br The consensus for administering adjuvant chemotherapy br
The consensus for administering adjuvant chemotherapy
for gastric cancer was mainly reached through reports from meta-analysis32,33 before 2007. The Adjuvant
Figure 2 Overall survival curves according to the prognostic model.
LNR-based prognostic model for gastric cancer 91
Chemotherapy Trial of TS-1 for Gastric Cancer (ACTS-GC) was the first well-designed phase III study that showed the significant benefit of adjuvant chemotherapy for Japanese patients with stage II and III gastric cancer who had un-dergone D2 surgery.19 Subsequently, another positive result of a phase III study examining the benefit of capecitabine and oxaliplatin combination treatment (CLASSIC study) in patients with stage II to III gastric cancer after D2 resection was reported in 2012.20 Based on these two positive phase
III studies, the clinical benefit of adjuvant chemotherapy was agreed upon. Unfortunately, the ACTS-GC trial
included patients categorized using the Japanese Gastric Cancer staging system,34 which categorized N-classification according to anatomic location of metastatic lymph nodes rather than the numbers of lymph nodes metastases like that in the AJCC staging system. In the CLASSIC trial, pa-tients with metastasis to more than 15 regional lymph nodes were excluded. As such, the clinical value of adju-vant chemotherapy in patients with more than 15 meta-static lymph nodes was uncertain. In real-world practice, our study identified adjuvant chemotherapy as an inde-pendent prognostic factor in this disease. D2 surgery alone was inadequate to achieve long-term survival for gastric cancer patients with metastasis to more than 15 regional lymph nodes, as the median OS and DFS was only 10.9 and 7.8 months, respectively. Based on our study, adjuvant chemotherapy should be mandatory after radical surgery for gastric cancer patients with metastasis to more than 15 regional lymph nodes.
Our model also identified T-classification, CEA level, and ECOG PS to be independent prognostic factors for advanced
gastric cancer patients who underwent curative surgery. High T-classification35 and elevated preoperative CEA level36 reflected advanced tumor status, whereas ECOG PS indi-cated the patient’s Latrunculin A for radical surgery.37 Patients with better ECOG PS may have speedy recovery after sur-
gery, can receive adjuvant chemotherapy on schedule, and tolerate the adverse events of adjuvant chemotherapy.38 Using these well-documented clinical variables in conjunc-tion with LNR and adjuvant treatment, we developed a prognostic model and demonstrated its superior perfor-mance to using single variables in terms of monotonicity, homogeneity, and high capability of discrimination.
In the current study, the survival outcome can be further stratified into 3 different prognostic groups based on clini-copathologic variables of each patient. To our knowledge, this is the first study to construct a prognostic model for predicting survival outcome after D2 surgery in gastric can-cer patients with metastasis to more than 15 regional lymph nodes. However, this study has several limitations. First, a selection bias might exist because of the retrospective na-ture of the study. In our series, a high ratio of total gas-trectomy (46%) was noted, which was due to the large size and bulk of the tumors (70.5% of patients had T-4 classifi-cation in our cohort) found preoperatively rather than a high prevalence of several body tumors. To achieve free and adequate length of gross sectional margin, a higher number of total gastrectomy was performed compared with other earlier stage gastric cancer patients. Even with aggressive surgical resection, the margin-positive rate is as high as 22.3% in stage IIIc gastric cancer patients, which is much higher than the value in our usual practice in less locally advanced stage patients.36 Second, we included adjuvant treatment as a prognostic factor; as such there was selection bias regarding which patients were offered the treatment. The decision of receiving adjuvant chemotherapy and which regimens to administer might vary and depend on the pa-tient’s postoperative fitness and economic status, avail-ability of adjuvant medications, and physician’s preference. However, adjuvant chemotherapy still affected survival after adjustment for other confounding factors in the multivariate model. Third, because of the lack of a standard duration for image follow-up in all of the patients, the actual DFS might be overestimated in our study. Finally, and most importantly, even though our prognostic model showed good accuracy to predict postoperative survival outcomes, all data were available from a single medical center and internally validated. An external validation of the model is essential before it can be widely used.