Primary Tumor Resection in Patients With Unresectable Metastatic Colorectal Cancer: A Systematic Review and Meta‐Analysis

ABSTRACT

Aim

The benefit of primary tumor resection (PTR) in patients with asymptomatic, unresectable metastatic colorectal cancer (mCRC) remains uncertain. Previous systematic reviews often included symptomatic cases or cohorts with imbalanced chemotherapy use, weakening comparisons between upfront PTR followed by chemotherapy and chemotherapy alone. This analysis focused exclusively on studies involving asymptomatic patients with adequate chemotherapy administration in both groups.

Methods

A systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses guidelines. Eligible studies explicitly excluded symptomatic cases and reported modern chemotherapy regimens. Overall survival (OS) was the primary outcome. Secondary outcomes included progression‐free survival (PFS), perioperative complications, 30‐day mortality, and tumor‐related emergency interventions in the non‐resection (NR) group.

Results

Seventeen studies involving 9317 patients met inclusion criteria: 5581 underwent PTR (PR group), and 3736 did not (NR group). A pooled analysis of 16 studies showed no significant OS benefit with PTR (hazard ratio [HR] = 1.10, 95% confidence interval [CI]: 0.84–1.44). No improvement in PFS was observed (HR = 0.89, 95% CI: 0.68–1.16). Postoperative morbidity and 30‐day mortality in the PR group were 19.8% and 2.0%, respectively. Emergency surgery was required in 6.4% of NR group patients, with a 30‐day mortality rate of 5.3%.

Conclusion

Upfront PTR did not improve survival outcomes in asymptomatic patients with unresectable mCRC and introduced substantial perioperative risks. Routine PTR is not advisable in this population. Clinical decision‐making should be individualized based on patient risk factors and tumor behavior.

Trial Registration

PROSPERO: CRD420251065487

This systematic review and meta‐analysis evaluated the role of primary tumor resection (PTR) in asymptomatic patients with unresectable metastatic colorectal cancer. Across 17 eligible studies including 9317 patients, PTR did not improve overall or progression‐free survival compared with chemotherapy alone. Routine PTR is not recommended due to significant perioperative risks and the lack of survival benefit.

1. Introduction

Colorectal cancer (CRC) ranks as the third most commonly diagnosed malignancy worldwide, with approximately 1.92 million new cases and 930 000 deaths reported annually [1]. At the time of diagnosis, approximately 25% of patients with CRC present with stage IV disease involving distant metastases; among these, 75%–90% have unresectable distant metastases [2]. In recent years, advances in systemic chemotherapy—particularly regimens incorporating irinotecan, oxaliplatin, and targeted agents—have significantly improved the prognosis of stage IV CRC, with reported 5‐year overall survival (OS) rates ranging from 10% to 18% [3, 4].

Despite these therapeutic improvements, the clinical value of primary tumor resection (PTR) in patients with synchronous unresectable metastatic CRC (mCRC) remains a matter of debate. The benefit of PTR for symptomatic primary tumors is well established and widely accepted [5, 6], whereas the benefit of surgery for asymptomatic or minimally symptomatic primary tumors remains controversial.

Historically, upfront PTR followed by systemic therapy was a common strategy, particularly in the early 2000s. A large‐scale analysis by Temple et al. [7] showed that more than 65% of patients with stage IV CRC underwent PTR as their initial treatment. Proponents of this approach have argued that upfront PTR may prolong survival and prevent future complications associated with tumor progression, including bowel obstruction and bleeding [8, 9, 10]. However, accumulating evidence suggests that PTR does not necessarily translate into improved OS. Concerns have been raised regarding surgical morbidity, postoperative mortality, and delays in initiating systemic chemotherapy, all of which may negatively affect patient outcomes [11, 12]. Furthermore, several studies have reported that modern chemotherapy can adequately control the primary tumor locally even in patients with asymptomatic lesions; thereby, questioning the need for surgical intervention in such cases [11, 13].

Although some systematic reviews have examined the role of PTR in patients with unresectable mCRC, only a very limited number of studies have specifically focused on individuals with asymptomatic or minimally symptomatic disease. This study aimed to evaluate the validity of upfront PTR in this subset of patients based on the currently available literature.

2. Materials and Methods

2.1. Registration and Protocol

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines. The review protocol was developed based on the methodology recommended by the Cochrane Handbook for Systematic Reviews of Interventions [14]. The protocol was prospectively registered in the International Prospective Register of Systematic Reviews. Given the nature of the study design, ethical approval was not required.

2.2. Eligibility Criteria

Eligible studies compared OS between PTR and non‐PTR in unresectable mCRC among asymptomatic or minimally symptomatic patients. Exclusion criteria included: non‐adenocarcinoma histology, restriction to specific patient populations (e.g., elderly, obese), inclusion of metastasectomy, lack of modern chemotherapy, fewer than 10 cases, or non‐English publications.

2.3. Data and Literature Sources

A systematic literature search was conducted using the Cochrane Library (CENTRAL) and MEDLINE (via PubMed) to identify relevant studies published between January 2012 and June 2025. The final search was performed on June 15, 2025. The search terms included “colorectal cancer,” “metastatic,” “primary tumor resection,” and “asymptomatic.” A manual search supplemented the electronic search. All articles were independently reviewed before study selection.

