Efficacy and safety of chemoradiotherapy versus chemotherapy for resectable gastric cancer: a systematic review and meta-analysis

Abstract

Background:

Gastric cancer is a leading cause of cancer-related deaths. Surgery combined with adjuvant treatments improves survival. This meta-analysis compares the efficacy of chemoradiotherapy (CRT) vs. chemoimmunotherapy (CT) in resectable gastric cancer.

Methods:

A comprehensive search across five databases from inception to December 2024 identified studies on efficacy of CRT vs CT in adult patients with resectable gastric cancer. Data were pooled as odds ratio (OR) or hazard ratio (HR) with 95% confidence intervals (CI). Analysis used Review Manager 5.4 with a random-effects model.

Results:

A total of 29 studies with 20 794 patients were included for further quantitative and qualitative analysis. CRT was associated with significantly higher recurrence-free survival (HR: 0.79; 95%CI: 0.69, 0.91; P = 0.004) and slightly improved overall survival (HR: 0.85; 95%CI: 0.73, 1; P = 0.05). The odds of locoregional metastases (OR: 0.60; 95%CI: 0.43, 0.82; P = 0.001) were significantly reduced following CRT. Adverse events between both treatments were comparable, although the risk of neutropenia (OR: 1.54; 95%CI: 1.29, 1.84; P<0.0001) and gastrointestinal side effects (OR: 1.30; 95%CI: 1.01, 1.69; P = 0.04) was significantly higher in the CRT arm.

Conclusion:

This meta-analysis concludes a higher efficacy of CRT over CT in improving survival and preventing recurrence, while limiting regional metastasis. Further studies are needed to assess modulation of radiotherapy dosage to reduce the adverse event while maintaining its efficacy.

Introduction

Gastric cancer remains a major global health challenge, ranking as the fifth most commonly diagnosed malignancy and the fifth leading cause of cancer-related death worldwide^[1]. Early-stage disease often presents with vague, non-specific symptoms, such as dyspepsia or mild epigastric discomfort that frequently contribute to delayed diagnosis and limit the efficacy of surgery alone^[2]. Consequently, multimodal strategies combining curative resection with systemic therapy are now standard, aiming to reduce both locoregional and distant recurrence.

HIGHLIGHTS

• Chemoradiotherapy was associated with higher recurrence-free survival (HR: 0.79, P = 0.004) and slightly improved overall survival (HR: 0.85, P = 0.05) compared to chemotherapy alone.

• Chemoradiotherapy was particularly superior in the postoperative setting (HR: 0.78), while no benefit was observed compared to perioperative chemotherapy (HR: 1.01).

• Chemoradiotherapy significantly reduced locoregional metastases (OR: 0.60, P = 0.001), indicating better local tumor control.

• The treatment completion rate was lower in the chemoradiotherapy group (OR: 0.80), though not statistically significant.

• Chemoradiotherapy was linked to higher rates of neutropenia (OR: 1.54, P < 0.0001) and gastrointestinal side effects (OR: 1.30, P = 0.04).

Current practice patterns vary geographically. In Eastern Asia and parts of Europe, a D2 lymphadenectomy followed by adjuvant chemotherapy (CT) (most commonly fluoropyrimidine- and platinum-based regimens) is widely adopted based on landmark trials and regional guidelines^[3]. By contrast, North American and some Western European centers typically perform a less extensive D1 resection and follow with postoperative chemoradiotherapy (CRT) to address residual microscopic disease^[3]. The pivotal MAGIC trial demonstrated that perioperative ECF CT improved overall survival (OS) compared with surgery alone^[4], while the INT 0116 study showed that adding postoperative CRT to surgery yielded a significant survival benefit over surgery alone^[5].

Despite consistent reductions in locoregional recurrence with CRT, long-term survival advantages have been inconsistent when directly compared with CT alone. For example, the international CRITICS trial found no additional OS benefit with postoperative CRT versus adjuvant CT following D2 resection and preoperative CT ^[6]. These conflicting results likely reflect heterogeneity among studies in radiation dose and fields, CT regimens, timing of CRT (preoperative vs. postoperative), and patient selection criteria.

Given the limitations of adjuvant CRT, attention has turned toward neoadjuvant approaches. Preoperative CRT is now standard in locally advanced esophageal cancer due to its ability to downstage tumors, increase R0 resection rates, and improve tolerability before surgery. Early phase II studies in gastric cancer suggest that neoadjuvant CRT is feasible and may further reduce micrometastatic spread. However, whether adding radiotherapy to systemic CT in the pre- or post-operative setting confers a definitive survival advantage, without unacceptable toxicity, remains to be established.

Therefore, we performed an updated systematic review and meta-analysis to compare the efficacy (OS, disease-free survival) and safety (grade ≥3 toxicities, postoperative complications) of CRT versus CT alone in patients undergoing curative resection for gastric cancer, stratified by timing of radiotherapy (neoadjuvant vs. adjuvant) and extent of lymphadenectomy.

Methods

This systematic review and meta-analysis were conducted per the Cochrane Handbook for Systematic Reviews of Interventions and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines^[7]. It was prospectively registered on PROSPERO.

