| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From The Queen’s Cancer Institute (SAH), Honolulu, Hawaii; St. Vincent’s Comprehensive Cancer Center (JSM), New York, New York; Southwest Oncology Group Statistical Center (JB), Seattle, Washington; and The Tripler Army Medical Center (TF), Honolulu, Hawaii.
Correspondence: Address correspondence and reprint requests to: Scott A. Hundahl, MD, FACS, Medical Director, The Queen’s Cancer Institute, 1301 Punchbowl St., Honolulu, HI 96813; Fax: 808-537-7080; E-mail: shundahl{at}queens.org
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods: We prospectively captured complete surgical information, including the treatment of various lymph node stations, for 553 of the 556 eligible participants in this trial. Before any survival analysis, we coded D level by using the Japanese general rules and used the Maruyama program to estimate the likelihood of disease in undissected regional node stations, defining the sum of these estimates as the Maruyama Index of Unresected Disease (MI). We analyzed survival with Cox multivariate regression.
Results: Fifty-four percent of participating patients underwent D0 lymphadenectomy. The median MI was 70 (range, 0–429). In contrast to D level, MI proved to be an independent prognostic factor, even with adjustment for the potentially linked variables of T stage and number of positive nodes. We detected no significant interaction between surgical or pathologic variables and the favorable effect of adjuvant treatment, but the power to detect such interaction was generally low.
Conclusions: MI, a measure of unresected regional nodal disease in gastric cancer, proved an independent predictor of survival. Surgical undertreatment, as observed in this trial, clearly undermined survival.
Key Words: Gastric cancer • Stomach cancer • Surgery • Lymphadenectomy • Quality • Survival
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Building on observations that concomitant systemic fluorouracil (5-FU) and regional radiotherapy can result in 12% to 20% long-term survival for patients with unresected local-regional disease,3,4 the Southwest Oncology Group (SWOG) and the Gastric Intergroup have, over the last decade, conducted a two-armed, prospective, randomized trial of adjuvant postoperative chemoradiotherapy for patients with gastric adenocarcinoma surgically resected to negative margins. Positive results for this trial, Intergroup 0116 (SWOG 9008), have been published previously.5
Notwithstanding its somewhat vague definition, surgical quality warrants consideration in any postoperative adjuvant trial. Expert surgeons remain divided concerning the optimal extent of surgical treatment for gastric cancer. This was particularly the case in 1990, when Intergroup 0116 (SWOG 9008) was conceived. Speculating that varying the extent of surgical treatment, particularly varying the extent of lymphadenectomy, might affect survival, the impact of adjuvant treatment, or both, detailed surgical information was collected and coded before any survival analysis. In this article, we explore findings with respect to these surgical variables.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
All cases needed to meet the following eligibility criteria: (1) adenocarcinoma of the stomach or gastroesophageal junction (including those tumors involving the distal esophagus), confirmed by central review by SWOG pathologists; (2) complete resection of the gastric neoplasm, defined as resection performed with curative intent and resulting in resection of all tumor with resection margins negative for carcinoma (and confirmed by SWOG review of eligibility checklists, pathology reports, surgical checklists, and operative reports); (3) stage IB or higher, M0 disease (with 1988 third edition American Joint Committee on Cancer [AJCC] definitions)6; (4) good performance status (i.e., 0, 1, or 2 by SWOG definitions); (5) adequate major organ function, defined as creatinine 
1.25 times the institutional upper limits of normal (ULN), normal hemogram, bilirubin 1.5 times the ULN, aspartate aminotransferase 5 times the ULN, and alkaline phosphatase 5 times the ULN; (6) enteral caloric intake >1500 calories per day; (7) registration between days 20 and 41 after surgery, with treatment beginning within 7 working days of registration; and (8) proper informed consent according to institutional and federal guidelines.