2.4. Study Selection and Data Extraction

Two reviewers independently screened titles and abstracts, followed by full‐text review. Discrepancies in judgment were resolved through discussion.

Data were extracted using a standardized form, including:

• Study characteristics: author, year, country, study design (randomized controlled trial [RCT], prospective cohort study [PCS], or retrospective cohort study [RCS]), study period, and number of participating institutions.

• Patient characteristics: age, sex, tumor site, symptomatic status, chemotherapy regimens and proportion of patients receiving chemotherapy.

• Survival outcomes:

  • ○
    OS: hazard ratio (HR) and 95% confidence interval (CI).
  • ○
    Progression‐free survival (PFS): HR and 95% CI.
• Safety outcomes in the primary tumor resection group (PR group):

  • ○
    Presence, number, and types of postoperative complications.
  • ○
    Postoperative mortality within 30 days, along with specific causes of death.
• Primary tumor‐related complications in the non‐primary tumor resection group (NR group):

  • ○
    Frequency and causes of interventions, including emergency surgery and non‐surgical procedures such as stent placement, for managing primary tumor‐related complications.
  • ○
    Associated complication and mortality rates following such interventions.

The primary outcome was OS. Secondary outcomes included PFS, postoperative complication rate, 30‐day postoperative mortality, and the incidence of interventions for tumor‐related complications in the NR group. For both complications and mortality, the number of events and the number of patients at risk were recorded; CIs were calculated when possible.

2.5. Quality Assessment

Study quality was assessed using appropriate tools. Non‐RCTs were evaluated with the Newcastle‐Ottawa Scale (NOS), which scores selection, comparability, and outcome [15]. The maximum score is 9 points, and studies scoring more than 6 were considered high quality.

RCTs were assessed with the Cochrane Risk of Bias 2.0 (RoB 2.0) tool, covering five bias domains [16]. Each domain was rated as having “low risk,” “some concerns,” or “high risk” of bias. An overall risk‐of‐bias judgment was also made for each study. Two reviewers independently performed the assessments, and any disagreements were resolved through discussion.

2.6. Statistical Analysis

This meta‐analysis assessed the effect of PTR by pooling HRs and corresponding 95% CIs reported in each study for OS and PFS. For binary outcomes, such as postoperative complication rates, 30‐day mortality, and the incidence of emergency surgery in the NR group, event rates with 95% CIs were calculated.

Statistical heterogeneity was assessed with the I ² statistic; I ² ≥ 50% indicated substantial heterogeneity [17]. Given the anticipated clinical and methodological diversity across studies, a random‐effects model based on the DerSimonian–Laird method was used for all pooled analyses, including OS, PFS, and safety outcomes [18].

Subgroup analyses were conducted according to study design (RCT vs. non‐RCT, PCS vs. RCS), chemotherapy receipt balance (100% or ≤ 5% difference vs. others), tumor location (sidedness and rectal cancer proportion), RAS mutation status, targeted therapy use, and metastatic burden (multi‐organ vs. single‐organ). Sensitivity analyses were also performed to evaluate the influence of individual studies on OS [19].

Publication bias was assessed through visual inspection of funnel plots and Egger's linear regression test, with a p‐value < 0.1 considered indicative of significant bias [20].

All statistical analyses were performed using EZR (version 1.68) [21]. Statistical significance was set at p < 0.05.

3. Results

3.1. Study Selection

As shown in Figures 1 and 3, 405 articles were identified. After screening, five RCTs, one PCS, and 11 RCSs were deemed eligible for inclusion in the analysis. The list of included studies is presented in Table 1 [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38]. Moreover, 23 studies were excluded because they did not report the proportion of patients undergoing symptomatic resection. An additional four studies were excluded because the proportion of symptomatic resections was reported but deemed excessively high (ranging from 27% to 93%). Furthermore, three studies were excluded due to a markedly low chemotherapy administration rate in the NR group (ranging from 27.8% to 46.2%).

FIGURE 1.

Flow chart of the literature search according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta‐Analyses) statement.

FIGURE 3.

Forest plot and meta‐analysis of the impact of primary tumor resection on overall survival. CI, confidence interval; SE, standard error.

TABLE 1.

Summary of the included studies.

