Comparison of Efficacy of Aspirin Plus EOX vs. EOX Alone in Patients with Locally Advanced and Metastatic Gastric Cancer: a Randomized Clinical Trial

The role of aspirin in cancer prevention has been well defined; the last decade revealed its therapeutic role with improved efficacy when aspirin was added to capecitabine in heavily pre-treated metastatic colorectal cancer. Aspirin affects tumour growth through the PI3K pathway, which regulates apoptosis and autophagy. The objective was to compare the efficacy of aspirin plus epirubicin, oxaliplatin, capecitabine (EOX) chemotherapy versus EOX alone in locally advanced and metastatic gastric cancer. All patients with advanced gastric cancer reporting to the Department of Medical oncology between March 2017 and May 2019 were screened for study eligibility. They were randomly assigned to standard EOX with or without aspirin at a daily dose of 150 mg. Tumour measurements were assessed at baseline and after 3–4 cycles by an independent blinded radiologist according to RECIST criteria 1.1. Toxicity profiles were recorded as per CTCAE v 4.03. Per-protocol group was identified as 70 patients. The primary endpoint was overall response rates in the per-protocol group (defined as patients who received a minimum of 3 cycles and had an evaluable response after randomization). The secondary endpoints included toxicity analysis, progression-free survival, and overall survival. Ninety-five patients who fulfilled the study inclusion and exclusion criteria were randomized to group 1 EOX (50) or group 2 EOX plus aspirin (45). Seventy patients were included for the per-protocol analysis. The overall response rate in group 1 was 27% compared to group 2, which was 42%, P = 0.176. The median duration of follow was 29 (18.56–39.45) months. The median overall survival (n = 95) of group 1 versus group 2 was 11 (8.58–13.42) months and 10 (6.86–13.14) months, respectively, P = 0.90. There was no statistical significance in the overall survival per-protocol analysis (n = 70) between group one 12 (8.75–15.25) months versus group two 12 (6.21–17.79) months, P = 0.50. There was no improvement in the response rates, progression-free survival, and overall survival on adding aspirin to EOX chemotherapy in locally advanced and metastatic gastric cancer in an unselected population. A further role of PI3K mutation as a biomarker needs to be evaluated in this setting.


Introduction
Gastric cancer ranks 5th in incidence among all sites and is the 3rd most common cause of cancer death worldwide. It is responsible for an estimated 1 out of every 12 cancer deaths in the world. Gastric cancer continues to be a major cause of cancer-related mortality despite multimodality treatment [1].
Chemotherapy is the mainstay of treatment in locally advanced inoperable and metastatic gastric cancer. Epirubicin, oxaliplatin, cisplatin, 5-fluorouracil, capecitabine, docetaxel, and irinotecan are active drugs in combination. A phase 3 trial showed the best median overall survival of 11.2 months with EOX compared to ECF, ECX, and EOF [2]. However, despite these advances, the outcome for patients with advanced gastric or gastric and gastro-oesophageal junction is poor. Targeted drugs like Her2 neu inhibitors, VEGF inhibitors, and immunotherapy have recently been approved with marginal improvement in survival. Liao et al. found that colorectal cancer patients who used aspirin regularly had a better overall survival if they had PI3KCA mutation [3]. Giampiere et al. investigated the role of aspirin in addition to capecitabine in metastatic colorectal cancer who were heavily pre-treated in a small non-randomized study and found a significant difference in the disease control rate (80% vs. 30%), P = 0.000377. The median progression-free survival in patients treated with aspirin was 6.5 months vs. 3.3 months (HR-0.48; CI (0.30-0.0042). The study also showed a significant median overall survival benefit in patients with aspirin (14.7 months vs. 8.7 months), P = 0.0023 [4].
The phosphatidylinositol 3-kinase (PI3K) signaling pathway plays an important role in carcinogenesis. Patients with colorectal malignancy treated with aspirin having mutated PIK3CA had significantly better survival than the wild-type PIK3CA [3]. PIK3CA mutation is one of the multiple mutations associated with gastric cancer. A study by Polom et al. on PIK3 CA mutation in gastric cancer showed a frequency of 8%, ranging from 4 to 13%, and had a strong association with MSI molecular subgroup [5]. There is limited data available on the addition of aspirin to standard chemotherapy in patients with gastric cancer.

