Citations

ABSTRACT Background Evaluating fairness in artificial intelligence (AI) is essential before clinical implementation to prevent unintended harm across diverse patient populations. This study assesses performance differences of an image-based surgical site infection (SSI) detection model across age, gender, geography, race, and skin color. Methods Model development/validation study using retrospective, multicenter data across surgical subspecialties from adults who submitted an image of their incision via the electronic health record within 30 days of their procedure between 2019-2024. Area under receiver operating curve (AUROC) was compared by age (<55 vs. ≥55 years old), gender (male vs. female), geography (Rochester, Minnesota vs. Jacksonville, Florida vs. Scottsdale, Arizona vs. regional Minnesota/Wisconsin areas), race (White/Caucasian vs. non-White/Caucasian), and skin color (Fitpatrick I-III vs. IV-VI and Monk A-E vs. F-J). Results Training data contained 13,702 images from 4,274 patients between 2019-2022 (50.1% ≥55 years old; 63.2% female; 91.6% White/Caucasian; 58.5% Rochester, 17.3% Florida, 10.5% Arizona; 96.3% light skin tone), of which 1,348 images (9.8%) had an SSI. No differences in SSI detection existed for age, gender, race, or skin color (AUROC 0.83-0.87, p>0.09), however geographical differences were significant: lowest AUROC 0.83 (95% CI 0.77-0.88) vs. highest AUROC 0.87 (95% CI 0.84-0.89, p=0.01). Temporal validation using 4,242 images from 1,353 patients between 2022-2023 showed intact fairness across subgroups (ΔAUROC ≤0.06) except geography (ΔAUROC ≤0.16). Conclusions The model behaved fairly across age, gender, race, and skin color, however geographical differences persisted. These findings have practical implications underscoring the importance of evaluating bias in clinical AI before deployment.

ABSTRACT Background ACS‐NSQIP began capturing operative approach in 2022. With increased adoption of minimally invasive (MIS) techniques for malignancy, a contemporary review of benchmark outcomes for open, laparoscopic and robotic gastrectomy is warranted. Methods We performed a retrospective cohort study of adults in ACS‐NSQIP (2022–2024) undergoing partial or total gastrectomy (CPT 43631‐4 and 43620‐2) for cancer. Outcomes included operative time, length of stay (LOS), unplanned conversion‐to‐open, and 30‐day morbidity, mortality, and readmission. Multivariable logistic regression evaluated patient and operative factors associated with morbidity. Results Among 1,459 patients (median age 66, 60.4% male), 66.8% underwent partial gastrectomy and 33.2% total gastrectomy. Approaches were 61.2% open, 19.5% laparoscopic, and 19.3% robotic. Thirty‐day mortality (1.8%), major complications (18.2%), and readmissions (10.3%) did not differ by approach. Unplanned conversion‐to‐open occurred in 14.0% of MIS. Robotic had longer median operative time of +1.2 and +1.7 h than laparoscopic and open and median −1 and −2 days shorter LOS respectively ( p < 0.001). Hypoalbuminemia (OR 3.10), CKD ≥ 3 (OR 1.75), and anemia (OR 1.48) were independently associated with major complications ( p < 0.05). Conclusion Operative approach influences operative time and LOS but not short‐term morbidity and mortality. The newly available ACS‐NSQIP operative approach variable enables approach‐specific benchmarks for cancer‐related gastrectomy.

10043 Background: Tumor resection is a cornerstone of therapy for neuroblastoma (NB) and Wilms tumor (WT). Minimally invasive surgical approaches such as laparoscopic or robotic tumor resection hold promise to reduce surgical morbidity while maintaining excellent oncologic outcomes. This study aims to determine the utilization and outcomes of laparoscopic and robotic resection of NB and WT among children in a national cohort, hypothesizing that utilization is increasing over time and outcomes are not inferior to open resection. Methods: A retrospective cohort analysis of patients <18 years old in the 2020-2023 National Cancer Database who underwent resection of NB or WT was performed. Surgical approach trends were assessed via Cochran-Armitage tests. Conversions to open surgery, length of stay (LOS), and readmissions were compared between approaches using Kruskal-Wallis, Chi-square, and Fisher’s exact tests. Results: Among 1,262 children who underwent resection of NB (n=581) or WT (n=681), open procedures were most common (83.4%), followed by laparoscopic (13.9%), and robotic (2.7%). Utilization of each approach was stable over time aside from increased robotic resection of NB from 0% to 4% (p=.024). Approach did not vary by patient age (Table 1). Median WT size and lymph node yield were similar between approaches (both p>.05). In NB, laparoscopic procedures were performed on smaller tumors than open or robotic (p<.001). There were no robotic conversions to open; 11.4% of laparoscopic procedures were converted without change over time or difference between NB and WT (both p>.05). Median LOS was greater for open procedures than laparoscopic or robotic (p<.001). Readmission rates (p=.64) and margin status (p=.63) were similar between approaches. Outcomes did not vary by diagnosis. Conclusions: Utilization of minimally invasive approaches for NB or WT resection remains low despite noninferior or improved outcomes. Efforts to safely increase utilization are needed, especially for robotics, so that survivorship can be improved through reduced surgical morbidity. Differences in patient characteristics and outcomes by surgical approach. Characteristic or Outcome Open Laparoscopic Robotic p value Age in years (median, IQR) 3 (1, 4) 3 (1, 5) 3 (1, 7) 0.08 % female 53% 53% 44% 0.59 NB tumor size in mm (median, IQR) 72 (50, 101) 47 (35, 71) 70 (51, 86) <0.001 WT tumor size in mm (median, IQR) 110 (75, 140) 115 (93, 150) 93 (54, 125) 0.10 % metastatic disease 33% 30% 15% 0.07 % neoadjuvant therapy 35% 27% 22% 0.014 % conversion to open n/a 11% 0% 0.05 Negative margins 86% 83% 96% 0.63 # Lymph nodes assessed in WT (median, IQR) 5 (3, 8) 6 (3, 10) 4 (3, 9) 0.62 LOS in days (median, IQR) 5 (4, 7) 4 (2, 6) 4 (1, 5) <0.001 % readmission 15% 14% 9% 0.64