Author	Year	Country	Study design	Study setting	Leading specialty	PR (n, %)	NR (n)	Symptomatic resections (%)	CTx received (%)		Chemotherapy regimens
									PR	NR
Kruijssen [22]	2024	Netherlands, Denmark	RCT	Multi	Onco	101 (49.5%)	103	0	100	100	FOLFOX/FOLFIRI/CAPOX/CAPIRI±bevacizumab
Rahbari [23]	2024	Germany, Spain, Austria	RCT	Multi	Surg	187 (47.6%)	206	0	75.9	93.6	FOLFOX/FOLFIRI±bevacizumab/cetuximab
Huang M [24]	2022	China	RCT	Single	Onco	42 (48.8%)	44	0	97.4	92.5	FOLFOX/CAPOX
Kanemitsu [25]	2021	Japan	RCT	Multi	Surg	81 (49.1%)	84	0	83	94	CAPOX±bevacizumab
Doah [26]	2021	Korea	RCS	Single	Surg	98 (67.1%)	48	0	81.6	85.4	FOLFOX/FOLFIRI/CAPOX±bevacizumab/cetuximab
Huang Y [27]	2020	China	RCS	Single	Surg	150 (50.0%)	150	0	89.3	85.3	FOLFOX/FOLFIRI/CAPOX±bevacizumab/cetuximab
Park [28]	2020	Korea	RCT	Multi	Surg	23 (47.9%)	25	0	100	100	FOLFIRI±bevacizumab/cetuximab
Mehta [29]	2017	USA	RCS	database	Surg	4152 (65.2%)	2216	0	60.5	100	NR
Zhang [30]	2017	China	RCS	Single	Surg	125 (64.4%)	69	0	80	94	FOLFOX/FOLFIRI/CAPOX/CAPIRI±bevacizumab/cetuximab
Wang [31]	2016	China	PCS	Single	Onco	118 (61.8%)	73	0	100	100	FOLFOX/FOLFIRI/CAPOX+bevacizumab
Niitsu [32]	2015	Japan	RCS	Single	Surg	42 (73.7%)	15	0	100	100	FOLFOX/FOLFIRI/CAPOX±bevacizumab/cetuximab
Miyamoto [33]	2015	Japan	RCS	Single	Surg	68 (51.9%)	63	0	100	100	FOLFOX/FOLFIRI±bevacizumab/cetuximab
Yun [34]	2014	Korea	RCS	Single	Surg	218 (52.4%)	198	0	65.6	100	FOLFOX/FOLFIRI/CAPOX/CAPIRI±bevacizumab/cetuximab
Matsumoto [35]	2014	Japan	RCS	Single	Surg	41 (46.6%)	47	0	85.4	100	FOLFOX/FOLFIRI/CAPOX±bevacizumab/cetuximab
Kim [36]	2014	Korea	RCS	Single	Surg	72 (22.2%)	252	0	100	100	FOLFOX/FOLFIRI±cetuximab
Watanabe [37]	2014	Japan	RCS	Single	Surg	46 (29.1%)	112	0	100	100	FOLFOX/FOLFIRI±bevacizumab
Boselli [38]	2013	Italy	RCS	Single	Surg	17 (35.4%)	31	0	64.7	100	FOLFOX±bevacizumab

Note: % indicates the proportion of patients undergoing primary tumor resection among the total number of included patients in each study.

Abbreviations: CTx, chemotherapy; Database, database study; Multi, multicenter study; NR, no primary tumor resection; NR, not reported; Onco, medical oncology department‐led study; PCS, prospective cohort study; PR, primary tumor resection; RCS, retrospective cohort study; RCT, randomized controlled trial; Single, single center study; Surg, surgical department‐led study.

3.2. Study and Patient Characteristics

A total of 9317 patients were included, comprising 5581 in the PR group and 3736 in the NR group. Details are shown in Tables 1 and 2.

TABLE 2.

Clinical characteristics of included studies.

Author	Age (median)		Sex (man, %)		Primary site		RAS mut (%)		Multi‐organ		Median OS (mo)		OS		PR complications (%)	PR 30‐day mortality (%)	Intervention for TRC (%)	Emergency surgery for TRC (%)
					(right/rectum, %)				Metastases (%)
	PR	NR	PR	NR	PR	NR	PR	NR	PR	NR	PR	NR	HR	95% CI
Kruijssen [22]	66	66	65	67	52/23	48/27	NR	NR	62	63	20.1	18.3	1.1	0.84–1.43	65	1.98	20	13
Rahbari [23]	69	69	67.4	66	44.9/0	45.9/0	NR	NR	40.9	38.5	16.7	18.6	0.944	0.74–1.21	NR	NR	NR	NR
Huang M [24]	NR	NR	NR	NR	38/NR	39/NR	47.6	38.6	60	63	24	26.3	0.95	0.55–1.62	NR	NR	9.1	6.8
Kanemitsu [25]	65	65	56	54	NR/7	NR/7	NR	NR	NR	NR	25.9	26.7	1.11	0.78–1.58	38	4	13	13
Doah [26]	69	66.5	50	60.4	29.6/30.6	20.8/43.8	NR	NR	43.9	43.8	18	15	0.614	0.40–0.94	3.1	0	27.1	27.1
Huang Y [27]	57	56	58.7	59.3	NR/19.7	NR/37.5	NR	NR	NR	NR	17	7	0.43	0.42–0.44	NR	NR	NR	NR
Park [28]	61	61	65	65	26.9/26.9	31.8/13.6	NR	NR	23.1	45.5	25.9	26.7	1.2	0.65–2.21	19.2	3.8	NR	NR
Mehta [29]	76	76	50	50	NR/16.5	NR/33.3	NR	NR	NR	NR	16.7	13.9	0.68	0.44–1.04	NR	NR	3.9	3.9
Zhang [30]	58	59	69.6	68.1	27.2/0	23.2/0	NR	NR	0	0	22	14	1.034	0.97–1.10	NR	NR	NR	NR
Wang [31]	57.5	57.5	NR	NR	NR/38.1	NR/42.5	39	38.4	56.8	58.9	22.5	17.8	0.59	0.38–0.92	5.1	NR	21.8	6.8
Niitsu [32]	61.5	63	19	46.7	NR/42.9	NR/26.7	42.9	60	61.9	60	23.9	13.4	1.3	0.6–2.73	42.9	NR	20	20
Miyamoto [33]	61.6	64.4	62	67	28/42	34/30	NR	NR	29	49	30.4	24.1	NR	NR	17.6	0	25.4	25.4
Yun [34]	59	60	64.6	60.2	24.8/30.7	26/35.4	NR	NR	40.4	63.6	17.2	14.4	1.16	0.89–1.52	9.6	0.5	4.5	4.5
Matsumoto [35]	67	62	61	70.2	19.5/29.3	23.4/23.4	NR	NR	51.2	46.8	23.9	22.6	0.72	0.42–1.25	19.5	1.7	19.3	19.3
Kim [36]	58.6	60.9	54.2	58.7	NR/25	NR/32.1	NR	NR	45.8	53.2	17.2	13.6	0.64	0.45–0.89	11.6	1.9	14	8
Watanabe [37]	63	60	54.3	63.4	NR/15.2	NR/21.4	NR	NR	56.5	63.4	19.9	19	0.81	0.53–1.19	17	0	16	16
Boselli [38]	70	73	NR	NR	23.5/35.3	22.6/58	NR	NR	NR	NR	4	5	2.1	1.06–4.5	35.3	35.3	NR	NR

