Prognosis of Laparoscopic Surgery for Colorectal Cancer in Middle-Aged Patients: A Propensity Score-Matched Analysis


 Background: The prognosis of middle-aged patients with CRC treated by laparoscopic resection (LR) is unclear. This study aimed to evaluate the survival outcomes of LR compared with open resection (OR) for patients with CRC and 45-65 years of age.Methods: This retrospective cohort study used the data from a database of all consecutive colorectal resections performed between January 2009 and December 2017. Propensity score matching (PSM) was done to handle the selection bias based on age, gender, body mass index (BMI), tumor location, AJCC stage, and admission year. Univariate and multivariate COX regression model was used to identify risk factors of overall survival (OS) and progression-free survival (PFS).Results: After PSM, 217 patients were included in each group. There were no differences in OS and PFS between the two groups (all P>0.05). There was less blood loss for LR (P<0.001), but the other complications were similar between the two groups. The multivariate analysis showed that high histological grade (hazard ratio [HR]=2.262, 95%CI: 1.334-3.836, P=0.002), stage III (HR=1.744, 95%CI: 2.360-25.406, P=0.001), stage IV (HR=47.905, 95%CI: 14.430-159.033, P<0.001), and adjuvant therapy (HR=0.547, 95%CI: 0.358-0.838, P=0.006) were independently associated with OS. High preoperative CEA (HR=1.585, 95%CI: 1.049-2.394, P=0.029), high histological grade (HR=2.128, 95%CI: 1.272-3.558, P=0.004), stage III (HR=5.562, 95%CI: 1.980-15.624, P=0.001), and stage IV (HR=26.338, 95%CI: 9.090-76.315, P<0.001), were independently associated with OS. LR was not associated with OS and PFS.Conclusions: In middle-aged patients with CRC, OR and LR have similar survival outcomes and complications.

Therefore, this study aimed to evaluate the survival outcomes of laparoscopic surgery compared with open surgery for Chinese patients with CRC and 45-65 years of age.

Study design and patients
This retrospective cohort study used the data from a database of all consecutive colorectal resections performed at the Department of General Surgery of Mianyang Central Hospital, Sichuan Province, China, between January 2009 and December 2017. The study followed the ethical standards of the Declaration of Helsinki, as revised in 2013. This study was approved by the ethics committee of Mianyang Central Hospital. The requirement for individual consent was waived by the committee because of the retrospective nature of this study.
The inclusion criteria were: 1) radical (R0) D3 lymphadenectomy (TNM stage I-III) or palliative surgery (TNM stage IV); 2) histopathologically proven primary CRC; 3) 45-65 years of age (based on the quartiles of onset age of CRC in our database) (13); 4) no history of cancer before the CRC included in this study; 5) histological type included adenocarcinoma, signet ring cell carcinoma, and mucinous adenocarcinoma; and (6) American Society of Anesthesiologists (ASA) class I-III.
The exclusion criteria were: 1) emergency surgery; 2) neoadjuvant treatment; 3) synchronous tumors; 4) familial adenomatous polyposis; 5) in ammatory bowel diseases; 6) primary tumor unresected; or 7) for patients with stage IV cancers, the primary tumors were not resected, or both primary and metastatic lesions were removed (including concurrent and secondary metastasis resection).

Surgery and grouping
The patients were grouped according to laparoscopic resection (LR) or open resection (OR). Patients for whom the LR was converted into OR were analyzed as part of the OR group. All surgical procedures were performed by an experienced surgeon (> 50 colorectal cancer operations/year for more than 5 years) using standardized techniques. All operations were performed or assisted by one of the team's six colorectal surgeons. The choice of the procedure was based on the patient's preference after an informed, comprehensive discussion.
Conversion to an open procedure was de ned as using an abdominal incision larger than necessary for specimen retrieval. According to the Japanese Society for Cancer of the Colon and Rectum, some patients underwent D3 lymphadenectomy, and others did not (17). All laparoscopic procedures were performed through a standardized medial-to-lateral approach (18). Total mesorectal excision (TME) was performed for patients with rectal cancer below the peritoneal re ection.

