ORIGINAL ARTICLE Clinical Outcome of Primary Gastric Lymhoma Treated with Chemotheray Alone or Surgery Followed by Chemotheray Ming-Chih Chang, 1,2 * Ming-Jer Huang, 1 Ying-Wen Su, 1 Yi-Fang Chang, 1 Johnson Lin, 1 Ruey-Kuen Hsieh 1 Background: The role of surgical resection in the treatment of rimary gastric lymhoma (PGL) remains unclear. This retrosective study evaluated the clinical outcome of PGL treated with chemotheray alone or surgery followed by chemotheray. Methods: During 1986 2003, 59 atients with PGL (other than mucosa-associated lymhoid tissue tye lymhoma) were identified from hosital files. The medical records, athologic sections, radiograhic images and treatment modalities of these atients were reviewed. Patients were categorized into localized (stage IE and IIE-1) and advanced (stage IIE-2 or beyond) stage grous. Survival was estimated by the Kalan-Meier method. Results: The study included 55 atients who received treatment at the same institute. Among them, 32 had localized PGL (15 stage IE, 17 stage IIE-1) and 23 had advanced disease. The median survival of the localized stage grou was not reached during a mean follow-u of 168.1 ± 16.7 months (95% confidence interval [CI], 135.4 200.8 months), while that of the advanced stage grou was 33.0 ± 6.8 months (95% CI, 19.7 46.5; < 0.001, log-rank test). Among atients with localized PGL, the 5-year overall survival rate of those receiving chemotheray alone (n = 19) or combination theray (surgery followed by chemotheray, n = 13) was 73.4% and 87.5%, resectively ( = 0.229). The 5-year disease-free survival was 68.4% and 84.6%, resectively ( = 0.540). However, ost-chemotheray life-threatening hemorrhage occurred in five of the 32 atients (15.6%) in the localized stage grou: four in the chemotheray-alone grou, and one in the combination theray grou, all of whom had failed to achieve comlete resonse. Conclusion: The clinical outcome of localized PGL treated by chemotheray alone is similar to that treated by surgery followed by chemotheray in terms of tumor resonse, disease-free survival and overall survival, suggesting that surgery be reserved for those with residual tumors after chemotheray. [J Formos Med Assoc 2006;105(3):194 202] Key Words: chemotheray, gastric hemorrhage, gastric erforation, localized rimary gastric lymhoma, surgery Over the last two decades, the management of gastric lymhoma has continued to evolve, and the role of surgical resection of the stomach aears to be diminishing, and may no longer even be defined as a routine treatment modality in large rosective randomized trials. Primary gastric lymhoma (PGL) is not a common malady, accounting for erhas 10 15% of all non- Hodgkin s lymhoma cases. 1 Our review of the literature found that most of the relevant studies of the treatment and outcome of PGL featured small numbers of atients and were 2006 Elsevier & Formosan Medical Association 1 Deartment of Internal Medicine, Mackay Memorial Hosital and 2 Mackay Medicine, Nursing and Management College, Taiei, Taiwan. Received: Aril 29, 2005 Revised: June 10, 2005 Acceted: August 2, 2005 *Corresondence to: Dr. Ming-Chih Chang, Deartment of Hematology-Oncology, Mackay Memorial Hosital, 92, Section 2, Chung-Shan North Road, Taiei, Taiwan. E-mail: mmhdonald@yahoo.com.tw 194 J Formos Med Assoc 2006 Vol 105 No 3
