Using AI and computer vision to analyze technical proficiency in robotic surgery.

BACKGROUND: Intraoperative skills assessment is time-consuming and subjective; an efficient and objective computer vision-based approach for feedback is desired. In this work, we aim to design and validate an interpretable automated method to evaluate technical proficiency using colorectal robotic surgery videos with artificial intelligence. METHODS: 92 curated clips of peritoneal closure were characterized by both board-certified surgeons and a computer vision AI algorithm to compare the measures of surgical skill. For human ratings, six surgeons graded clips according to the GEARS assessment tool; for AI assessment, deep learning computer vision algorithms for surgical tool detection and tracking were developed and implemented. RESULTS: For the GEARS category of efficiency, we observe a positive correlation between human expert ratings of technical efficiency and AI-determined total tool movement (r = - 0.72). Additionally, we show that more proficient surgeons perform closure with significantly less tool movement compared to less proficient surgeons (p < 0.001). For the GEARS category of bimanual dexterity, a positive correlation between expert ratings of bimanual dexterity and the AI model's calculated measure of bimanual movement based on simultaneous tool movement (r = 0.48) was also observed. On average, we also find that higher skill clips have significantly more simultaneous movement in both hands compared to lower skill clips (p < 0.001). CONCLUSIONS: In this study, measurements of technical proficiency extracted from AI algorithms are shown to correlate with those given by expert surgeons. Although we target measurements of efficiency and bimanual dexterity, this work suggests that artificial intelligence through computer vision holds promise for efficiently standardizing grading of surgical technique, which may help in surgical skills training.

Defining the Role of Minimally Invasive Proctectomy for Locally Advanced Rectal Adenocarcinoma.

OBJECTIVE: National examination of open proctectomy (OP), laparoscopic proctectomy (LP), and robotic proctectomy (RP) in pathological outcomes and overall survival (OS). BACKGROUND: Surgical management for rectal adenocarcinoma is evolving towards utilization of LP and RP. However, the oncological impacts of a minimally invasive approach to rectal cancer have yet to be defined. METHODS: Retrospective review of the National Cancer Database identified patients with nonmetastatic locally advanced rectal adenocarcinoma from 2010 to 2014, who underwent neoadjuvant chemoradiation, surgical resection, and adjuvant therapy. Cases were stratified by surgical approach. Multivariate analysis was used to compare pathological outcomes. Cox proportional-hazard modeling and Kaplan-Meier analyses were used to estimate long-term OS. RESULTS: Of 6313 cases identified, 53.8% underwent OP, 31.8% underwent LP, and 14.3% underwent RP. Higher-volume academic/research and comprehensive community centers combined to perform 80% of laparoscopic cases and 83% of robotic cases. In an intent-to-treat model, multivariate analysis demonstrated superior circumferential margin negativity rates with LP compared with OP (odds ratio 1.34, 95% confidence interval 1.02-1.77, P = 0.036). Cox proportional-hazard modeling demonstrated a lower death hazard ratio for LP compared with OP (hazard ratio 0.81, 95% confidence interval 0.67-0.99, P = 0.037). Kaplan-Meier analysis demonstrated a 5-year OS of 81% in LP compared with 78% in RP and 76% in OP (P = 0.0198). CONCLUSION: In the hands of experienced colorectal specialists treating selected patients, LP may be a valuable operative technique that is associated with oncological benefits. Further exploration of pathological outcomes and long-term survival by means of prospective randomized trials may offer more definitive conclusions regarding comparisons of open and minimally invasive technique.

Is Fecal Diversion Needed in Pelvic Anastomoses During Hyperthermic Intraperitoneal Chemotherapy (HIPEC)?