Adjuvant therapy
According to the patient's postoperative general condition or compliance and the physicians' experience, the patients with T3-4 or stage III-IV disease were considered for 5-uorouracil-based chemotherapy. Patients with rectal cancer were considered for radiotherapy using a four-eld box technique to the pelvis consisting of 45-50.4 Gy in 25-28 fractions (19).

Postoperative complications
Postoperative complications were those occurring within 30 days after surgery. Anastomosis-related complications (leakage, stenosis, or intraluminal bleeding) were con rmed by X-ray, endoscopy, or angiography. Intra-abdominal collections and abscesses were proven by ultrasound or computed tomography scans and concomitant systemic in ammatory response lasting ≥ 24 h. Postoperative hemorrhage was de ned as a blood loss of > 300 mL according to the drainage volume. The severity of postoperative complications was assessed according to the Clavien-Dindo classi cation (20).

Data collection
Preoperative variables (age, gender, tumor location, and preoperative CEA), intraoperative data (operation time, blood loss, and the number of lymph nodes harvested), and postoperative data (largest tumor diameter, American Joint Committee on Cancer (AJCC) stage, histological grade, postoperative complications, and length of postoperative hospital stay) were recorded.
The tumor was staged according to the seventh American Joint Commission on Cancer (AJCC) TNM classi cation (21). Tumor location was classi ed as right-sided tumors (cecum, ascending colon, hepatic exure, and transverse colon), left-sided tumors (splenic exure, descending colon, and sigmoid colon), and rectal cancer.

Follow-up
The patients were routinely followed at 3-month intervals for the rst 2 years, at 6-month intervals for the next 3 years, and then annually. The last follow-up was in December 2020. Local recurrence was de ned as a recurring tumor limited to the previous tumor resection site or adjacent organs. Distant recurrence was de ned as any tumor recurrence outside the primary site of the disease. Overall survival (OS) was measured from the date of surgery to death from any cause. Progression-free survival (PFS) was measured from the time of operation to the date of progression or death from any cause. Lost to follow-up was de ned as the complete impossibility of seeing or contacting the patients, either by phone, mail, or e-mail.
Propensity score matching (PSM) To mitigate the selection bias due to the retrospective nature of this study, PSM was done to handle the selection bias and to estimate the prognosis of the two groups. Propensity scores were obtained by logistic regression analysis of age, gender, body mass index (BMI), tumor location, AJCC stage, and admission year. One-to-one pair matching without replacement and a nearest-neighbor matching algorithm with calipers less than 0.1 were performed using SPSS 26 (IBM, Armonk, NY, USA).

Statistical analysis
Data are presented as means ± standard deviations for continuous variables and as numbers and percentages for categorical variables. The characteristics of the two groups were compared using the independent t-test for continuous variables and either the chi-square test or Fisher's exact test for the categorical variables. Survival analysis was performed using the Kaplan-Meier method and compared using the log-rank test. The patients lost to follow-up were censored at their last visit. Variables with P < 0.05 in the univariate analyses were entered into a multivariate Cox regression analysis using the forward method. Statistical analyses were performed using SPSS 26.0 (IBM, Armonk, NY, USA). Two-sided P-values < 0.05 were considered statistically signi cant.

Results
Clinicopathological characteristics Figure 1 presents the owchart of the patients' enrollment. A total of 997 patients were identi ed as potentially eligible. Patients with emergency surgery (e.g., acute intestinal obstruction or perforation and acute bleeding) (n = 184), familial adenomatous polyposis (FAP) (n = 4), synchronous tumors (n = 19), preoperative adjuvant treatment (n = 121), in ammatory bowel diseases (n = 5), other types of cancers (n = 5) (except for adenocarcinoma, signet ring cell carcinoma, and mucinous adenocarcinoma), without resection of the primary tumor (n = 4), and patients who underwent resection of the primary tumor and metastasis in two different surgical sessions (n = 18) were excluded. Finally, 637 patients were included.