Search strategy and information sources

The literature search was performed systematically in PubMed, Cochrane Library Database, Clinicaltrials.gov, Scopus and Embase databases by two independent authors for articles related to the efficacy of CRT in comparison to CT in patients with resectable gastric cancer. The following keywords were used to search the relevant studies: “stomach neoplasms or gastric carcinoma,” “chemotherapy,” “chemoradiotherapy or radiochemotherapy or radiotherapy,” “resectable or surgical resection.” The search, conducted using appropriate keywords, included studies published up until December 24th, 2024. The line-by-line search strategy for each database is presented in Supplementary Digital Content files, available at: http://links.lww.com/MS9/A890.

Inclusion and exclusion criteria

Studies were included in this systematic review and meta-analysis if they met the following criteria: (1) patients with resectable gastric cancer; (2) age ≥18 years; (3) studies with clinic-pathological factors; (4) studies that compared at least two treatment strategies; (5) survival outcomes of grouped patients should be presented; and (6) studies published in English language.

Exclusion criteria were as follows: (1) studies including targeted therapy or immunotherapy or intraperitoneal CT; (2) conference abstracts, reviews, case reports, studies without useable data, and studies published before 1990; (3) studies published in languages other than English; and (4) conference presentations and grey literature.

Study selection process

The retrieved references were screened based on their titles, abstracts, and relevant outcomes, and assessed for eligibility according to predefined inclusion and exclusion criteria by two independent authors. Any discrepancies were resolved by a third independent author. The conference presentations and grey literature were not included in the literature search.

Screening and data extraction

Two reviewers extracted data from the included studies, focusing on data characteristics, interventions, comparisons, and outcomes. Key outcomes of interest included OS (OS), recurrence-free survival (RFS), distant metastasis, loco-regional metastasis, treatment completion rate, gastrointestinal (GI) side effects, hematological side effects, nausea and vomiting, and neutropenia. The extracted data were compiled into a standardized Excel sheet. A third independent reviewer was involved in resolving any discrepancies.

Risk of bias assessment & quality assessment

The risk of bias in the included studies was evaluated across five domains-selection bias, performance bias, detection bias, attrition bias, and reporting bias-using the RoB 2.0 tool from the Cochrane Collaboration.

The quality of the studies was assessed by The Newcastle-Ottawa scale,^[8] and studies were grouped according to their own scores: ≥ 8 were considered high quality; 7-8 were medium quality; less than 7 were low quality.

Outcomes and statistical analysis

The efficacy endpoints included OS, RFS, distant metastasis, loco-regional metastasis, and treatment completion rate. The safety endpoints of this meta-analysis were the incidence of any GI side effects, hematological side effects, nausea and vomiting, and neutropenia. Pooled outcomes estimate for distant metastasis, loco-regional metastasis, treatment completion rate, GI side effects, hematological side effects, nausea and vomiting, and neutropenia were expressed as odds ratio (OR) with 95% confidence intervals (CI). OS, and RFS were measured as hazards ratio (HR) with 95% CI. Interstudy heterogeneity was addressed using the random-effects model, with the Inverse Variance (IV) method applied for continuous outcomes. Heterogeneity was quantified using the I² statistics, where I² values <50% indicated low heterogeneity, and >50% indicated significant heterogeneity. We used funnel plots to assess the potential publication bias. All statistical analyses were performed using Review Manager 5.4.1 (The Cochrane Collaboration), with statistical significance defined as a P-value of <0.050. In- addition, Sub-group analysis was run to examine the effectiveness of intervention across different subgroups of population, based on preoperative or perioperative timing of CRT, follow up duration and RCTs and observational studies.

Results

Our meta-analysis includes 29 studies^[6,9-36], consisting of 18 observational studies and 11 randomized control trials (RCTs), with a total of 20 794 patients being included to assess the safety and efficacy of preoperative CRT versus CT alone for resectable gastric cancer (Fig. 1). The studies varied in design, geographic regions, and patient characteristics, which contributed to the observed heterogeneity in several outcomes.

Figure 1.

PRISMA flow diagram depicting article selection and screening process.

Baseline characteristics

The included studies spanned multiple countries, including the USA, China, Korea, Greece, Brazil, Canada, Tunisia, and the Netherlands. Most studies involved patients with TNM stage IB-IV gastric cancer, with the majority undergoing D1 or D2 lymphadenectomy with R0 resection. The median age of patients ranged from 19 to 89 years across studies. The median follow-up period varied from 24 to 86.7 months.

The sample sizes for each study ranged from as low as 61 to as high as 5058 patients. Common CT regimens in both arms included fluoropyrimidine-platinum combinations, such as capecitabine + cisplatin (XP), oxaliplatin-based regimens (FOLFOX, XELOX), and triplet therapies like epirubicin + cisplatin + fluorouracil (ECF/ECX). Radiotherapy doses in the CRT group typically ranged from 45 to 50.4 Gy (Table 1). The risk of bias assessment for observational studies is available in Supplementary Digital Content Table 2, available at: http://links.lww.com/MS9/A890 and randomized clinical trials is in Supplementary Digital Content Fig. 1, available at: http://links.lww.com/MS9/A890.