Chemotheray for rimary gastric lymhoma conducted retrosectively during the 1980s and 1990s. 2 5 Moreover, many investigators continue to advocate surgical resection to debulk the mass, accurately stage the lesion, and revent erforation. 6 8 During the early 1990s, some researchers suggested that chemotheray alone was effective and may also be sufficient to revent morbidity associated with gastrectomy. 2,4 This led to increased suort for gastric reservation in the treatment of subsequent residual disease. Other researchers, however, recommended surgery as curative for PGL, articularly for stage IE disease. 9,10 Those researchers who advocated surgery either with or without associated chemotheray focused on accurate tumor staging and the association of this treatment with less hemorrhage or stomach erforation; although with the advent of endoscoic ultrasound, laarotomy biosy is no longer warranted. 9 Most studies have revealed a rather low incidence of severe hemorrhage or erforation, accounting for 2.1% and 1.7%, resectively, of those individuals treated with chemotheray alone, and 2.2% and 0.9%, resectively, of surgically-treated individuals. 11,12 Such evidence suggests that the role of surgery in the treatment of PGL may be less imortant than reviously considered. Whether or not a PGL atient is referred to a medical oncologist or a surgeon thus aears to deend on the gastroenterologist s oint of view. 13 This study retrosectively examined the imact of different theraeutic modalities on the treatment outcome for PGL erformed at a single medical institution in Taiwan between 1986 and 2003. Methods Patients The records of all atients with newly diagnosed PGL in our hosital from July 1986 to June 2003 were reviewed. The inclusion criteria were: histologically roven malignant lymhoma that was considered to originate from the stomach because of redominant gastric lesions (for stage IE IIE-2 diseases) or resenting with uer gastrointestinal symtoms and signs (for stage IIIE IVE diseases); adequate athologic material for review (according to the Revised Euroean American Lymhoma/ World Health Organization classification 14 ); comlete clinical information including romt staging rocedures; rimary treatment with adequate follow-u in our hosital. Patients with mucosaassociated lymhoid tissue (MALT) lymhoma were excluded. Clinical evaluations Staging work-u included detailed hysical examination, insection of Waldeyer s ring, comuted tomograhy (CT) of the abdomen, smallbowel series, barium enema study of the colon and rectum, gallium-67 scan, and bone marrow asiration and biosy. 13 Chest CT was not routinely erformed in all atients. Patients received followu endoscoy with biosy and CT scan of the abdomen after every 2 3 cycles of chemotheray to evaluate tumor resonse. Classification of the tumor stage was according to Musshoff s modification of the Ann Arbor staging system. 15 Stage IE is when the tumor remains confined within the stomach; stage IIE-1 is when there is erigastric nodal involvement only; stage IIE-2 is when there is more distant nodal involvement below the diahragm; stage IIIE denotes nodal involvement above and below the diahragm; and stage IV denotes multile visceral organ involvement. Grading of treatment toxicity as well as tumor resonse were evaluated according to the criteria defined by the World Health Organization. 16 All tumors were evaluated by abdominal/chest CT and endoscoy. Comlete resonse (CR) was defined as the disaearance of all evidence of tumor(s) for a duration of at least 4 weeks. Partial resonse (PR) was defined as > 50% reduction in the sum of the roducts of the longest erendicular diameters of all measurable lesions in radiograhic images, with the reduction lasting at least 4 weeks. Stable disease (SD) was defined as < 50% reduction or < 25% increase in the sum of the roducts of the longest erendicular diameters of all measurable lesions, lasting > 4 weeks. Patients with rogressive lesions were not classified as having J Formos Med Assoc 2006 Vol 105 No 3 195