Survival
With a median follow-up of 46 months (range: 3-139) in the total cohort (before matching), 11 patients (3.0%) in the OR group and 14 patients (5.3%) in the LR group were lost to follow-up (P = 0.136).
Compared with the OR group in the total cohort, there were better survival outcomes in the LR group for 5-year OS (P = 0.009) and 5-year PFS (P = 0.003) (Fig. 2); these differences were not observed after PSM (P = 0.458 and P = 0.309) (Fig. 2). Subgroup analyses revealed no signi cant difference between the two groups for stages I, II, and III diseases (both before and after PSM) for 5-year OS (all P > 0.05) and PFS (all P > 0.05) (Figs. 3 and 4). For stage IV disease, the 5-year OS (P = 0.005) and PFS (P = 0.018) were signi cantly better in the LR group than in the OR group before PSM, but the differences were not signi cant after PSM (P = 0.052 and P = 0.183) (Figs. 3  and 4). No signi cant differences were observed between the two groups in subgroup analysis for right-sided and left-sided tumors regarding 5-year OS (all P > 0.05) and PFS (all P > 0.05) (Figs. 5 and 6). For rectal cancer, the 5-year OS (P = 0.010) and PFS (P = 0.005) were signi cantly better in the LR group than in the OR group before PSM, but the differences were not signi cant after PSM (P = 248 and P = 0.137) (Figs. 5 and 6).

Prognostic factors
The  In this study of middle-aged patients with CRC, there were no differences in OS and DFS in all patients after matching with age, gender, BMI, tumor location, AJCC stage, and admission year. LR has become an accepted therapeutic option for CRC patients, providing similar OS and DFS compared with OR in stages I-III disease (22)(23)(24)(25)(26)(27)(28)(29), supporting the present results. Previous studies in patients with CRC (irrespective of age, but most studies included older patients) showed that LR was at least comparable to OR, as shown by two meta-analyses (7,30), supporting the present study. Jayne et al. (23) reported that LR maximized the short-term outcomes compared with OR without comprising the long-term outcomes. Long-term results by Green et al. (31) also support the use of LR for CRC. On the other hand, Fabio et al. (32) reported that the 5-year cancer-related survival after LR was signi cantly higher than that after OR. Similar results were reported by Law et al. (33). Park et al. (34) reported that similar outcomes between LR and OR in T4 tumors, but a tendency for better outcomes with LR in tumors < 4 cm. Hence, contradictory results can be observed in the literature. Middle-age patients are tter for surgery and have fewer comorbidities than older patients, which could explain why there were no differences between LR and OR in the present study. LR is known to minimize the surgical stress and systemic in ammatory response after surgery (8), which is conducive to reduce complications and improve recovery. Hypotheses have been proposed to explain this bene cial oncological role of LR in the treatment of CRC. One possible reason is that the numbers of LNs harvested by laparoscopy were higher than by OR (32,33), but in the present study, the mean number of LNs harvested was not different between the two groups. Still, the number of harvested LNs was small but supported by previous studies (35)(36)(37)(38)(39).
In the present study, LR showed an advantage over OR in stage IV CRC before PSM, but the differences were lost after PSM. The present study is supported (at least before PSM) by Day et al. (40), who reported 665 resections (457 LR and 208 OR) for CRC, and the 5-year OS for non-stage IV disease in the LR group was signi cantly better than in the open group (79.4% vs. 74.0%, P = 0.03). The role of LR in stage IV CRC patients with unresectable metastases has not been su ciently evaluated. The National Comprehensive Cancer Network (NCCN) guidelines recommend that asymptomatic patients with metastatic CRC should receive chemotherapy and that patients with symptomatic or curable metastatic disease should be considered for surgery (41). Verheijen et al. (42) reported that short-term outcomes after laparoscopic surgery for stage IV colorectal cancer in selected patients are equivalent to those for stage I. Previous studies reported that LR for stage IV CRC had equivalent long-term outcomes to OR (43,44), supporting the present study after PSM, but Hida et al. (45) reported that the OS after LR was better than that after OR, but no difference was apparent in the multivariable analysis.
The surgical approach was not independently associated with better OS and PFS in the multivariable analyses. These results are in contradiction with studies in elderly patients (46, 47), but results in middle-aged patients are scarce, and the comorbidities probably play roles in this discrepancy. One study suggests that LR achieved similar outcomes among elderly and middle-aged patients with rectal cancer (48). The present study is supported by the literature because it is known that middle-aged patients have speci c risk factors (14,15) and better prognosis because of better access to treatments and fewer comorbidities (16).
In this study, LR showed an advantage over OR in rectal cancer before PSM, but the differences disappeared after PSM. In rectal cancer, some studies revealed that LR and OR have similar effects on long-term survival (49,50). On the other hand, Nonaka et al. (51) reported that DFS was better in the LR group than in the OR group. Ng et al. (52) reported a trend toward a lower recurrence rate at 10 years after LR than after OR among patients with stage III rectal cancer (P = 0.078). These points will have to be examined in future studies because the numbers of stage IV and rectal cancer patients were small in this study.
The complications were similar between LR and OR in this study, except for less blood loss with LR, supported by the general principle of LR (4-8) and by a previous meta-analysis (53) and a cohort study (54). A recent study showed that LR and OR was safe both in elderly and non-elderly patients (55). This lack of difference in complications might play a role in the lack of difference in survival. The conversion rate was 3%, similar to that of a previous study (2%) (29).
There are some limitations that are inherent to retrospective studies (such as missing data and various biases), but the present study used the PSM approach, which allows for more robust analysis by reducing confounders (56), but without completely mitigating all sources of bias of non-randomized studies. This is particularly useful when a randomized controlled trial is not possible due to practical, medical, or ethical issues. Moreover, this study was based on a single-institution series, and the generalizability of the results might be limited. Performing a prospective observational study at multiple centers would be needed to con rm the conclusions of this study. In addition, the patients who received neoadjuvant chemotherapy (n = 121 in the present study) were excluded because of the possibility that the neoadjuvant therapy affected staging (41) and prognosis (57