Study Design and Participants
The study was a prospective open-label phase II randomized controlled trial of adding aspirin to epirubicin, oxaliplatin, capecitabine (EOX) in locally advanced and metastatic gastric cancer from November 2017 to May 2019 with a minimum follow-up period of 6 months from the date of randomization. The patients included in the study were inoperable locally advanced and metastatic gastric cancer of 18 to 70 years with a performance status (ECOG-PS ≤ 2) who could take oral medications. Patients with significant organ dysfunction such as those with ejection fraction less than 50%, serum creatinine > 2 mg/dl, serum bilirubin > 1.5 mg/ dl, AST /ALT > 3 X ULN, ALP > 5 X ULN, thrombocytopenia less than 100,000/μl, previous history of aspirin intake, and aspirin hypersensitivity were excluded from the study.
The primary objective of the study was to compare the efficacy of the combination of aspirin plus epirubicin, oxaliplatin, capecitabine (EOX) versus EOX alone in terms of response rates in patients with locally advanced and metastatic inoperable gastric cancer. The secondary objectives were to study the adverse effects, progression-free survival, and overall survival among both the groups. The primary endpoint was overall response rates in the per-protocol group (defined as patients who received a minimum of 3 cycles and had an evaluable response after randomization). The secondary endpoints included toxicity analysis, progression-free survival, and overall survival.

Sample Size Calculation and Randomization
The sample size was estimated using the statistical formula to compare 2 independent proportions. The minimum expected difference in the proportion of patients concerning response rates between the two groups was 0.3 (30%). The sample size was estimated at a 5% level of significance and 80% power. Accounting for a dropout rate of 10%, a final sample size of 110 was planned, with 55 in each arm.
Patients were randomly assigned to a 1:1 ratio to receive standard EOX (group 1) or EOX with aspirin (group 2). This was an open-label study. Randomization was done using a computer-generated fixed block size. Allocation concealment was done using sequentially numbered, opaque sealed envelopes.

Statistical Analysis
The continuous data were expressed as mean with S.D. or median with range. The categorical data comparison was carried out using the chi-square or Fisher's exact test. The continuous data between groups was analysed by independent t-test and Mann-Whitney U test; Kaplan-Meier product estimates were used to analyse survival functions, and the log-rank test was used to compare survival function between the groups. Covariates associated with survival function were analyzed by using Cox proportional hazards model. Statistical analysis was carried out using IBM SPSS Statistics for Windows, version 19 (IBM Corp., Armonk, NY, USA).

Baseline Characteristics
Patient data were collected using a standardized case record proforma thorough medical history, detailed physical examination, a complete blood count, routine biochemistry (renal and liver function tests), viral markers (HIV I and II, anti-HCV, HBsAg), 2D ECHO, ECG, upper G.I. endoscopy and biopsy, and baseline imaging by contrast-enhanced CT scan thorax, abdomen, and pelvis.

Chemotherapy
Group1 -EOX in a standard schedule (epirubicin 50 mg/ m 2 iv bolus, D1 oxaliplatin 130 mg/m 2 iv infusion over 2 h D1 and capecitabine 625 mg/m 2 BD orally days 1-21 q 3 weekly for 8 cycles). Group 2 patients received EOX as mentioned above plus aspirin 150 mg O.D. daily until disease progression [6]. The new cycle was started when the absolute neutrophil count was greater than 1500/mm 3 and platelet count greater than 100,000/mm 3 . Aspirin dose was withheld if the platelet counts were less than 50,000/mm 3 .

Methodology and Definition of Endpoints
Tumour measurements were performed at baseline and after 3-4 cycles of chemotherapy and then every 3 months until progression by the radiologist who was blinded to treatment arms according to RECIST 1.1 criteria [7]. Clinical benefit rate (CBR) or disease control rate (DCR) was taken as the percentage of patients in each arm who had a complete response (C.R.), partial response (P.R.), or stable disease (S.D.). Overall response rate (ORR) was taken as that percentage of patients with a complete or partial response. Progression-free survival (PFS) was defined as the time from randomization to first progression or death. Overall survival (O.S.) was defined as the time from randomization to death due to any cause or date of last follow-up (censored) if the patient was lost to follow-up. The follow-up status was updated (clinical visit or telephonic call) as on December 2021.
The study was approved by the Institute's ethical committee (JIP/IEC/2016/1069). Written informed consent was obtained from each participant in their vernacular language (Tamil, English). The trial was registered in CTRI (Clinical trial registry of India) (CTRI no. CTRI/2017/11/010651).