PR or SD. Progressive disease (PD) was defined as the aearance of new lesions or > 25% increase in the area(s) of original measurable disease. Statistical analysis Comarisons between discrete variables were erformed using Student s t test. In the analysis of overall survival, an event was defined as death from any cause. Disease-free survival was calculated from the date of comlete remission or surgery to the date of tumor relase defined by the results of imaging studies or endoscoic biosy. The survival distribution of disease-free survival and overall survival were estimated by the method of Kalan and Meier, 17 and comared by log-rank test. The association of rognostic factors with survival was estimated with Cox s roortional hazards model and exressed as relative risk with 95% confidence interval (CI). 18 Model selection for identifying the variables with an effect on survival was based on a forward stewise rocedure, with values of 0.05 and 0.06 being entered or removed, resectively. The tested variables for localized disease were gender, age (< 60, 60 years old), international rognostic index (IPI) score 19 1 vs. > 1, stage of disease (IE vs. IIE-1), extent of gastric involvement (< 5 cm vs. 5 cm), and calendar year of diagnosis (before 2000 vs. 2000 or beyond). All values were two-tailed and a value < 0.05 was considered to be statistically significant. SPSS version 10.0 (SPSS Inc, Chicago, IL, USA) was used for all statistical analyses. Results Fifty-nine consecutive atients who fulfilled the broad-sectrum diagnostic criteria for PGL, excluding those with MALT lymhoma, were identified by a review of the files. The diagnosis was made by uer gastrointestinal endoscoic biosy in 52 atients and by laaroscoy in three. Four atients who did not receive active treatment or had romt loss of follow-u after the diagnosis were subsequently excluded from the analysis. Of the 55 atients (26 men, 29 women) included in the study, 38 who received chemotheray alone were categorized into the chemotheray grou, and the other 17 who received surgery (13 subtotal gastrectomy lus D1 nodal dissection, 4 total gastrectomy) followed by chemotheray were categorized into the combined treatment grou. The chemotheray regimen was mainly CHOP (cyclohoshamide, doxorubicin, vincristine, rednisone; n = 32) or CEOP (cyclohoshamide, eirubicin, vincristine, rednisone; n = 20), which consisted of intravenous injection of cyclohoshamide 750 mg/m 2, doxorubicin 50 mg/m 2 or eiubicin 60 g/m 2, and vincristine 1.4 mg/m 2 (maximum 2 mg) on day 1, and rednisone 60 mg/m 2 orally on days 1 5. The treatment was given every 21 28 days as one cycle, for a total of 4 8 cycles. Other chemotheray regimens were m-bacod (methotrexate, bleomycin, cyclohoshamide, vincristine, dexamethasone) in two atients and COP (cyclohoshamide, vincristine, rednisone) in one. Palliative radiotheray was given in one atient with advanced PGL in the combined treatment grou. The demograhic and clinicoathologic characteristics of the two grous of atients are listed in Table 1. Diffuse large B-cell lymhoma was the most common (80%) histologic tye. Among the 32 atients with stage IE IIE-1 disease, 19 received chemotheray alone and 13 received subtotal gastrectomy followed by chemotheray. Diagnosis was made before year 2000 in 11 atients in the combined treatment grou; surgical intervention for stage IE/IIE-1 disease was erformed after 2000 in only two atients with an initial diagnosis of oorly differentiated carcinoma by histologic examination of endoscoic biosy secimen. Among the 23 atients with stage IIE-2 disease or beyond, 19 received chemotheray alone while four underwent debulking surgery before chemotheray because of uncontrolled uer gastrointestinal tract bleeding (2 atients), gastric outlet obstruction (1) and for diagnostic uroses (1). Adverse events Grade 3 4 neutroenia occurred in 13.2% of a- 196 J Formos Med Assoc 2006 Vol 105 No 3
