Can serious postoperative complications in patients with Crohn’s disease be predicted using machine learning?

Objective: To investigate the incidence of postoperative complications in Chinese patients with gastric or colorectal cancer, and to evaluate the risk factors for postoperative complications. Methods: This was a national, multicenter, prospective, registry-based, cohort study of data obtained from the database of the Prevalence of Abdominal Complications After Gastro- enterological Surgery (PACAGE) study sponsored by the China Gastrointestinal Cancer Surgical Union. The PACAGE database prospectively collected general demographic characteristics, protocols for perioperative treatment, and variables associated with postoperative complications in patients treated for gastric or colorectal cancer in 20 medical centers from December 2018 to December 2020. The patients were grouped according to the presence or absence of postoperative complications. Postoperative complications were categorized and graded in accordance with the expert consensus on postoperative complications in gastrointestinal oncology surgery and Clavien-Dindo grading criteria. The incidence of postoperative complications of different grades are presented as bar charts. Independent risk factors for occurrence of postoperative complications were identified by multifactorial unconditional logistic regression. Results: The study cohort comprised 3926 patients with gastric or colorectal cancer, 657 (16.7%) of whom had a total of 876 postoperative complications. Serious complications (Grade III and above) occurred in 4.0% of patients (156/3926). The rate of Grade V complications was 0.2% (7/3926). The cohort included 2271 patients with gastric cancer with a postoperative complication rate of 18.1% (412/2271) and serious complication rate of 4.7% (106/2271); and 1655 with colorectal cancer, with a postoperative complication rate of 14.8% (245/1655) and serious complication rate of 3.0% (50/1655). The incidences of anastomotic leakage in patients with gastric and colorectal cancer were 3.3% (74/2271) and 3.4% (56/1655), respectively. Abdominal infection was the most frequently occurring complication, accounting for 28.7% (164/572) and 39.5% (120/304) of postoperative complications in patients with gastric and colorectal cancer, respectively. The most frequently occurring grade of postoperative complication was Grade II, accounting for 65.4% (374/572) and 56.6% (172/304) of complications in patients with gastric and colorectal cancers, respectively. Multifactorial analysis identified (1) the following independent risk factors for postoperative complications in patients in the gastric cancer group: preoperative comorbidities (OR=2.54, 95%CI: 1.51-4.28, P<0.001), neoadjuvant therapy (OR=1.42, 95%CI:1.06-1.89, P=0.020), high American Society of Anesthesiologists (ASA) scores (ASA score 2 points:OR=1.60, 95% CI: 1.23-2.07, P<0.001, ASA score ≥3 points:OR=0.43, 95% CI: 0.25-0.73, P=0.002), operative time >180 minutes (OR=1.81, 95% CI: 1.42-2.31, P<0.001), intraoperative bleeding >50 mL (OR=1.29,95%CI: 1.01-1.63, P=0.038), and distal gastrectomy compared with total gastrectomy (OR=0.65,95%CI: 0.51-0.83, P<0.001); and (2) the following independent risk factors for postoperative complications in patients in the colorectal cancer group: female (OR=0.60, 95%CI: 0.44-0.80, P<0.001), preoperative comorbidities (OR=2.73, 95%CI: 1.25-5.99, P=0.030), neoadjuvant therapy (OR=1.83, 95%CI:1.23-2.72, P=0.008), laparoscopic surgery (OR=0.47, 95%CI: 0.30-0.72, P=0.022), and abdominoperineal resection compared with low anterior resection (OR=2.74, 95%CI: 1.71-4.41, P<0.001). Conclusion: Postoperative complications associated with various types of infection were the most frequent complications in patients with gastric or colorectal cancer. Although the risk factors for postoperative complications differed between patients with gastric cancer and those with colorectal cancer, the presence of preoperative comorbidities, administration of neoadjuvant therapy, and extent of surgical resection, were the commonest factors associated with postoperative complications in patients of both categories.

