Clinical stage after preoperative chemoradiation is a better predictor of patient outcome than the baseline stage for localized gastric cancer

Authors


Abstract

BACKGROUND.

For patients with localized gastric cancer (LGC) who are receiving preoperative chemoradiation (CTRT), the postsurgical pathologic stage predicts overall survival (OS) better than the baseline stage. The authors hypothesized that presurgical (postCTRT) stage would also correlate better with patient outcome than the baseline stage.

METHODS.

The authors analyzed 74 LGC patients treated with preoperative CTRT receiving similar treatment. Patients were staged with baseline endoscopic ultrasonography (EUS) and laparoscopy. Patients received induction chemotherapy, then CTRT (45 Gy), and had an attempted surgery. After CTRT, patients had complete preoperative staging including EUS in 35 patients.

RESULTS.

Thirty-five had all 3 sets of staging, baseline, presurgical, and postsurgical. Baseline stage did not associate with OS (P = .16) nor disease-free survival (DFS; P = .13). However, presurgical stage was associated with OS (P = .01), and DFS (P = .05). OS was also associated with postsurgical stage and was longer for stages 0 and I than for stages III and IV (P = .01 and .04, respectively). Similarly, DFS was longer in postsurgical pathologic stages 0, I, and II than in stage III or IV (P < .001, <.001, and <.01, respectively). Baseline staging did not correlate with the proportion of patients alive at 4 years; however, presurgical staging did. Patients with stage I or II survived longer than those with stage III or IV (81% vs 25%; P < .01).

CONCLUSIONS.

Heterogeneity in clinical biology of LGC is best reflected after CTRT in presurgical and postsurgical pathologic stages rather than by the baseline stage. Correlation of outcome with presurgical staging may facilitate strategies to individualize therapy for LGC. Cancer 2007. © 2007 American Cancer Society.

Although its incidence is decreasing, gastric cancer is rarely detected early in the western hemisphere, thus accounting for its poor prognosis. In approximately half of newly diagnosed cases, the carcinoma is advanced beyond its original locoregional boundaries.1 Surgery remains the primary therapy for localized gastric cancer.2–6 Nevertheless, the rate of curative R0 resection is achieved in <50% in patients treated with surgery alone.7 Distant metastases and locoregional recurrence can also be a significant problem, the latter reaching as high as 54% after surgical resection in some series.8, 9 However, for patients with localized gastric carcinoma, the strategy of preoperative chemoradiation therapy (CTRT) is emerging as a promising mode of treatment.10–12 In a recently published study, we reported that postsurgical pathologic stage as determined by the American Joint Commission on Cancer (AJCC) is a better predictor of overall survival (OS) and disease-free survival (DFS) than baseline clinical stage (before CTRT) because of changes in cancer stage brought about by treatment.13 However, it is still undetermined whether presurgical stage (post-CTRT) is also a better predictor of OS than baseline clinical stage. Such a finding would have immense implications, perhaps allowing patients who may have poor OS and/or DFS, as determined by presurgical stage, to be spared unnecessary surgery. We hypothesized that presurgical stage, and not the baseline clinical stage, would be the best prognosticator of OS and DFS. To test this hypothesis, we examined OS and DFS in relation to baseline, presurgical, and postsurgical pathologic stage by using data from 2 prospectively conducted preoperative CTRT trials.10, 11

MATERIALS AND METHODS

This study examined patient populations from 2 prospective preoperative CTRT trials.10, 11 Both protocols were similar in terms of patient selection and therapeutic strategy. The chemotherapy combination administered before and during CTRT varied only to some extent; however, this slight variation in chemotherapy in 2 trials is not considered fundamentally different.

Patient Selection and Evaluation

Patients with localized, histologically confirmed gastric or gastroesophageal adenocarcinoma were eligible for either trial. The bulk of cancer was in the stomach, although the gastroesophageal junction may have been involved. As part of the staging workup, patients had a chest radiograph, computed tomography (CT) scan of the abdomen (pelvis and chest were scanned if needed [CT pelvis was often done in women and CT chest in patients with gastroesophageal junction involvement]), upper gastrointestinal barium radiographs, esophagogastroduodenoscopy with endoscopic ultrasonography (EUS), electrocardiogram, SMA-12 serum chemistry panel, complete blood count, and measurement of electrolyte and carcinoembryonic antigen levels. Patients with T2 to T3 carcinoma with (N+) or without lymph node involvement (N0) and patients with T1N1 carcinoma were eligible to be in these clinical trials. T-classification was determined by EUS only. Laparoscopic staging and J-tube placement were performed in all patients. Nutritional counseling was provided as needed. All patients provided written informed consent after the above evaluation was completed and after they had recovered from laparoscopic staging. A multidisciplinary evaluation was also performed before patients participated in the clinical trials.