Availability of data and materials
All data generated or analyzed during this study are included in this manuscript and its supplemental les.

Competing interests
The authors have no con icts of interest to declare. Authors' contributions BF and SJY carried out the studies, participated in collecting data, and drafted the manuscript. WLR, DZG, XCH and LGQ performed the statistical analysis and participated in its design. ZX, LW, WD and XYJB participated in acquisition, analysis, or interpretation of data and draft the manuscript. All authors read and approved the nal manuscript. Figure 1 Enrollment owchart. Propensity score matching (PSM) was done using age, gender, body mass index (BMI), tumor location, AJCC stage, and admission year.

Figure 2
The overall survival (OS) and progression-free survival (PFS) curves of the two groups before and after propensity score matching (PSM). The survival in the laparoscopic resection (LR) group was signi cantly better than that of the open resection (OR) group in OS and PFS before PSM (A and C), but not after PSM (B and D). The overall survival (OS) curves of the two groups according to cancer stage before and after propensity score matching (PSM). The laparoscopic resection (LR) group was signi cantly better than the open resection (OR) group in OS (G; P=0.005) in stage IV disease before PSM but not after (H). There were no differences between LR and OR for stage I-III cancer (A-F; all P>0.05).

Figure 4
The progression-free survival (PFS) curves of the two groups according to cancer stage before and after propensity score matching (PSM). The laparoscopic resection (LR) group was signi cantly better than the open resection (OR) group in PFS (G; P=0.018) in stage IV disease before PSM but not after (H). There were no differences between LR and OR for stage I-III cancer (A-F; all P>0.05).

Figure 5
The overall survival (OS) curves of the two groups according to cancer location before and after propensity score matching (PSM). The laparoscopic resection (LR) group was signi cantly better than the open resection (OR) group in OS (E; P=0.010) in rectal cancer before PSM but not after (F). There were no signi cant differences for right-(A and B) and left-sided colon cancer (C and D). The progression-free survival (PFS) curves of the two groups according to cancer location before and after propensity score matching (PSM). The laparoscopic resection (LR) group was signi cantly better than the open resection (OR) group in OS (E; P=0.005) in rectal cancer before PSM but not after (F). There were no signi cant differences for right-(A and B) and left-sided colon cancer (C and D).