Persistent descending mesocolon as a vital risk factor for anastomotic failure and prolonged operative time for sigmoid colon and rectal cancers

Persistent descending mesocolon (PDM) is a congenital disease first reported by Morgenstern in 1960 [1]. The incidence of PDM varies from 1.3% to 4.0% [2‐4]. For patients with PDM, the descending colon is often medially located toward the failed fusion with the dorsal abdominal wall. The sigmoid colon is shifted to the right side of the abdomen with adhesion to the intestine mesentery. Some clinical complications, such as colonic volvulus, primary intestinal obstruction, and internal hernia, were reported in patients with PDM [5]; however, most patients are asymptomatic. PDM diagnosis mainly relies on imaging examinations, such as barium enema and abdominal CT scans [6‐8]; moreover, it can also be identified incidentally during surgery.

The current PDM and colorectal surgery research is mostly case-by-case [9‐11]. Although several retrospective studies are available, they have not been able to present the effect of PDM on anterior resection or low anterior resection (AR/LAR) due to left hemicolectomy enrollment. In addition, whether the variant characteristics of PDM affect short-term outcomes and colorectal anastomosis still lacks data.

Hence, this study focused on patients with PDM undergoing radical surgery for sigmoid and rectal cancers (SRCs). The anatomic relationships of the inferior mesenteric artery (IMA), the inferior mesenteric vein (IMV), the left colic artery (LCA), and the colon were evaluated using multiplanar reconstruction (MPR) and maximum intensity projection (MIP) techniques of CT post-processing [12]. In addition, the effects of PDM on short-term surgical outcomes and anastomotic outcomes of SRCs were further evaluated.

Data relating to patient demographics, clinicopathological features, radiological findings, and perioperative characteristics were reviewed retrospectively, including age; gender; body mass index (BMI); ASA Physical Status (ASA); tumor location, TNM staging and carcinoembryonic antigen (CEA) level; receiving of neoadjuvant chemoradiotherapy (nCRT), LCA retention and indocyanine green (ICG) angiography. Furthermore, radiological findings included the IMA bifurcating patterns; the lengths between the IMA and IMV, IMA and LCA, IMA and marginal arch, and IMA and the colon; and the length of IMA and IMA shift (right or left). Perioperative characteristics included operative time, blood loss, surgical procedure, intraoperative accident, number of lymph nodes harvested, conversion to laparotomy, anastomotic failure, postoperative complications, mortality, time-to-first anal exhaust, time-to-first oral intake, and the length of hospital stay.

Radiological findings are presented in the total and matched cohorts are summarized in Table 2. Before matching, the frequency of types I, II, III, and IV in the PDM and non-PDM groups was 37.5%/31.3%/28.1%/3.1% and 47.1%/30.1%/17.6%/5.1%, respectively (Fig. 4A-D). The PDM group had higher rates of IMA originating from the right side of the aorta than did the non-PDM group (40.0% vs. 1.0%, p < 0.001).The lengths between IMA and IMV (1.6 vs. 2.5 cm, p < 0.001), IMA and LCA (1.5 vs. 1.6 cm, p = 0.063), IMA and the marginal artery arch (2.7 vs. 8.4 cm, p < 0.001), and IMA and the colon (3.5 vs. 9.8 cm, p < 0.001) were significantly shorter in the PDM group than those in the non-PDM group.

After PSM, the frequency of types I, II, III, and IV in the PDM and non-PDM groups was 40.7%/29.7%/25.9%/3.7% and 40.0%/34.3%/18.1%/6.6%, respectively (Fig. 4A-D). The PDM group had higher rates of IMA originating from the right side of the aorta than did the non-PDM group (33.3% vs. 0.0%, p = 0.001). The lengths between IMA and IMV (1.6 vs. 2.5 cm, p = 0.001), IMA and LCA (1.4 vs. 1.6 cm, p = 0.083), IMA and the marginal artery arch (2.7 vs. 8.4 cm, p = 0.001), and IMA and the colon (3.3 vs. 10.2 cm, p = 0.001) were significantly shorter in the PDM group than those in the non-PDM group.