Patients ineligible for inclusion in the clinical trials were those with T4, M1, or T1N0 carcinoma, peritoneal carcinomatosis (gross or microscopic), or uncontrolled medical conditions (eg, diabetes, hypertension, NY Heart Association class III or IV, or psychiatric illness). Patients who could not comprehend or comply with the study requirements were also ineligible.

Design of Clinical Trials

Patients were to receive 1 or 2 cycles of induction chemotherapy and then CTRT. The second induction chemotherapy cycle was given if the cancer had not progressed and was skipped if the cancer did progress. Patients were taken off protocol if distant metastases developed. Clinical responses were judged by using radiographic or endoscopic techniques.

If a patient subsequently underwent an R0 resection, no further therapy was given. In the case of R1 resection (microscopically positive margin), R2 resection (only partial resection), or discovery of M1 carcinoma, the patient was given palliative care (systemic chemotherapy or only best supportive care was offered depending on the circumstances). Each patient was assessed at 3, 6, and 12 months, and then every 6 months for 5 years or until the patient's death.

Step 1: Chemotherapy

In both trials,10, 11 patients received 5-fluorouracil (5-FU)-based chemotherapy. In 1 trial,10 patients received 5-FU, cisplatin, and leucovorin, and in the second trial,11 patients received 5-FU, cisplatin, and paclitaxel. If the cancer had not progressed, then the second cycle was to be repeated 28 days after the first cycle began. Complete blood counts were performed weekly. Serum chemistries were monitored before each course.

Step 2: Chemoradiotherapy

In both trials,10, 11 patients received 5-FU with radiation therapy; however, in the second trial,11 patients also received weekly paclitaxel. The interval between Day 1 of the last induction chemotherapy cycle and the first day of CTRT was ≥28 days. Radiation fields included the entire stomach, perigastric extension (if present), and draining lymph nodes (gastric, celiac, porta hepatis, gastroduodenal, splenic-suprapancreatic, and retropancreaticoduodenal). For lesions involving the cardia or gastroesophageal junction, a 5-cm margin of esophagus was included, and for distal lesions at or near the gastroduodenal junction, a5-cm margin of duodenum was included. Esophagoscopy, barium-swallow radiographs, and CT scan of the chest and abdomen were used to determine the maximal extent of disease relative to the primary tumor and the nodal groups. Idealized fields were modified as needed to shield at least two-thirds of 1 kidney. For proximal lesions, cardiac shielding was recommended, along with evaluation of lateral fields for a component of treatment. Linear accelerators delivered a dose of 45 Gy (25 fractions of 1.8 Gy) over 5 weeks by using either 15-MV or 18-MV photons or a 3-dimensional conformal radiotherapy technique. Typically, anterior and posterior fields were sometimes positioned slightly obliquely to avoid the spinal cord and the right kidney.

Step 3: Surgery

The type of surgery performed depended on the location and extent of the primary cancer. The cancer was resected along with a luminal gastric margin of ≥5 cm when feasible. Also when feasible, a 2-cm duodenal margin was obtained for distal cancers, and a 3-cm esophageal margin was obtained for proximal cancers. In both cases, frozen-section confirmation of a negative margin was sought. For distal cancers, a subtotal gastrectomy was considered adequate; total gastrectomy was at the discretion of the surgeon. For proximal cancers, total gastrectomy or total esophagogastrectomy was performed. An en bloc resection of adjacent organs was performed when their involvement was questionable. The spleen was preserved when possible. An attempt was made to perform a D2-type nodal dissection. During surgery, a J-tube was placed for temporary nutritional support.

Tumor Response

Upper gastrointestinal barium radiographs were taken after each 28-day cycle of induction chemotherapy just before surgery and just before esophagogastroduodenoscopy. Connecticut scan of the abdomen, chest radiography, and all blood tests were repeated before surgery. Endoscopic ultrasonography just before surgery was optional and was performed in only 35 patients. Previously described criteria for response evaluation were used.11 Briefly, pathologic complete response (pathCR) was defined as an absence of carcinoma cells in the primary site, and pathologic partial response (pathPR) was defined as <10% residual carcinoma cells in the specimen.

Staging

Baseline and presurgical (postCTRT) staging were performed by endoscopic ultrasound (EUS) and radiologic imaging studies. Endoscopy and radiologic imaging studies were repeated a few days before surgery. Finally, post-treatment pathologic stage was determined by examination of the resected specimens.

Statistical Analyses

With data from the 2 clinical trials, 2 sets of analyses were performed. First, chi-square and Fisher exact test were used to examine whether baseline and/or presurgical N-classification, T-classification, and/or overall stage were associated with incidence of death and/or relapse (locoregional recurrence or distant metastasis). These tests were also used to determine associations between baseline and/or presurgical stage and incidence of pathCR or R0 resection. Log-rank tests were used to compare overall survival and disease-free survival among various baseline, presurgical, and postsurgical stages. Kaplan-Meier curves were generated and plotted for the patients stratified by baseline and presurgical T, N, and EUS stage. All statistical tests were 2-sided, and significance was set at the 0.05 level. Overall survival was calculated from the date of first treatment on the protocol, and disease-free survival was calculated from the date of surgery.

