Palliative resection versus palliative stenting for intestinal obstruction in patients with metastatic left-sided colonic cancer: a propensity score–matched analysis Authors: Ruby Siu Ting Lau, Sophie Sok Fei Hon, Man Fung Ho, Simon Chu, Dennis CK Ng, Simon SM Ng Published: 2025 ABSTRACT AbstractPurposePalliative resection and palliative stenting are established options for managing obstruction in patients with metastatic left-sided colonic cancer. This retrospective study investigated the long-term outcomes and survival associated with each treatment modality.MethodsPatients with left-sided colon cancer complicated by intestinal obstruction and unresectable metastatic lesions were included. Propensity score matching was conducted to balance demographic characteristics. The primary outcome was long-term survival. Secondary outcomes included short-term morbidity, length of hospital stay, clinical success rate, stoma formation rate, and number of readmissions due to tumor-related complications.ResultsInitially, 131 patients who underwent palliative resection or stenting between 2015 and 2022 were included. After propensity score matching, 98 patients remained (49 in each group). Survival was significantly better among patients receiving palliative resection compared to stenting (median, 19.6 months vs. 9.6 months; P=0.003). However, subgroup analysis for patients older than 70 years demonstrated no statistically significant survival benefit (median, 11.5 months vs. 10.2 months; P=0.240). The resection group experienced significantly higher rates of stoma formation and longer postoperative hospital stays. Readmission rates were similar. Cox regression analysis identified low carcinoembryonic antigen levels, tumor resection, chemotherapy, and targeted therapy as independent predictors of longer survival.ConclusionFor metastatic colon cancer patients presenting with intestinal obstruction, palliative resection may offer a survival advantage. However, this benefit diminishes in patients over 70 years of age. Additionally, resection is associated with a higher rate of stoma formation. Therefore, individualized treatment decisions are warranted when choosing between palliative resection and palliative stenting in metastatic colonic cancer patients.Keywords: Colectomy, Colon neoplasms, Secondary, Self expandable metallic stents INTRODUCTION Colorectal cancer is the third most common cancer worldwide and the second most common in Hong Kong [1, 2]. Oncological therapy is the first-line treatment for patients with unresectable metastatic colorectal cancer. It has been reported that 10% to 18% of patients with colorectal cancer present with obstruction at initial diagnosis [3], while others develop obstructive symptoms during treatment. In metastatic colorectal cancer, both palliative resection and self-expanding metallic stents are possible treatment options for large bowel obstruction [4]. Emergency surgery for malignant colonic obstruction is associated with morbidity rates of 40% to 60% and mortality rates of 3% to 11%, as well as a high rate of stoma formation [5]. European guidelines recommend colonic stenting for the palliation of malignant colonic obstruction [6]. Previous studies have shown that palliative stenting is effective, with high success rates in relieving obstruction, shorter hospital stays (5 days vs. 12 days), lower stoma rates, and earlier commencement of chemotherapy (4 weeks vs. 7 weeks) compared to surgery [7–9]. Some studies have reported that emergency surgery following failed stenting may be associated with worse complications than upfront emergency surgery [10]. While short-term outcomes may be promising, there is limited and controversial data regarding long-term results and survival impact [4]. In addition to survival, patients who undergo palliative stenting may experience other tumor-related complications such as bleeding, infection, and perforation, which can require repeated readmissions and pose significant physical and psychological burdens. This retrospective study, including data from 2 high-volume centers in Hong Kong, aims to evaluate the long-term outcomes among patients receiving palliative stenting and palliative resection. METHODS Ethics statement The study protocol was reviewed and approved by the Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee (No. 2025.102). The requirement for informed consent was waived