Background: ASCO recommends extended venous thromboembolism (VTE) prophylaxis with low-molecular-weight heparin (LMWH) or direct oral anticoagulants (DOACs) following cancer surgery. However, these recommendations were based on trials using routine venography, which detects asymptomatic VTEs, and real-life efficacy is not well studied. We aimed to assess the association of extended VTE prophylaxis with the incidence of clinically significant VTE after elective cancer surgery. Patients and Methods: In this retrospective, population-based cohort study, we identified patients who underwent surgical resection for lung, breast, esophageal, gastric, pancreas, and colon cancer from 2010 to 2017 using the SEER–Medicare linked dataset. Patients were divided into 1 of 2 main groups for comparative analyses: those who received extended VTE prophylaxis after cancer surgery and those who did not. Patients who received extended VTE prophylaxis after cancer surgery were identified as those prescribed prophylactic doses of LMWH or a DOAC within 7 days of discharge, as captured in Medicare Part D files. Results: A total of 113,739 Medicare beneficiaries were identified, of whom 1,570 (1.4%) received extended VTE prophylaxis. There was a year-over-year increase in the use of extended VTE prophylaxis during the study period, rising from 0.6% in 2010 to 3.3% in 2017. The overall incidence of VTE was 1.6% at 1 month, 4.3% at 6 months, 5.7% at 12 months, with significant differences observed across cancer types ( P <.001). However, patients who received extended VTE prophylaxis had a higher 1-year cumulative incidence of postoperative VTE compared with those who did not (11.4% vs 5.3%; P <.001), an association that was consistent across cancer sites. In adjusted analyses, extended VTE prophylaxis was independently associated with an increased risk of VTE (hazard ratio, 1.6; 95% CI, 1.3–2.0; P <.001). Extended prophylaxis was also associated with a higher 1-year incidence of bleeding events compared with patients who did not receive prophylaxis (12.0% vs 8.5%; P <.001). Conclusions: The utilization of extended VTE prophylaxis after cancer surgery remains low, although it has increased every year. Paradoxically, patients who received extended VTE prophylaxis had higher incidences of clinically significant VTE and bleeding compared with those who did not, likely reflecting patient selection and surveillance bias after surgery. In light of these findings, national guidelines recommending postoperative extended VTE prophylaxis, and its use as a surrogate for quality of care, should continue to be re-examined.

Objective We sought to characterize factors associated with postoperative complications in patients undergoing primary cleft lip repair (CLR). Design The National Surgical Quality Improvement Program-Pediatric database was queried to identify patients under 12 months old undergoing CLR using CPT codes. Multivariable analysis was conducted to characterize complications and associated comorbidities. Setting A multicenter sample. Patients/Participants A total of 11 957 cases (81.9% unilateral CLR and 18.1% bilateral CLR) were identified. Interventions None. Main Outcomes/Measures The postoperative outcomes of significance were surgical site infection (SSI), wound dehiscence, and unplanned reintubation within 30 days of CLR. Results The cohort had a mean age of 4.7 ± 2.0 months and a median American Society of Anesthesiologists classification of 2. SSIs occurred in 0.6% of cases and were independently associated with bilateral cleft anatomy (OR 2.16, P = .003) and longer anesthesia time (OR per minute: 1.004, P = .009). Wound dehiscence occurred in 0.4% of patients and was independently associated with bilateral cleft anatomy (OR 4.03, P < .001) but was associated with shorter anesthesia duration (OR per minute: 0.991, P < .001). Unplanned reintubation occurred in 0.2% of cases, with these patients having significantly higher rates of multiple comorbidities. Conclusion Wound dehiscence, SSI, and unplanned reintubation were rare complications following CLR. Patients with a bilateral cleft lip had increased odds of both SSI and wound dehiscence.