Chemotheray for rimary gastric lymhoma Table 1. Demograhic and clinicoathologic characteristics of atients with localized and advanced rimary gastric lymhoma treated with chemotheray alone or surgery followed by chemotheray (combined treatment) Stage IE / IIE-1 (n = 32) Stage IIE-2 IV (n = 23) Total (n = 55) (n = 19) (n = 13) ( n = 19) (n = 4) ( n = 38) (n = 17) Gender, n (%) 0.60 0.05 0.23 Male 07 (36.8) 06 (46.2) 13 (68.4) 00 20 (52.6) 06 (35.3) Female 12 (63.2) 07 (53.8) 06 (31.6) 4 (100) 18 (47.4) 11 (64.7) Age, yr 0.06 0.56 0.11 Median 59 45 60 69 60 46 95% CI 41.02 72.13 36.85 58.21 45.25 72.32 49.58 76.92 43.26 72.10 38.28 64.18 Gastric lesion, n (%) 0.13 0.48 0.23 Size 5 cm 12 (63.2) 09 (69.2) 06 (31.6) 2 (50.0) 18 (47.4) 11 (64.7) Size > 5 cm 07 (36.8) 04 (30.8) 13 (68.4) 2 (50.0) 20 (52.6) 06 (35.3) Diagnosis, n (%) 0.03 0.19 0.03 Before 2000 09 (47.3) 11 (84.6) 13 (68.4) 4 (100) 22 (57.9) 15 (88.2) 2000 or after 10 (52.7) 02 (15.4) 06 (31.6) 00 16 (42.1) 2 (11.8) Histology, n (%) 0.70 0.29 0.51 DLBCL 15 (78.9) 11 (84.6) 14 (73.7) 4 (100) 29 (76.3) 15 (88.2) Diffuse mix 03 (15.8) 02 (15.4) 1 (5.3) 00 04 (10.5) 02 (11.8) Others 1 (5.3) 00 04 (21.0) 00 05 (13.2) 00 IPI score, n (%) 0.40 0.97 0.08 0 1 14 (73.7) 12 (92.3) 03 (15.8) 00 17 (44.7) 12 (70.6) 2 05 (26.3) 1 (7.7) 16 (84.2) 4 (100) 21 (55.3) 05 (29.4) CI = confidence interval; DLBCL = diffuse large B cell lymhoma; IPI = international rognostic index. tients in the chemotheray and 17.6% of atients in the combined treatment grou; thrombocytoenia occurred in 2.6% and 5.9% of atients, resectively, in the two grous. Three of 32 stage IE/IIE-1 atients develoed sesis, bacteremia or neumonia. Nine of 32 stage IIE-2 IV atients develoed serious infection (including the 5 atients with gastric erforation), all of whom had resulting fatal outcome. Six of the nine stage IIE-2 IV atients with serious infectious comlications were aged over 65 years. Severe uer gastrointestinal tract hemorrhage occurred in 15.6% (5/32) of stage IE/IIE-1 atients, including four in the chemotheray grou and one in the combined treatment grou. The hemorrhage events were associated with fatality in three atients in the chemotheray grou, with a median survival of 3 months. Nine of stage IIE-2 IV atients suffered severe hemorrhagic comlications during chemotheray; five of them died after develoing erforation. The tumor resonse and treatment-related comlications are summarized in Table 2. Tumor resonse and survival Among those evaluable for chemotheray resonse, CR was achieved in 84.2% (16/19 in the chemotheray grou) of stage IE/IIE-1 atients and in 30.4% (7/23 of both grous) of stage IIE-2 IV atients. The treatment resonse was durable in early-stage atients, with 5-year disease-free survival of 68.4%. On the contrary, four of seven advanced-stage comlete resonders had tumor recurrence within 4 years, with a 5-year diseasefree survival of 15.8% (3/19). At a median follow-u of 42 months (range, 1 128 months), the median overall survival had not been reached in the chemotheray grou (median follow-u, 34.5 months; 95% CI, 2.25 J Formos Med Assoc 2006 Vol 105 No 3 197
Table 2. Tumor resonse and treatment-related comlications of rimary gastric lymhoma Stage IE / IIE-1 (n = 32) Stage IIE-2 IV (n = 23) Total (n = 55) (n = 19) (n = 13) ( n = 19) (n = 4) ( n = 38) (n = 17) Tumor resonse, n (%) 0.50 0.48 0.29 CR 16 (84.2) 12 (92.3) 06 (31.6) 1 (25.0) 22 (57.9) 13 (76.5) PR 03 (15.8) 1 (7.7) 09 (47.4) 1 (25.0) 12 (31.6) 02 (11.8) SD / PD 00 00 04 (21.0) 2 (50.0) 04 (10.5) 02 (11.8) Anti-ulcer theray during chemo, n (%) 19 (100). 11 (85.4) 0.31 19 (100). 4 (100). 38 (100). 