The treatment consisted of one cycle of 5-FU (425 mg/m2) and leucovorin (LV) (20 mg/m2) in a daily x 5 regimen followed the next month by 4500 cGy (180 cGy/day) of external beam radiotherapy given with 5-FU/LV (400 and 20 mg/m2) on days 1 through 4 and on the last 3 days of radiation. One month after completion of radiation, two cycles of daily x five 5-FU/LV (425 and 20 mg/m2) were given at monthly intervals.
The trial mandated a negative-margin resection (i.e., simple negative margin with no tumor detected at resection margin with standard hematoxylin and eosin stains). Beyond this, the extent of organ resection was not specified. Similarly, the choice of gastrointestinal reconstruction was left to the operating surgeon. The trial protocol did recommend D2 lymphadenectomy, as defined by the Japanese Research Society for the Study of Gastric Cancer (JRSGC) in the Japanese general rules,7 and appropriate instructions were provided in the protocol. Unfortunately, the postoperative nature of this trial, with registration 20 to 41 days after surgery, generally thwarted communication of such recommendations to treating surgeons. It became apparent with compilation of early descriptive data that the surgical treatment offered to patients in this trial reflected current community practice rather than adherence to any particular recommendations.
Surgical treatment information was captured by using a checklist completed by the operating surgeon or a member of the operative team. In some instances, someone other than a named member of the operative team signed the checklist, and in such instances, the accuracy of the checklist was cross-checked with the operative note (and the pathology report, if sufficiently detailed). If there was conflict, the surgeon’s dictated report took precedence. The use of this surgical checklist both enhanced the quality of surgical data and highlighted ineligible cases, which were excluded. Review of pathology reports alone, which represented the major surgical quality-control method used by SWOG before this trial, would have been insufficient. Pathology information in this trial was abstracted from institutional pathology reports, and this information was checked and supplemented through central pathology review. Institutional pathology reports only rarely referred to the nodal stations listed on the surgical checklist.
The surgical checklist contained items designed to confirm the absence of any extraregional disease and the absence of any residual grossly detectable local-regional disease. Additionally, the checklist captured information concerning the status and the treatment of each of the 16 lymph node stations defined by the JRSGC. These stations were illustrated on the checklist and were also named according to their anatomical locations. Stations 1 to 6 consisted of perigastric node stations; stations 7 to 12 consisted of left gastric, common hepatic, celiac, splenic hilar, splenic artery, and hepatoduodenal nodes, respectively; and stations 13 to 16 were extraregional stations. Disease in these extraregional node stations rendered the case ineligible. The checklist required that the status (i.e., grossly involved or not, dissected or not) of each of the 16 node stations be described.
Before analysis of any survival information, the following surgical information was coded: (1) location of the tumor (i.e., proximal, middle, or distal); (2) type of gastrectomy (i.e., proximal gastrectomy, distal gastrectomy, or total gastrectomy); (3) D level of the lymphadenectomy, as defined by the JRSGC7; and (4) a factor defined by one of the authors (S.A.H.) relating to the likelihood of unresected microscopic nodal disease, termed the Maruyama Index of Unresected Disease (MI). This fourth item warrants further explanation.
The Maruyama computer program basically sorts a 3843-patient database of gastric cancer patients treated by D2 or D2+ lymphadenectomy at the National Cancer Center Hospital in Tokyo. The program sorts cases on the basis of similarity with seven specified variables: age (±5 years), sex, Borrmann type of the tumor, greatest dimension of the tumor as measured on the luminal surface (±2.5 cm), location of the tumor, estimated tumor depth, and histology. On the basis of actual patient experience, the program reports the percentage likelihood of disease at each of the 16 JRSGC-defined lymph node stations around the stomach. This output is designed to be used prospectively, to predict which node stations are at risk for disease in particular cases. The performance characteristics of this program have been evaluated in both Japan and Germany, and it has generally proven to be quite accurate.8,9 In a German series, for example, the accuracy of predictions for disease at the perigastric node stations (1–6) and at other regional node stations (7–12) was 83% and 89%, respectively.9
For the patients in this trial, we used the Maruyama program to predict the percentage likelihood of disease in the regional lymph node stations left undissected by the surgeon. We termed the sum of these predictions the MI. Maruyama program inputs were based on patient demographics, information from pathology and operative reports, information from the surgical checklist, and the report of the pathology reviewer.
