Inter-ischial spine distance is a simple index of narrow pelvis that may predict diculty during laparoscopic low anterior resection: a retrospective study

A narrow pelvis makes laparoscopic rectal resection dicult. This study aimed to evaluate whether a simple measurement on computed tomography can predict procedural diculty. A total of 62 patients with low rectal cancer underwent conventional laparoscopic low anterior resection. The inter-ischial spine (IS) distance (i.e., distance between the ischial spines) was measured on an axial computed tomography slice. Operative time, blood loss, and time from insertion of linear staplers to completion of clamping on the distal end of the rectum (clamp time) were compared.


Abstract Background
A narrow pelvis makes laparoscopic rectal resection di cult. This study aimed to evaluate whether a simple measurement on computed tomography can predict procedural di culty.

Methods
A total of 62 patients with low rectal cancer underwent conventional laparoscopic low anterior resection. The inter-ischial spine (IS) distance (i.e., distance between the ischial spines) was measured on an axial computed tomography slice. Operative time, blood loss, and time from insertion of linear staplers to completion of clamping on the distal end of the rectum (clamp time) were compared.

Results
Overall, 42 men and 20 women with low rectal cancer were assessed. The mean tumor size was 34.5 mm. Total or tumor-speci c mesorectal excisions were performed in all cases; high ligation and resection of the inferior mesenteric arteries were carried out in 92% of patients. The mean operative time, and blood loss were 206 min, and 15 mL respectively. Four patients (6.5%) experienced postoperative complications, including two anastomotic leaks (3.2%). The mean IS distance was 93.3 mm. With simple linear regression, shorter IS distance correlated with longer operative time (R 2 = 0.08, P = 0.030) and clamp time (R 2 = 0.07, P = 0.046). Using an receiver operating characteristic curve, a narrow pelvis was de ned as IS distance < 94.7 mm. Multivariate regression analysis revealed that IS distance < 94.7 mm (odds ratio, 3.51; P = 0.04) was independently associated with a longer clamp time.

Conclusions
The IS distance is a simple and useful measurement for predicting the di culty of laparoscopic low anterior resection.

Background
Owing to the superior short-term results [1,2] and non-inferior long-term oncologic outcomes [3,4] relative to laparotomy, laparoscopic surgery is becoming a standard surgical treatment for colon cancer worldwide. The safety and superior short-term results of laparoscopic low anterior resection performed by expert surgeons for rectal cancer have also been reported [5,6]. In Japan, more than 30% of patients with low rectal cancer undergo low anterior resection using the laparoscopic approach [7]. However, its shortterm safety in the general setting, and the long-term safety, are not well-established [8-11]. One of the reasons why developing the evidence base for rectal surgery is taking longer than it did for colon surgery could be the difference in the technical di culty of the two procedures. It is well-known that the narrow pelvis makes laparoscopic rectal resection di cult, making the operative time longer [12,13] and the handling more di cult [14]. A narrow pelvis also increases the risk of anastomotic leakage [15]. Additionally, the de nition of the narrow pelvis remains unclear in terms of surgical anatomy. Several indexes that predict surgical di culty have been reported using sagittal computed tomography (CT) imaging, 3D-CT, and magnetic resonance imaging (MRI) [12][13][14][15][16]. However, the indexes are quite complicated in terms of the measurement sites, with some requiring calculations. A novel and simple index is required.
The present study aimed to establish a simple CT-derived index for the narrow pelvis that can predict the likelihood of experiencing di culty during laparoscopic low anterior resection.

Methods
A total of 252 patients with rectal cancer underwent laparoscopic low anterior resection between January 2009 and December 2018 at Hokkaido University Hospital. After excluding cases of upper rectal cancer, we retrospectively assessed the short-term outcomes of 62 conventional laparoscopic low anterior resections for low rectal cancers that were electively performed and evaluated whether the distance measured on axial CT slices can predict procedural di culty. The distance between the iliopsoas muscles at the pelvic inlet (IM distance) and the distance between the ischial spines (IS distance) were measured ( Fig. 1). The size of mesorectum was de ned as the sum of the transverse and anterior-posterior diameters of the mesorectum at the S4 level. The operative time, blood loss, and time from the insertion of linear staplers to the completion of clamping on the distal end of the rectum (clamp time) were set as the outcomes of procedural di culty.
Postoperative complications were evaluated according to the Clavien-Dindo classi cation [17]. The study protocol was reviewed and approved by the hospital ethics committee of Hokkaido University Hospital, and informed consent was obtained from all patients through the opt-out method in accordance with the guidelines of the Japanese Ministry of Health, Labour and Welfare.