RESULTS

Sample Characteristics

From the 74 total patients who were enrolled in the 2 clinical studies, baseline and presurgical EUS staging were determined for 35 patients. Thus, results are based on analysis of 35 patients. The median age was 58 years (range, 30–69 years), and all patients had an Eastern Cooperative Oncology Group (ECOG) performance score of 1. Baseline and presurgical characteristics of these patients are shown in Table 1. Most patients were male (n = 24, 69%), and the cancers of most patients were poorly differentiated (n = 29, 83%). Location of tumor was evenly distributed between distal and proximal, (51% and 49%, respectively). Most cancers were classified as baseline EUS stage III (n = 17, 49%), T3 (n = 30, 86%), and/or N1 (n = 21, 60%). Regarding presurgical staging, most cancers were classified as stage II (n = 15, 43%), T3 (n = 20, 57%), and/or N0 (n = 24, 69%). Most cancers were down-staged after CTRT (n = 19; 54%).

Table 1. Patient Characteristics at Baseline (N = 35)
CharacteristicPatients
No.%
  • *

    None of the cancers was classified as baseline T1.

  • All stage I patients had N1 cancer by baseline EUS.

  • None of the cancers was classified as Presurgery stage IV.

  • §

    Down-staged from baseline stage to presurgery stage.

Sex
 Men2469
 Women1131
Pathologic differentiation
 Well13
 Moderate411
 Poor2983
 Unknown13
Tumor location
 Distal1851
 Proximal1749
Baseline T classification*
 T2514
 T33086
Baseline N classification
 N01440
 N12160
Baseline EUS stage
 Stage I411
 Stage II1337
 Stage III1749
 Stage IV13
Presurgery T classification
 T013
 T1/T21440
 T32057
Presurgery N classification
 N02469
 N11131
Presurgery EUS stage
 Stage 013
 Stage I1131
 Stage II1543
 Stage III823
Down-staged§
 Yes1954
 No1646

Baseline Clinical Stage Versus PresurgicalStage as Predictor of OS

Among the 35 patients, the correlation between baseline EUS stage and overall survival (OS) was not statistically significant (P = .16; Fig. 1a). However, presurgical stage correlated with OS. Patients with lower presurgical stage (stages I-II) survived significantly longer than patients with presurgical stages III-IV (median time not reached vs 15.2 months; P = .01; Fig. 1b). As in our previous study, postsurgical pathologic stage continued to be associated with OS. Patients with postsurgical stages 0 and I (median survival time not reached for each stage group) survived significantly longer than patients with postsurgical stages III and IV (P = .01 and 0.04, respectively). Patients with postsurgical stage II (median time not reached) also tended to survive longer than patients with postsurgical stages III-IV, although the association was not statistically significant (P = .09; Fig. 1c).

Figure 1.

Kaplan-Meier curves comparing overall survival (OS) among patients stratified by (a) baseline stage, (b) presurgical stage, and (c) postsurgical stage. There was no significant difference between postsurgical stages 0 versus I, 0 versus II, and I versus II. P-values were obtained from log-rank tests.

There was no correlation between baseline EUS stage and proportion of patients alive at the median follow-up of 38+ months (stages I and II = 76% vs stages III and IV = 55%, P = .12); however, a greater proportion (81%) of patients with presurgical stages I and II cancer than patients with presurgical stages III and IV cancer (25%) were alive at follow-up (P < .01; Table 2). There was no association between OS baseline N-classification (N0 = 43% vs N1 = 57%, P = .24); however, there was a significant association between OS and presurgical N-classification (N0 = 83% vs N1 = 17%, P = .02; Table 3). There was no association between OS and T-classification at baseline T-classification (T2 = 17% vs T3 = 83%, P = .33) or presurgical (T1–T2 = 52% vs T3 = 49%, P = .07; Table 4).

Table 2. Correlation Between Overall Survival (OS) and Presurgery Stage Versus Baseline Stage
 Patients alive at follow-up median 38+ mo; No. (%)P*
  • *

    Two-sided Fisher exact test.

  • Percentage of patients with similar stage subcategory.

  • None of the cancers was classified as presurgery stage IV.

Baseline stage .12
 I–II13 (76)
 III–IV10 (55)
Presurgery stage <.01
 I–II21 (81)
 III2 (25)
Table 3. Comparison of Correlation Between Overall Survival (OS) and Presurgery N Classification Versus Baseline N Classification
 Patients alive at follow-up, No. (%)P*
  • *

    Two-sided Fisher exact test.