Abbreviations: 95% Cl, 95% confidence interval; HR, hazard ratio; mo, months; Multi‐organ metastases, multiple organ metastases; NR, no primary tumor resection; NR, not reported; OS, overall survival; PR, primary tumor resection; RAS mut, RAS mutation status; TRC, tumor‐related complications.

All studies reported 0% symptomatic resections. The proportion of patients receiving chemotherapy was significantly higher in the NR group (100.0% vs. 89.3%, p < 0.001). Among the 16 studies that provided detailed chemotherapy regimens, all employed oxaliplatin or irinotecan, and 15 studies also included molecular targeted therapies such as bevacizumab or cetuximab. One study [29] did not specify the regimen in detail; however, as it was a retrospective cohort study based on the United States database from 2001 to 2009, it was presumed that the regimens included capecitabine, oxaliplatin, and irinotecan.

The median age ranged from the 60s to 70s, with no significant intergroup difference observed in any of the studies. Sex ratio was approximately equal, with no notable intergroup differences. Two studies [23, 30] included only patients with colon cancer, whereas the remaining studies included mixed populations of colon and rectal cancer. The proportion of rectal cancer across the studies was generally around 20%–30%, indicating that rectal cancer cases constituted a relatively small proportion of the total study population.

3.3. Bias Assessment

3.3.1. Risk of Bias in Included Studies

The methodological quality and risk of bias for each included study are summarized in Table 3. All five RCTs provided high‐certainty evidence. Of these, four were judged to have a low risk of bias according to the RoB 2 tool, whereas one study [23] was rated as having some concerns. In this particular study, the open‐label design meant that both patients and clinicians were aware of the group assignments, potentially influencing post‐intervention behaviors such as delays or discontinuation of chemotherapy. However, the use of a strict intention‐to‐treat analysis and the consistency of results across sensitivity and subgroup analyses supported the study's validity. Additionally, the primary outcome—overall survival—was objectively measurable, and missing data were minimal. Despite the RoB 2 rating of “some concerns,” this study was still considered to provide high‐certainty evidence.

TABLE 3.

Methodological quality and risk of bias of the included studies.

Author	Assessment method	Result	Certainty
Kruijssen [22]	RoB2	Low risk of bias	High
Rahbari [23]	RoB2	Some concerns	High
Huang [24]	RoB2	Low risk of bias	High
Kanemitsu [25]	RoB2	Low risk of bias	High
Doah [26]	NOS	8	High
Huang [27]	NOS	9	High
Park [28]	RoB2	Low risk of bias	High
Mehta [29]	NOS	8	High
Zhang [30]	NOS	8	High
Wang [31]	NOS	8	High
Niitsu [32]	NOS	8	High
Miyamoto [33]	NOS	8	High
Yun [34]	NOS	9	High
Matsumoto [35]	NOS	9	High
Kim [36]	NOS	9	High
Watanabe [37]	NOS	9	High
Boselli [38]	NOS	8	High

Abbreviations: NOS, Newcastle–Ottawa Scale; RoB2, Cochrane risk‐of‐bias tool for randomized trials version 2.

The 12 non‐randomized studies were assessed using the NOS. All studies scored either 8 or 9 points, reflecting high methodological quality. These scores supported the classification of all 12 studies as providing high‐certainty evidence.