Table 1.

Baseline Characteristics of included studies and treatment regimens of gastric cancer patients

Study	Country, study type, year	Median follow-up (range, months)	Age (median or range)	Tumor stage	Surgery	Treatment group	Sample size (CRT vs CT)	Treatment regimens
Park SH et al	Korea, 2020	47	27–85	TNMII-III, N +	D2, R0	S + CRT vs S + CT	364 (183 vs 181)	CT: SOX:
								RT: 45 Gy
Park et al	USA, 2015	84	22-77	N0-3; M0; TNMIB-IV	D2, R0	S + CRT vs S + CT	453 (227 vs 226)	CT: XP
								RT: 45 Gy
Bamias et al	Greece, 2010	53.7 (0.1–77.8)	32-79	T1-4; N0-3; M0/1; TNM IB-IV	D0 or D1 + D2, R0	S + CRT vs S + CT	141 (71 vs 70)	CT: docetaxel + cisplatin
								RT: 45 Gy
Cats et al	Multicentered, 2018	61.4 (43.3–82.8)	54-69	T1-4; N0-3; M0/1; TNM IB-IV	D0, D1, D2 or D3, R0/R1	CT + S + CRT vs CT + S + CT	788 (395 vs 393)	CT: epirubicin + cisplatin;
								Or oxaliplatin + orally capecitabine
								RT: 45 Gy
Mansouri et al	Tunisia, 2021	38.48 (4–139)	59.21	T1-4; N0-3; M0/1; TNM IB-IV	D1, D1.5, and D2	S + CRT vs S + CT	80 (53 vs 27)	CT: LV5FU2; or ELF; or FOLFOX4
								3D-RT: 45–50.4 Gy
Datta et al	USA, 2016	80.3	65	TNM IB-III	R0 (83.4%), R1	S + CRT vs S + CT	2538 (1869 vs 669)	-
Jabo et al	USA, 2018	76	-	T1-4, TNM IB-IV	-	S + CRT vs S + CT	1043 (1031 vs 462)	-
Yu et al	Korea, 2018	65.4 (3.9–141.7)	22-84	T1-4, N1-3b	D2, R0	S + CRT vs S + CT	1633 (909 vs 724)	CT: XP
								RT: 45 Gy
Zaidi et al	Canada, 2021	-	63 (56–71)	TNM I-III	-	S + CRT vs CT + S + CT	88 (67 vs 21)	CT: epirubicin + cisplatin; oroxaliplatin + FU; or capecitabine
								RT: 45 Gy
Zhu et al	China, 2012	42.5	38-75	TNM Ib-IV, M0	D2, R0	S + CRT vs S + CT	351 (186 vs 165)	CT: FU + LV
								RT: 45 Gy
Yu et al	China, 2012	-	18-70	T3/4, N +	D1, D2, R0	S + CRT vs S + CT	68 (34 vs 34)	CT: FU + LV
								RT: 45 Gy
Kwon et al	Korea, 2010	77.2 (24–92.8)	23-70	TNM IIIA, IIIB, IV(M0)	D2, R0	S + CRT vs S + CT	61 (31 vs 30)	CT: 5-FU + cisplatin
								RT: 45 Gy
Stumpf et al	USA, 2017	47 (1–128)	62 (19–90)	T1-4; N0-3; M0; TNM I-III	D0, D1, D2, R0/1	S + CRT vs CT + S + CT	3656 (1772 vs 1884)	-
Kim et al	Korea, 2012	86.7 (60.3–116.5)	-	T2-4, N0-3, TNM IIIA/ IIIB/IV	D2, R0	S + CRT vs S + CT	90 (46 vs 44)	CT: FU + LV
								RT: 45 Gy
Girardi et al	Brazil, 2018	0.2–61.3	30-80	TNM IB-IIIC	D1, D2, R0/1	S + CRT vs S + CT	309 (227 vs 82)	CT: fluoropyrimidine-platinum doublet; or XELOX; or XP
								RT: -
Peng et al	China, 2016	41.1 (14–111.1)	18-75	TNM IIA-IIIC	D2, R0	S + CRT vs S + CT	337 (124 vs 213)	CT: FOLFOX
								RT: 45 Gy
Wang et al	China, 2021	27.1 (2–116)	28-84	T2-4; N0-3; TNM IA-IIIC	D2, R0	S + CRT vs S + CT	188 (94 vs 94)	CT: capecitabine or S-1
								RT: 45 Gy
Han et al	China, 2020	-	20-75	TNM IB-III	D2, R0	S + CRT vs S + CT	207 (73 vs 134)	CT: FOLFOX
								RT: 45 Gy
Ma et al	China, 2019	41.1 (7.0–104.2)	43-66	T2-4, N0-3, TNM IIIA-C	D2, R0	S + CRT vs S + CT	415 (135 vs 280)	CT: FU-based regimens
								RT: 45 Gy
Li et al	China, 2014	30 (2–63)	19-77	T1-4; N0-3; TNM IB-IIIC	D2, R0	S + CRT vs S + CT	186 (93 vs 93)	CT: 5-FU; or capecitabine; or tegafur gimeracil oteracil potassium capsule
								RT: 45 Gy
Turanli et al	Ankara, 2014	30 (8–112)	60 (25–77)	T2-4, N1-3, TNM III	D2, R0	S + CRT vs S + CT	92 (71 vs 21)	CT: FU + LV; or ECF
								RT: 45 Gy
Stiekema et al	Netherlands, 2014	-	21-89	T1-4, N0-3	At least D1, R1	S + CRT vs S + CT	409 (40 vs 369)	CT: capecitabine + cisplatin
								RT: 45 Gy
Zhou et al	China, 2018	30/24	-	T3-4, N0-3	D1, D2, R1	S + CRT vs S + CT	114 (33 vs 81)	CT: 5-FU-based regimens
								RT: 45 Gy
Zhou et al	China, 2019	38/32.4	19-80	T1-4, N3, TNM IIB, IIIA-C	D2, R0	S + CRT vs S + CT	540 (175 vs 365)	CT: 5-FU + capecitabine; or S-1
								RT: 45–50.4 Gy
Yekedu¨z et al	Turkey, 2021	38.6 (20.3–68.5)	47-65	T1-4, N0-3, TNM I-III	D2, R0	S + CRT vs S + CT	230 (166 vs 64)	CT: 5-FU-based regimens
								RT: 45 Gy
Ejaz et al	USA, 2014	28	62.3 (53.9–70.1)	T1-4; N +, TNM I-IV	D1, D2; R0/R1	S + CRT vs CT + S + CT	505 (294 vs 211)	CT: epirubicin + cisplatin + 5-FU
								RT: -
Fitzgerald et al	Australia, 2017	24/31	63	N0-3, TNM II,III	R0/1	S + CRT vs CT + S + CT	5058 (536 vs 4522)	-
Fan et al	China, 2016	36 (4–88)	19-82	T1-4, N1-3, TNM IB-IIIC	D2, R0	S + CRT vs S + CT	276 (138 vs 138)	CT: 5-FU; or capecitabine
								RT: 45 Gy
Leong et al	Multicentered, 2024	67	61 (50-72)	T1-4; N0-3; M0/1; TNM IB-IV	D1 + or D2, R0	S + CRT vs S + CT	574 (286 vs 288)	CT: ECF/ ECX; or FLOT
								RT: 45 Gy

