Treatment Outcomes of Stage II or III Gastric Cancer Treated with Adjuvant Chemotherapy with TS-1 or XELOX after Radical Surgery

Hong Kong J Radiol 2023 Jun;26(2):91-9 | Epub 8 Jun 2023

 

 

Department of Clinical Oncology, Pamela Youde Eastern Hospital, Hong Kong SAR, China

 

  Full Article in PDF

Gastric cancer was the sixth commonest cancer in Hong Kong, accounting for 3.7% of all new cancer cases in 2019.[1] Although the incidence has been gradually declining, compatible with global trends due to efficacious Helicobacter pylori eradication therapy,[2] gastric cancer remains more prevalent in Asian countries than in the West.

 

Clear surgical resection with D2 lymphadenectomy and chemotherapy is considered the standard of care for resectable locoregionally advanced gastric cancer nowadays,[3] and this has been advocated in several international guidelines.[4] [5] Adjuvant chemoradiotherapy (45 Gy over 25 fractions concurrent with 5-fluorouracil and leucovorin) had once been widely adopted, but was later criticised for the inclusion of a high proportion of patients with D1 lymphadenectomy in the study recommending it.[6]

 

The choice of chemotherapy regimen significantly differs among different parts of the world. In European countries, perioperative chemotherapy, such as the combination of epirubicin, cisplatin, and 5-fluorouracil[7] or 5-fluorouracil, leucovorin, oxaliplatin, and docetaxel,[8] is frequently used, whereas in Hong Kong, clinicians tend to use adjuvant chemotherapy as in most Asian countries. The two most commonly used regimens of adjuvant chemotherapy after radical surgery are capecitabine plus oxaliplatin (XELOX) and tegafur/gimeracil/oteracil (TS-1, also known as ‘S-1’). They both demonstrated significant benefits when compared with surgery alone in randomised clinical trials[9] [10] conducted in Asian countries. Despite the two regimens having been widely used, there are no prospective randomised clinical trials directly comparing their efficacy.

 

Regarding the prognostic stratification of patients with resected gastric cancer, several clinical and pathological parameters have long been adopted to predict the recurrence of gastric cancer including age, comorbidities, tumour size, differentiation status, and presence of lymphovascular or perineural invasion.[11] [12] [13] [14] In recent years, the clinical utility of the peripheral neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) as systemic inflammatory markers has been addressed. In relation to cancer prognosis, several meta-analyses showed that elevated NLR and PLR correlated with tumour progression and poor survival in a number of gastrointestinal cancers.[15] [16] However, what the same observation connotes in the adjuvant setting remains uncertain.

 

This retrospective study was conducted with three aims: to compare the efficacy of adjuvant XELOX with TS-1 chemotherapy for patients with stage II or III gastric cancer who received radical surgery in our locality; to investigate important clinical factors that may impact on the risk of disease recurrence; and to explore the prognostic value of NLR and PLR as potentially useful and easily available clinical parameters.

 

Patients who received adjuvant treatment (XELOX: n = 40; TS-1: n = 25) for gastric cancer following radical surgical resection from January 2016 to December 2020 at the Department of Clinical Oncology, Pamela Youde Eastern Hospital, Hong Kong were included in the study. Patients with metastatic disease at presentation (including small-volume peritoneal metastasis) or double primary cancers were excluded. Patients who received adjuvant radiotherapy were also excluded. Relevant clinical and pathological parameters were captured from clinical notes and the Clinical Management System of Hospital Authority.

 

XELOX consists of oral capecitabine (1,000 mg/m² twice daily on days 1-14 of each cycle) plus intravenous oxaliplatin (130 mg/m² on day 1 of each cycle) up to 8 cycles. TS-1 is oral chemotherapy (daily dose according to body surface area [BSA]: patients with BSA <1.25 m² received 80 mg daily, those BSA ranging from ≤1.25 m² to 1.50 m² received 100 mg daily, and those with BSA ≥1.50 m² received 120 mg daily) given for 4 weeks followed by 2 weeks of rest for a total of 9 cycles.

 

In practice, patients of an advanced age, borderline Eastern Cooperative Oncology Group (ECOG) performance status and pre-existing neuropathy would be more likely to be given TS-1, as it is a non–self-financed item under the institution.

 

Doses and schedule modifications were conducted based on patients’ ECOG performance status, organ functions, and toxicities by clinicians’ decisions. Dose reduction of chemotherapy was conducted in a stepwise manner (75%-85% of the initial dose for 1st dose reduction, then 60%-70% for the 2nd dose reduction). The relative total dose intensity (RTDI) is the ratio of the delivered actual dose intensity (ATDI) to the standard planned dose intensity (PTDI) for a chemotherapy regimen, which is calculated as follows:

 

 

Patients were seen by doctors prior to each cycle of chemotherapy, when tolerance of chemotherapy and results of blood tests would be recorded in the Clinical Management System. Patients who had completed the adjuvant chemotherapy would be followed up at an interval of 3 to 6 months. Computed tomography was performed if there was clinical suspicion of disease relapse. Disease relapse was defined as any radiological and/or histological confirmation of recurrence. Elevated tumour markers alone were not considered as relapse without proof of recurrent disease.