Tumor histology, pathology, and microscopic extent of disease were coded on the basis of both institutional pathology reports and the results of central review. To facilitate comparisons with modern series, we report fifth edition AJCC stage designations.10
The SWOG statistical office performed data quality checks on the basis of consistency between surgical data coded by the authors and that independently coded by pathology reviewers. All discrepancies were resolved before any actual analysis.
Follow-up for all cases occurred every 3 months for 2 years, every 6 months for 3 years, and then yearly. Follow-up consisted of physical examination, hemogram, liver function tests, chest x-ray, and computed tomography scans as clinically indicated. Site and date of first recurrence and date of death were recorded.
The Kaplan-Meier method was used to estimate survival curves. Analysis of survival by surgical factor (e.g., by D level, type of gastrectomy, or MI) was performed with the Cox regression model, with adjustment for treatment and other factors as appropriate.11
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
There were 556 patients registered and eligible for this study. Two hundred seventy-five were on the observation arm of surgery alone, and 281 were on the therapy arm. It is important to note that at the time of registration, patients in this study exhibited very good performance status, with 94% of patients being fully active or ambulatory with mild symptoms but capable of light work.
Demographic and baseline characteristics of the patients were well balanced between the groups. The median age was 60 years, 72% were men, 74% were white, 16% were black, and 6% were Asian. Only 15% of the patients in this trial were node negative; 41% had one to three positive nodes, and 43% had four or more positive nodes. Tumor depth was T1 in 6.5% of cases, T2 in 24.5% of cases, T3 (penetrating serosa) in 61%, and T4 (invading adjacent organs; resected) in 7%.
Figure 1 highlights the observation that cases with more advanced-stage disease are overrepresented in this trial, at least compared with what one would predict by applying the same eligibility criteria used for this trial to a relatively contemporaneous cohort of gastrectomy-treated patients accessioned to the National Cancer Data Base from 1985 to 1996.12
|
|
15 nodes were positive), and breakdowns for this stage are not included.
|
|
5. Achieving a low MI with a D0 lymphadenectomy was not impossible, but it required a low-stage tumor with consistently favorable prognostic factors.
|
|
At the onset of this project, we speculated that cases with MI <5 would have higher survival than cases with MI 5. As seen in the overall survival curves in Fig. 3 and the relapse-free survival curves in Fig. 4, this indeed proved to be the case. For the group of 62 cases with MI <5, median overall survival has yet to be reached, whereas median overall survival for the cases with MI 5 is 27 months (hazard ratio, 1.77; P = .005). Median relapse-free survival has yet to be reached for the cases with MI <5, and it is 20 months for the cases with MI 5 (hazard ratio, 1.9; P < .002).
|
|
135). The relationship between MI and stage is complicated. If one hypothetically holds the extent of lymphadenectomy constant, higher stage, more advanced tumors will tend to have a higher likelihood of disease in undissected regional node stations and, therefore, a higher MI. For this reason, adjusting for T stage and N stratum potentially biases the analysis against significance for MI. However, after adjusting for T stage and N stratum, the relationship between MI and outcome remained significant (.036 for overall survival and .008 for relapse-free survival).
The a priori selection of 5 as a cut point for the MI was based on our belief that any adverse survival effect related to 5% chance of residual nodal disease would not be detectable. To further explore the choice of cut point, we used recursive partitioning to assess the optimal cut point for MI, on the basis of a technique that maximizes survival differences between MI categories by using the log-rank test. It is of interest that this technique identified 5 as the optimal cut point whether survival or relapse-free survival was used. A second cut point at 140 was also selected. With these cut points, the estimated hazard ratios for survival, with respect to the <5 category, were 1.6 for the 5 to 140 category and 2.3 for the MI >140 category (P < .001). The corresponding hazard ratios for relapse-free survival are 1.7 and 2.5, respectively (P < .001).
