%)

37

13

140 (100%)

NU

MA ED

PT AC CE

ed

SC

1 or 2

Spearman’s coefficient test

Ulcerat

cosal tumor

EGD, esophagogastroduodenoscopy; FL, follicular lymphoma 1)

Submu

T

Granul

RI P

EGD findings

P value1

0.667

ACCEPTED MANUSCRIPT 29 / 29

Table 6: Correlation between gastroduodenal and bone marrow involvement in 556 FLs whose primary sites were not gastroduodenal Positive

Negative

Total

χ2 test

RI P

Involvement Bone marrow involvement 157 (28%)

Negative

38 (7%)

333 (60%)

Total

66 (12%)

490 (88%)

185 (33%)

SC

28 (5%)

371 (67%)

556 (100%)

NU

Positive

Of 556 FL cases, 29 (5%) were up-staged by gastroduodenal-positive results.

PT

ED

MA

FL, follicular lymphoma

AC CE

P value of

T

Gastroduodenal

0.095