15 (88.2) 0.17 Grade 3 4 AE, n (%) 0.89 0.58 0.58 Hematologic Neutrohils 02 (10.5) 1 (7.7) 03 (15.8) 2 (50.0) 05 (13.2) 03 (17.6) Platelets 00 00 1 (5.3) 1 (25.0) 1 (2.6) 1 (5.9) Infection 02 (10.5) 1 (7.7) 06 (31.6) 3 (75.0) 08 (21.0) 04 (23.5) Tumor-related Bleeding 04 (21.0) 1 (7.7) 06 (31.6) 3 (75.0) 10 (26.3) 4 (23.5) Perforation 00 00 05 (26.3) 00 5 (13.2) 00 Follow-u, mo 0.054 0.295 < 0.001 Median 39 1080 21 14 34.5 62 95% CI 4.71 87.44 40.45 163.39 1.79 58.84 4.18 32.57 2.25 74.14 17.4 147.12 CR = comlete remission; PR = artial remission; SD = stable disease; PD = rogressive disease; AE = adverse event. 74.14), while it was 62 months (95% CI, 17.40 147.12) in the combined treatment grou ( < 0.001, log-rank test). The 5-year overall survival rate of stage IE/IIE-1 atients in the chemotheray grou was 73.4%, and it was 87.5% in the combined treatment grou ( = 0.229, log-rank test; Figure 1). The 5-year disease-free survival rate in the corresonding atient grous was 68.4% and 84.6%, resectively ( = 0.540, log-rank test). The disease-free survival curves of stage IE/IIE-1 atients are shown in Figure 2. The overall survival curves of both localized (stage I IIE-1) and ad- 1.0 1.0 Cumulative survival 0.9 0.8 0.7 Surgery followed by chemotheray Chemotheray alone Cumulative survival 0.9 0.8 0.7 0.6 0.5 0.4 0.3 Surgery followed by chemotheray Chemotheray alone 0.6 0.2 0.5 = 0.229, log-rank test 0 100 200 300 Months 0.1 0.0 = 0.540, log-rank test 0 60 120 180 240 Disease-free survival (mo) Figure 1. Overall survival of localized rimary gastric lymhoma. Figure 2. Disease-free survival of localized rimary gastric lymhoma. 198 J Formos Med Assoc 2006 Vol 105 No 3
Chemotheray for rimary gastric lymhoma vanced (stage IIE-2 IV) PGL, regardless of treatment modality, are shown in Figure 3. Among stage IIE-2 IV atients, the median survival of the combined treatment grou was not significantly different from that of the chemotheray grou: 14 months (95% CI, 4.2 32.6 months) versus 21 months (95% CI, 1.8 58.8 months; = 0.295, log-rank test). Univariate analysis by Cox s roortional hazards model showed that time to treatment (odds ratio, 1.529; = 0.046) and IPI (odds ratio, 1.384; = 0.051) were significant redictors of overall survival for stage IE/IIE-1 atients; while time of treatment (before 2000 vs. 2000 or after) was the only indeendent redictor by multivariate analysis. Gastric lesion size, age and gender were not significantly associated with overall survival in the univariate analysis. Discussion The otimal treatment for localized PGL remains to be established. In the 1980s, gastrectomy was used to treat PGL because of its low surgeryrelated mortality (2 5%). 20 24 A study from the Mayo Clinic reorted the 5-year-survival rate of 75% with curative resection, and 32% with alliative resection. 4 For stage IE PGL, the cure rate Cumulative survival 1.0 0.8 0.6 0.4 0.2 0.0 Localized PGL Advanced PGL < 0.001 0 50 100 150 200 250 Time Figure 3. Overall survival of advanced comared with localized rimary gastric lymhoma (PGL), regardless of treatment modality. was as high as 80%. However, because the success of surgical management of PGL deends on tumor size, the deth of its enetration into gastric tissue, and the involvement of regional lymh nodes, 9,24,25 some investigators began using chemotheray, mostly CHOP and its related regimens, to control the tumors and revent ostoerative morbidity. 2,4,12 Maor et al showed that the 6-year overall survival of atients treated with chemotheray alone was 76%. 4 However, for bulky tumors, the advantage of chemotheray is overshadowed by the otential for tumor bleeding and gastric erforation. Therefore, some investigators suggested that debulking surgery followed by chemotheray might offer better tumor control with reduced comlication rates. 