due to the use of deidentified data and the retrospective nature of the study. Patient selection Clinical data were retrospectively collected for patients who underwent palliative stenting or palliative resection for metastatic colorectal cancer at 2 hospitals in Hong Kong (Prince of Wales Hospital and North District Hospital) between 2015 and 2022. Patients were included if they had intestinal obstruction caused by underlying unresectable metastatic left-sided colorectal cancer. Left-sided colonic cancer was defined as tumors arising from the splenic flexure or beyond. The diagnosis of intestinal obstruction and metastatic status was confirmed by computed tomography. Tumor resectability was determined by a specialist in general surgery. Exclusion criteria included the following: right-sided colonic tumors; low rectal tumors not suitable for stenting (defined as tumors located within 5 cm from the anal verge or palpable low rectal tumors as assessed by a colorectal surgery specialist); other tumor complications such as perforation, abscess formation, or closed-loop obstruction with bowel ischemia, in which stenting was contraindicated; patients who were not surgical candidates and thus unable to undergo palliative resection; and cases in which only diversion stoma, rather than resection, was performed. Data collection Preintervention demographic data (age, sex, tumor location, site of metastasis, American Society of Anesthesiologists [ASA] physical status, medical comorbidities, albumin, carcinoembryonic antigen [CEA], and hemoglobin levels), as well as short- and long-term outcomes (technical and clinical success rates, 30-day mortality, length of hospital stay, postoperative complications, overall survival), subsequent chemotherapy or targeted therapy, and number of readmissions due to tumor complications were collected. All data were extracted from the hospitals’ electronic health records systems. Statistical analysis Categorical variables were compared using the chi-square test or Fisher exact test, as appropriate. Continuous variables were compared using Student t-test or Mann-Whitney U-test. Survival was analyzed using the Kaplan-Meier method, and differences in survival were assessed with the log-rank test. Variables with a P-value of <0.1 in univariate analysis were included in a multivariate Cox proportional hazards regression to identify risk factors for poor outcomes and survival. A P-value of <0.05 was considered statistically significant. All analyses were performed using IBM SPSS ver. (IBM Corp). Propensity score matching was performed between the stenting and resection groups to minimize bias related to patient background and tumor characteristics. Nearest neighbor matching was conducted without replacement at a 1:1 ratio, using a caliper calculated as 0.2 times the standard deviation of the logit of the propensity score. The matched baseline variables included sex, age, ASA physical status, use of chemotherapy or targeted therapy, location of primary and metastatic tumors, and CEA level. Differences in outcomes between the matched cohorts were assessed using the chi-square test to confirm that there were no statistically significant differences between the 2 groups. RESULTS Clinical characteristics of patients During the study period from 2015 to 2022, data from a total of 131 patients were identified according to the inclusion and exclusion criteria. After propensity score matching, there were 49 patients in each group. There were no statistically significant differences in baseline demographics, including sex, age, site of primary tumor, site of metastasis, ASA physical status, albumin, hemoglobin, and CEA levels (Table 1). There was also no difference in the percentage of patients receiving palliative chemotherapy or targeted therapy (including both anti–epidermal growth factor receptor [EGFR] and anti–vascular endothelial growth factor therapy [VEGF] agents) between the groups. However, a higher proportion of patients in the resection group received anti-VEGF therapy (22.4% vs. 8.2%, P=0.050). All patients were followed regularly until death. Table 1.Baseline demographics of patientsCharacteristicBefore propensity score matchingAfter propensity score matchingResection (n=70)Stenting (n=61)P-valueSMDResection (n=49)Stenting (n=49)P-valueSMDAge (yr)64.5±12.170.3±13.70.011*0.45164.6±12.868.3±13.60.6800.278Sex0.005*0.5110.2230.248 