Importance As gender-affirming care expands, hysterectomies for gender dysphoria are increasing, but data comparing outcomes to benign hysterectomies remain limited. Objectives The objectives of this study were to compare perioperative characteristics and 30-day outcomes between transgender and gender diverse individuals (TGDI) undergoing gender-affirming hysterectomy (GAH) and patients undergoing hysterectomy for benign indications, by surgical approach. Study Design This was a retrospective cohort study using the American College of Surgeons National Surgical Quality Improvement Program data (October 2015–December 2022). TGDI undergoing GAH were identified by gender dysphoria diagnosis and hysterectomy Current Procedural Terminology codes. Approaches included abdominal hysterectomy (AH), laparoscopic/robotic hysterectomy (L/RH), laparoscopically assisted vaginal hysterectomy (LAVH), and vaginal hysterectomy (VH). A matched cohort was created based on demographics and comorbidities. Primary outcomes were operative time, length of stay (LOS), and 30-day complications. Results A total of 975 TGDI undergoing GAH were included; 82% of GAHs were minimally invasive, mostly L/RH. GAH represented 0.7% of all benign hysterectomies, increasing over time ( P <0.001). TGDI (970) were matched to controls. In L/RH, TGDI had longer operative times (113 vs 97 min, P <0.001) but shorter LOS (0 vs 1 d, P <0.001). Similar patterns were seen in VH: longer operative times (103 vs 62 min, P <0.001) and shorter LOS (0 vs 1 d, P =0.018). In AH, TGDI had shorter LOS (1 vs 2 d, P =0.006) with similar operative times. No significant differences were found in LAVH or complications across groups. Conclusion GAH is safe with perioperative outcomes comparable to benign hysterectomy. TGDI experience shorter hospital stays and low complication rates. Promoting minimally invasive and vaginal approaches can enhance equitable, high-quality care for TGDI.

Abstract Objectives Society guidelines discourage biopsy of thymic epithelial tumours (TETs) due to the perceived risk of local seeding or pleural dissemination. However, few studies have evaluated this concern, and the true risk of preoperative biopsy remains unknown. Methods We queried the United States’ National Cancer Database for adults with localized TETs who underwent upfront R0 thymectomy 2004-2015. Use of preoperative biopsy over time was assessed with the Cochran-Armitage test. Postoperative outcomes were compared with chi-square and Wilcoxon rank sum tests. Overall survival (OS) was analysed with Kaplan-Meier and Cox proportional hazard methods. Results We identified 1056 patients. The median age was 61 years; 47% (492) were male. Use of preoperative percutaneous biopsy increased 2006-2015 from 18% to 33% (P &lt; .001). Patients who underwent biopsy had larger tumours [median 6.6 vs 5.0 cm], more open operations [72% (150) vs 62% (331)] and regional lymphadenectomies [44% (118) vs 33% (250)], and a longer length of hospitalization (median 4 vs 3 days). Thirty-day mortality was similar between the groups. There was no difference in 10-year OS between the biopsy and no-biopsy groups [75% (95% CI 69%-82%) vs 76% (95% CI 73%-80%), P = .8]. After adjusting for age, sex, primary payor, histology, and tumour size, OS was similar between the biopsy and no-biopsy groups [AHR 0.97 (95% CI 0.72-1.29), P = .9]. Conclusions In this analysis of patients with localized TETs undergoing upfront R0 thymectomy, preoperative biopsy is not associated with worse OS. More study is needed to further explore whether biopsy increases risk of recurrence.

Surgical outcomes and perioperative factors associated with pediatric cranioplasty remain incompletely understood. The authors queried the National Surgical Quality Improvement Program-Pediatric (NSQIP-P) database to evaluate outcomes and intraoperative/postoperative blood transfusions in patients undergoing cranioplasty between 2019 and 2023. A total of 249 patients (mean age 9.1 ± 5.4 y) were included. Most cranioplasties (70.7%) utilized alloplastic materials compared with autologous. Blood transfusion was required in 24.1% of cases, predominantly on the day of surgery (86.7%). Multivariable analysis identified age (OR: 0.86 per year, 95% CI: 0.80–0.93, P < 0.001) and cranioplasty size >5 cm (versus ≤5 cm) (OR: 2.10, 95% CI: 1.01–4.50, P = 0.049) as predictors of transfusion. The 30-day unplanned readmission rate was 5.3% at a mean of 14.6 ± 7.9 days from surgery, most commonly for a surgical site infection (33.0%). The 30-day unplanned reoperation rate was 2.9%, at a mean of 10.2 ± 5.6 days. Autologous cranioplasty was not associated with higher odds of readmission (OR: 0.79, 95% CI: 0.25–2.13, P = 0.66) or reoperation (OR: 5.04, 95% CI: 0.96–36.99, P = 0.07) compared with alloplastic cranioplasty. Rates of readmission and reoperation following pediatric cranioplasty are comparable to other complex craniofacial procedures. Younger age and larger cranioplasty size are associated with higher rates of blood transfusion. Our results show that autologous reconstruction is not associated with higher short-term complication rates compared with alloplastic reconstruction, although future studies are still needed to further the understanding of the optimal cranioplasty material.