26 31 In the resent series, all five atients who develoed gastric erforation received chemotheray alone, and eventually died of this comlication. In contrast, none of the atients who received combination theray had this comlication, suggesting that surgery will continue to lay an imortant role in PGL management. Of note, gastrointestinal bleeding develoed at the time of disease rogression in five atients (4 in the chemotheray grou, 1 in the combined treatment grou). The overall surgical morbidity (15.4%, 2/13) and mortality (7.7%, 1/13) in this study are similar to the findings of Rackner et al. 32 The notion that the rate of surgical comlications might offset the benefit exected from tumor control is further suorted by the results of Salles et al, who showed that for localized PGL, surgical resection rior to chemotheray did not affect the comlete resonse rate, survival rate, or disease-free survival. 2 Similar results have also been demonstrated in several studies in Asian atients and in the Groue d Etude des Lymhomes Digestifs and Groue d Etude des Lymhomes de l'adulte studies. 31,33 37 Although recent studies of PGL surgery aear to suggest its diminished role, it might enhance the effects of chemotheray in stage IE PGL. 38,39 Patients treated with conservative surgery followed by three cycles of chemotheray had a better 5-year survival rate than atients treated with J Formos Med Assoc 2006 Vol 105 No 3 199
chemotheray alone. 23,24,26,29 31 Similarly, this study found that atients treated with surgery followed by chemotheray tended to fare better than those treated with chemotheray alone. The lack of a statistical difference between the two grous might be due to the small grou sizes in this study. This study also found an imroved outcome in atients who received treatment after the year 2000, which may be attributable to better suortive care and a more widesread use of granulocyte colony stimulating factor for the treatment of febrile neutroenia. A subgrou of atients was identified who might be more likely to develo chemotheray-induced life-threatening comlications. Often treated in an outatient setting, these atients failed to achieve a comlete resonse to chemotheray. To avoid such severe comlications, we recommend re-evaluating atients by endoscoy after two cycles of chemotheray. At the same time, atients should be warned that comlications such as gastric erforation and bleeding are ossible, and awareness rograms involving comrehensive education should be art of the treatment rocess. In this series, overall survival was significantly lower for atients with advanced PGL than localized PGL, signifying again that stage is an imortant rognostic factor in the treatment of PGL. In the chemotheray grou, three of seven advanced PGL atients with comlete resonse (3/ 19, 15.8%) were long-term survivors (disease-free survival > 6 years), while all four atients in the combined-modality grou died. The outcome of combined theray was dismal due to the higher incidence of treatment-related tumor bleeding (75%) and febrile neutroenia (50%) than in the chemotheray grou (31.6% and 15.8%, resectively). Analysis of median overall survival showed that atients in the chemotheray grou survived longer than those in the combined treatment grou, with median overall survival for advanced PGL of 21 months and 14 months, resectively ( = 0.295). However, the four nonsurvivors in the combined treatment grou all had IPI 2 and were treated before 2000. One of the four atients with advanced PGL who received combined theray achieved comlete resonse and survived for 39 months. The role of surgery was limited in the treatment of advanced PGL in this series. It is noteworthy that no atients in this series received stem cell translantation as rimary or salvage treatment. While combination chemotheray may currently be regarded as the standard of