The clinical effect and oncological results of preserving inferior mesenteric artery during laparoscopic complete mesocolic excision for patients with descending colon cancer: a propensity score-matched analysis.

Introduction
One of the primary causes of fatalities due to cancer globally is colorectal cancer [1]. During the previous decades, with the development of advanced surgical techniques such as CME with central vascular ligation (CVL) and D3 lymph node (LN) dissection, significant progress has been made in the treatment of colorectal cancer. Recent studies have shown that CME combined with CVL has important prognostic importance for colorectal cancer [2–4]. However, CME currently still has many controversies [5]. Whether to preserve or ligate the IMA is one of the issues. Theoretically, IMA ligation may affect blood supply, thereby impacting postoperative GIF and complications. To our knowledge, there are not enough research reports focusing on this problem during CME surgery.
According to a speculation by Mihara, preserving the IMA and its branches for SCC may help maintain blood supply to the anastomotic site, thereby potentially accelerating postoperative GIF recovery and reducing postoperative complications [6]. Moreover, some investigations have shown that preserving the IMA may not affect the oncological safety of the surgery [6, 7]. Additionally, with improved understanding of surgical anatomy and the maturation of laparoscopic surgical techniques, related studies show that laparoscopic surgery is widely used in colorectal cancer surgery [8–10]. Laparoscopic CME is safe and feasible [11, 12].
Therefore, considering the current lack of relevant evidence on whether the IMA should be preserved in laparoscopic CME surgery for SCC, more research findings and clinical data are needed to address this issue. This research attempts to investigate the clinical effect and oncological results in patients who underwent laparoscopic CME surgery, applying 1:1 PSM analysis to compare the impact of IMA-P versus IMA-L.

Patients and methods

Patients
We collected the medical information of individuals who underwent CME surgery at Xiamen University Attached Zhongshan Hospital from January 2012 to August 2021. The inclusion criteria were (1) age 28–87 years, (2) pathological T stage of T1–T4, (3) patients who had not received antitumor treatments such as neoadjuvant chemotherapy, immunotherapy, or targeted therapy before surgery, and (4) all patients completed the surgery with laparoscopic assistance. Exclusion criteria were (1) pathological stage IV, (2) patients with severe comorbidities before surgery, (3) patients who could not tolerate surgery, (4) patients who underwent open surgery, (5) patients who had undergone gastrointestinal surgery or had other major organ resections, (6) patients with multicentric cancer, (7) patients undergoing emergency surgery, and (8) patients with severe coagulation disorders.
This research employed a single-center database to conduct a retrospective examination, and the ethics committee of Xiamen University Attached Zhongshan Hospital approved the research (no. 2025-018). As this is a retrospective study, it was exempted from the requirement to obtain patient informed consent, and the research followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies.

Surgical technique
We first fully mobilized the mesocolon. Subsequently, during the CVL procedure, in patients with IMA-P, one needs to preserve the IMA, while patients with IMA-L had the IMA transected 1–2 cm away from the aorta. Finally, according to the “10 cm” principle, the bowel and the corresponding mesocolon were cut, and the distal and proximal colon were anastomosed end-to-side. All procedures were carried out by skilled gastrointestinal surgeons in accordance with the principles of CME [13].

Design and outcome
We categorized patients with IMA-P as the observation group and those with IMA-L as the control group. By collecting information from the hospital’s medical record system and following up with patients postoperatively, we obtained the following data as our assessment indicators: clinical characteristics (including age, gender, body mass index (BMI), American Society of anesthesiologists (ASA) physical status (PS) score, and postoperative chemotherapy), pathological characteristics (including location of tumor, maximum tumor diameter, differentiated degree, pathological T stage, pathological N stage, tumor–node–metastasis (TNM) stage, r0 resectability, lymphatic invasion, vascular invasion, and neural invasion), surgical results (including operating time, blood loss, number of harvested total LNs, resected bowel length, distal margin, proximal margin, hospital stay following surgery, Intake, Feeling nauseated, Emesis, Examination, and Duration (I-FEED) score, first postoperative gas passage, first postoperative defecation, and postoperative complications), and mid-term survival results (including 3-year OS and 3-year DFS).
Our definition of the descending colon follows the standard established by the International Federation of Associations of Anatomists. We used the eighth edition of the Union for International Cancer Control (UICC) TNM stage criteria for conducting the TNM stage [14]. We utilized the I-FEED scoring system, which is a standardized scoring tool designed to objectively assess GIF recovery in patients following abdominal surgery, to evaluate the postoperative GIF recovery in patients, and defined patients with a score of 3 or higher as having postoperative gastrointestinal intolerance (POGI), and those with a score of 6 or higher as having postoperative gastrointestinal dysfunction (POGD) [15]. We used the Clavien–Dindo grading system to classify postoperative complications and considered complications with grade IIIb or higher as severe postoperative complications [16]. All postoperative complications were determined through imaging findings and clinical diagnosis.