3.3.2. Assessment of Publication Bias

Publication bias was assessed using Egger's linear regression test. The funnel plot for OS showed symmetry, and Egger's test indicated no significant publication bias (p = 0.77). This finding supported a low risk of publication bias in the meta‐analysis (Figure 2).

FIGURE 2.

Funnel plot for assessing publication bias in overall survival analysis. HR, hazard ratio.

3.4. Survival Outcomes

3.4.1. Overall Survival

A meta‐analysis was conducted using 16 studies that reported HRs and 95% CIs for OS. The pooled analysis showed no OS benefit from PTR (pooled HR = 1.10, 95% CI: 0.84–1.44; Figure 3).

Heterogeneity was high (I ² = 98.1%), indicating substantial variability across the included studies. This high heterogeneity may be attributable to differences in patient characteristics, treatment strategies, and analytical methods among studies.

A leave‐one‐out sensitivity analysis demonstrated no substantial change in the pooled HR. This result supported the robustness of the overall findings.

3.4.2. Subgroup Analysis: Impact of Study Design on the Effect of PTR on Overall Survival

3.4.2.1. Subgroup Analysis by Study Design

A subgroup analysis was conducted to evaluate the impact of study design on the effect of PTR on OS. In RCTs, PTR did not show a significant survival benefit, with a pooled HR of 1.10 (95% CI: 0.94–1.20). In contrast, among non‐RCTs, PCSs demonstrated a significant OS benefit from PTR, with an HR of 0.59 (95% CI: 0.38–0.92). Although RCSs did not show a statistically significant benefit (HR = 0.83, 95% CI: 0.56–1.24), a trend toward improved OS was observed.

Heterogeneity analysis revealed no heterogeneity in the RCT subgroup (I ² = 0%), indicating consistency across trials. However, the non‐RCT subgroup exhibited substantial heterogeneity (I ² = 95.4%), reflecting notable variability across the studies. Although non‐RCTs suggested a potential survival benefit of PTR, the reliability of these findings was considered lower than that of RCTs. These results suggest that differences in study design may influence the estimated effect of PTR on OS.

3.4.2.2. Subgroup Analysis by Chemotherapy Receipt Balance

Chemotherapy receipt rates were assessed in 17 studies comparing the PR and NR groups. In the PR group, 1846 of 5581 patients (33.1%) did not receive chemotherapy, compared to 54 of 3736 (1.5%) in the NR group. This difference was statistically significant (p < 0.001; p = 5.0 × 10⁻³⁰¹), indicating significantly lower chemotherapy administration in the PR group. Such an imbalance may have influenced key outcomes, including OS; thereby, confounding the effect of PTR. Regarding chemotherapy‐related adverse events, four studies provided relevant data; however, none showed significant differences between the PR and NR groups.

To evaluate the impact of chemotherapy receipt imbalance on OS, subgroup analyses were performed in studies with balanced chemotherapy administration between the PR and NR groups.

First, in six studies where all patients in both groups received chemotherapy (PR group = 402 patients, NR group = 580 patients), PTR did not confer a survival benefit (pooled HR = 0.87, 95% CI: 0.66–1.14, I ² = 35%). Sensitivity analyses yielded consistent results, confirming no significant effect of PTR.

Next, nine additional studies were analyzed, including three in which the difference in chemotherapy receipt rates between groups was ≤ 5% (PR group = 695 patients, NR group = 1013 patients). This analysis also did not demonstrate a significant OS benefit with PTR (pooled HR = 0.77, 95% CI: 0.54–1.09). Although a trend favoring PTR was observed, the difference was not statistically significant.

Overall, these findings indicate that even after accounting for chemotherapy receipt imbalance, PTR did not provide a clear OS benefit. The results remained consistent when studies with ≤ 5% differences in chemotherapy receipt rates were included, offering no evidence to support PTR efficacy.

3.4.2.3. Subgroup and Meta‐Regression Analysis by Clinical and Molecular Characteristics

Subgroup analyses based on tumor location, including sidedness and the proportion of rectal cancer, did not demonstrate a significant OS benefit from PTR in either subgroup. Analyses restricted to studies reporting RAS mutation rates also failed to show a statistically significant effect of PTR. No eligible studies provided data on BRAF mutation or MSI status; therefore, these markers could not be evaluated. Furthermore, subgroup analyses stratified by the proportion of patients with multi‐organ metastases did not reveal any OS advantage with PTR. Collectively, these findings suggest that PTR does not confer a clear survival benefit regardless of tumor location, molecular characteristics, or metastatic burden.