 

Statistical analysis was performed using SPSS (Windows version 22; IBM Corp, Armonk [NY], United States). Clinical and pathological data were retrospectively reviewed and analysed by descriptive statistics. Pearson’s Chi squared test was used for testing any significant correlations and differences between groups.

 

Treatment outcomes, including disease-free survival (DFS, the time from surgery to disease relapse) and overall survival (OS, the time from diagnosis of disease to death from any cause) were analysed by the Kaplan-Meier method and the difference between groups were tested with the log-rank test. Different clinical parameters were tested for their impact on DFS and OS by Cox regression analysis.

 

In order to have an accurate assessment of the baseline NLR and PLR of our patients, the complete blood counts right before the administration of first cycle of chemotherapy were recorded. This is to minimise the effect due to postoperative inflammation and chemotherapy on peripheral blood counts.

 

Using all-cause mortality as an endpoint for NLR and PLR, the optimal cut-off values were determined by receiver operating curve analysis as shown in Figure 1. The area under the curve of NLR and PLR was 0.653 and 0.575, respectively. The optimal cut-off values determined by the Youden’s index for NLR and PLR were 1.9 and 169, respectively.

 

Figure 1. Receiver operating curve (ROC) analysis for optimal cut-off values of neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR).

 

Sixty-five patients were identified and included in the analysis. Forty patients received XELOX and 25 received TS-1. The median follow-up time for this study was 33.7 months (range, 6.5-78.9).

 

Patient baseline characteristics are summarised in Table 1. The median age of the entire cohort was 66.0 years. The mean and median age in the XELOX group were 57.7 and 59.0 years respectively, compared to 69.4 and 71.0 years in the TS-1 group. Patients who received TS-1 were significantly older, with 72.0% of them ≥66 years compared to 37.5% in XELOX group (p = 0.007). All included patients had an ECOG performance status score of ≤2. There were significantly more patients with ECOG performance status score ≤1 in the XELOX group (97.5%) than in TS-1 group (60.0%) [p < 0.001].

 

Table 1. Patient baseline characteristics of XELOX and TS-1 groups.

 

The median OS was 38.9 months for the XELOX group and 22.9 months for TS-1 group. The observed OS and DFS in the XELOX group were significantly longer than those in the TS-1 group (Figure 2) [p = 0.037 and 0.012, respectively]. However, it should be interpreted carefully as the baseline patients’ characteristics suggested a bias towards prescribing TS-1 in the older age-group and less fit patients. These factors likely confound the survival analysis.

 

Figure 2. Kaplan-Meier analysis of (a) overall survival (p = 0.037, log-rank test) and (b) disease-free survival (p = 0.012, log-rank test) in all patients with respect to treatment with TS-1 (tegafur/gimeracil/oteracil) or XELOX (capecitabine plus oxaliplatin).

 

Overall survival analysis showed that patients with high NLR (≥1.9) before adjuvant chemotherapy had shorter OS than those with low NLR (<1.9), although the difference was marginally significant (p = 0.051; Figure 3). The same analysis also demonstrated that patients with high PLR (≥169) before adjuvant chemotherapy had significantly shorter OS than those with low PLR (<169) [p = 0.036; Figure 4].

 

Figure 3. High neutrophil-to-lymphocyte ratio (≥1.9) versus low neutrophil-to lymphocyte-ratio (<1.9) on overall survival (p = 0.051).

 

Figure 4. High platelet-to-lymphocyte ratio (≥169) versus low platelet-to-lymphocyte ratio (<169) on overall survival (p = 0.036).

 

In relation to clinical characteristics, patients with elevated NLR correlated with female gender (borderline p value of 0.049) and elevated PLR was associated with more advanced disease (p = 0.012) [Tables 2 and 3].

 

Table 2. Clinical characteristics of patients with high and low neutrophil-to-lymphocyte ratio.

 

Table 3. Clinical characteristics of patients with high and low platelet-to-lymphocyte ratio.

 

As shown in Table 4, univariate Cox regression analysis showed that ECOG performance status score of 2, nodal stage of N2 to N3, and elevated PLR (≥169) were adverse prognostic factors for OS, while ECOG performance status score of 2, nodal stage of N2 to N3, and RTDI of chemotherapy <70% were adverse factors associated with disease relapse.

 

Table 4. Cox regression analysis on multiple clinical and pathological parameters on overall survival.