Propensity score-matched analysis
We employed IBM SPSS Statistics 27 to calculate the propensity scores for each patient, with confounding factors including age, gender, BMI, ASA-PS score, postoperative chemotherapy, location of tumor, maximum tumor diameter, differentiated degree, pathological T stage, pathological N stage, and TNM stage. To accomplish 1:1 matching, we utilized the nearest-neighbor matching technique with a matching tolerance of 0.02.

Statistics
For all data calculations, IBM SPSS Statistics 27 was applied. The median (range) is used to characterize continuous variables. Mann–Whitney U tests were used to evaluate non-normally distributed continuous values, and two-sample t-tests were utilized to assess regularly distributed continuous data. All categorical and ordinal variables are described using case numbers and percentages. The Fisher exact test or the chi-square test was adopted to compare two sets of variables, while rank-sum tests were performed to analyze two ordinal variables. We used the Kaplan–Meier technique to draw survival curves and examine statistical differences using the log-rank test. When p was under 0.05, all results were considered to be significantly different.

Results

Clinical and pathological characteristics
Figure 1 shows the flowchart of this research. In conclusion, we enrolled 81 patients on the basis of the inclusion and exclusion criteria, with 26 in the IMA-L group and 55 in the IMA-P group. After PSM, the experimental group and the control group included 22 patients each. A comparison of the two groups’ clinical and pathological traits is displayed in Table 1. Before PSM, the median (range) BMI in the IMA-P group was 25.19 kg/m2 (17–35.4 kg/m2), and in the IMA-L group was 25.06 kg/m2 (18.7–29.2 kg/m2). The difference between the two groups was statistically significant (p = 0.031). However, after PSM, the BMI was well balanced between the two groups (24.85 (20.0–35.0) kg/m2 versus 25.1 (18.7–29.2) kg/m2, p = 0.677). No statistical significant variations were found in the remaining indicators before and after PSM.

Surgery-related features
Table 2 presents the surgical results of the two groups. Before PSM, the blood loss of the IMA-L group was noticeably higher compared with the IMA-P group (70 (10–500) versus 10 (10–200) ml, p = 0.008). After PSM, although the difference was reduced, it remained statistically significant (70 (10–500) versus 15 (10–200) ml, p = 0.048). Nevertheless, the IMA-P group had a substantially quicker time to first postoperative gas flow versus the IMA-L group (p = 0.041). In addition, no statistically significant differences were found in operating time, total number of LNs harvested, resected bowel Length, distal margin, proximal margin, hospital stay following surgery, I-FEED score, and first postoperative defecation.

Postoperative complications
Before PSM, postoperative complications occurred in three cases in the IMA-P group and seven cases in the IMA-L group (p = 0.004), among which severe complications occurred in one case and four cases, respectively (p = 0.035). After PSM, no patients in the IMA-P group experienced postoperative complications, while postoperative complications occurred in seven individuals in the IMA-L group (p = 0.021), among which three cases were severe complications (p = 0.036). Therefore, the probability of postoperative complications was substantially lower in the IMA-P group in comparison with the IMA-L group (Table 2).
Patients with impaired postoperative GIF and nonsevere complications improved after dietary adjustment or conservative drug therapy. Patients with severe small bowel obstruction obtained relief after surgical treatment. Patients with severe anastomotic leakage received treatment under endoscopy. Patients with cerebral infarction were transferred to the vascular surgery department for further treatment. None of the patients developed other complications such as enteritis, lymphatic leakage, or hematoma.