CT chemotherapy, CRT chemoradiotherapy, S surgery, FU fluorouracil, LV leucovorin, SOX S-1 + oxaliplatin, XP capecitabine + cisplatin, LV5FU2 leucovorin + 5-FU, ELF folinic acid + etoposide + 5-FU, FOLFOX4 oxaliplatin + leucovorin + 5-FU, XELOX capecitabine + oxaliplatin, ECF epirubicin + cisplatin + FU, ECX Epirubicin + Cisplatin + Capecitabine, FLOT Fluorouracil + Leucovorin + Oxaliplatin + Docetaxel

Overall survival

Eighteen studies, with a total of 7143 patients, reported data on OS, including 13 observational studies and 5 randomized controlled trials (RCTs). In the pooled analysis of observational studies, CRT was associated with a significant improvement in OS compared to CT, with a hazard ratio (HR) of 0.79 (95% CI: 0.64–0.97; P = 0.02). But the heterogeneity among these studies was substantial (I² = 90%). In contrast, the pooled analysis of RCTs showed no significant difference in OS between CRT and CT (HR = 1.00, 95% CI: 0.89–1.13; P = 0.96), with no observed heterogeneity (I² = 0%) (Fig. 2).

Figure 2.

Forest plot comparing overall survival (OS).

When all studies were combined, the pooled HR was 0.85 (95% CI: 0.73–1.00; P = 0.05), suggesting a barely statistically significant benefit of CRT. High heterogeneity was observed (I² = 86%, P = 0.00001), which could be attributed to differences in radiation protocols, CT regimens, and baseline patient characteristics across studies. Test for subgroup differences revealed a statistically significant association with study design and effect size (P = 0.05), indicating that the benefit was primarily observed in observational studies.

Overall survival subgroup by postoperative vs perioperative setting

In the sub group analysis CRT significantly improved OS as compared to CT administered in postoperative setting with a statistically significant HR of 0.78 (0.69-0.90) while the CRT showed no benefit in OS when compared to peri operative CT (HR = 1.01 (0.74-1.38)). This shows CRT is particularly superior to CT in situations where postoperative radiotherapy is advised (Fig. 3).

Figure 3.

Forest plot comparing overall survival (OS) subgrouping by preoperative and perioperative therapy.

Overall survival based on follow-up duration

A total of 13 studies were included to assess the impact of follow-up duration on OS in patients receiving preoperative CRT versus CT alone. The analysis was subgrouped by median follow-up time, one subgroup with studies with less than 40 months of follow up period and the other subgroup with studies with more than 40 months of follow up period.