Male46 (65.7)25 (41.0)30 (61.2)24 (49.0) Female24 (34.3)36 (59.0)19 (38.8)25 (51.0)Site of primary tumor0.7840.0410.1020.044 SF/descending colon13 (18.6)12 (19.7)7 (14.3)8 (16.3) Sigmoid44 (62.9)35 (57.4)32 (65.3)29 (59.2) RSJ/rectum13 (18.6)14 (23.0)10 (20.4)12 (24.5)Site of metastasis0.5960.0430.4040.067 Liver25 (35.7)20 (32.8)18 (36.7)15 (30.6) Lung7 (10.0)2 (3.3)4 (8.2)1 (2.0) Peritoneum5 (7.1)5 (8.2)5 (10.2)3 (6.1) Multiple23 (32.9)23 (37.7)16 (32.7)20 (40.8) Other10 (14.3)11 (18.0)6 (12.2)10 (20.4)ASA physical status0.1410.0800.7330.032 I10 (14.3)6 (9.8)6 (12.2)5 (10.2) II38 (54.3)27 (44.3)26 (53.1)24 (49.0) III22 (31.4)25 (41.0)17 (34.7)19 (38.8) IV0 (0)3 (4.9)0 (0)1 (2.0)Past medical history Cardiac disease9 (12.9)12 (19.7)0.2890.1856 (12.2)8 (16.3)0.5640.117 Renal disease2 (2.9)2 (3.3)0.8890.0242 (4.1)2 (4.1)>0.9990 Pulmonary disease2 (2.9)8 (13.1)0.027*0.3852 (4.1)6 (12.2)0.1400.302 Diabetes16 (22.9)18 (29.5)0.3860.15110 (20.4)12 (24.5)0.6280.098 Hypertension16 (22.9)27 (44.3)0.009*0.46514 (28.6)20 (40.8)0.2030.259Biochemical marker Albumin (g/L)33.0±5.730.5±7.30.028*0.38633.2±7.330.8±7.30.2190.365 Hemoglobin (g/dL)11.4±2.111.4±1.80.9890.00511.3±2.211.5±1.90.3670.077 CEA (µg/L)38 (7.1–250.8)81.5 (14.5–494.8)0.0550.33435 (6.1–208.0)57 (9.7–194.5)0.3480.173 Chemotherapy46 (65.7)24 (39.3)0.003*0.54730 (61.2)22 (44.9)0.1050.332 Targeted therapy29 (41.4)10 (16.4)0.002*0.57418 (36.7)10 (20.4)0.0740.367 Anti-VEGF20 (28.6)4 (6.6)0.001*0.60411 (22.4)4 (8.2)0.0500.404 Anti-EGFR9 (12.9)6 (9.8)0.7840.0957 (14.3)6 (12.2)0.7660.060Values are presented as mean±standard deviation, number (%), or median (interquartile range). Percentages may not total 100 due to rounding.SMD, standardized mean difference; SF, splenic flexure; RSJ, rectosigmoid junction; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen; VEGF, vascular endothelial growth factor therapy; EGFR, epidermal growth factor receptor.*P<0.05. Baseline demographics of patients Values are presented as mean±standard deviation, number (%), or median (interquartile range). Percentages may not total 100 due to rounding. SMD, standardized mean difference; SF, splenic flexure; RSJ, rectosigmoid junction; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen; VEGF, vascular endothelial growth factor therapy; EGFR, epidermal growth factor receptor. P<0.05. Short-term outcomes Short-term outcomes, including clinical success, postoperative complications, 30-day mortality, and length of hospital stay, are summarized in Table 2. Clinical success was defined as the return of gut function, evidenced by tolerance of an oral diet (soft diet or above) and the presence of bowel opening after the primary intervention during the same hospital stay. The clinical success rate was high for both cohorts: 100% in the resection group and 89.8% in the stenting group. Although there was no significant difference between the groups (P=0.056), there was a trend toward a higher success rate in the resection group. Among the 5 unsuccessful stenting cases, 3 (6.1%) experienced stent-related perforation, 1 stent (2.0%) failed to deploy, and 1 stent (2.0%) migrated distally. All 5 patients underwent tumor resection and end colostomy as salvage procedures. Table 2.Short-term outcomes of palliative resection and palliative stentingOutcomeBefore propensity score matchingAfter propensity score matchingResection (n=70)Stenting (n=61)P-valueResection (n=49)Stenting (n=49)P-valueClinical success69 (98.6)53 (86.9)0.008*49 (100)44 (89.8)0.056Operation time (min)203.5±66.2--203.5±65.3--Postoperative complication18 (25.7)12 (19.7)a0.41211 (22.4)8 (16.3)0.443 Anastomotic leak0 (0)0 (0)0 (0)0 (0) Pneumonia2 (2.9)6 (9.8)1 (2.0)4 (8.2) Urinary tract infection2 (2.9)3 (4.9)1 (2.0)2 (4.1) Abdominal collection1 (1.4)1 (1.6)1 (2.0)1 (2.0) Ileus10 (14.3)2 (3.3)6 (12.2)0 (0) Surgical site infection3 (4.3)1 (1.6)2 (4.1)1 (2.0)Stent complication-8 (13.1)--5 (10.2)- Failure to deploy2 (3.2)1 (2.0) Stent migration1 (1.6)1 (2.0) Perforation5 (8.2)3 (6.1)Clavien-Dindo classification19 (27.1)--12 (24.5)-- Grade I6 (8.6)4 (8.2) Grade II11 (15.7)7 (14.3) Grade III0 (0)0 (0) Grade IV1 (1.4)1 (2.0) Grade V1 (1.4)0 (0)Surgical reintervention1 (1.4)7 (11.5)0.028*0 (0)5 (10.2)0.059Length of stay (day)10 (8–16.5)8 (6–12.5)0.009*10 (8–17)8 (6–12.5)0.005*30-Day mortality2 (2.9)6 (9.8)0.0961 (2.0)3 (6.1)0.617Values are presented as number (%), mean±standard deviation, or median (interquartile range).