BACKGROUND: The role of fecal diversion with pelvic anastomosis during cytoreductive surgery (CRS) with hyperthermic intraperitoneal chemotherapy (HIPEC) is not well defined. METHODS: A retrospective review of patients who underwent CRS and HIPEC between 2009 and 2016 was performed to identify those with a pelvic anastomosis (colorectal, ileorectal, or coloanal anastomosis). RESULTS: The study identified 73 patients who underwent CRS and HIPEC at three different institutions between July 2009 and June of 2016. Of these patients, 32 (44%) underwent a primary anastomosis with a diverting ileostomy, whereas 41 (56%) underwent a primary anastomosis without fecal diversion. The anastomotic leak rate for the no-diversion group was 22% compared with 0% for the group with a diverting ileostomy (p < 0.01). The 90-day mortality rate for the no-diversion group was 7.1%. The hospital stay was 14.1 ± 8.0 days in the diversion group compared with 17.9 ± 12.5 days in the no-diversion group (p = 0.12). Of those patients with a diverting ileostomy, 68% (n = 22) had their bowel continuity restored, 18% of which required a laparotomy for reversal. Postoperative complications occurred for 50% of those who required a laparotomy and for 44% of those who did not require a laparotomy (p = 0.84). CONCLUSION: Diverting ileostomies in patients with a pelvic anastomosis undergoing CRS and HIPEC are associated with a significantly reduced anastomotic leak rate. Reversal of the diverting ileostomy in this patient population required a laparotomy in 18% of the cases and had an associated morbidity rate of 50%.

Volume and outcomes relationship in laparoscopic diaphragmatic hernia repair.

BACKGROUND: There is no published data regarding the relationship between hospital volume and outcomes in patients undergoing laparoscopic diaphragmatic hernia repair. We hypothesize that hospitals performing high case volume have improved outcomes compared to low-volume hospitals. MATERIALS AND METHODS: We reviewed the National Inpatient Sample (NIS) database between 2008 and 2012 for adults with the diagnosis of diaphragmatic hernia who underwent elective laparoscopic repair of diaphragmatic Hernia and/or Nissen fundoplication. Pediatric, emergent, and open cases were excluded. Main outcome measures included logistic regression analysis of factors predictive of in-hospital mortality and outcomes according to annual hospital case volume. RESULTS: A total of 31,228 laparoscopic diaphragmatic hernia operations were analyzed. The overall in-hospital mortality was 0.14%. Risk factors for higher in-hospital mortality included renal failure (AOR: 6.26; 95% CI: 2.48-15.78; p < 0.001), age>60 years (AOR: 5.06; 95% CI: 2.38-10.76; p < 0.001), and CHF (AOR: 3.80; 95% CI: 1.39-10.38; p = 0.009) while an incremental increase in volume of 10 cases/year (AOR: 0.89; 95% CI: 0.81-0.98; p = 0.019) and diabetes (AOR: 0.34; 95% CI: 0.12-0.93; p = 0.036) decreases mortality. There was a small but significant inverse relationship between hospital case volume and mortality with a 10% reduction in adjusted odds of in-hospital mortality for every increase in 10 cases per year. Using 10 cases per year as the volume threshold, low-volume hospitals (≤10 cases/year) had almost a twofold higher mortality compared to high-volume hospitals (0.23 vs. 0.12%, respectively, p = 0.02). CONCLUSIONS: There was a small but significant inverse relationship between the hospitals' case volume and mortality in laparoscopic diaphragmatic hernia repair.

Improved survival with adjuvant chemotherapy in locally advanced rectal cancer patients treated with preoperative chemoradiation regardless of pathologic response.