In studies with a follow-up of less than 40 months (4 studies), the pooled HR was 0.82 (95% CI: 0.62–1.08; P = 0.60), indicating no statistically significant difference in OS between CRT and CT groups. High heterogeneity was observed in this subgroup (I² = 95%, P < 0.0001).

In contrast, studies with a follow-up of more than 40 months (9 studies) demonstrated a significant survival benefit with CRT, with a pooled HR of 0.87 (95% CI: 0.77–0.97; P = 0.01), indicating a 13% reduction in the risk of death. Moderate heterogeneity was observed in this subgroup (I² = 63%, P = 0.006).

The overall pooled analysis across all studies yielded an HR of 0.88 (95% CI: 0.73–1.05; P = 0.16), suggesting a statistically insignificant benefit of CRT. No significant difference in treatment effect was observed between the subgroups (χ² = 0.01, P = 0.93, I² = 0%), indicating that the observed survival benefit with CRT was consistently insignificant regardless of follow-up duration (Fig. 4).

Figure 4.

Forest plot comparing overall survival (OS) subgrouping by followup timing.

Overall survival based on lymphadenectomy extent (D1 vs D2)

A total of six studies including 1140 patients assessed the impact of lymphadenectomy extent (D1 vs D2) on OS. The pooled analysis demonstrated that D2 dissection significantly improved OS compared to D1, with a combined HR of 0.71 (95% CI: 0.56–0.90, P = 0.005), suggesting a 29% reduction in the risk of death with D2. Moderate heterogeneity was observed (I² = 63%, P = 0.02).

Subgroup analysis revealed that the survival benefit was consistent across both D1 and D2 groups. For the D1 group, the HR was 0.54 (95% CI: 0.35–0.83), indicating a 46% reduction in mortality risk (P = 0.005). Among the D2 subgroup (five studies), the pooled HR was 0.74 (95% CI: 0.57–0.97, P = 0.03), further supporting the advantage of D2 dissection (Fig. 5).

Figure 5.

Forest plot comparing overall survival (OS) subgrouping by D1 and D2 resection.

There was no statistically significant subgroup difference between D1 and D2 cohorts (χ² = 1.55, P = 0.21, I² = 35.3%), implying a consistent trend favoring D2 dissection regardless of the baseline extent. These findings suggest that more extensive lymphadenectomy results in increased OS in patients undergoing surgery for gastric cancer.

Recurrence-free survival (RFS)

A total of eleven studies, including 5 observational studies and 6 randomized controlled trials (RCTs), reported data on RFS in 4508 patients. In the pooled analysis of observational studies, CRT was associated with a significant improvement in RFS compared to CT, with a HR of 0.67 (95% CI: 0.55–0.81; P < 0.0001). Moderate heterogeneity was observed among the observational studies (I² = 56%).

In the analysis of RCTs, the pooled HR was 0.92 (95% CI: 0.82–1.03; P = 0.16), indicating statistically insignificant difference in RFS between the CRT and CT groups. Heterogeneity among RCTs was low (I² = 6%).

The overall pooled HR across all studies was 0.79 (95% CI: 0.69–0.91; P = 0.0008), indicating a statistically significant benefit in the CRT group with 21% reduction in the risk of recurrence or death. Moderate heterogeneity was observed in the overall pool (I² = 63%). Tests for subgroup differences demonstrated a significant interaction between study design and treatment effect (P = 0.005, I² = 87.5%), suggesting that the benefit observed in RFS in favour of CRT group was primarily indicated by observational studies (Fig. 6).

Figure 6.

Forest plot comparing recurrence-free survival (RFS).

Recurrence-free survival subgroup by postoperative vs perioperative setting

Significant difference was found when comparing postoperative CRT to postoperative CT with a HR of 0.74 (0.64-0.85) while the results were insignificant between peri operative CRT and peri operative CT (HR = 0.99 (0.86-1.14)) as shown by the subgroup analysis (Fig. 7).

Figure 7.

Forest plot comparing recurrence-free survival (RFS) subgrouping by preoperative and perioperative therapy.

Recurrence-free survival based on follow-up duration

Eleven studies were analyzed to evaluate the impact of follow-up duration on RFS in patients treated with preoperative CRT versus CT alone. Studies were divided into two subgroups based on median follow-up time. One with less than 40 months of follow up duration and the second one with more than 40 months of follow up duration.

In studies with less than 40 months of follow-up (2 studies), CRT was associated with a significant improvement in RFS compared to CT, with a pooled HR of 0.72 (95% CI: 0.55–0.94; P = 0.005). No heterogeneity was observed within this subgroup (I² = 0%, P = 0.67), indicating consistent findings.

In the more than 40 months follow-up group (9 studies), CRT also demonstrated a significant benefit, with a pooled HR of 0.85 (95% CI: 0.74–0.96; P = 0.01), corresponding to a 15% reduction in the risk of recurrence or death. Moderate heterogeneity was noted in this subgroup (I² = 53%, P = 0.04).

The overall pooled analysis across all studies yielded an HR of 0.82 (95% CI: 0.73–0.92; P = 0.0009), confirming a statistically significant benefit in RFS with CRT. Subgroup analysis showed no significant difference in treatment effect between shorter and longer follow-up durations (χ² = 1.66, P = 0.20, I² = 39.8%), suggesting that the observed advantage of CRT was consistent regardless of follow-up time (Fig. 8).