*P<0.05.aThe sum of individual complications may exceed the total number of patients because some patients experienced more than one complication. Short-term outcomes of palliative resection and palliative stenting Values are presented as number (%), mean±standard deviation, or median (interquartile range). P<0.05. The sum of individual complications may exceed the total number of patients because some patients experienced more than one complication. Postoperative complications within 30 days occurred in 11 patients (22.4%) in the resection group and in 8 patients (16.3%) in the stenting group, with no significant difference between groups (P=0.443). Most postoperative complications were Clavien-Dindo classification grade I or II, except for 1 patient who suffered a grade IV complication due to pneumonia requiring intensive care unit admission; he subsequently recovered fully and was discharged home. One patient (2.0%) in the resection group and 3 (6.1%) in the stenting group died within 30 days, but the difference was not statistically significant (P=0.617). The patient in the resection group died of sudden cardiac arrest during rehabilitation. Among the stenting group, 1 patient died of disease progression, while 2 died due to chest infection. The median length of hospital stay was 10 days for the resection group and 8 days for the stenting group, the latter being significantly shorter (P=0.005). Long-term outcomes Overall survival in the resection group was significantly better than in the stenting group, with a median survival of 19.6 and 9.6 months, respectively (P=0.003, log-rank test) (Fig. 1). However, in the subgroup analysis of patients older than 70 years (n=43, 24 received stenting and 19 received surgery), survival between the groups was comparable, with a median survival of 11.5 and 10.2 months, respectively (P=0.240, log-rank test) (Fig. 2). In the Cox proportional hazards regression analysis, chemotherapy and targeted therapy were independent protective factors, while high CEA level and stenting were adverse factors affecting overall survival (Table 3). Survival analysis for patients of all ages. Survival analysis for patients aged 70 or above. Table 3.Prognostic factors for overall survivalFactorUnivariate analysisMultivariate analysisHazard ratio (95% CI)P-valueHazard ratio (95% CI)P-valueSex Male1 (Reference)1 (Reference) Female0.656 (0.42–1.03)0.066a0.63 (0.39–1.02)0.061Age (yr)1.01 (0.99–1.03)0.095a0.98 (0.98–1.02)0.766Primary tumor site RSJ/rectum1 (Reference)- SF/descending colon1.53 (0.75–3.11)0.246 Sigmoid1.45 (0.83–2.51)0.734Site of metastasis Other1 (Reference)- Liver1.02 (0.53–1.94)0.963 Lung0.20 (0.03–1.52)0.120 Peritoneum1.10 (0.46–2.64)0.831 Multiple1.54 (0.81–2.93)0.191ASA physical status IV1 (Reference)- I0.37 (0.04–3.06)0.354 II0.78 (0.11–5.71)0.807 III0.73 (0.10–5.38)0.755Cardiac disease1.77 (0.96–3.27)0.066a1.12 (0.56–2.15)0.733Renal disease1.18 (0.43–3.22)0.753Pulmonary disease1.46 (0.70–3.05)0.311Diabetes1.01 (0.61–1.65)0.985Hypertension1.14 (0.73–1.77)0.563Albumin (g/dL)0.98 (0.95–1.91)0.137Hemoglobin (g/dL)0.97 (0.85–1.08)0.601CEA (µg/L)1.00 (1.00–1.00)0.007a,*1.00 (1.00–1.00)0.012*Chemotherapy0.41 (0.26–0.64)<0.001a,*0.58 (0.34–0.99)0.048*Targeted therapy0.39 (0.23–0.64)<0.001a,*0.48 (0.25–0.90)0.022*Intervention Resection1 (Reference)-1 (Reference) Stenting1.93 (1.25–2.98)0.003a,*1.77 (1.10–2.83)0.018*CI, confidence interval; RSJ, rectosigmoid junction; SF, splenic flexure; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen.aVariables with a P-value of <0.1 in univariate analysis were entered into a multivariate Cox proportional hazards regression model.*P<0.05 Prognostic factors for overall survival CI, confidence interval; RSJ, rectosigmoid junction; SF, splenic flexure; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen. Variables with a P-value of <0.1 in univariate analysis were entered into a multivariate Cox proportional hazards regression model. P<0.05 Cumulative stoma formation was significantly higher in the resection group compared to the stenting group (P<0.001). In the resection group, 35 patients (71.4%) underwent Hartmann procedure, while 7 patients (14.3%) in the stenting group required stoma formation during the disease course. Among the 7 patients, 5 had failed stenting and underwent stoma formation during the index admission. One