Procedure
The procedure was performed under pneumoperitoneum with 10 mmHg CO 2 . Five ports were inserted at the umbilicus, on both sides of the hypogastric region, and at the ank. In particular, the right hypogastric port was inserted just laterally or medially to the root of the inferior epigastric artery. First, in the Trendelenburg position, the whole jejunum and ileum were ipped up to the cranial side. The mesocolon of the sigmoid colon was dissected from the retroperitoneum and mobilized using the medial-to-lateral approach. The left gonadal vessels and ureter were preserved. High ligation and resection of the inferior mesenteric artery (IMA) in advanced cancer were performed. The inferior mesenteric vein and the peripheral part of the left colic artery were resected at the same level as the root of the inferior mesenteric artery. After the lateral side of the sigmoid colon was mobilized, total or tumor-speci c mesorectal excision was performed with dissection along the circumference of the fascia propria of the rectum. In mobilizing the rectum from the pelvic organs, both sides of the ureter, hypogastric nerves, pelvic plexus, and pelvic splanchnic nerves were all preserved. The Denonvilliers fascia was preserved unless the tumor was located at the anterior wall of the rectum. The mesorectum 2 cm distal to the tumor was dissected vertically towards the rectum, and the entire circumference of the rectal wall was exposed. Subsequently, the site of the rectum was enclosed by clip-shaped forceps, the lumen of the rectum was washed with 1 L of physiological saline and 5% of povidone iodine. A linear stapler was inserted from the right hypogastric region port, and the rectum was clamped with the stapler on the distal side of the clip-shaped forceps. The rectum was transected by single or double use of the linear stapler. A small incision was added at the umbilical port, and the mesocolon and the sigmoid colon were extracorporeally cut at 10 cm on the oral side from the tumor. Reconstruction of the intestine was intracorporeally performed using the double-stapling technique with a circular stapler. The anterior part of the intersected stapler was reinforced by a full layer of sutures. A diverting ileostomy was created for patients with anastomosis located in the anal canal, patients who received preoperative therapy, or patients with a positive air leak from the anastomosis.

Statistical analysis
All continuous data are presented as mean and standard deviation. Descriptive statistics for patient characteristics were calculated. All statistical tests were performed using an alpha level of 0.05 (twosided). Chi-squared and Student's t-tests were used to analyze categorical and non-normal continuous data, respectively. Simple linear regression was performed to assess the relationship between IS/IM distance and the size of the mesorectum and clamp time, operative time, and blood loss during all rectal surgeries. To determine factors that could affect longer clamp time, univariate and multivariate logistic regression analysis with covariates based on preoperative and intraoperative values was also performed.
Factors with P-value below 0.2 in the univariate analysis were entered into the multivariate analysis. All statistical analyses were performed using JMP Pro version 13.1.0 (SAS Institute, Inc., Cary, NC, USA).