OBJECTIVE: The aim of this study is to examine the effect of postoperative chemotherapy on survival in patients with stage II or III rectal adenocarcinoma who undergo neoadjuvant chemoradiation (CRT) and surgical resection. METHODS: A retrospective review of the National Cancer Database (NCDB) from 2006 to 2013 was performed. Cases were analyzed based on pathologic complete response (pCR) status and use of adjuvant therapy. The Kaplan-Meier method was used to estimate overall survival probabilities. RESULTS: 23,045 cases were identified, of which 5832 (25.31%) achieved pCR. In the pCR group, 1513 (25.9%) received adjuvant chemotherapy, and in the non-pCR group, 5966 (34.7%) received adjuvant therapy. In the pCR group, five-year survival probability was 87% (95% CI 84%-89%) with adjuvant therapy and 81% (95% CI 79%-82%) without adjuvant therapy. In the non-pCR group, five-year survival probability was 78% (95% CI 76%-79%) with adjuvant therapy and 70% (95% CI 69%-71%) without adjuvant therapy. In the non-pCR and node-negative subgroup (ypN-), five-year survival probability was 86% (95% CI 84%-88%) with adjuvant therapy and 76% (95% CI 74%-77%) without adjuvant therapy. In the non-pCR and node-positive subgroup (ypN+), five-year survival probability was 67% (95% CI 65%-70%) with adjuvant therapy and 60% (95% CI 58%-63%) without adjuvant therapy. CONCLUSIONS: Adjuvant chemotherapy in stage II or III rectal adenocarcinoma is associated with increased five-year survival probability regardless of pCR status. We observed similar survival outcomes among non-pCR ypN- treated with adjuvant chemotherapy compared with patients achieving pCR treated with adjuvant chemotherapy.

Lymph Node Positivity in Appendiceal Adenocarcinoma: Should Size Matter?

BACKGROUND: The management algorithm for appendiceal adenocarcinoma is not well defined. This study sought to determine whether tumor size or depth of invasion better correlates with the presence of lymph node metastases in appendiceal adenocarcinoma, and to compare these rates with colon adenocarcinoma. STUDY DESIGN: A retrospective review of the National Cancer Database was performed to identify patients with appendiceal or colonic adenocarcinoma from 2004 to 2013 who underwent surgical resection. Cases were categorized by tumor size and by T stage. Rates of lymph node metastases were examined as a function of size and T stage. RESULTS: A total of 3,402 appendiceal and 314,864 colonic cases were identified. For appendiceal adenocarcinoma, larger tumor size was associated with higher T stage: Pearson correlation of 0.41 (95% CI 0.408 to 0.414; p < 0.001). Lymph node metastases were present in 19.1%, 27.8%, 39.6%, 39.4%, 42.4% and 39.1% for tumor sizes <1 cm, >1 to 2 cm, >2 to 3 cm, >3 to 4 cm, >4 to 5 cm, and >5 cm, respectively. Lymph node metastases were present in 0%, 11.2%, 12.3%, 35.5%, and 40.0% for in situ, T1, T2, T3, and T4 tumors, respectively. There was no difference in the rates of lymph node metastases between appendiceal and colonic adenocarcinoma for tumor sizes <3 cm, or for in situ and T1 tumors. Rates of lymph node metastases are higher in colonic adenocarcinoma for tumor sizes >3 cm and for T2, T3, and T4 tumors (p < 0.01). CONCLUSIONS: In appendiceal adenocarcinoma, the rate of lymph node metastases is substantial, even for small tumors. Tumor size should play no role in the decision of whether to perform a hemicolectomy. Appendectomy alone does not produce an adequate lymph node sample. Right hemicolectomy should be performed for all appendiceal adenocarcinomas.

Analysis of Endoscopic Retrograde Cholangiopancreatography after Positive Intraoperative Cholangiogram: Is It Necessary?

The necessity of routine endoscopic retrograde cholangiopancreatography (ERCP) after positive intraoperative cholangiogram (IOC) during laparoscopic cholecystectomy is not well defined. We aimed to examine the incidence of positive IOC among patients who undergo IOC during cholecystectomy and the rate of subsequent ERCP stone extraction. The Nationwide Inpatient Sample database was reviewed for all patients undergoing cholecystectomy with IOC from 2002 to 2012. Patients were then analyzed for ERCP and stone extraction. A total of 73,508 patients who underwent cholecystectomy with IOC for a diagnosis of acute cholecystitis and found to have a bile duct stone were identified. Of these patients, 5915 underwent subsequent ERCP. In the patients that underwent subsequent ERCP, 1478 had a documented stone extraction during ERCP. The rate of stone extraction in the ERCP subset is 25 per cent, which is 2 per cent of all patients who had a positive IOC. The rate of stone extraction after positive IOC is low. Positive IOC may not warrant a routine postoperative ERCP. Our results suggest that clinical monitoring of patients with positive IOC is reasonable, as the majority of patients with a positive IOC ultimately have no stone extraction.