patient experienced stent blockage after 21 months and required transverse colostomy. Another patient had stent blockage after 7 months, underwent re-stenting, but developed another stent blockage 4 months later and ultimately required transverse colostomy. Patients could have multiple subsequent hospital admissions due to tumor-related or non–tumor-related causes. Only tumor-related causes, including tumor perforation, anemia requiring blood transfusion, intestinal obstruction, tumor pain, and tumor infection, were included in the analysis. Readmissions related to metastases, such as ascites, obstructive uropathy, or malignant biliary obstruction, were also analyzed. Overall, 21 patients after surgery and 30 patients after stenting required readmission for various reasons. There was no statistically significant difference between the groups (P=0.069). However, there was a higher incidence of readmission due to anemia requiring blood transfusion in the stenting group (18.4% vs. 2.0%), although this difference did not reach statistical significance. DISCUSSION Palliative stenting is a potential solution for patients with left-sided metastatic colorectal cancer who present with obstruction and wish to avoid stoma formation. Four systematic reviews or meta-analyses have compared colonic stenting and surgery for the palliation of malignant colonic obstruction [4, 11–13]. The technical success rate of colonic stenting ranged from 88% to 100%. Benefits of endoscopic colonic stenting in patients with metastatic large bowel obstruction include shorter hospital stay, lower intensive care unit admission rate, earlier initiation of chemotherapy, and lower rates of stoma formation. No significant differences in overall morbidity have been observed between stent and surgery groups [12, 13]. Our study supports previous literature, showing that palliative stenting provides better short-term outcomes, such as shorter hospital stays and lower stoma rates. However, prior studies have been heterogeneous and controversial regarding long-term outcomes after palliative stenting. Therefore, our study’s primary endpoint was to evaluate long-term survival in this patient group. In our cohort, long-term survival was better in the palliative resection group, even though a comparable number of patients in both groups received systemic oncological treatment. Based on these findings, we suggest that palliative resection may be a better option for managing metastatic colorectal cancer patients with obstructing symptoms, as it is associated with longer survival in our setting. In addition to survival, there was also a trend toward lower transfusion requirements in the resection group. Notably, the survival benefit diminished in patients older than 70 years. Although we could not demonstrate age as an independent prognostic factor, we believe overall survival in elderly patients may be more limited by comorbidities than by malignancy itself—an important consideration in the shared decision-making process. Nevertheless, a larger sample size is needed before establishing a definitive age cutoff for patients who may benefit from palliative resection. Chemotherapy is a well-established positive prognostic factor for survival [14]. Previous studies have favored palliative stenting for enabling earlier initiation of chemotherapy. In our study, however, the shorter hospital stay in the stenting group was not clinically significant—the difference was only 2 days, compared to a 1-week difference in prior international studies [8]. Moreover, the timing of chemotherapy initiation was not documented in our cohort and was influenced by various logistical factors in local hospitals. As such, we cannot conclude that patients in the stenting group received chemotherapy earlier than those in the resection group, nor that this affected outcomes. Furthermore, the duration and choice of chemotherapy were not included in the study, which could impact patient survival. For example, patients willing to receive second- or third-line chemotherapy might achieve longer survival. Several retrospective studies have also indicated an increased risk of perforation when anti-VEGF agents are used [15]. Thus, leaving the primary tumor in situ may limit the use of targeted therapy and potentially affect survival. From our results, chemotherapy and targeted therapy were protective factors for longer survival, while