care, the outcome of advanced PGL might be imroved if stem cell translantation could be integrated into the treatment, esecially in atients with chemotheray-sensitive relase. In summary, this study suggests that the clinical outcome of localized PGL treated by chemotheray alone is comarable to that treated by surgery combined with chemotheray in terms of tumor resonse, disease-free survival and overall survival. We may resume, however, that organ function is better reserved by chemotheray alone than surgery. The role of surgery for advanced PGL has become less imortant than chemotheray. However, surgery could be reserved for those with residual tumors after chemotheray. Acknowledgments The authors would like to thank Dr. Fan-Jay Lin and Dr. Shee-Chan Lin for their kind assistance with the statistical analysis of the data. References 1. d Amore F, Brinker H, Gronbak J, et al. Non-Hodgkin s lymhoma of the gastrointestinal tract: a oulation-based analysis of incidence, geograhic distribution, clinicoathological resentation, features, and rognosis. J Clin Oncol 1994;12:1673 84. 2. Salles G, Herbrecht R, Tilly H, et al. Aggressive rimary gastrointestinal lymhomas: review of 91 atients treated with the LNH-84 regimen. A study of the Groue d Etude des Lymhomes Agressifs. Am J Med 1991;90:77 84. 3. Raderer M, Valencak J, Osterreicher C, et al. Chemotheray for the treatment of atients with rimary high grade gastric B-cell lymhoma of modified Ann Arbor Stages IE and IIE. Cancer 2000;88:1979 85. 4. Maor MH, Velasquez WS, Fuller LM, et al. Stomach 200 J Formos Med Assoc 2006 Vol 105 No 3
Chemotheray for rimary gastric lymhoma conservation in stages IE and IIE gastric non-hodgkin s lymhoma. J Clin Oncol 1990;8:266 71. 5. Shiu MH, Nisce LZ, Pinna A, et al. Recent results of multimodal theray of gastric lymhoma. Cancer 1986;58: 1389 99. 6. Sheerd FA, Evans WK, Kutas G, et al. Chemotheray following surgery for stage IE and IIE non-hodgkin s lymhoma of the gastrointestinal tract. J Clin Oncol 1988;6:253 60. 7. Gosodarowicz MK, Sutcliffe SB, Clark RM, et al. Outcome analysis of localized gastrointestinal lymhoma treated with surgery and ostoerative irradiation. Int J Radiat Oncol Biol Phys 1990;19:1351 5. 8. Hockey MS, Powel J, Crocker J, et al. Primary gastric lymhoma. Br J Surg 1987;74:483 7. 9. Schwarz RJ, Conners JM, Schmidt N, et al. Diagnosis and management of stage IE and stage IIE gastric lymhomas. Am J Surg 1993;165:561 5. 10. Durr ED, Bonner JA, Strickler JG, et al. Management of stage IE rimary gastric lymhoma. Acta Haematol 1995;94: 59 68. 11. Yoon SS, Coit DG, Portlock CS, et al. The diminishing role of surgery in the treatment of gastric lymhoma. Ann Surg 2004;240:28 37. 12. Aviles A, Nambo MJ, Neri N, et al. The role of surgery in rimary gastric lymhoma: results of a controlled clinical trial. Ann Surg 2004;240:44 50. 13. de Jong D, Aleman BMP, Taal BG, et al. Controversies and consensus in the diagnosis, work-u and treatment of gastric lymhoma: an international survey. Ann Oncol 1999; 10:275 80. 14. Harris NL, Jaffe ES, Diebold J, et al. World Health Organization classification of neolastic diseases of the hematooietic and lymhoid tissues: reort of the Clinical Advisory Committee meeting Airlie House, Virginia, November 1997. J Clin Oncol 1999;17:3835 49. 15. Musshoff K. Clinical staging classification of non-hodgkin s lymhomas. Strahlentheraie 1977;153:218 21. [In German] 16. Miller AB, Hoogstraten B, Staquet M, et al. Reorting results of cancer treatment. Cancer 1981;47:207 14. 17. Kalan EL, Meier P. Nonarametric estimation from incomlete observations. J Am Stat Assoc 1958;53:457 81. 18. Cox DR. Regression model and life tables. J R State Soc B 1972;34:187 200. 