Hand-Assisted Laparoscopic Donor Nephrectomy in Complete Situs Inversus.

Complete situs inversus is a rare congenital anomaly characterized by transposition of organs. We report a case of renal transplantation using a kidney from a living complete situs inversus donor. The recipient was a 59-year-old female with end-stage renal disease because of type 2 diabetes mellitus. The donor was the 56-year-old sister of the recipient with complete situs inversus. CT angiogram of the abdomen and pelvis showed complete situs inversus and an otherwise normal appearance of the bilateral kidneys with patent bilateral single renal arteries and longer renal vein in the right kidney. The patient was taken to the operating room for a hand-assisted laparoscopic right donor nephrectomy. The patient tolerated the procedure well and was discharged home in good condition on postoperative day 1. The recipient experienced no episodes of acute rejection or infection, with serum creatinine levels of 0.8-1.2 mg/dL. Laparoscopic donor nephrectomy in a patient with complete situs inversus remains a technically feasible operation and the presence of situs inversus should not preclude consideration for living kidney donation.

Colorectal Surgery in Patients with HIV and AIDS: Trends and Outcomes over a 10-Year Period in the USA.

BACKGROUND: HIV has become a chronic disease, which may render this population more prone to developing the colorectal pathologies that typically affect older Americans. METHODS: A retrospective review of the Nationwide Inpatient Sample was performed to identify patients who underwent colon and rectal surgery from 2001 to 2010. Multivariate analysis was used to evaluate outcomes among the general population, patients with HIV, and patients with AIDS. RESULTS: Hospital admissions for colon and rectal procedures of patients with HIV/AIDS grew at a faster rate than all-cause admissions of patients with HIV/AIDS, with mean yearly increases of 17.8 and 2.1 %, respectively (p < 0.05). Patients with HIV/AIDS undergoing colon and rectal operations for cancer, polyps, diverticular disease, and Clostridium difficile were younger than the general population (51 vs. 65 years; p < 0.01). AIDS was independently associated with increased odds of mortality (OR 2.11; 95 % CI 1.24, 3.61), wound complications (OR 1.53; 95 % CI 1.09, 2.17), and pneumonia (OR 2.02; 95 % CI 1.33, 3.08). Risk-adjusted outcomes of colorectal surgery in patients with HIV did not differ significantly from the general population. CONCLUSION: Postoperative outcomes in patients with HIV are similar to the general population, while patients with AIDS have a higher risk of mortality and certain complications.

Racial Disparities in Access and Outcomes of Cholecystectomy in the United States.

Disparities in access to health care between white and minority patients are well described. We aimed to analyze the trends and outcomes of cholecystectomy based on racial classification. The Nationwide Inpatient Sample database was reviewed for all patients undergoing cholecystectomy from 2009 to 2012. Patients were stratified as white or non-white. A total of 243,536 patients were analyzed: 159,901 white and 83,635 non-white. Non-white patients had significantly higher proportions of Medicaid (25% vs 9.3%), self-pay (14% vs 7.1%), and no-charge (1.8% vs 0.64%). Non-white patients had significantly higher rates of emergent admission (84% vs 78%) compared with the white patients. Multivariate analysis revealed that non-whites had a significantly longer length of stay [mean difference of 0.14 days, 95% confidence interval (CI) 0.08-0.20] and higher total hospital charges (mean difference of $6748.00, 95% CI 5994.19-7501.81) than whites, despite a lower morbidity (odds ratio 0.94, 95% CI 0.90-0.98). Use of laparoscopy and mortality were not different. These differences persisted on subgroup analysis by insurance type. These findings suggest a gap in access to and outcomes of cholecystectomy in the minority population nationwide.