Results
A total of 280 patients with gastric cancer were screened for eligibility, of which 185 were excluded. A total of 95 patients were randomized to two groups, viz., group 1 (50) and group 2 (45) patients. Seventy patients were found to be evaluable for the per-protocol analysis as defined as completion of a minimum of 3 cycles of chemotherapy and an evaluable response. The consort diagram is shown in Fig. 1 The baseline characteristics in both the groups were matched in the intention to treat and per-protocol analysis is shown in Table 1. The median age of the entire cohort was 53 (21-70) years, with 59% metastatic. The most common primary site was antropyloric, followed by gastrooesophageal junction with the liver and peritoneum as the commonest sites for metastasis. Intestinal histology was seen in 52% of the patients. The performance status (ECOG)-2 was seen in 24% of the patients. Twenty-eight patients underwent palliative surgery, of which 18 were in the EOX arm, and 10 were in the aspirin arm. Seven patients underwent definitive surgery, of which all except 1 had a D2 gastrectomy. Two patients received palliative radiotherapy.

Chemotherapy Delivery and Toxicity (Intention to Treat)
The median number of cycles in the EOX arm was 5 (1)(2)(3)(4)(5)(6)(7)(8), and in EOX plus aspirin arm was 5.5 (1)(2)(3)(4)(5)(6)(7)(8). Eighty-four patients received at least 3 cycles of EOX (EOX 45, EOX plus aspirin 39). Patients who completed the planned 8 cycles of chemotherapy were 8 in the EOX arm and 9 in aspirin plus EOX arm. No significant differences in grade ¾ toxicities or treatment-related issues between the 2 groups, as shown in Table 3. There were 2 reported cases of non-life-threatening bleeding in the aspirin arm, with one patient who had gum bleeding and the other had disease progression leading to hematemesis which was attributed to tumour bleed.
Treatment discontinuation/change to single or double agent due to toxicity and poor tolerability was noted in 4 patients in group 1 and 3 in group 2. Treatment discontinuation due to default was recorded in 3 patients in each group. Dose modifications were done in 2 patients in group 1 and 5 patients in group 2. Treatment-related death occurred in 1 patient in aspirin plus EOX arm due to lifethreatening febrile neutropenia,

Overall Survival Data
The data was analyzed for overall survival in the intentionto-treat and perprotocol groups. The median duration of follow of the entire cohort was 29 (18.56-39.45) months. The overall survival in the ITT population was 11 (9.07-12.93) months. The median O.S. in the ITT population was 11 (8.58-13.42) months in group 1 versus 10 (6.86-13.14) months in group 2, and there was no significant difference between the 2 arms, P = 0.90 as shown in Fig. 2a. The overall survival (O.S.) in the per-protocol group (n = 70) was 12 (9.51-14.49) months. There was no significant change in the O.S. between group one 12 (8.75-15.25) months versus group two 12 (6.21-17.79) months, as shown in Fig. 2b. Univariate and multivariate analyses were used to compare the prognostic impact of different factors (age, gender, histology, site, performance status, stage, and intervention on overall survival). The univariate analysis of the per-protocol group (n = 70) revealed that site and stage were significant predictors of O.S. The disease site was the only significant predictor of O.S. on multivariate analysis (Table 4).