On the basis of a test of interaction between treatment and each surgical variable (i.e., type of gastrectomy, D level, or MI), we were unable to detect differing effects of the adjuvant treatment.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Through the surgical checklist, we captured detailed information concerning the extent of surgical treatment, particularly with respect to lymphadenectomy. We anticipated that the extent of lymphadenectomy in this trial would vary and speculated that such variation might affect survival, the effect of adjuvant treatment, or both. This is the first nationwide US study to actually assess, by using JRSGC definitions, the extent of surgical treatment typically rendered to US gastric cancer patients. The finding that 54% of patients in this trial underwent D0 lymphadenectomy frankly shocked us, particularly because many of the registered cases were treated at academic teaching hospitals. When one of the authors (S.A.H.) presented data documenting this at a 1995 SWOG Surgical Subcommittee meeting, it was hoped that the very high proportion of D0 resections might decrease with progressive accrual. An independent analysis of more cases, subsequently published as a preliminary report,13 described similar findings. The only other US study attempting to monitor the extent of surgical treatment for gastric cancer, an American College of Surgeons Commission on Cancer Patient Care Evaluation Study of patients treated in 1982 and 1987, documented that dissection of common hepatic nodes occurred in 6% of cases, of celiac nodes in 13% of cases, and of splenic nodes in 8% of cases. In this 1982/1987 study, the dissection of perigastric node stations was specifically documented in operative reports, pathology reports, or both and could be unequivocally coded in only 45% of cases.14,15 The approximately 10% incidence of D2 dissections and the 54% incidence of D0 dissections that we observe in this trial is not inconsistent with these findings. This observation tends to refute the view that this trial may have preferentially attracted those patients whose surgical treatment was questionable relative to current community standards in the United States.
Comparison with National Cancer Data Base information as described in Fig. 1 reveals that cases with a more advanced stage are overrepresented in this trial. This is not surprising. We suspect that surgeons and medical oncologists were more likely to consider or recommend participation in this trial—or that patients were more likely to consent—when pathology revealed advanced disease. Also, medical oncologists who participate in clinical trials may have a referral pattern that favors cases with more advanced disease.
As documented by the high proportion of D0 resections, many patients received what many experts consider suboptimal lymphadenectomy. For the past 20 years, a great international debate concerning the optimal extent of lymphadenectomy—D1 versus D2 (a.k.a. R1 versus R2 lymphadenectomy in literature submitted before the end of 1993)—has raged.16–20 Two large, prospective, randomized European lymphadenectomy trials,21,22 as well as others looking at total versus subtotal gastrectomy with or without extended node dissection,23,24 have attempted to settle the issue, albeit perhaps unsuccessfully given the apparent lack of surgical quality control in some trials,22 the documented contamination and noncompliance in others,21,25 and the powerfully adverse effect of D2-related pancreatic-splenic resection.21,22,26,27 Such pancreatic-splenic resection is no longer deemed a necessary component of modern D2 or D2+ lymphadenectomy. Although the aforementioned factors may well have confounded results in both of the larger trials,28 all of these trials have failed to show a convincing survival difference between D1 lymphadenectomy (or D1 subtotal gastrectomy) and more radical surgery. Our study makes it clear, however, that the practical issue in US community hospitals is not D1 versus D2 lymphadenectomy, but, rather, D0 versus D1 lymphadenectomy. We see this as a key finding with important quality implications.
The D level of a lymphadenectomy is based on JRSGC definitions. The JRSGC staging system identifies four levels of node involvement (regional node levels N1 and N2 and extraregional node levels N3 and N4). The node stations comprising these levels vary according to the location of the primary tumor in the stomach. When all JRSGC-defined N1 nodes are cleared but one or more N2 stations are left behind, the lymphadenectomy is designated D1. When all N1 and all N2 stations are cleared, but not all N3 stations, the dissection is designated D2. The status of only one node station can spell the difference between a D1 and a D2 dissection or the difference between a D0 and a D1 dissection. Sometimes, the risk of disease in nodes within that station is negligible, but proper application of the JRSGC definitions requires that a distinction be made. For this reason, the D level designation seems, to us, a fairly blunt descriptive tool. In this study, D-level designation failed to significantly correlate with survival, but power to detect a significant survival effect was low.