Meta‐regression was then conducted to investigate potential sources of heterogeneity in OS outcomes. Variables included tumor location (proportion of right‐sided, left‐sided, and rectal cancer), RAS mutation rate, use of targeted agents, and the proportion of multi‐organ metastases. Study design and chemotherapy receipt rates were addressed in prior subgroup analyses; therefore, the meta‐regression focused on these clinical and molecular factors. None of the examined variables were significantly associated with the effect of PTR (right‐sided proportion: β = −0.0008 [95% CI: −0.0066 to 0.0050], p = 0.78; left‐sided proportion: β = 0.0008 [−0.0050 to 0.0066], p = 0.78; rectal proportion: β = −0.0016 [−0.0063 to 0.0032], p = 0.50; RAS mutation rate: β = −0.0014 [−0.0061 to 0.0033], p = 0.52; targeted therapy use: β = −6.9927 [−918.82 to 904.83], p = 0.99; multi‐organ metastasis proportion: β = 0.0036 [−0.0047 to 0.0119], p = 0.37). These results indicate that the substantial heterogeneity observed in this study cannot be explained by individual clinical or molecular variables; however, it is more likely attributable to a combination of differences in patient populations and treatment strategies across studies.

3.4.3. Progression‐Free Survival

PFS was analyzed in four studies (three RCTs, one PCS; n = 646). The PR group did not show a significant improvement in PFS compared to the NR group (HR = 0.89, 95% CI: 0.68–1.16, p = 0.39).

The statistical heterogeneity among the studies was moderate (I ² = 42%, p = 0.16), which was considered acceptable. A sensitivity analysis including only RCTs also failed to show a significant benefit of PTR for PFS (HR = 0.95, 95% CI: 0.71–1.26, I ² = 0%).

These findings suggested that PTR may not confer a clear advantage in terms of PFS in patients with unresectable mCRC (Figure 4).

FIGURE 4.

Forest plot and meta‐analysis of the impact of primary tumor resection on progression‐free survival. CI, confidence interval; SE, standard error.

3.5. Surgical Safety

3.5.1. Perioperative Morbidity in the PR Group

Postoperative complications in the PR group were reported in 13 studies, involving a total of 967 patients. Among them, 191 patients experienced postoperative complications, corresponding to a complication rate of 19.8% (95% CI: 17.4%–22.4%). Seven studies described specific complications, including surgical site infection (27 cases), anastomotic leakage (13 cases), postoperative ileus or bowel obstruction (15 cases), intra‐abdominal abscess (two cases), pneumonia (two cases), sepsis (one case), and urinary tract infection (one case).

3.5.2. 30‐Day Mortality in the PR Group

Thirty‐day perioperative mortality following PTR was evaluated in 10 studies, comprising 765 patients. A total of 15 deaths were reported, corresponding to a mortality rate of 2.0% (95% CI: 1.2%–3.2%). Five studies described the causes of death. The most common cause was liver failure (five cases), followed by pneumonia (three cases), heart failure (two cases), sepsis due to anastomotic leakage (one case), and multiple organ failure (one case).

3.6. Emergency Intervention in the NR Group

Interventions (e.g., surgery or stent placement) for primary tumor‐related complications were reported in 12 studies, involving 241 of 3255 patients in the NR group (7.4%, 95% CI: 6.6%–8.4%). Among these, emergency surgery was performed in 163 cases (6.4%, 95% CI: 5.6%–7.3%), as reported in the same 12 studies. Seven studies described the causes of these emergency surgeries, with bowel obstruction being the most common (56 cases), followed by perforation (16 cases) and bleeding (nine cases). Additional surgeries were performed due to tumor‐related pain or intra‐abdominal abscesses.

Postoperative complications following emergency surgery were described in three studies, with seven complications among 41 patients, resulting in a complication rate of 17.1% (95% CI: 8.5%–31.3%). Two studies provided details of specific complications, including surgical site infection (three cases), anastomotic leakage (two cases), bowel obstruction (two cases), pneumonia (one case), intra‐abdominal abscess (one case), and sepsis (one case).

Postoperative mortality after emergency surgery was reported in two studies, with two deaths among 38 patients (mortality rate: 5.3%). However, the causes of death were not described.

4. Discussion

This meta‐analysis evaluated the validity of upfront PTR for asymptomatic CRC with unresectable distant metastases. Despite several limitations, current evidence suggests that upfront PTR does not improve OS in asymptomatic unresectable mCRC.

The appropriateness of PTR in asymptomatic mCRC remains controversial. Although many previous studies have been conducted, results have shown inconsistency. Some studies have reported that PTR can prevent tumor‐related complications during chemotherapy and improve OS [39]. In contrast, other reports suggest that PTR may delay the initiation of chemotherapy and worsen physical condition due to postoperative complications, potentially accelerating tumor progression [24, 33]. Similarly, conclusions from previous systematic reviews and meta‐analyses on this topic have been inconsistent. Even when limited to studies published since 2012, a total of 16 systematic reviews and meta‐analyses have been reported [40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55]. Among them, only three concluded that PTR did not confer an OS benefit [41, 53, 55], whereas 11 reported a significant OS advantage associated with PTR [40, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52], and the remaining two suggested a favorable trend [42, 54]. Consequently, when the existing literature is viewed as a whole, the interpretation that PTR contributes to improved OS appears to be predominant. However, the majority of these reviews have methodological limitations, and their results should therefore be interpreted with caution. The present meta‐analysis re‐evaluates this issue by explicitly addressing these limitations and aims to provide clinically relevant insight.