With the gradual unification and recognition of PDM diagnostic criteria, more colorectal surgeons have realized the influence of PDM on laparoscopic colorectal surgery [16]. The research on PDM mainly focuses on articles published by scholars in recent years. Wang et al. reviewed the records of 2775 patients who underwent laparoscopic radical colorectal cancer surgery at a single center [17]. Sixty patients (2.1%) were diagnosed with PDM: five were diagnosed preoperatively, and the others were diagnosed during surgery. Hanaoka et al. investigated PDM's frequency and radiological features for left-sided colorectal cancer with a ratio of 2.3% [18]. In our study, the frequency of PDM in the 845 cases was 3.8%. The lengths from IMA to IMV, IMA to the marginal arch, and IMA to the colon were significantly shorter in the PDM group (p < 0.001). PDM was an independent risk factor for prolonged operative time (OR = 3.205, p = 0.004) and anastomotic failure (OR = 7.601, p = 0.003). To the best of our knowledge, this study is the first to reveal that PDM was an independent risk factor for increased operative time and anastomotic failure for SRCs being subjected to radical surgery. However, the radiological diagnostic criterion has not completely reached a consensus. Porky et al. first described the imaging manifestations of PDM using barium and air-contrast enema examinations in 1966 [19]. Barium enema revealed that the sigmoid colon was longer and moved to the right side of the abdominal cavity, adjacent to the ileocecal mesentery, and the descending colon moved to the middle of the abdominal cavity. The CT axial scan revealed PDM when the right margin of the descending colon was located medially to the left renal hilum. Hanaoka et al. further revealed that the median lengths between the IMA and IMV and the descending colon in PDM cases were 14.8 mm and 17.2 mm, respectively, significantly shorter than those in non-PDM cases. In this study, more variables, including the length and shift of IMA; IMA type; and the lengths between the IMA and IMV, IMA and LCA, IMA and the marginal arch, and IMA and the descending colon were measured using MPR and MIP in 845 cases. Short lengths from IMA to IMV, IMA to LCA, and IMA to the colon [20]. Besides, the IMA to marginal arch length was significantly shorter in the PDM than that in the non-PDM group (2.8 vs. 8.4 cm, p < 0.001), which is important imaging evidence that the marginal artery is more prone to injury after IMA ligation. However, caution should be taken not to damage the marginal arch, and if necessary, intraoperative ICG fluorescence angiography can be used to determine vessel distributions [21] (Fig. 5A, B).

PDM was an independent risk factor for prolonged operative time (OR = 3.205, p = 0.004); nevertheless, some factors may increase the difficulty for PDM patients. First, the midline shift of the descending colon and extensive intraperitoneal adhesions are the main features of PDM. The sigmoid colon adhesions in most PDM patients are in the right lower abdomen but not the left. Some adhesion to the right pelvic wall and the ileocecal mesentery block IMA exposure [22]. Therefore, if the operator stands on the right side, it increases the difficulty and leads to more time being needed to separate the adhesions in the counter-direction views. Second, the mesentery of the sigmoid colon was not fully expanded, and the length from IMA to the colon was significantly shorter in the PDM group (3.5 vs. 9.8 cm, p < 0.001). After completing the separation of congenital adhesions, we proposed the sequence of dissociating the rectum prior to IMA ligation. The right yellow-white junction line cuts into the retrorectal space and is fully exposed to the opposite side. After joining the left and right sides, a sling was used to bind the rectum and mesentery, lift it from the caudal to the cranial direction to expand the retrorectal space, and finally reach the IMA root. Third, vascular IMA variations and their branches increase the difficulty of IMA, especially during LCA preservation. The lengths from IMA to IMV (1.6 vs. 2.5 cm, p < 0.001) and IMA to the marginal arch (2.8 vs. 8.4 cm, p < 0.001) were significantly shorter in the PDM group. The increased incidence of marginal vascular arch injuries and splenic flexure dissection will inevitably increase the surgery difficulty and prolong the operation time. Wang et al. revealed that the operative time of the PDM group was significantly longer (217.7 vs. 176.2 min, p = 0.003), and the volume of blood loss was higher (32.3 vs. 16.7 mL, p = 0.03). However, Hanaoka et al. did not observe similar results (176.0 vs. 171.5 min, p = 0.755).

Our multivariate analyses results revealed that PDM was an independent risk factor for anastomotic failure (HR = 7.601, p = 0.003). Furthermore, AL incidence in the PDM group displayed an increasing tendency (3.7% vs. 0.9%), though the difference was not significant. Due to the variation of the mesentery and IMA in PDM, the blood supply at the proximal end of the anastomosis is easily weakened. Mesenteric contracture can cause increased anastomotic tension, thereby affecting AL occurrence. In contrast, the descending colon is often medially located. Therefore, when the marginal artery was injured, the splenic flexure was freed, the proximal bowl extension was limited, and the Hartmann procedure had to be performed. Wang et al. revealed that rates of postoperative AL did not differ in both groups. LCA retention had no impact on proximal specimen margins or AL in patients with PDM. Hanaoka et al. also presented similar postoperative complications (8.0% vs. 10%, p = 0.346) but with a small sample size.

The limitations of this study lie in the retrospective study design and patient selection bias. Furthermore, some patients with congenital malformation of the descending colon that does not meet the diagnostic criteria for PDM might cause deviations in the significance of the results, and the cases of PDM observed during operation might cause deviations in the significance of the results. Moreover, the relatively small PDM sample and lack of long-term oncological outcomes affected the results. However, our study used the largest PDM series to date with detailed radiological data and operative imaging and was the first to evaluate CT imaging using MPR and MIP [23]. Meanwhile, more prospective and multicenter studies need to be done.

In conclusion, PDM with mesenteric and IMA variation is an independent risk factor for prolonged operative time and anastomotic failure. Therefore, preoperative evaluation using MPR and MIP can help surgeons handle this rare congenital variant better.