Figure 8.

Forest plot comparing recurrence-free survival (RFS) subgrouping by followup timing.

Recurrence-free survival based on lymphadenectomy extent (D1 vs D2)

Seven studies assessed RFS in patients who underwent either D1 or D2 lymphadenectomy as part of their treatment. No data were available for statistical estimation in the D1 subgroup alone, as no eligible studies reported recurrence outcomes exclusively in this group.

In the D2 subgroup (7 studies), CRT was associated with a significant improvement in RFS compared to CT alone, with a pooled HR of 0.77 (95% CI: 0.64–0.92; P = 0.004). This indicates a 23% reduction in the risk of recurrence or death. Moderate heterogeneity was present (I² = 63%, P = 0.01), suggesting variability in treatment protocols or patient characteristics across studies. Subgroup comparison between D1 and D2 could not be performed due to lack of estimable data for the D1 group (Fig. 9).

Figure 9.

Forest plot comparing recurrence-free survival (RFS) subgrouping by D1 and D2 resection.

These findings support the role of CRT in enhancing RFS among patients treated with D2 lymphadenectomy.

Distant metastasis

Fifteen studies, comprising 9 observational studies and 6 randomized controlled trials (RCTs), reported on the occurrence of distant metastasis in a total of 3859 patients. In the pooled analysis of observational studies, there was no statistically significant difference between the CRT and CT groups (OR = 0.91; 95% CI: 0.73–1.13; P = 0.40), with very low heterogeneity (I² = 9%).

Similarly, the pooled analysis of RCTs showed no significant difference in the risk of distant metastasis between CRT and CT (OR = 0.88; 95% CI: 0.68–1.15; P = 0.35), with no observed heterogeneity (I² = 0%).

The overall pooled OR across all studies was 0.90 (95% CI: 0.77–1.04; P = 0.15), confirming a statistically insignificant reduction in the risk of distant metastasis with CRT. No heterogeneity was observed across the combined dataset (I² = 0%, P = 0.60). No significant subgroup differences were found between observational studies and RCTs (P = 0.76), suggesting consistent findings across both study designs (Fig. 10).

Figure 10.

Forest plot comparing distant metastasis rate.

Locoregional metastasis

Seventeen studies, including 11 observational studies and 6 randomized controlled trials (RCTs), assessed the incidence of locoregional metastasis in a total of 5633 patients. In the pooled analysis of observational studies, CRT significantly reduced the risk of locoregional metastasis compared to CT alone, with an OR of 0.63 (95% CI: 0.40–0.99; P = 0.05). High heterogeneity was observed among these studies (I² = 76%).

The pooled analysis of RCTs demonstrated a statistically significant benefit of CRT as well, with an OR of 0.54 (95% CI: 0.34–0.86; P = 0.006) and moderate heterogeneity (I² = 42%).

When all studies were combined, the overall pooled OR was 0.60 (95% CI: 0.43–0.82; P = 0.001), indicating a 40% reduction in the risk of locoregional metastasis with CRT. Heterogeneity across the combined dataset was slightly high (I² = 69%, P < 0.0001). No significant subgroup differences were found between observational studies and RCTs (P = 0.80), suggesting consistency in the treatment effect across study designs (Fig. 11).

Figure 11.

Forest plot comparing loco-regional recurrence.

Treatment completion rate

Thirteen studies, comprising 8 observational studies and 5 randomized controlled trials (RCTs), evaluated treatment completion rates in a total of 3719 patients. In the pooled analysis of observational studies, there was no significant difference in treatment completion between CRT and CT groups (OR = 0.85; 95% CI: 0.56–1.16; P = 0.35), with moderate heterogeneity (I² = 42%).

Among the RCTs, the pooled analysis suggested insignificantly lower completion rates in the CRT group (OR = 0.63; 95% CI: 0.32–1.24; P = 0.18), with very high heterogeneity (I² = 80%).

The overall pooled OR across all studies was 0.80 (95% CI: 0.59–1.09; P = 0.16), indicating a 20% lower likelihood of treatment completion in the CRT group, though this difference was not statistically significant. Moderate heterogeneity was observed across all studies (I² = 63%, P = 0.001). No significant subgroup differences were identified between observational studies and RCTs (P = 0.27), suggesting consistent trends across study designs. The reduced completion rate in the CRT group may reflect factors such as increased toxicity, extended treatment duration, and reduced patient tolerance (Fig. 12).

Figure 12.

Forest plot comparing treatment completion rates in patients undergoing (CRT) versus chemotherapy (CT).

Safety outcomes

Gastrointestinal (GI) side effects

Six studies, including four observational studies and two randomized controlled trials (RCTs), reported on the incidence of GI side effects in a total of 1552 patients. In the pooled analysis of observational studies, CRT was associated with a non-significant increase in GI toxicity compared to CT (OR = 1.41; 95% CI: 0.98–2.05; P = 0.07), with no heterogeneity observed (I² = 0%).

