The success rate of robotic natural orifice intracorporeal anastomosis and transrectal extraction (NICE procedure) in a large cohort of consecutive unselected patients.

Consecutive patients who presented with benign or malignant disease requiring left-sided colorectal resection and anastomosis between May 2018 and June 2021 were evaluated. All patients underwent Robotic Natural orifice Intracorporeal anastomosis with transrectal Extraction of specimen (NICE Procedure) consisting of either Anterior Resection (AR) or Low Anterior Resection (LAR). Patients who received neoadjuvant radiation with planned diverting loop ileostomy were excluded. All the cases were performed by a single surgeon (EMH) and the colorectal surgery team in the Texas Medical Center. All patients underwent preoperative workup with physical examination, colonoscopy, imaging, and laboratory exams as required. All patients were recovered on a standardized enhanced recovery program (ERP) including multimodal post-operative pain control. For benign cases three 8 mm robotic ports and a 5 mm Air Seal port (Optiview, Inc., Jacksonville, FL, USA) are utilized. Once placed, the patient is positioned with 8-degree left side elevation and 18–20 degrees of Trendelenburg and the robot is docked. A lateral to medial mesenteric sparing approach is performed [ 11 , 13 ]. The dissection commences by incising the white line of Toldt along the left colon and exposing the paracolic gutter by retracting on the colonic mesentery. Splenic flexure mobilization is then performed when indicated for a tension free anastomosis. Attention is then drawn to the pelvis and the lateral attachments along the intersigmoid fold are released. The right and left rectal reflection is then incised as is the anterior reflection to release the peritoneal attachments of the rectum. The proximal level of resection is then identified and divided using the vessel sealer extend (VSE) with minimal use of energy. A mesenteric sparing dissection is performed staying close to the bowel, preserving the superior rectal artery and avoiding the retroperitoneal plane. Dissection is carried to the distal level of resection along the rectum which is then transected with the VSE. Sizers are then passed through the rectal cuff followed by placement of the small Alexis wound retractor (Applied Medical, Rancho Santa Margarita, CA, USA). Transrectal extraction of the specimen is then performed. In benign cases in which there is a bulky specimen, a thinning maneuver is performed in which the mesentery is detached from the surface of the bowel wall to facilitate transrectal extraction. The retractor is then removed in preparation for the end-to-end anastomosis. For malignant disease, the port site configuration is the same as benign disease except a 12 mm robotic port (instead of an 8 mm port) is placed in the right lower quadrant to facilitate the use of a robotic stapler. On occasion an additional 8 mm port is placed in the left lower quadrant for deep dissection in a narrow pelvis. The linear stapler (SureFormTM60) is used to divide and close the specimen to avoid intraabdominal contamination in the presence of malignant pathology. A medial to lateral approach is performed with high ligation of the pedicle and oncologic lymph node dissection which has been previously described [ 12 ]. Following the dissection and mobilization, the robotic linear stapler is used to divide the bowel at the proximal and distal level of resection thereby keeping the specimen intact. The staple line along the proximal colon and rectum is then excised and removed. The small Alexis retractor is placed across the rectum and the specimen is extracted. The retractor is then removed in preparation for the end-to-end anastomosis. In cases in which there is a concern for trauma to the rectum due to a bulky specimen, we complete the anastomosis and thereafter extract the specimen through a Pfannenstiel incision. The anvil to the circular stapler (CDH 29 mm or 31 mm circular stapler, Ethicon, Somerville, NJ, USA) is delivered through the rectum into the pelvis. A 6-inch 3.0 V-Loc suture on a v20 needle (V-Loc 180™, Covidien; Mansfield, MA, USA) is used to place a purse-string suture around the cut edge of the proximal bowel. The anvil is inserted, and the purse-string suture is tightened. Next a vicryl Endoloop® (Endoloop®, Ethicon, Somerville, NJ, USA) is