Discussion
This randomized phase II study could not demonstrate the superiority of adding aspirin to standard EOX chemotherapy in locally advanced and metastatic gastric cancer. Gastric cancer continues to be a significant cause of morbidity and mortality. The prognosis of patients with advanced inoperable and metastatic gastric cancer remains guarded, with a median survival ranging from 7 to 10 months in most larger clinical studies [8].
Aspirin has been commonly used to prevent cardiovascular diseases for decades with excellent long-term tolerance. NSAID's have been successfully used in the prevention trials of colorectal malignancy. The pivotal study by Giampieri et al. [4], where aspirin was used in heavily pre-treated patients with metastatic colorectal cancer receiving capecitabine, showed an improved disease control rate of 80% versus 30% favoring the aspirin group. They also demonstrated a significant improvement in the median overall survival in aspirin users (14.7 months versus 8.7 months). This was a retrospective study on patients who were incidentally on aspirin for cardiovascular disease. However it was only a proof of concept study and required a randomized controlled study to determine the efficacy of aspirin.
There are a number of small inadequate randomized trials reported, and pooling of results suggests a 9% reduction in cancer death with a H.R. (0.91; 95%CI: 0.79, 1.04) [9][10][11] with most of the studies related to colorectal, breast, and prostate cancer.
A systematic review and meta-analysis of 118 observational studies of aspirin and 18 cancer types suggested a considerable reduction in mortality of 20%, and the benefit appears not to be restricted to one or a few cancers. The hazard ratios of cancer-specific mortality and all-cause mortality of all cancers combined were 0.79 (0.73-0.84) and 0.80 (0.74-0.86), respectively [12]. The authors also looked at the risk of bleeding, with many studies reporting an excess GI bleed in patients on aspirin,  but fatal bleed was very uncommon. There was a study looking at a reduction in bleeding in patients on proton pump inhibitors and aspirin [13]. There were few concerns about these studies regarding classification of patients whether on continuous aspirin intake, or non-use of aspirin by the controls, and the associated comorbidities.
There are varied definitions of use of low-dose aspirin for example. some studies used it for minimum 1 year after treatment initiation. The authors also concluded that further research on aspirin and cancer, including observational and randomized trials, is required, especially on previously less-focused cancers.
Our study used aspirin in locally advanced and metastatic disease postdiagnosis rather than in an adjuvant setting or in patients who were already on aspirin for comorbidity for a prolonged period where most of the data are available. Although the dose of aspirin was 150 mg in our study higher than other studies which have used 75 mg, the drug was well tolerated with no increased bleeding risks. The duration of aspirin was limited till disease progression, after which patients were discontinued on the aspirin, which might have had an impact on overall survival. Most of the survival advantage has been seen in breast, colorectal, and prostate cancers, and it is not clear if tumour types play an important role in the efficacy of aspirin. There have been contradictory reports of the use of aspirin in gastric and oesophageal cancers. A meta-analysis by Spence et al. [14] on newly diagnosed oesophageal and gastric cancers from cancer registries in England and Scotland over a period of 10 to 14 years suggested that low-dose aspirin did not increase survival. The proportions of patients surviving 1 year were similar in aspirin users versus non-users in oesophageal (48% vs. 50% in England and 49% vs. 46% in Scotland, respectively) and gastric cancer 58% vs. 57% in England and 59% vs. 55% in Scotland, respectively.
A meta-analysis by Lin et al. [15] on the relationship between aspirin use for oesophageal, gastric cancer, and colorectal cancer showed no differences between postdiagnosis aspirin use and overall survival for gastric and oesophageal cancers in comparison with colorectal cancers, which suggested an improved overall and cancerspecific survival. This again reiterates that aspirin could be beneficial in certain tumour types. There was no survival benefit for colorectal cancer patients with prediagnosis aspirin use. The survival benefit of postdiagnosis aspirin appeared to be confined to patients with mutated PIK3CA tumours (HR = 0.78, 95%CI (0.50), 0.99) and PTGS2 (COX-2) expression. Our study was in an unselected population, and the role of PIK3CA and COX2 needs to be evaluated as a biomarker for studies related to aspirin.

Limitations of the Study
This was a randomized phase II study with response rate as the primary end point. A difference in survival would not be expected with this small sample size. A per-protocol analysis was performed in 70 patients, with 30% not reaching the primary endpoint due to chemotoxicity or default. Aspirin was not continued beyond progression.

Strengths of the Study
This is one of the first randomized controlled studies on the effect of postdiagnosis aspirin in locally advanced and metastatic gastric cancer.

Conclusion
This study could not demonstrate a superior response rate or survival in patients with add-on aspirin to chemotherapy in locally advanced and metastatic gastric cancer. There is a need to use robust molecular biomarkers to identify the population who might benefit from the drug. A large phase III randomized controlled study will be required before we can dismiss the use of aspirin in gastric cancer.