19. The international non-hodgkin s lymhoma rognostic factors roject. A redictive model for aggressive non- Hodgkin s lymhoma. N Engl J Med 1993;329:984 94. 20. Rosen C, Heerden JA, Martin LK, et al. Is an aggressive surgical aroach to the atient with gastric lymhoma warranted? Ann Surg 1987;205:634 40. 21. Paulson S, Sheehan RG, Stone MJ, et al. Large cell lymhoma of the stomach: imroved rognosis with comlete resection of the all intrinsic gastrointestinal disease. J Clin Oncol 1983;1:263 9. 22. Taal BG, Burgers JM. Primary non-hodgkin s lymhoma of the stomach: endoscoic diagnosis and the role of surgery. Scand J Gastroenterol Sul 1991;188:33 7. 23. Gobbi PG, Dionigi P, Barbieri F, et al. The role of surgery in the multimodal treatment of rimary gastric non-hodgkin s lymhoma: a reort of 76 cases and review of the literature. Cancer 1990;65:2528 36. 24. Bellesi G, Alterini A, Messori A, et al. Combined surgery and chemotheray for the treatment of rimary gastrointestinal intermediate- or high-grade non-hodgkin s lymhomas. Br J Cancer 1989;60:244 8. 25. Tondini C, Giardini F, Bezztti P, et al. Combined modality treatment for rimary gastrointestinal non-hodgkin s lymhoma: The Milan Cancer Institute exerience. Ann Oncol 1993;4:831 7. 26. Sheridan WP, Medley G, Brodie GN. Non-Hodgkin s lymhoma of the stomach: a rosective ilot study of surgery lus chemotheray in early and advanced disease. J Clin Oncol 1985;3:495 500. 27. Sheherd FA, Evans WK, Kutas G, et al. Chemotheray following surgery for stages IE and IIE non-hodgkin s lymhoma of the gastrointestinal tract. J Clin Oncol 1988; 6:253 60. 28. Cooer DC, Doria R, Salloum E, et al. Primary gastrointestinal lymhomas. Gastroenterologist 1996;4:54 64. 29. Maor MH, Maddux B, Osborne BM, et al. Stage IE and IIE non-hodgkin s lymhoma of the stomach. Cancer 1984; 54:2330 7. 30. Miller TP, Dahlberg S, Cassady JR, et al. Chemotheray alone comared with chemotheray lus radiotheray for localized intermediate- and high-grade non-hodgkin s lymhoma. N Engl J Med 1998;339:21 6. 31. Ruskone-Fourmestraux A, Aegerter P, Delmer A, et al. Primary digestive tract lymhoma: a rosective multicentric study of 91 atients. Groue d Etude des Lymhomes Digestifs. Gastroenterology 1993;105:1662 71. 32. Rackner VL, Thirlby RC, Ryan JA Jr. Role of surgery in multimodality theray for gastrointestinal lymhoma. Am J Surg 1991;161:570 5. 33. Liu HT, Hsu C, Chen CL, et al. Chemotheray alone versus surgery followed by chemotheray for stage I/IIE large-cell lymhoma of the stomach. Am J Hematol 2000;64:175 9. 34. Coiffier B, Gisselbrecht C, Herbrecht R, et al. LNH-84 regimen: a multicentric study of intensive chemotheray in 737 atients with aggressive malignant lymhoma. J Clin Oncol 1989;7:1018 26. 35. Nakamura S, Matsumoto T, Iida M, et al. Primary gastric lymhoma in Jaan: a clinicoathologic analysis of 455 atients with secial reference to its time trends. Cancer 2003;97:2462 73. 36. Ho CL, Hsieh AT, Dai MS, et al. Non-Hodgkin s lymhoma of the stomach: treatment outcomes for 57 atients over a 20-year eriod. J Chin Med Assoc 2005;68:11 5. 37. Binn M, Rouskone-Fourmestraux A, Leage E, et al. Surgical resection lus chemotheray versus chemotheray alone: comarison of two strategies to treat diffuse large B-cell J Formos Med Assoc 2006 Vol 105 No 3 201
lymhoma. Ann Oncol 2003;14:1751 7. 38. Vaillant JC, Ruskone-Fourmestraux A, Aegerter P, et al. Management and long-term results of surgery for localized gastric lymhomas. Am J Surg 2000;179:216 22. 39. Koch P, del Valle F, Berdel WE, et al. Primary gastrointestinal non-hodgkin s lymhoma: II. Combined surgical and conservative or conservative management only in localized gastric lymhoma results of the rosective German Multicenter Study GIT NHL 01/92. J Clin Oncol 2001;19: 3874 83. 202 J Formos Med Assoc 2006 Vol 105 No 3