placed about the neck of the tissue to further secure the anvil. Attention is then drawn to the divided edge of the rectum and a second V-Loc pursestring suture is placed. The spike of the circular stapler is then advanced and the pursestring suture is tightened about the spike. The anvil and the spike are seated, and an anastomosis is fashioned. In the majority of cases of malignant disease, the rectal cuff is closed using the linear stapler. When performed, a handsewn end-to-end anastomosis was achieved with the use of two or more 3.0 V-Loc suture on a v20 in a single layer. The integrity of the anastomosis is assessed with both direct endoscopic visualization as well as air insufflation test. Demographic data including age, sex, and Body Mass Index (BMI) as well as clinical data, including American Society of Anesthesiology (ASA), previous abdominal surgery, and diagnosis were abstracted. Diverticulitis was classified as recurrent or complicated which consisted of the presence of abscess, fistula or stricture. The intervention data included type of surgery, operative time, estimated blood loss, intraoperative transfusion, splenic flexure takedown, site of specimen extraction, anastomosis type, anastomosis location, diverting loop ileostomy, intraoperative complications, and conversion to open or other minimally invasive approach. The surgical pathology data were abstracted for the subset of patients with malignant disease. The surgical outcomes analyzed were return of bowel function, length of hospital stay, 30-day postoperative complications, 30-day unplanned readmissions, 30-day unplanned reoperations, and 30-day mortality. The primary outcomes consisted of success rate of intracorporeal anastomosis (ICA), success rate of transrectal extraction of specimen (TRSE) and conversion rate. Summary statistics were presented for all data. Categorial variables were presented as count and percentages. Continuous variables were presented as median or mean, as appropriate, and range. This study was approved by the Houston Methodist Hospital Internal Review Board (study protocol 00012111).

There was a total of 306 patients who underwent Robotic NICE left-sided resection of which 54% were female. The median age was 59 years-old (range, 19–88) and the median BMI was 27.8 kg/m2 (range, 17.2–55.6) of which 34.3% (105 patients) presented with BMI ≥ 30 kg/m 2 . The majority of patients were ASA II (59.8%) and just over half (52.9%) had a previous abdominal surgery. Diverticulitis was the main diagnosis (64%) followed by colorectal neoplasm (27%) and other (9%). In the subgroup of patients with diverticulitis, 47.4% had complicated disease (abscess, fistula and/or stenosis). Demographic and clinical features are presented in Tables 1 and 2 . Table 1 Demographic and clinical features of patients undergoing Robotic NICE procedure All cases n  = 306 (100%) Diverticulitis n  = 196 (64.1%) Neoplasm n  = 83 (27.1%) Other n  = 27 (8.8%) Age in years, median (IQR) 59 (19–88) 59 (22–84) 61 (31–88) 39 (19–83) Sex  Female 166 (54.2) 111 (56.6) 32 (38.6) 23 (85.2)  Male 140 (45.8) 85 (43.4) 51 (61.4) 4 (14.8) BMI kg/m 2 , median (IQR) 27.8 (17.2–55.6) 28.5 (18.3–55.6) 27.3 (18.2–45) 26.1 (17.2–38.4) ASA  II 183 (59.8) 120 (61.2) 42 (50.6) 21 (77.8)  III 120 (39.2) 75 (38.3) 39 (47) 6 (22.2)  IV 3 (1) 1 (0.5) 2 (2.4) 0 (0) Previous abdominal surgery 162 (52.9) 111 (56.6) 32 (38.6) 19 (70.4) All values are count (percentages), unless specified otherwise NICE Natural orifice intracorporeal anastomosis with transrectal extraction, BMI Body Mass Index, ASA American society of anesthesiology Table 2 Diagnosis n  = 306 (100) Recurrent diverticulitis 103 (33.6) Complicated Diverticulitis 93 (30.4)  Abscess 36 (11.7)  Fistula 45 (14.8)  Stricture 12 (3.9) Colorectal neoplasm 83 (27.1) Other  Endometriosis 15 (4.9)  Rectal prolapse 6 (2)  Crohn’s/colitis 6 (2) All values are count (percentages), unless specified otherwise Demographic and clinical features of patients undergoing Robotic NICE procedure All values are count (percentages), unless specified otherwise NICE Natural orifice intracorporeal anastomosis with transrectal extraction, BMI Body Mass Index, ASA American society of anesthesiology Diagnosis All values are count (percentages), unless specified otherwise In terms of the procedure type, the majority were anterior