Survival outcomes
The median follow-up period for both groups was 36 months (p = 0.969). Figures 2 and 3 illustrate the 3-year OS and DFS rates. Before PSM, we estimated the 3-year OS rate in the IMA-P group at 83.6% and in the IMA-L group at 73.1%, with no discernible difference (p = 0.392). Similarly, there was no apparent difference in the 3-year DFS (70.9% versus 73.1%, p = 0.974). After PSM, the 3-year OS rate still did not demonstrate a noticeably difference (86.4% versus 77.3%, p = 0.615) and the 3-year DFS rates showed the same results (63.6% versus 72.7%, p = 0.619).

Discussion
This study systematically evaluated patients with SCC who underwent laparoscopic CME with CVL surgery, exploring both the short- and the long-term prognosis of the IMA-P strategy. It should be pointed out that Sato’s report is the latest article on this issue [7]. However, his study adopted D3 dissection. Considering that D3 dissection has similar oncological prognosis to CME surgery, our study extends this evaluation to CME surgery, providing an important supplement to the existing research.
Our research is the first to compare the clinical effect and oncological results of CME surgery for SCC between IMA-P and IMA-L. While prior studies suggest that the IMA-P approach may reduce complication rates—potentially improving long-term survival by mitigating systemic inflammatory responses, immune suppression, and physiological reserve depletion caused by chronic inflammation [17–21], and that DFS may correlate with mesenteric resection area and bowel length [22, 23]—our findings indicate no significant difference in survival outcomes between the two groups. Thus, this association warrants further investigation.
With respect to GIF, apart from the IMA-P group having a shorter time to first postoperative gas flow (3 (1–6) versus 4 (2–7) days, p = 0.041), no other results displayed considerable changes. This phenomenon may be due to several reasons: (1) the I-FEED score may not be sensitive enough, (2) retrospective study records may not capture subtle differences in GIF recovery, (3) our sample size was relatively small, especially in the matched cohort, which may have led to some actual differences not reaching statistical significance, and (4) IMA-P may be considered a more physiologically appropriate surgical approach, thereby reducing the need for personalized interventions, but for those with POGI or POGD, early intervention by doctors may have decreased the differences in study results.
Although this research revealed no obvious difference in the primary outcomes, the IMA-P group demonstrated a clear advantage in the secondary outcomes. Firstly, the IMA-P group had reduced risk of bleeding. This is consistent with the findings of Mihara’s study [24]. The possible reason is that high ligation requires manipulation at the origin of the aorta, where there are many anatomical variations in vascular branches and a venous plexus, which can easily be injured and cause bleeding during dissection. However, the reasons need to be further proven by additional research. Additionally, our research results on postoperative complications are similar to those of Zeng and Yin’s meta-analysis: IMA-P may reduce the risk of anastomotic leakage [25, 26]. This is probably because ligation of the IMA leads to insufficient blood supply to the anastomosis. However, the study by Munechika reported that, with high ligation of the IMA under the guidance of indocyanine green (ICG) fluorescence imaging, postoperative complications related to anastomosis were not observed. This difference may stem from intraoperative ICG fluorescence imaging being useful for determining the transection line and observing the blood supply to the oral and anal sides of the colon in laparoscopic colorectal surgery, which can reduce anastomotic leakage [27]. Additionally, we discovered that the resected bowel length in the experimental group was less than that in the control group (17.50 cm versus 23.20 cm, p = 0.051). Although this difference did not reach the traditional level of statistical significance, it showed a trend that may have clinical importance—IMA-P can better ensure blood supply, thus avoiding excessive resection.
This research does still have certain limitations. First, the quality of evidence in a single-center retrospective study is poorer than that in prospective group studies or randomized controlled trials. Furthermore, we are unable to carry out multivariate Cox regression analysis owing to the insufficient sample size. Only 44 individuals were included in this investigation after PSM, which might have distorted the findings. Finally, we failed to evaluate long-term quality of life. These long-term results have an important impact and should be taken into consideration in future studies.

Conclusions
For patients with SCC, although the two methods are similar in terms of early GIF recovery and oncological outcomes, the surgical approach that preserves the IMA demonstrated a lower risk of blood loss and incidence of postoperative complications. Therefore, preserving the IMA may be a better decision in CME.