 

Multivariable Cox regression analysis demonstrated that ECOG performance status score of 2 and nodal stage of N2 to N3 were the two independent adverse prognostic factors for OS (Table 4). For DFS, ECOG performance status score of 2, nodal stage of N2 to N3, and RTDI of chemotherapy <70% were the three independent factors associated with disease relapse (Table 5).

 

Table 5. Cox regression analysis on multiple clinical and pathological parameters on disease-free survival.

 

Our study revealed that the XELOX group had more favourable oncological outcomes (both DFS and OS) than the TS-1 group. However, it should be noted that patients included in the TS-1 group in out centre were older (p = 0.007) and of worse ECOG performance status (p < 0.001). This is largely due to the fact that the institutional guideline recommends TS-1 as the treatment of choice for older patients with anticipated poor tolerance to XELOX and that under such circumstances, only the drug costs of TS-1 would be covered by the institution. There has not been any randomised controlled trial comparing the efficacy of the two regimens. Retrospective studies[17] [18] [19] [20] did not demonstrate statistically significant differences in DFS between adjuvant TS-1 and XELOX (Table 6). In the subgroup analysis, one study[17] demonstrated the use of XELOX in stage II disease was associated with better OS while another study[18] suggested the same but in stage IIIB/C disease only. Apart from XELOX, combination chemotherapy with more than three agents has shown superior treatment outcomes in recent years. Combination of TS-1 with oxaliplatin[21] or docetaxel[22] is considered a preferred option for high-risk patients and is increasingly recognised as a new standard of care.

 

Table 6. Summary of selected retrospective studies comparing XELOX and TS-1 regimens.

 

In our cohort, elevated PLR is associated with inferior OS after curative surgery and there was a similar trend for NLR despite not reaching statistical significance (p = 0.051). NLR and PLR are important parameters indicating systemic inflammation. It is observed that a chronic inflammatory state confers unfavourable oncological outcomes.[23] Several meta-analyses revealed that elevated NLR and PLR were associated with tumour progression and poor survival in gastrointestinal cancers.[15] [16] Microscopically, various inflammatory cytokines and growth factors in the tumour microenvironment are known to dampen hosts’ anti-tumour immune response. In tumour models, inflammatory cytokines such as interleukin 6 (IL-6), IL-8 and IL-11 are associated with chemotherapy resistance in gastric cancer through mechanisms such as inhibition of apoptosis pathways, increasing efflux of chemotherapeutic agents, and evasion of DNA damage.[24] [25] [26] We therefore postulated that in an adjuvant setting, the persistent inflammatory state after curative surgery possibly led to tumour evasion from immunosurveillance and enhanced chemoresistance of micrometastases.[27] NLR and PLR are two readily accessible clinical parameters and may serve as simple prognostic tools in addition to performance status, stage, and age.

 

Our study revealed that the RTDI is an independent prognostic factor for disease recurrence. Inadequate chemotherapy dose intensity is either attributed to excessive dose reduction or failure to complete scheduled cycles within the planned time interval. It is noteworthy that severe adverse events of chemotherapy (≥ Grade 3) have been shown to be quite uncommon (≤6%) with TS-1 in a large-scale clinical trial,[28] although these patients were generally frailer and older. For elderly patients who may be more vulnerable to chemotherapy toxicity, proper geriatric assessments (such as comorbidity and frailty indices) are needed, as biological age is not a reliable indicator for chemotherapy dose adjustment, and an adaptive dose optimisation approach is recommended based on patients’ tolerance of each cycle.

 

This study has several limitations. First, it is only a single-centre retrospective study in which the small sample size limits its statistical power. Second, there is imbalance between the baseline characteristics of the two groups of patients. Similar to the Korean studies,[17] [18] [19] [20] patients in the TS-1 group were generally older and had a worse ECOG performance status. There is a tendency for clinicians to prescribe a more conservative chemotherapy dosage in this group of patients, which may explain the lower dose intensity of TS-1 than XELOX. Propensity score matching should be performed in a larger cohort to reduce the bias due to these confounding variables. Third, a much large sample size is needed to further evaluate the prognostic power of NLR and PLR on OS and DFS in the adjuvant setting. In our cohort, high PLR appeared to correlate with patients with more advanced disease (stage III), which is an important confounding factor.

 

In conclusion, we compared the OS and DFS between adjuvant XELOX and TS-1 in our local gastric cancer patients. Clinical outcomes were statistically better with XELOX- than TS-1–treated patients. However, the results should be viewed with caution because of the limited sample size and obvious imbalance in baseline characteristics. ECOG performance status score of 2 and advanced nodal stage of N2 to N3 are independent adverse prognostic factors associated with poor OS and a higher rate of disease recurrence. NLR and PLR are readily available markers that may be further explored as prognostic markers for gastric cancer after radical surgery. We also speculated that the RTDI of chemotherapy of <70% might affect the risk of disease relapse.