resection (54.2%) followed by low anterior resection (45.8%). The overall median operative time was 219 min (range, 99–510) and the median estimated blood loss was 50 ml (range, 0–600). The median operative time for those with neoplasm was 230 min, for complicated diverticulitis was 222 min, and for recurrent diverticulitis was 186 min. Splenic flexure mobilization was performed in 83.7% of the patients. Intraoperative outcomes are summarized in Table 3 . Table 3 Intraoperative variables of patients undergoing Robotic NICE procedure All cases n  = 306 (100) Diverticulitis n  = 196 (64.1) Neoplasm/cancer n  = 83 (27.1) Others n  = 27 (8.8) Da Vinci model Xi 306 (100) 196 (100) 83 (100) 27 (100) Type of surgery  AR 166 (54.2) 121 (61.7) 35 (42.2) 10 (37)  LAR 140 (45.8) 75 (38.3) 48 (57.8) 17 (63) Operative time in minutes, median (range) 219 (99–510) 204.5 (99–449) 230 (118–510) *203.5 (110–353) Estimated blood loss in ml, median (range) 50 (0–600) 50 (0–300) 50 (10–200) 50 (25–600) Intraoperative transfusions 0 (0) 0 (0) 0 (0) 0 (0) Diverting loop ileostomy 40 (13.1) 27 (13.8) 11 (13.3) 2 (7.4) Splenic flexure takedown 256 (83.7) 182 (92.9) 69 (83.1) 5 (18.5) Intraoperative complications** 5 (1.6) 0 (0) 0 (0) 0 (0) All values are count (percentages), unless specified otherwise NICE Natural orifice intracorporeal anastomosis with transrectal extraction, AR Anterior resection, LAR Low anterior resection *Operative time not reported in cases involving more than one specialty **Three patients had rectal tear during extraction, two patients had a splenic capsular tear Intraoperative variables of patients undergoing Robotic NICE procedure All values are count (percentages), unless specified otherwise NICE Natural orifice intracorporeal anastomosis with transrectal extraction, AR Anterior resection, LAR Low anterior resection *Operative time not reported in cases involving more than one specialty **Three patients had rectal tear during extraction, two patients had a splenic capsular tear Transrectal extraction of specimen was successfully achieved in 292 patients (95.4%). In 14 patients (4.5%), an abdominal incision was required due to inability to extract a bulky specimen through the rectum. Of these cases, 13 had malignant disease. Of the 223 patients with benign disease, 36.3% ( n  = 81) underwent a thinning maneuver to facilitate transrectal extraction. There was a total of 5 (1.6%) intraoperative complications. In three patients, trauma during the extraction process resulted in a rectal tear that required proctorrhaphy. In two patients a splenic capsular tear was encountered during splenic flexure takedown. In all cases (100%), an intracorporeal end-to-end colorectal anastomosis was successfully achieved either with a circular stapler (93.5%) or handsewn technique (6.5%). There were no conversions to open or other MIS approaches. Table 4 . Table 4 Success rate of TRSE and ICA All cases n  = 306 (100) Diverticulitis n  = 196 (64.1) Neoplasm/cancer n  = 83 (27.1) Others n  = 27 (8.8) Specimen extraction  Transrectal 292 (95.4) 195 (99.5) 70 (84.3) 27 (100)  Transabdominal 14 (4.6) 1 (0.5) 13 (15.7) 0 (0) Anastomosis location  Upper rectum 166 (54.2) 121 (61.7) 35 (42.2) 10 (37)  Mid rectum 77 (25.2) 51 (26) 20 (24.1) 6 (22.2)  Low rectum 63 (20.6) 24 (12.2) 28 (33.7) 11 (40.7) Anastomosis  ICA 306 (100) 196 (100) 83 (100) 27 (100)  ECA 0 (0) 0 (0) 0 (0) 0 (0) Converted to open or other MIS 0 (0) 0 (0) 0 (0) 0 (0) All values are count (percentages), unless specified otherwise ICA Intracorporeal anastomosis, ECA Extracorporeal anastomosis, MIS Minimally invasive surgery Success rate of TRSE and ICA All values are count (percentages), unless specified otherwise ICA Intracorporeal anastomosis, ECA Extracorporeal anastomosis, MIS Minimally invasive surgery The median time to return of bowel function was 20 h (range, 4–96) and the mean length of stay was 2.4 days (1–15). There overall postoperative complications rate was 12.4%. Eight patients (2.6%) experienced postoperative ileus. There were no superficial or deep surgical site infection (SSI). Eleven patients (3.6%) developed organ SSI space including 5 patients with intra-abdominal abscess and 4 patients with anastomotic leak. There was one mortality (0.3%) due to toxic megacolon from resistant Clostridium difficile. The 30-day reoperation rate was 2.9% ( n  = 