The pooled analysis of RCTs also showed a non-significant increase in GI side effects with CRT (OR = 1.21; 95% CI: 0.84–1.74; P = 0.31), with no heterogeneity (I² = 0%).

When all studies were combined, CRT was associated with a statistically significant increase in the risk of GI side effects compared to CT (OR = 1.30; 95% CI: 1.01–1.69; P = 0.04). Heterogeneity across the included studies was absent (I² = 0%, P = 0.80), indicating consistent findings (Fig. 13).

Figure 13.

Forest plot comparing gastrointestinal (GI) side effects.

Hematological side effects

Six studies, including four observational studies and two randomized controlled trials (RCTs), evaluated the incidence of hematological side effects in a total of 1552 patients. In the pooled analysis of observational studies, CRT was associated with a non-significant increase in hematological toxicity compared to CT (OR = 1.51; 95% CI: 0.61–3.77; P = 0.38), with high heterogeneity observed (I² = 75%).

Similarly, the pooled analysis of RCTs showed no significant difference in hematological side effects between the two groups (OR = 1.14; 95% CI: 0.83–1.57; P = 0.42), with no observed heterogeneity (I² = 0%).

The overall pooled OR across all studies was 1.32 (95% CI: 0.77–2.26; P = 0.32), indicating a statistically insignificantly higher likelihood of hematological side effects in the CRT group. Moderate heterogeneity was present across all studies (I² = 67%, P = 0.009), likely reflecting differences in CT regimens, toxicity grading, and reporting practices (Fig. 14).

Figure 14.

Forest plot comparing hematological side effects.

Nausea and vomiting

Fourteen studies, including 10 observational studies and 4 randomized controlled trials (RCTs), assessed the incidence of nausea and vomiting in a total of 5,778 patients. In the pooled analysis of observational studies, there was no significant difference in nausea and vomiting between the CRT and CT groups (OR = 0.94; 95% CI: 0.52–1.71; P = 0.85), with high heterogeneity (I² = 69%).

Similarly, the pooled analysis of RCTs also showed no significant difference (OR = 1.44; 95% CI: 0.53–3.87; P = 0.47), with low heterogeneity (I² = 26%).

The overall pooled OR across all studies was 1.03 (95% CI: 0.63–1.67; P = 0.91), indicating no statistically significant difference in the incidence of nausea and vomiting between the two treatment groups. Moderate heterogeneity was observed (I² = 63%, P = 0.0007) (Fig. 15).

Figure 15.

Forest plot comparing nausea and vomiting in patients undergoing chemoradiotherapy (CRT) versus chemotherapy (CT).

Neutropenia

Fourteen studies, including 10 observational studies and 4 randomized controlled trials (RCTs), reported on the incidence of neutropenia in a total of 5,807 patients. The pooled analysis of observational studies showed that CRT was associated with a significantly higher risk of neutropenia compared to CT (OR = 1.70; 95% CI: 1.36–2.13; P < 0.00001), with no observed heterogeneity (I² = 0%).

Similarly, the pooled analysis of RCTs indicated a statistically insignificantly increased risk of neutropenia with CRT (OR = 1.29; 95% CI: 0.96–1.73; P = 0.09), with no heterogeneity (I² = 0%).

The overall pooled OR across all studies was 1.54 (95% CI: 1.29–1.84; P < 0.00001), indicating a significantly higher risk of neutropenia in the CRT group. The absence of heterogeneity across all studies (I² = 0%, P = 0.67) shows consistent findings across study designs (Fig. 16).

Figure 16.

Forest plot comparing neutropenia.

Discussion

The purpose of this meta-analysis was to evaluate the impact of CRT on patients suffering from resectable gastric cancer. Although there is abundant literature present to correlate and establish CT with the improvement of the prognosis^[37,38], the significance of CRT is yet to be explored. This updated meta-analysis comprises 29 studies with a total of 20 794 patients that led us to the following observations. In comparison to surgery alone, preoperative CRT was associated with a significant impact in OS, RFS, locoregional metastasis, GI side effects and neutropenia. However, no significant effect was observed in distant metastasis, treatment completed rate, hematological side effects and nausea and vomiting. These findings suggest a potential survival benefit with CRT but highlight concerns regarding treatment tolerance and toxicity. Our updated meta-analysis will mitigate the previous limitations by incorporating recent data and refining the analyses to provide a more comprehensive insight in CRT efficacy.

According to the pooled analysis, CRT was associated with a 15% reduction in the risk of mortality, aligning our observations with the previous studies of improved OS. However, it is crucial to compare these findings to landmark studies, most notably the INT-0116 and CRITICS trials. INT-0116 on one hand showed a significant OS benefit with adjuvant CRT following D0/D1 resections, whereas on the other hand CRITICS trial, which included D2 resections and a perioperative approach, failed to show OS improvement^[4,6]. Additionally, the CI range in this meta-analysis indicated statistical significance that can be described by variations in treatment protocols, patient selection criteria, and follow-up durations. According to Matuscheck et al adjuvant CRT demonstrated improved disease-free survival (DFS) however, no significant improvement of OS was recorded.^[39] Moreover, heterogeneity advocates for different factors such as radiation field size, CT regimens, and surgical approaches. In the previous meta-analyses, Matuscheck et al was not able to record any improvement in OS and Lu et al was able to observe an improvement trend, though it had borderline significance^[2].