Two major reasons may explain these discrepancies. The first involves variation in chemotherapy regimens across study populations. Positive outcomes associated with PTR were more frequently reported before 2010. The demonstrated efficacy of modern chemotherapeutic agents, including irinotecan, oxaliplatin, and molecularly targeted therapies, has likely contributed substantially to improved treatment outcomes for mCRC. Considering the influence of modern chemotherapy, this analysis limited the screening to articles published from January 2012 onward. Although oxaliplatin, capecitabine, and bevacizumab were approved in 2004–2005, many studies published before 2012 included patients treated before 2004, making it difficult to reflect the true impact of modern chemotherapy. Restricting the search period ensured better alignment with current clinical practice. This time frame also coincides with the widespread implementation of anti‐angiogenic therapy. The second concern involves the proportion of symptomatic cases and the chemotherapy receipt rate in the NR group. Several studies included in previous systematic reviews failed to clearly specify whether symptomatic cases had been excluded, despite the reviews targeting asymptomatic mCRC. Moreover, some of those studies included NR groups with substantially lower chemotherapy receipt rates, even though the reviews aimed to compare outcomes between upfront PTR and chemotherapy. To minimize such bias, this analysis applied strict inclusion criteria: only studies that explicitly excluded symptomatic patients and showed a balanced chemotherapy receipt rate were included. Ultimately, 30 studies were excluded during the full‐text review, 23 of which had been included in previous reviews. All 17 included studies explicitly stated that the rate of symptomatic resection was 0%, and the chemotherapy non‐receipt rate in the NR group was only 1.5%.

A closer examination of the previous reviews revealed that all 11 reviews reporting an OS benefit from PTR [40, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52] shared the following two characteristics: (1) inclusion of patient cohorts that encompassed earlier therapeutic eras, in which modern systemic chemotherapy was not adequately represented, and (2) inclusion of patient populations in which asymptomatic status was not explicitly stated. Although six systematic reviews [50, 51, 52, 53, 54, 55] have been published since 2023, three of these [50, 51, 52] are included among the aforementioned 11 reviews and thus have similar limitations regarding study selection and patient characteristics. In contrast, some recent reviews have reported no OS benefit associated with PTR. However, the meta‐analysis by Zhang et al. [53] included only 10 studies and, despite applying eligibility criteria comparable to those of the present study, did not capture several relevant publications. Similarly, the meta‐analysis by Huang et al. [55] did not incorporate all available RCTs, whereas the present analysis includes two additional RCTs. Taken together, the apparent OS benefit of PTR reported in most previous reviews may be attributable to the inclusion of studies conducted in earlier treatment eras and/or those involving symptomatic patients. Notably, all five RCTs included in the present meta‐analysis consistently demonstrated no OS benefit from upfront PTR. The discrepancy between these RCT findings and the conclusions of prior reviews may be attributable to such methodological differences. By strictly limiting eligibility to asymptomatic patients treated with contemporary systemic chemotherapy and incorporating all available RCT evidence, the present meta‐analysis provides a more rigorous re‐evaluation of this controversial issue and offers a perspective that prompts reconsideration of the conventional interpretation.

The pooled analysis demonstrated no improvement in OS or PFS from PTR. Therefore, routine PTR should not be recommended. Although this meta‐analysis does not provide a basis for establishing absolute indications, several clinical insights can be drawn. Doah et al. [26] reported a median interval of 29 days (IQR: 21–36) from PTR to initiation of chemotherapy, whereas Miyamoto et al. [33] noted that this interval was significantly prolonged in patients with postoperative complications (45 vs. 29 days, p = 0.035), suggesting that complications disrupted treatment schedules. In the PR group, postoperative morbidity reached 19.8%, and 30‐day mortality was 2.0%, confirming that PTR represents an invasive procedure associated with a measurable risk. For patients judged to be at high risk of postoperative complications, chemotherapy should be prioritized over PTR. Although several studies reported comparable rates of chemotherapy‐related adverse events between PR and NR groups [22, 23, 24, 28], Kanemitsu et al. [25] found that both grade ≥ 2 and grade ≥ 3 adverse events occurred significantly more often in the PR group, and Park et al. [28] reported a higher incidence of severe (grade IV) adverse events. These findings suggest that postoperative physical deterioration may affect the tolerability of chemotherapy, a factor that should be considered when evaluating the appropriateness of PTR.

Conversely, 7.4% of patients in the NR group required emergency interventions for primary tumor‐related complications, and 6.4% underwent surgical procedures. The complication rate of emergency surgery was 17.1%, and the 30‐day mortality was 5.3%, indicating a high risk. Given the frequency and risk of emergency surgery, PTR may have a preventive role in avoiding such emergencies. However, this analysis relied on aggregated data from published studies rather than individual patient data, which precluded the application of competing risk methods such as the Fine–Gray model to account for death as a competing event. Therefore, the reported proportions reflect crude cumulative rates and may underestimate the true clinical risk, as deaths occurring before tumor‐related events were not incorporated. The NSABP C‐10 trial also emphasized this limitation [56], underscoring the need for future investigations using individual patient data with competing risk analyses. Thus, the present findings suggest that the potential prevention of tumor‐related emergencies alone may not justify routine upfront PTR in asymptomatic patients, particularly in the absence of any demonstrable survival benefit. Nevertheless, Miyamoto et al. [33] identified circumferential tumors as an independent risk factor for perforation or obstruction in the NR group. Thus, even among asymptomatic or mildly symptomatic patients, certain subgroups may benefit from upfront PTR. Patients with bulky tumors might warrant consideration for upfront PTR despite being asymptomatic.