9) including six patients presenting with organ space SSI and three patients for postoperative obstruction at diverting loop ileostomy site. There were no complains of fecal incontinence at the time of the last postoperative follow up in any patient. However, formal anatomical and functional pelvic floor analysis was not performed in this cohort. Postoperative outcomes are in Tables 5 and 6 . Table 5 Postoperative outcomes of patients undergoing robotic NICE procedure All cases n  = 306 (100) Diverticulitis n  = 196 (64.1) Neoplasm/cancer n  = 83 (27.1) Others n  = 27 (8.8) Return of bowel function in hours, median (range) 20 (4–96) 19 (5–96) 21.5 (4–60) N/p Length of Hospital Stay in days, mean (range) 2.4 (1–15) 2.5 (1–14) 2.0 (1–15) N/p Postoperative ICU 1 (0.3) 0 (0) 1 (1.2) 0 (0) 30-day readmissions 21 (6.9) 12 (6.1) 5 (6) 4 (14.8) 30-day reoperations 9 (2.9) 8 (4.1) 1 (1.2) 0 (0) 30-day mortality 1 (0.3) 1 (0.5) 0 (0) 0 (0) N/p Not reported due to combined specialty surgery All values are count (percentages), unless specified otherwise NICE Natural orifice intracorporeal anastomosis with transrectal, ICU Intensive care unite, SSI Surgical site infection Table 6 Postoperative complications All cases n  = 306 (100) Diverticulitis n  = 196 (64.1) Neoplasm/cancer n  = 83 (27.1) Others n  = 27 (8.8) Overall postoperative complications 38 (12.4) 27 (13) 7 (8.4) 4 (14.8) Atrial fibrillation 1 (0.3) 1 (0.5) 0 (0) 0 (0) Thrombosis 1 (0.3) 0 1 (1.2) Stroke 1 (0.3) 1 (0.5) 0 (0) 0 (0) Blood transfusion 1 (0.3) 1 (0.5) 0 (0) 0 (0) Dehydration/High output ileostomy 3 (1) * 3 (1.5) * 0 (0) 0 (0) Urinary tract infection/pyelonephritis 2 (0.7) 2 (1) 0 (0) 0 (0) Anal fissure 1 (0.3) 1 (0.5) 0 (0) 0 (0) Crohn’s flare 1 (0.3) 1 (0.5) 0 (0) 0 (0) C difficile colitis 3 (1) 3 (1.5) 0 (0) 0 (0) Perianal abscess drainage 1 (0.3) 0 (0) 1 (1.2) 0 (0) Postoperative ileus 8 (2.6) 6 (3) 2 (2.4) 0 (0) SSI  Superficial 0 (0) 0 (0) 0 (0) 0 (0)  Deep 0 (0) 0 (0) 0 (0) 0 (0)  Organ Space 11 (3.6) 6 (3) 1 (1.2) 4 (14.8) Obstruction at diverting ileostomy site 3 (1) 2 (1) 1 (1.2) 0 (0) Respiratory insufficiency 1 (0.3) 0 (0) 1 (1.2) 0 (0) All values are count (percentages), unless specified otherwise SSI Surgical site infection *Percentage reported based on entire cohort. When calculated based on the subset of 40 ileostomy patients, there was a 7.5% rate of high ileostomy output dehydration Postoperative outcomes of patients undergoing robotic NICE procedure N/p Not reported due to combined specialty surgery All values are count (percentages), unless specified otherwise NICE Natural orifice intracorporeal anastomosis with transrectal, ICU Intensive care unite, SSI Surgical site infection Postoperative complications All values are count (percentages), unless specified otherwise SSI Surgical site infection *Percentage reported based on entire cohort. When calculated based on the subset of 40 ileostomy patients, there was a 7.5% rate of high ileostomy output dehydration Table 7 depicts the subset of patients ( n  = 83) with colorectal neoplasm. The median tumor size was 3 cm (range, 1–9). The mesorectal grade was complete in all cases. Distal and proximal margins were negative in all cases. One patient had positive microscopic disease at circumferential margin (pT4 stage). This patient had known distant disease and had received neoadjuvant chemotherapy. The median examined lymph nodes were 19 (range, 6–36) and 37.3% patients had node positive disease. Recurrence and survival outcomes are not reported due to insufficient follow up interval. Table 7 Subset of patients with neoplasm or cancer undergoing Robotic NICE procedure n  = 83 (%) Preoperative chemotherapy 7 (8.4) Abdominal site extraction due to bulky specimen 13 (15.7) Surgical pathology  Adenoma 3  No lesion or no residual cancer 6  AJCC stage 0 3  AJCC stage 1 17  AJCC stage 2 15  AJCC stage 3 32  AJCC stage 4 7 Examined lymph nodes, median (range) 19 (6–36) Presence of positive lymph nodes 31 (37.3) Circumferential Resection Margins  Compromised (≤ 1 mm) please see below 1 (1.2)  Uncompromised (> 1 mm) 82 (98.8) Clear distal margins (> 2 cm) 83 (100) Clear proximal margins (> 2 cm) 83 (100) Tumor size in cm, median (range) 3.0 (0.7–9.0) All values are count (percentages), unless specified otherwise NICE Natural orifice intracorporeal anastomosis with transrectal extraction, AJCC American joint committee on cancer Subset of patients with neoplasm or cancer undergoing Robotic NICE procedure All values are count (percentages), unless specified otherwise NICE Natural orifice intracorporeal anastomosis with transrectal extraction, AJCC American joint committee on cancer

