Long-term results of extended intersphincteric resection for very low rectal cancer: a retrospective study

Sixty-two patients with pathologically proven rectal adenocarcinoma between January 2001 and December 2010 were selected as subjects from the database of Department of Surgery, Pusan National University Hospital. Selection criteria were patients with low rectal cancer with well or moderately differentiation located less than 4 cm from the anal verge to inferior margin of the tumor. Exclusion criteria for ISR were poorly differentiated or signet ring cell adenocarcinoma diagnosed by biopsy, suspected tumor invasion of levator ani, impaired fecal continence preoperatively. Abdominoperineal resection (APR) was performed in these patients with resectable primary tumor.

Staging work-up included digital rectal examination, colonoscopy with biopsy, abdominal CT, MRI, and anorectal manometry for performing anorectal function assessment.

Fifty-three patients with locally advanced rectal cancer (LARC) received the neoadjuvant chemoradiotherapy. It composed of total irradiation dose of 40.5–50.4 Gy delivered in conventional fractionation (daily fractions of 1.8 Gy in weekdays for over 5–6 weeks) and capecitabine or 5-fluorouracil with low-dose folinic acid as the radiosensitizer. Capecitabine was administered orally at a dose of 825 mg/m² twice a day throughout the radiation therapy. Fluorouracil bolus (425 mg/m²) synchronously with folinic acid (20 mg/m²) was delivered on five consecutive days in the first and last week. For standardization of the assessment, preoperative T staging after neoadjuvant chemoradiation was determined by MRI, for the sixty-two patients as follows: Group I (yT2N0, n = 19; yT2N1, n = 5), Group II (yT3N0, n = 22; yT3N1, n = 13; yT3N2, n = 3).

All surgical procedures, open surgery, were performed or supervised by an experienced surgeon. For the abdominal approach, the same procedure as the low anterior resection was carried out after midline abdominal incision in modified supine position. The inferior mesenteric artery was ligated at the origin from the abdominal aorta. The splenic flexure was completely mobilized to mandate tension-free coloanal anastomosis and pelvic dissection was performed while preserving the superior hypogastric plexus. The rectum was completely dissected free from the pelvic floor and levator ani. A colonic J-pouch was made using the proximal colon after completion of abdominal phase of the procedure.

After pelvic dissection with an abdominal approach, intersphincteric resection was performed by perineal approach. For T2 (Group I) lesions, standard ISR was performed. After an incision was placed on the intersphincteric groove level, the gap between the internal and external sphincters was incised and dissection continued along the levator ani in the intersphincteric plane to connect with the pelvic dissection from the abdominal approach. In this process, the intersphincteric plane was dissected while the pelvis was illuminated from the abdomen (Fig. 1a). For the patient with T3 (Group II) lesions, extended ISR was performed. To ensure the safe radial resection margin after dissecting the anal sphincter plane, the deep part of the external sphincter was incised and removed, superficial part and subcutaneous part was preserved at least (Fig. 1b). Macroscopic findings of the resected specimen in group I and II are shown in Fig. 2.

Then, a colonic J-pouch - anal anastomosis was performed with 3–0 Vicryl sutures. In this procedure, 3-point interrupted sutures were performed, which the cut edge of the J-pouch was anastomosed to the cut edge of the anoderm connecting fixed levator ani muscle, to invaginate the anastomotic area and prevent the prolapse of the colon. Temporal diverting ileostomy was carried out for fecal detour, and then, the abdomen was closed. Unless there were special circumstances, ileostomy closure was performed in 6 months after surgery.

During the first 6 months after surgery, outpatient follow-up was carried out every 3 months. Clinical symptoms were observed during the follow-up interview and a clinical laboratory test (CBC and CEA) and a radiographic examination (chest radiography) were conducted. In 6 months postoperatively, abdominal CT and colonoscopy were performed.

The patients with restored ileostomy were followed using a standard protocol every 6 months for 5 years. The assessment of defecatory function was evaluated through personal interviews about stool frequency in a 24-h period and incontinence status assessed by Kirwan’s classification [13] and Wexner’s score [14] at the 12^th to 24^th month after ileostomy closure.

To evaluate local recurrence anddistant metastasis, abdominal computed tomography (CT) every 6 months for 2 years, and then once a year and colonoscopy were performed every 1 years. Anorectal manometry was not routinely performed to evaluate postoperative sphincter function.

Statistical analysis was performed using MedCalc Statistical Software version 14.10.2 (MedCalc Software bvba, Ostend, Belgium). Differences in distribution were calculated using the t-test for continuous variables, and chi-square test or Fisher’s exact test for categorical variables. Overall survival (OS) rates were analysed using the Kaplan–Meier method, and statistical significance of the survival rates was evaluated using the log-rank test. A P-value of < 0.05 was considered statistically significant.