higher CEA levels and stenting were adverse predictors. Therefore, for long-term survival, patients at low surgical risk, especially those younger than 70 years, with lower CEA levels, and who are motivated to undergo prolonged chemotherapy or targeted therapy, should be counseled toward palliative resection to optimize survival. Conversely, stoma care remains a challenge and burden for many patients. Many report a decreased quality of life due to stoma formation [16]. As in previous studies, our palliative resection group had a higher rate of stoma formation. Therefore, patients’ acceptance of a stoma should also be respected during treatment decision-making. Patients may also experience complications from retaining the primary tumor in situ. Repeated admissions for various reasons impose both physical and psychological burdens on patients and their families. However, the sample size in our study was too small to demonstrate statistical differences in readmission rates. Limitations There are several limitations to our study. First, as a retrospective study, there is potential selection bias between the 2 groups. Although propensity score matching reduced the standardized mean difference, small-to-moderate residual imbalance in some covariates remains (Table 1). Completely balancing the groups is impossible; for example, quantifying the total disease and metastasis burden is challenging and may affect survival. We matched patients using metastasis location and CEA level to minimize differences, but the median CEA was still lower in the resection group than the stenting group (35 μg/L vs. 57 μg/L, P=0.348; standardized mean difference, 0.173), possibly reflecting greater disease burden in the latter. As CEA level is an independent adverse prognostic factor, this may partially explain the worse survival in the stenting group. Another factor that could not be balanced was administration of targeted therapy, as leaving the tumor in situ may limit options, especially anti-VEGF agents. More patients in the resection group received anti-VEGF therapy (P=0.05), which could have contributed positively to survival. Second, the sample size was small: after propensity score matching, only 49 patients remained in each group, limiting the ability to detect significant differences in postoperative morbidity, readmission, and transfusion rates. Third, patients’ functional status was not documented, even though it is an important preoperative consideration. Large-scale prospective studies focused on long-term outcomes in such heterogeneous groups may not be feasible. Ultimately, long-term survival in metastatic colorectal cancer is multifactorial. Conclusions Our study suggests that both resection and stenting are viable options for patients with acute left-sided colonic obstruction, although each has drawbacks. Survival is better after resection, while stenting results in a lower stoma formation rate. However, stenting can limit the use of targeted therapy and carries risks of reocclusion and tumor-related complications in the long term. Patient fitness, expectations, tumor burden, and tumor biology must all be considered. Patients with reasonable surgical risk and relatively younger age may achieve better survival with palliative resection. For those who are older, have a higher disease burden, or are less motivated for oncological treatment, stenting remains a reasonable management option. Patient acceptance of stoma is also an important factor in decision-making. Based on our results, patients can be better counseled regarding optimal palliative treatment choices. Footnotes Conflict of interest Simon SM Ng is an editorial board member of this journal, but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflict of interest relevant to this article was reported. Funding None. Acknowledgments The authors thank the contributions of participating medical centers, research teams, technology providers, and patient volunteers who made this study possible. Author contributions Conceptualization: RSTL, MFH; Data curation: RSTL, MFH; Formal analysis: RSTL; Investigation: RSTL; Methodology: RSTL, SSFH; Project administration: RSTL; Resource: SSFH, SC, DCKN, SSMN; Software: RSTL; Supervision: SSMN; Validation: SSFN; Visualization: RSTL; Writing–original draft: RSTL; Writing–review & editing: all authors. All authors read and approved the final manuscript. REFERENCES