Franklin et al. were the first to describe laparoscopic sigmoid resection with intracorporeal anastomosis and transrectal extraction of specimen, in 1993 [ 5 ]. Over the next two decades, several case series reporting feasibility and safety of laparoscopic natural orifice specimen extraction (NOSE) procedures were published [ 14 – 19 ]. In two randomized control trials, one including malignant and the other including both benign and malignant left-sided disease, the laparoscopic NOSE arm had less postoperative pain, lower use of opioids, lower rates of wound infections, and better cosmesis compared to the conventional laparoscopic arm [ 8 , 10 ]. In spite of the enthusiasm with natural orifice techniques in left-sided colorectal resections, adaptation has been limited chiefly due to technical barriers and there has been a general consensus that these procedures should be reserved for selected patients only [ 20 , 21 ]. In 2018 we published the first report describing the application of daVinci Xi® robotic technology for this natural orifice total intracorporeal surgery and termed it the NICE Procedure [ 11 , 12 ]. We thereafter developed a stepwise technique with ten defined steps to facilitate reproducible results in an expanded patient population [ 13 ]. As such the NICE procedure has become our preferred approach regardless of patient characteristics such as age, ASA, BMI or disease type. In this report, we present the largest case series of robotic NICE procedure on consecutive and unselected cohort of patients undergoing left-sided colorectal resections for benign and malignant disease. This was an intention to treat cohort and we therefore excluded patients in which we planned to perform a diverting loop ileostomy such as those who had undergone neoadjuvant radiation for rectal cancer. In these patients, we use the ileostomy site to extract the specimen. In our series however, we had a subset of 40 patients in which a decision was made to perform a diverting loop ileostomy intraoperatively. In most of these cases, the decision was made following the natural orifice extraction and based on the integrity of the anastomosis as well as clinical features of the patient. In this subset of 40 patients, all underwent successful ICA and all but 2 cases, in which bulky malignant disease was present, had transrectal extraction of the specimen. We feel that transrectal extraction in these cases still serves several advantages and avoids manipulation and extension of the incision at the ileostomy site. Although BMI greater than 30–35 kg/m2 has been reported to be a limitation for successful transrectal extraction of specimen, [ 10 , 18 , 20 – 23 ] we have not found this to be the case. In our cohort of unselected patients 34.3% of the patients had BMI OVER 30 kg/m 2 which is representative of a typical patient population in Texas as reported by the CDC. [ 24 ] Obesity was not associated with inability to extract the specimen or perform an ICA. In fact, the formation of an abdominal extraction incision is typically more cumbersome in those with large abdominal girth. Additionally, the extracorporeal portions of the anastomosis may be fraught with difficulty due to limitations of the tissue reaching above the abdominal wall. Since obesity is associated with increased risks of SSI as well as higher rates of incisional hernia formation, obese patients might benefit the most from avoidance of an abdominal wall incision [ 25 – 30 ]. It has been suggested that ICA and TRSE during robotic colorectal resection is associated with increased operative time [ 6 ]. However, these studies were performed on the Si daVinci® platform which has limitations resulting from instrument clashing and the propensity for re-docking to gain access to the left upper quadrant for splenic flexure mobilization. The Xi daVinci® platform was chiefly utilized during our study and we found the medium operative time of 219 min fared well when compared to published operative times ranging from 130 to 256 min for robotic colorectal left-sided resections with ECA and transabdominal extraction.