The study protocol was approved by the Institutional Review Board of Pusan National University Hospital (Reference No.: PNUH IRB E-2015122). Written informed consent forms concerning this procedure were obtained for all patients.

Patients who had undergone operations were found to have early complications, such as temporary urinary retention (group I; 10 cases: group II; 15 cases), postoperative paralytic ileus (group I; 2 cases: group II; 3 cases), perineal abscess (group I; 1 case: group II; 1 case), and anastomotic leakage (group I; 1case: group II; 1 case). The patients who developed a perineal abscess healed with incision and drainage. The patients who developed an anastomotic leakage healed with conservative treatment (antibiotic therapy and TPN) without surgical intervention.

Late complications, such as anastomotic stricture (group I; 6 cases: group II; 10 cases), rectovaginal fistula (group I; 0case: group II; 1 case) after stoma closure were observed (Table 2). The patients who developed anastomotic stricture, however, the degree of narrowing was not severe, and most patients improved after laxative administration. Two patients recovered with endoscopic balloon dilatation. One patient developed a rectovaginal fistula, which was endoscopically closed with fibrin glue injection. There was no postoperative mortality in both groups.

All patients received potentially curative R0 ISR. The average distance measured to the distal resection margin during surgery was 1.69 cm in group I and 1.58 cm in group II. The lateral resection margin was negative in all cases. For the distance from the tumor to the resection margin, 5 cases were 1–3 mm in group I and 5 cases were the same in group II. The distance was greater than 3 mm in 19 cases in group I and 32 cases in group II (Table 1).

In postoperative pathologic tumor stages were as follows: group I (ypT2N0, n = 19; ypT3N0, n = 1; ypT2N1, n = 4), group II (ypT2N0, n = 2; ypT3N0, n = 26; ypT3N1, n = 7; ypT3N2, n = 3) (Table 3). In group I, one patient corresponded to T2 in the preoperative clinical stage was confirmed to be T3 after postoperative pathologic examination. Conversely, clinicopathological downstaging was observed by neoadjuvant chemoradiotherapy; two patients were downstaged to be T2 from T3 after pathological examination in group II. One case was observed from stage III to stage I in group I and 6 cases were observed from stage III to stage II in group II (Tables 4 and 5).

During the median follow-up period of 115 (range, 34–172) months after surgery, two patients with local recurrence and six patients with distant metastasis were observed. For distant metastasis, 2 cases of liver metastasis occurred in group I and 3 cases of liver metastasis and 1 case of both liver metastasis and lung metastasis occurred in group II.

Five-year overall survival (OS) and recurrence-free survival were 95.8 % and 87.5 %, respectively, in group I and 94.7 % and 86.8 %, respectively, in group II (Fig. 3a and b).

Bowel functions were evaluated and compared at the 12^th and 24^th months after ileostomy closure in both groups. The mean frequency of defecation per day was 3.54 times in group I and 4.29 times in group II at the 12^th month (p < 0.05). At the 24^th month, the mean frequency was decreased in group I by 2.21 times and in group II by 2.39 times on average, demonstrating no significant difference between both groups. All patients who had more than three instances of defecation per day showed stool fragmentation and four patients with two instances of defecation also showed the same symptoms.

In the comparison based on Kirwan’s classification, group I showed grade I continence (perfect continence) in 14 patients, grade II continence (flatus incontinence: not refraining from flatulence or mucus soiling) in six patients, grade III continence (minor soiling: incontinence 1–2 times a week) in three patients, and grade IV continence (major soiling: almost daily incontinence) in one patient. Group II exhibited grade I continence in 22 patients, grade II continence in ten patients, grade III continence in three patients, and grade IV continence in three patients at the 12^th month. No patient exhibited grade V continence. After 24 months, group I showed grade I continence in 19 patients, grade II continence in three patients, grade III continence in one patient, and grade IV continence in one patient. Group II exhibited grade I continence in 25 patients, grade II continence in eight patients, grade III continence in three patients, and grade IV continence in two patients. In this comparison, the two groups showed no statistically significant difference at the 12^th and 24^th months. The comparison of Wexner score of both groups revealed that at the 12^th month, the average score was 7.33 ± 2.8 in group I and 8.18 ± 2.9 in group II, and at the 24^th month, the average score was 5.21 ± 1.7 in group I and 5.82 ± 1.9 in group II. There was no statistically significant difference between the two groups (Table 6).