CRT in comparison to CT was found favorable in enhancing RFS. Our results recorded a reduction of 21% in the risk of recurrence, further solidifying results of previous studies which indicated improved disease-free survival (DFS). This finding mirrors prior work by Lu et al and Matuschek et al, who were able to document improvement in DFS and reduced locoregional recurrence.^[2,39] The reason behind improved DFS might indicate CRT’s ability to eradicate micrometastatic disease and improve local control by targeting residual tumor cells prior to surgery. Nevertheless, the diverse treatment protocols and different baseline characteristics from the retrospective studies introduced heterogeneity^[2]. The moderate heterogeneity can be explained due to differences in tumor staging criteria and follow up durations.

Despite the fact that CRT does not significantly impact distant metastasis, it notably reduces risk of locoregional metastasis by 40%. This suggests that CRT is more inclined toward managing local disease progression than systemic spread prevention through radiation and surgery induced cytoreduction. Previous studies have mirrored the similar finding, highlighting that locoregional control tends to improve more than systemic outcomes with CRT^[40]. The absence of heterogeneity in distant metastasis emphasized consistent results throughout the studies whereas a substantial heterogeneity in locoregional metastasis may stem from variations in radiation protocols and patient subgroups benefiting most from CRT. This suggests that while CRT may not address systemic disease spread effectively, underlining the necessity of combination regimens and optimized systemic treatments.

The lower treatment completion rate is also affiliated with the intervention group, illustrating a noteworthy drawback: the toxicity of CRT. While the results do not show a substantial impact, previous studies such as ARTIST and CRITICS, have reported comparable trends.^[6,41]. This implies that patient compliance with the treatment depends on the patient’s tolerance and the availability of supportive therapies. This also emphasizes the need for supportive care measures.

The meta-analysis documented a significantly higher risk of GI side effects in the preoperative CRT group compared to the CT. The intervention group was associated with a 30% increase in the GI toxicity combined with the absence of heterogeneity, indicating the outcome to be consistent across the included studies. However, no substantial evidence was found in the following studies.^{[12,21,25–27]} GI toxicities commonly include nausea, vomiting, diarrhea, and mucositis that can eventually lead to treatment discontinuation, resulting in a low treatment completion rate. GI tract contains rapidly dividing epithelial cells that are susceptible to radiation induced damage. Radiation can trigger inflammation eventually resulting in edema and damage to the lining. This process can lead to GI side effects such as diarrhea vomiting, constipation and loss of appetite.^[40,42,43] The most frequently reported issue was diarrhea followed by others making supportive therapy a necessity.

Neutropenia is also a major concern for the intervention group, as a total increase of 71% was substantially highlighted. This can increase susceptibility to infections and can lead to dose reductions or treatment delays. The high heterogeneity suggests variations in CT regimens, radiation dose intensity, and supportive care strategies can impact the observed trends. Moreover, hematological toxicities were also higher replicating previous results but the values were not statistically significant.

Research comes with new knowledge and limitations. A major strength of this meta-analysis is the pool of 29 studies that included a total of 20,794 patients. This led to a comprehensive evaluation of the intervention (CRT) versus control (CT) group thereby enhancing the credibility of our results. On the other hand, we can observe high heterogeneity in three major outcomes; OS, locoregional metastasis, and hematological toxicity, which could be due to differences in study design, geographic variations, and treatment protocols. Moreover, durations of the follow-ups varied from 24 to 60 months (2-5 years) thereby, influencing the reported outcomes. Furthermore, potential bias can be introduced due to inclusion of retrospective and prospective studies, restricting the extent to which the results can be broadly applied.

The findings of meta-analysis suggests that preoperative CRT provides patients with improved OS, RFS, better locoregional control but with high toxicity thereby leading to lower treatment compliance. Whilst these findings are partially in line with the previous literature, they must be interpreted and evaluated in the context of clinical situation and trial heterogeneity. Further research is needed to clarify CRT’s role in long term outcomes with the aim of enhancing the efficacy and minimizing adverse effects.

Ethical approval

Ethics approval was not required for this review.

Consent

Not applicable.

Sources of funding

Not applicable.

Author contributions

A.K. contributed to the initial conceptualization and design of the research study; A.K., H.A.R.A. performed statistical analysis and contributed to the interpretation of the data; S.R., M.F., R.A., L.M., S.C., B.M. participated in the data collection process; A.K., M.F., M.N.J, S.A.W., F.N., M.S. writing original manuscript: M.A., S.R., A.K. reviewed the manuscript, provided critical feedback, and improved its clarity.

Conflicts of interest disclosure

The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Guarantor

Allahdad Khan.

Research registration unique identifying number (UIN)

This systematic review and meta-analysis is registered on PROSPERO with its protocol ID CRD420251007162.

Provenance and peer review

This was not invited paper

Data availability statement

The data presented in this study are available within the article and its Supplementary Materials