[ 31 – 34 , 34 – 40 ]. Furthermore, for benign disease, we perform mesenteric sparing dissection which is an efficient technique that avoids the need for exposure and access to the retroperitoneal structures [ 41 ]. Adhering to the stepwise process for completing the intracorporeal portions combined with the absence of creating and closing an abdominal wall extraction site results in the optimization of the operative time. Additionally, more recent advances in instrumentation such as the vessel sealer extend and upgrades in the energy generator yields faster division of the tissue, hemostasis, and other surgical maneuvers. When assessing success rates of the NICE procedure we found that all consecutive patients were able to undergo an intracorporeal anastomosis. We predominantly used the 31 mm circular stapler because the larger chamber can receive additional tissue that usually results from a double purse-string closure of the proximal bowel and rectum. In 20 patients we opted for a handsewn end-to-end anastomosis to preserve the bowel length in the presence of possible tension in those with colorectal cancer ( n  = 3) or in cases in which numerous large-mouth diverticula were present at the anastomotic site ( n  = 17). All but one patient (99%) with benign pathology had successful TRSE compared to those presenting with malignant disease (84.3%). This high success rate is consistent with reports that identify the benign disease as a favorable indicator of achieving TRSE. In about 25% of those with benign disease, the specimen was bulky and required reduction prior to extraction by thinning of the mesentery. It is important to note that we do not thin or manipulate the specimen in any patient in which there is a suspicion of malignancy based on preoperative evaluation or intraoperative findings. In the one patient with bulky benign disease, we opted not to shave the mesentery because pathology was in question and therefore, we performed an abdominal extraction. Thirteen patients with malignant disease required abdominal wall extraction due to large cancer size and/or bulky mesentery. Thinning of the mesentery is not an option in this cohort. Although it has been reported that the upper limit of tumor size is from 4 to 6 cm [ 8 , 9 , 22 , 42 ]. for TRSE, we found that size of tumor did not necessarily correlate with success rate because even small tumors may be associated with a very thick and bulky specimen and therefore not appropriate for natural orifice extraction. Factors other than BMI, presence of benign disease, and small tumor size that have been associated with successful natural orifice transrectal extraction include female gender and the location of the disease in the distal colon or rectum [20, 48]. We found similar tendencies in our series. The success rates of TRSE was greater in females and among the 14 patients in which TRSE was not successful, 9 (64%) were male. In addition, all patients in our series had distal disease in the left colon, sigmoid and rectum which likely contributed to the overall high success rates of TRSE. The median return of bowel function was 20 h, and the median LOS was 2.4 days. In published works reporting laparoscopic NOSE, the length of stay varied from a range of 1.2–3.5 days, and from 4.2 to 11.8 days [ 6 – 10 , 42 – 44 ]. These reports are across various techniques, patient selection and recovery protocols which could explain the variation in results [ 10 , 45 ]. We enter all patients into an multimodal opioid sparing, enhanced recovery protocol, what could contribute to our cohorts’ improved postoperative recovery. Only 2.6% patients experience post-operative ileus. Our overall complication rate was 12.4% including 3.6% with SSI. This compares favorably to SSI superficial and deep rates of 4.7–6.2% in conventional robotic ECA surgery [ 31 , 37 , 46 ]. There were no occurrences of superficial or deep SSI which is a direct reflection of avoiding an extraction incision. In conclusion, the NICE procedure is associated with a very high success rate in regard to both intracorporeal anastomosis and transrectal specimen extraction in a large cohort of unselected and consecutive patients. This stepwise approach using the robotic platform may facilitate widespread adoption of this natural orifice total intracorporeal surgery with numerous inherent patient benefits.