Patient characteristics and the measured pelvic distance
A total of 62 patients (mean age, 66.2 years; male-to-female ratio, 42:20) were included in this study. The mean body mass index was 22.8 kg/m 2 . Two patients (3.2%) had preoperative comorbidities (American Society of Anesthesiologists class III). Of these 62 patients, 63% had tumors located on Ra (de ned as the rectal segment between the second sacral joint and peritoneal re ection), 37% had tumors located on Rb (de ned as the rectal segment below the peritoneal re ection), 45% were diagnosed with locally advanced cancer, and 10% were identi ed to have distant metastasis. Low anterior rectal resection with high ligation of the IMA (D3 lymph node dissection) was performed in 92% of patients. The mean operative time and blood loss were 206 min and 15 mL, respectively. The mean time required for clamping the distal edge of the harvested rectum was 79.8 seconds. Diverting stomas were created in 42% of the patients. No patient required conversion to open surgery. Four patients (6.5%) experienced postoperative complications (Clavien-Dindo grade ≥ 3), including anastomotic leakage (n = 2), small bowel obstruction (n = 1), and cholecystitis (n = 1). Overall, 3%, 42%, 21%, 24%, and 10% of patients had a pathological stage of 0, I, II, III, and IV, respectively. In 97% of the patients, a diagnosis of differentiated adenocarcinoma was con rmed. The mean number of harvested lymph nodes was 16.8, with a mean tumor size of 34.5 mm, and histological R0 resection was achieved in all patients with stages 0-III. The mean distance between the iliopsoas muscles at the pelvic inlet was 100.2 mm, whereas the mean distance between the ischial spines was 93.3 mm; the size of the mesorectum measured 125.8 mm ( Table 1). Relationship between short-term outcomes and pelvic index Patient characteristics and surgical outcomes between patients with narrow and wide pelvises A longer operative time was de ned as 79.8 seconds (mean) and logistic regression analysis was performed. According to the receiver operating characteristic (ROC) curve, we de ned a narrow pelvis as that with an IS distance < 94.7 mm.
A total of 36 men and 1 woman with a mean age of 65.7 years were classi ed as having a narrow pelvis, and 19 men and 6 women with a mean age of 67.0 years were classi ed as having a wide pelvis. The IS distance was signi cantly shorter in patients with a narrow pelvis (86.3 mm vs. 103.5 mm; P < 0.001). However, no differences in body mass index (BMI), the frequency of preoperative co-morbidities with American Society of Anesthesiologists (ASA) grade III or above, or tumor location, size, or stage were observed between the two groups. The port locations and the frequency of D3 lymph node dissection or need for diverting stoma were not different between the two groups.
The time required for clamping the distal edge of the rectum (88.9 seconds vs. 65.8 seconds; P = 0.01) and the duration of surgery (217 min vs. 191 min; P = 0.01) were signi cantly longer in the narrow pelvis group. The number of linear staplers used was similar in both groups. The rates of postoperative complication and anastomotic leakage (grade III or above) were 8.1% and 5.4% in the narrow pelvis group and 4.0% and 0% in the wide pelvis group, respectively, albeit without signi cant difference. No differences in the conversion rate, blood loss, number of harvested lymph nodes, or pathological R0 rates were observed (Table 2).