MI, however, proved a significant predictor of survival. We should emphasize that MI is a variable primarily under surgical control and that an MI of 0 was possible for every eligible patient in this trial, regardless of stage. In our series, optimal survival was achieved when MI was <5. Survival was poorest when MI was 140. These actual cutoffs for MI may be population dependent (i.e., cutoffs may be different in a population with different stage distribution or different prognostic factors). As noted previously, in this study, cases with extraregional disease were excluded and advanced-stage cases are overrepresented, compared with US norms.
The survival curves for MI <5 seem similar to those for N0 cases in previous trials. Might MI simply reflect stage of disease? Are the majority of the MI <5 cases merely N0 cases? In fact, the median MI for the N0 subset in this study was 95, and the majority of MI <5 cases were not N0.
Might the correlation of MI with survival merely reflect a stage-migration phenomenon?29 Two observations tend to refute this view. First, compared with fifth edition AJCC nodal categories (i.e., 0, 1–6, 7–15, and >15 positive nodes),10 the nodal stratification used in this trial was finer: 0, 1 to 3, and 4 or more positive nodes. The degree of lymphadenectomy-related stage migration should therefore be less than one sees with fifth edition AJCC staging. Second, the magnitude of the stage-migration effect in US patients has been estimated for the fifth edition AJCC staging system. The maximal shift in observed 5-year stage-stratified survival attributable to node-related stage migration is 20%.12 The magnitude of the survival difference for MI <5 versus MI 5 cases in this trial considerably exceeds this. Thus, it is unlikely that stage migration alone can account for our findings.
We see the MI <5 survival curve as reflecting the course of apparent local-regional disease addressed by adequate local-regional surgical treatment. Adjuvant chemoradiotherapy is, itself, a local-regional modality. Both the described MI analysis and the overall results of this trial point to the same conclusion: better local-regional therapy can favorably affect survival in gastric cancer.
Specifically evaluating Maruyama program predictions for the patients in this trial treated by D2 lymphadenectomy, although desirable, was not feasible because of the limitations of institutional pathology reports. Our observation that MI is a significant prognostic factor in this US trial does lend further support to the Maruyama program’s international utility, however (i.e., it seems predictive for all gastric cancer patients and not just Japanese patients). A prior study by Bollschweiler et al.9 documented impressive accuracy in a German population.
Provided that tumor depth can be estimated, Maruyama predictions can be generated on the basis of information commonly available before surgical resection and can be used to plan the extent of lymphadenectomy. The program was designed for such use.8 A 3-fold caution is in order, however: (1) we do not favor any routine sacrifice of uninvolved spleen or pancreas simply to collect splenic nodes, because such resections have been shown to adversely influence mortality and morbidity,21,22,26–28 and pancreas-preserving approaches are available in any case30; (2) dissection of celiac-based node stations demands meticulous technique, and such dissection can be hazardous in inexperienced hands31; and (3) gastric cancer in the United States tends to affect older patients, and comorbid conditions, poor physiologic status, and so on can effectively preclude such dissection. Having said this, however, a relatively user-friendly version of the Maruyama program is now available,32 and we predict that this tool will be used increasingly in the future.
Observations from this trial support something most surgical oncologists hold as a dear truth: inadequate cancer surgery can negatively affect survival. Dr. Blake Cady, in an October 2000 American College of Surgeons Commission on Cancer Oncology Lecture, articulated a more nuanced paradigm: ". . .the therapeutic effect of cancer surgery is akin to that of a drug with a threshold or plateau effect: dose response up to a certain plateau, and then no further therapeutic effect beyond this point, only more complications."33 In this study, we believe we have documented the dose-response portion of Cady’s curve. MI <5 cases clearly have better survival than cases with MI 5, and survival is particularly poor when MI is >140. Notwithstanding trial-related criticisms directed at the aforementioned Dutch21 and Medical Research Council22 trials, these trials may well have delineated a threshold or plateau for the therapeutic effect of gastric cancer surgery. Cady’s paradigm will probably prove to be a durable concept in surgical oncology.