Furthermore, the proportion of patients undergoing PTR in the included studies was approximately 60%, which exceeded rates reported in population‐based or real‐world cohorts. Among the 17 studies analyzed, 14 were led by surgical departments, which may have preferentially enrolled fitter patients considered more suitable for surgery. This pattern suggests potential surgical selection bias, limiting the generalizability of the findings to the broader real‐world population. However, no statistically significant correlation was observed between leading specialty (surgery vs. medical oncology) and PTR rate (r = −0.21, p = 0.43). Factors such as institutional characteristics and patient recruitment practices may also have contributed to the higher PTR rates observed.

This study had several limitations. Firstly, a significant difference in chemotherapy receipt rates was observed between the two groups: 33.1% of patients in the PR group did not receive chemotherapy compared to only 1.5% in the NR group (p < 0.001). This discrepancy indicates that a substantial proportion of patients undergoing PTR did not proceed to chemotherapy postoperatively, potentially influencing the OS outcome. In real‐world practice, systemic chemotherapy is generally initiated promptly after PTR; therefore, this deviation must be considered when interpreting the OS findings. To address this concern, subgroup analyses restricted to studies with balanced chemotherapy administration between the PR and NR groups were performed. Even in studies where both groups uniformly received chemotherapy, PTR did not confer a survival advantage. Additional analyses, including studies with minimal (< 5%) differences in chemotherapy receipt rates, yielded consistent results. These findings suggest that chemotherapy imbalance does not account for the observed lack of OS benefit with PTR.

Secondly, prognostic factors such as primary tumor location (right, left, or rectum), RAS mutation status, and the sites or number of metastases were not consistently reported across studies. In contemporary practice, chemotherapy tailored by gene mutation status and primary tumor sidedness has become standard, and insufficient stratification by these variables may limit the applicability of the findings to current clinical settings. Subgroup analyses restricted to studies that provided these data were also performed; however, none demonstrated a significant OS benefit from PTR.

Finally, statistical concerns and potential biases warrant attention. To account for inter‐study heterogeneity, a random‐effects model using the DerSimonian–Laird method was applied for all pooled outcome analyses (OS, PFS, complication rates, etc.) [17]. Substantial heterogeneity was observed in the pooled OS analysis (I ² = 98.1%), along with moderate‐to‐high heterogeneity in other outcomes, validating the use of a random‐effects model. A meta‐regression was conducted to explore potential sources of heterogeneity; no specific factor emerged as a significant contributor. These findings indicate that the high heterogeneity likely reflects multiple overlapping influences, including variability in patient populations and treatment strategies, rather than a single determinant. Consequently, the present results require cautious interpretation in the context of such heterogeneity. Publication bias was evaluated using funnel plot visualization and Egger's weighted linear regression test. The funnel plot showed no significant asymmetry, and Egger's test revealed no statistical evidence of bias (p = 0.77). These findings suggest that publication bias did not markedly influence the OS results. However, the potential for bias cannot be entirely excluded, as meta‐analyses of observational studies often include a higher proportion of studies with positive results. Additionally, the relatively small number of included studies may have limited the power of the funnel plot and Egger's test to detect bias. Thus, caution should be exercised in interpreting these results despite their general reliability. Because this study was a published data‐based meta‐analysis, competing risk analyses using individual patient data could not be performed. As a result, the tumor‐related emergency intervention rate in the NR group was presented only as a crude cumulative proportion, without adjustment for death as a competing event. This limitation may have underestimated the true risk and imposed restrictions on the interpretation of the findings. Future studies incorporating individual patient data and competing risk analyses are warranted.

5. Conclusion

The role of PTR in asymptomatic patients with unresectable mCRC remains unclear. Careful patient selection based on clinical background and treatment responsiveness is essential. However, the present findings indicate that PTR does not contribute to improved OS, and current evidence does not support the routine implementation of PTR in this patient population.

Author Contributions

Junpei Takashima: conceptualization, methodology, data curation, investigation, writing – original draft, writing – review and editing, formal analysis, visualization. Hirotoshi Kobayashi: conceptualization, methodology, data curation, validation, writing – review and editing, formal analysis, project administration. Daisuke Fujimoto: data curation, validation, supervision, writing – review and editing, visualization. Fumihiko Miura: data curation, validation, supervision, writing – review and editing, project administration.

Funding

The authors have nothing to report.

Ethics Statement

This study was conducted as a systematic review of previously published literature and did not involve any new studies with human participants or animals performed by the authors. Therefore, ethical approval was not required. The review protocol was prospectively registered in the International Prospective Register of Systematic Reviews.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

This study is a systematic review based on previously published data. All data analyzed during this study are available in the original articles cited within this publication. No new data were generated by the authors.