Discussion
The IS distance is a simple index for the narrow pelvis that can predict the di culty of laparoscopic low anterior resection for rectal cancer. A narrow pelvis made the procedure di cult in terms of the time taken to clamp the distal end of the rectum, when we de ned the narrow pelvis as an IS distance < 94.7 mm. Moreover, operative times were longer in the presence of a narrow pelvis. Measuring the distance between the ischial spines using an axial CT image provides a simple de nition of a narrow pelvis.
Some previous studies have developed narrow pelvis indexes. Tsuruta et al. reported that a pelvic index > 13 in male patients was associated with anastomotic leakage [15]. However, complicated measurements were needed to calculate their pelvic index, because their index was de ned as the difference between the interspinous distance and the diameter of the mesorectum, divided by the depth of the cavity of the lesser pelvis. Hamabe et al. reported that sacral promontory shape was related to the insertion of an additional port in reduced-port surgery [14], and Targarona et al. reported that the pubic coccyx axis was associated with prolonged operative time [13]. These narrow pelvis indexes are based on angles, and intuitively judging surgical di culty is di cult. Shimada et al. reported the antero-posterior/transverse ratio of the pelvic inlet to be associated with prolonged operative time [12]. This index requires three-dimensional CT measurements because of the need to measure the pelvic inlet while correcting for the angle of the pelvic axis. All these previously published indexes require multiplanar measurements using MRI or reconstructed CT images. However, in the current study we found the IS distance to be a novice and simple index requiring only the axial slice of a CT scan to predict the di culty of performing a laparoscopic low anterior resection. The axial slice of a CT scan is simple to interpret and one of the most popular images that can be available for surgeons.
In the current study, we used the time taken to clamp the distal end of the rectum as an outcome, which may not be popular. However, many surgeons may experience di culty in managing to clamp the distal end of the rectum, especially in the pelvis at a level deeper than the peritoneal re ection. Therefore, when we designed the current study, we hypothesized that being able to clamp the distal end of the rectum may be more di cult in the presence of a narrow pelvis. As per the results of the simple linear regression and logistic regression, a shorter IS distance was signi cantly associated with a longer clamp time and operative time. Regarding the operative time, the narrow pelvis in our de nition was not associated with a longer operative time in multivariate regression analysis. This might be because many other factors in uence the operative time. Tsuruga et al. reported that the pelvic index was associated with a longer operative time but not merely the IS distance [15]. This could also be because several factors could in uence the operative time. The creation of a stoma was the only independent factor that was signi cantly associated with a longer operative time. The creation of the stoma itself takes several minutes. Furthermore, a diverting ileostomy was created in patients where the anastomosis was located in the anal canal, patients who had received preoperative therapy, and patients with positive air leakage from the anastomosis. These are generally di cult cases [18][19][20] and it is natural that these operations took longer to complete. With this in mind, the process to clamp the distal end of the rectum is a simple index that directly re ects the di culty of the pelvic aspect of the operation, and it is reasonable to use it as an index of the di culty. One thing to consider is that the clump time may be dependent on the technical experience of the operator. However, all of the operations in the current study were performed among surgeons who had been quali ed as skilled in Endoscopic Surgical Skill Quali cation System (ESSQS) to eliminate the possible bias of individual surgeons' skills. In Japan, the Japan Society for Endoscopic Surgery introduced the ESSQS [21-23] to maintain and improve the quality of laparoscopic surgery, and laparoscopic colorectal procedures performed with ESSQS-quali ed surgeons have been proven to improve postoperative results [24].
In the current study, the rate of R0 resection and postoperative complication rate were too small to be reliable outcomes for operative di culty. The cutoff point was calculated from the ROC curve. The area under the curve derived from the ROC curve was not high, and this might be because of the limited number of patients in this study. The difference in the pelvic shape between men and women is an important issue to discuss because the width of the pelvis could not be measured using the IS distance in a straightforward manner. However, multivariate regression revealed that the IS distance was an independent factor for predicting the di culty of low anterior resection, and IS distance can be a common factor regardless of a patient's sex.
Predicting the di culty of low anterior resection can be useful in deciding who the operator will be and in determining the surgical approach. In di cult cases, it may be best that an expert surgeon performs the operation. For instance, a certi cated surgeon by the ESSQS may be appropriate to perform di cult cases. Determining the surgeon according to the IS distance may result in better surgical outcomes because procedures beyond a surgeon's capacity may be avoided by that surgeon. Laparoscopic surgery using the robotic assistant and the trans-anal endoscopic approach could be another option to improve the surgical outcomes in di cult cases with a narrow pelvis [25,26], although the merit of these approaches has not been fully shown to date. IS distance may be useful in determining these appropriate surgical approaches.
The current study is weak in terms of small number, single faculty and retrospective evaluateion.
Therefore, future prospective studies with a large number of patients from multiple institutions are required.

Conclusions
The distance between the ischial spines is a simple and useful method to predict the di culty of handling the narrow pelvis in laparoscopic low anterior resection. Predicting the di culty using this index might lead to a safer rectal cancer resection when considering the speci c operator or the surgical approach.
Abbreviations ASA American Society of Anesthesiologists BMI body mass index CI con dence interval CT computed tomography ESSQS Figure 1 The de nitions of the distance between the ischial spines (IS distance) and the distance between the iliopsoas muscles (IM distance) (a) The IS distance was de ned as the distance between both sides of the ischial spines. (b) The IM distance was de ned as the distance between both sides of the iliopsoas muscle at the level of the pelvic inlet.