It is interesting and perhaps most important to note that we found no evidence of a significant treatment interaction for pathologic variables (i.e., T or N) or surgical variables (i.e., type of gastrectomy, D level, or MI) in this trial. In other words, we found no significant evidence that the adjuvant treatment in this trial, chemoradiotherapy, failed to work in any particular surgical or pathologic subgroup. Our power to detect such interaction was generally low, however. For example, there were only 18 stage IB cases in each arm of the trial, and routinely administering adjuvant chemoradiotherapy to all such cases has been questioned.34 In countries such as Japan, where more extensive regional nodal treatment is routine and where a substantial proportion of patients present with earlier-stage disease, the value of adjuvant chemoradiotherapy should be explored in the context of a large phase III trial similar to ours. At this point, however, in a typical US gastric cancer patient recovering from a typical US cancer operation, with good performance status and with good enteral intake, adjuvant postoperative chemoradiotherapy improves survival.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The MI did prove to be a powerful predictor of outcome. Survival was much higher for cases with MI <5. Surgically undertreated cases, identified through Maruyama computer program predictions as having a high likelihood of unresected local-regional nodal disease, fared significantly worse. We do not believe that this observation can be ascribed to staging bias or the stage-migration phenomenon. Our observation that MI relates to survival provides indirect evidence supporting the utility of Maruyama program predictions in US patients with gastric cancer.
For US patients who meet this trial’s eligibility criteria, the use of adjuvant chemoradiotherapy for stage IB or higher, M0 adenocarcinoma of the stomach or gastroesophageal junction has apparent value. Our analysis failed to detect a differing treatment effect across various strata of surgical or pathologic subgroups. The study’s power to detect such interaction is low, however, and further scrutiny of its value in settings of low-stage disease and more thorough surgery may be warranted.
|
Acknowledgments |
|---|
Received for publication June 15, 2001. Accepted for publication October 12, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Sasako, T. Sano, S. Yamamoto, Y. Kurokawa, A. Nashimoto, A. Kurita, M. Hiratsuka, T. Tsujinaka, T. Kinoshita, K. Arai, et al. D2 Lymphadenectomy Alone or with Para-aortic Nodal Dissection for Gastric Cancer N. Engl. J. Med., July 31, 2008; 359(5): 453 - 462. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W.-C. Liao, J.-T. Lin, C.-Y. Wu, S.-P. Huang, M.-T. Lin, A. S.-H. Wu, Y.-J. Huang, and M.-S. Wu Serum Interleukin-6 Level but not Genotype Predicts Survival after Resection in Stages II and III Gastric Carcinoma Clin. Cancer Res., January 15, 2008; 14(2): 428 - 434. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Sakuramoto, M. Sasako, T. Yamaguchi, T. Kinoshita, M. Fujii, A. Nashimoto, H. Furukawa, T. Nakajima, Y. Ohashi, H. Imamura, et al. Adjuvant Chemotherapy for Gastric Cancer with S-1, an Oral Fluoropyrimidine N. Engl. J. Med., November 1, 2007; 357(18): 1810 - 1820. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. P. Barbour, N. P. Rizk, M. Gonen, L. Tang, M. S. Bains, V. W. Rusch, D. G. Coit, and M. F. Brennan Lymphadenectomy for Adenocarcinoma of the Gastroesophageal Junction (GEJ): Impact of Adequate Staging on Outcome Ann. Surg. Oncol., February 1, 2007; 14(2): 306 - 316. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. W. Gee and D. W. Rattner Management of Gastroesophageal Tumors Oncologist, February 1, 2007; 12(2): 175 - 185. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Ohtsu, S. Yoshida, and N. Saijo Disparities in Gastric Cancer Chemotherapy Between the East and West J. Clin. Oncol., May 10, 2006; 24(14): 2188 - 2196. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Nitti, J. Wils, J. G. Dos Santos, G. Fountzilas, P. F. Conte, C. Sava, A. Tres, R. C. Coombes, D. Crivellari, A. Marchet, et al. Randomized phase III trials of adjuvant FAMTX or FEMTX compared with surgery alone in resected gastric cancer. A combined analysis of the EORTC GI Group and the ICCG Ann. Onc., February 1, 2006; 17(2): 262 - 269. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. N. Baxter and T. M. Tuttle Inadequacy of Lymph Node Staging in Gastric Cancer Patients: A Population-Based Study Ann. Surg. Oncol., December 1, 2005; 12(12): 981 - 987. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. D. Smith, R. R. Schwarz, and R. E. Schwarz Impact of Total Lymph Node Count on Staging and Survival After Gastrectomy for Gastric Cancer: Data From a Large US-Population Database J. Clin. Oncol., October 1, 2005; 23(28): 7114 - 7124. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Bouche, M. Ychou, P. Burtin, L. Bedenne, M. Ducreux, G. Lebreton, J. Baulieux, B. Nordlinger, C. Martin, J. F. Seitz, et al. Adjuvant chemotherapy with 5-fluorouracil and cisplatin compared with surgery alone for gastric cancer: 7-year results of the FFCD randomized phase III trial (8801) Ann. Onc., September 1, 2005; 16(9): 1488 - 1497. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. P.M. Jansen, H. Boot, M. Verheij, and C. J.H. van de Velde Optimal Locoregional Treatment in Gastric Cancer J. Clin. Oncol., July 10, 2005; 23(20): 4509 - 4517. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.A. Ajani, P.F. Mansfield, N. Janjan, J. Morris, P.W. Pisters, P.M. Lynch, B. Feig, R. Myerson, R. Nivers, D.S. Cohen, et al. Multi-Institutional Trial of Preoperative Chemoradiotherapy in Patients With Potentially Resectable Gastric Carcinoma J. Clin. Oncol., July 15, 2004; 22(14): 2774 - 2780. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.H. Hartgrink, C.J.H. van de Velde, H. Putter, J.J. Bonenkamp, E. Klein Kranenbarg, I. Songun, K. Welvaart, J.H.J.M. van Krieken, S. Meijer, J.T.M. Plukker, et al. Extended Lymph Node Dissection for Gastric Cancer: Who May Benefit? Final Results of the Randomized Dutch Gastric Cancer Group Trial J. Clin. Oncol., June 1, 2004; 22(11): 2069 - 2077. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. P. Khatri and H. O. Douglass Jr D2.5 Dissection for Gastric Carcinoma Arch Surg, June 1, 2004; 139(6): 662 - 669. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.C.M.J. Peeters and C.J.H. van de Velde Improving Treatment Outcome for Gastric Cancer: The Role of Surgery and Adjuvant Therapy J. Clin. Oncol., December 1, 2003; 21(90230): 272s - 273. [Full Text] [PDF]
|
|
|
|
|
|
M. Sasako Principles of Surgical Treatment for Curable Gastric Cancer J. Clin. Oncol., December 1, 2003; 21(90230): 274s - 275. [Full Text] [PDF]
|
|
|
|
|
|
D. Nitti, A. Marchet, M. Olivieri, A. Ambrosi, R. Mencarelli, C. Belluco, and M. Lise Ratio Between Metastatic and Examined Lymph Nodes Is an Independent Prognostic Factor After D2 Resection for Gastric Cancer: Analysis of a Large European Monoinstitutional Experience Ann. Surg. Oncol., November 1, 2003; 10(9): 1077 - 1085. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J.H. van de Velde and K. C.M.J. Peeters The Gastric Cancer Treatment Controversy J. Clin. Oncol., June 15, 2003; 21(12): 2234 - 2236. [Full Text] [PDF]
|
|
| ||||||||||||||||||||