  • Percentage of patients with similar stage subcategory.

Baseline N .24
 N010 (43)
 N113 (57)
Presurgery N .02
 N019 (83)
 N14 (17)
Table 4. Comparison of Correlation Between Overall Survival (OS) and Presurgery T Classification versus Baseline T Classification
 Patients alive at follow-up, median, 38+ mo No. (%)*P
  • *

    Percentage of patients with similar stage subcategory.

  • Two-sided Fisher exact test.

  • None of the cancers were classified as baseline T1.

Baseline T .33
 T24 (17)
 T319 (83)
Presurgery T .07
 T1–T212 (52)
 T311 (49)

Baseline Clinical Stage Versus Presurgical Stage as Predictor of DFS

Among the 35 patients, the association between baseline EUS stage and disease-free survival (DFS) was not statistically significant (P = .13; Fig. 2a). However, presurgical stage was associated with DFS. Patients with lower presurgical stage (stages I and II) survived significantly longer without relapse than patients with presurgical stages III and IV (median time not reached vs 5.8 months, P = .05; Fig. 2b). Postsurgical pathologic stage followed a similar pattern. Patients with postsurgical pathologic stages 0, I, and II (median survival time not reached for each stage group) survived significantly longer than patients with postsurgical stages III and IV (P < .001, <.001, and <.01, respectively; Fig. 2c).

Figure 2.

Kaplan-Meier curves comparing disease-free survival (DFS) among patients stratified by (a) baseline stage, (b) presurgical stage, and (c) postsurgical stage. There was no significant difference between postsurgical stages 0 versus I, 0 versus II, and I versus II. P-values were obtained from log-rank tests.

The association was not significant between either baseline or presurgical stage and proportion of patients without relapse at follow-up (P = .09 and 0.07, respectively). The same was true for baseline and presurgical N-classification (P = .13 and .14, respectively) and T-classification (P = .34 and .11, respectively).

Baseline Clinical Stage Versus Presurgical Pathologic Stage as Predictor of pathCR or R0

There was no significant correlation between either baseline or presurgical stage and proportion of patients with pathologic complete response (pathCR) (P = .17 and .34, respectively) and also with R0 resection (P = .11 and .16, respectively).

DISCUSSION

Locally advanced gastric cancer poses a challenging situation and yet a potential opportunity because it is a potentially curable condition. Most patients have either stage II or III cancer by baseline clinical stage. Even with the advent of the value of preoperative and postoperative chemotherapy compared with surgery,14 surgery continues to be used as primary therapy, and adjuvant postoperative chemoradiation is given to selected patients. Clinical biology of localized gastric cancer suggests that even if cancers are grouped in 1 particular stage (for example, stage II), not all can be expected to behave similarly, and outcomes can vary considerably. The response to therapy, DFS, OS, metastatic potential, and cure rates are different. Clearly, we do not have the tools to understand how to individualize therapy for our patients. Until we have a much deeper understanding of the molecular biology of localized gastric cancer, it may be difficult to individualize therapy. Until such time, we must explore the possible benefits of staging at different junctures of therapy.

Baseline clinical staging remains an important step in the evaluation of patients with localized gastric cancer. However, it does not seem to predict overall survival and disease-free survival. In some patients, the true stage of gastric cancer is altered by therapy (particularly chemoradiation therapy resulting in pathologic complete response or only microscopic residual cancer). In other patients who are receiving preoperative chemoradiation, the cancer can be nearly totally chemoradiation resistant.15 Chemoradioresistant cancer often has high metastatic potential, and the life span of such patients can be very short.16 Identification of a chemoradioresistant cancer before surgery (by presurgical staging, if reliable) could spare these patients the morbidity of surgery. This issue has not been fully discussed in literature. Although our data are based on a small number of patients and quite preliminary, they are also provocative. A confirmation of our observations is needed and should be followed by a validation study before a change in the clinical decision process is implemented.

In this study of 35 patients who had all 3 types of sophisticated staging, baseline, presurgical, and systematic evaluation of surgical pathology,16 our data suggest that preoperative staging correlates significantly with overall survival and disease-free survival. We already know that postoperative surgical staging definitely correlates with overall survival and disease-free survival.

In conclusion, presurgical stage, including EUS, appears to correlate better with overall survival and disease-free survival than does baseline stage. Similarly, presurgical stage correlates better with postsurgical pathologic stage than it does with the baseline stage. Further refinement of presurgical staging by EUS-directed fine-needle aspiration of suspicious nodes or positron emission tomography may be warranted. In addition to confirmation of our data, future prospects in individualization of therapy for this group of patients will require detailed analyses of tumor biology and patients' genetics.

Acknowledgements

We are grateful to Mr. Simon Lunagomez for his assistance in data analysis.

Ancillary