BackgroundEarly complications increase in-hospital stay and mortality after intestinal obstruction surgery. It is important to identify the risk of postoperative early complications for patients with intestinal obstruction at a sufficiently early stage, which would allow preemptive individualized enhanced therapy to be conducted to improve the prognosis of patients with intestinal obstruction. A risk predictive model based on machine learning is helpful for early diagnosis and timely intervention.ObjectiveThis study aimed to construct an online risk calculator for early postoperative complications in patients after intestinal obstruction surgery based on machine learning algorithms.MethodsA total of 396 patients undergoing intestinal obstruction surgery from April 2013 to April 2021 at an independent medical center were enrolled as the training cohort. Overall, 7 machine learning methods were used to establish prediction models, with their performance appraised via the area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, specificity, and F1-score. The best model was validated through 2 independent medical centers, a publicly available perioperative dataset the Informative Surgical Patient dataset for Innovative Research Environment (INSPIRE), and a mixed cohort consisting of the above 3 datasets, involving 50, 66, 48, and 164 cases, respectively. Shapley Additive Explanations were measured to identify risk factors.ResultsThe incidence of postoperative complications in the training cohort was 47.44% (176/371), while the incidences in 4 external validation cohorts were 34% (17/50), 56.06% (37/66), 52.08% (25/48), and 48.17% (79/164), respectively. Postoperative complications were associated with 8-item features: Physiological Severity Score for the Enumeration of Mortality and Morbidity (POSSUM physiological score), the amount of colloid infusion, shock index before anesthesia induction, ASA (American Society of Anesthesiologists) classification, the percentage of neutrophils, shock index at the end of surgery, age, and total protein. The random forest model showed the best overall performance, with an AUROC of 0.788 (95% CI 0.709-0.869), accuracy of 0.756, sensitivity of 0.695, specificity of 0.810, and F1-score of 0.727 in the training cohort. The random forest model also achieved a comparable AUROC of 0.755 (95% CI 0.652-0.839) in validation cohort 1, a greater AUROC of 0.817 (95% CI 0.695-0.913) in validation cohort 2, a similar AUROC of 0.786 (95% CI 0.628-0.902) in validation cohort 3, and the comparable AUROC of 0.720 (95% CI 0.671-0.768) in validation cohort 4. We visualized the random forest model and created a web-based online risk calculator.ConclusionsWe have developed and validated a generalizable random forest model to predict postoperative early complications in patients undergoing intestinal obstruction surgery, enabling clinicians to screen high-risk patients and implement early individualized interventions. An online risk calculator for early postoperative complications was developed to make the random forest model accessible to clinicians around the world.

The impact of vedolizumab, a gut-selective monoclonal antibody, on postoperative outcomes is unclear. This study aimed to assess the impact of preoperative vedolizumab treatment on the rate of postoperative complications in patients with inflammatory bowel disease [IBD] undergoing abdominal surgery. A systematic search of multiple electronic databases from inception until May 2017 identified studies reporting rates of postoperative complications in vedolizumab-treated IBD patients compared to no biologic exposure or anti-tumor necrosis factor (anti-TNF) treated IBD patients. Outcomes of interest included postoperative infectious complications and overall postoperative complications. Pooled risk ratios and 95% confidence intervals were estimated using the random-effects model. Five studies comprising 307 vedolizumab-treated IBD patients, 490 anti-TNF-treated IBD patients and 535 IBD patients not exposed to preoperative biologic therapy were included. The risk of postoperative infectious complications (risk ratio [RR] 0.99, 95% confidence interval [CI] 0.37-2.65) and overall postoperative complications [RR 1.00, 95% CI 0.46-2.15] were not significantly different between vedolizumab-treated patients and those who received no preoperative biologic therapy. In addition, the risk of postoperative infectious complications [RR 0.99, 95% CI 0.34-2.90] and overall postoperative complications [RR 0.92, 95% CI 0.44-1.92] were not significantly different between vedolizumab-treated vs anti-TNF-treated patients. Preoperative vedolizumab treatment in IBD patients does not appear to be associated with an increased risk of postoperative infectious or overall postoperative complications compared to either preoperative anti-TNF therapy or no biologic therapy. Future prospective studies which include perioperative drug level monitoring are needed to confirm these findings.

BackgroundSeveral studies have assessed various clinical variables to identify risk factors for postoperative complications in patients with acute appendicitis. However, few studies have focused on the relationships between systemic inflammatory variables and postoperative complications in patients with acute appendicitis. We investigated the relationships between postoperative complications and systemic inflammatory variables, and assessed the clinical utility of these variables as predictors of postoperative complications in patients with acute appendicitis.MethodsWe retrospectively reviewed 181 patients who underwent immediate appendectomy for acute appendicitis. All postoperative complications were classified as infectious or noninfectious, and we evaluated the relationships between postoperative complications and clinical factors including the preoperative neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio.ResultsIn total, 28 patients (15.5%) had postoperative Clavien-Dindo grade II–IV complications; 17 patients (9.4%) and 11 patients (6.1%) were categorized as the infectious and noninfectious complication groups, respectively. The cutoff value of the preoperative neutrophil-to-lymphocyte ratio for all complications was 11.3, and multivariate analysis revealed that the preoperative neutrophil-to-lymphocyte ratio was an independent predictor of any postoperative complication (odds ratio: 4.223, 95% confidence interval: 1.335–13.352; P = 0.014). The cutoff value of the preoperative neutrophil-to-lymphocyte ratio for infectious complications was 11.4, and multivariate analysis revealed that the preoperative neutrophil-to-lymphocyte ratio was an independent predictor of infectious complications (odds ratio: 4.235, 95% confidence interval: 1.137–15.776; P = 0.031).ConclusionsIn patients with acute appendicitis, the preoperative neutrophil-to-lymphocyte ratio may be a useful predictor of all postoperative complications, especially infectious complications.

Malnutrition, immune deficiency, and skeletal muscle loss are associated with a risk of postoperative complications in patients with various types of cancer. This study evaluated whether malnutrition, immunological deficiencies, and skeletal muscle loss during neoadjuvant chemotherapy (NAC) predict postoperative complications in patients with esophageal cancer. We retrospectively reviewed 123 patients with esophageal squamous cell carcinoma treated with NAC and esophagectomy at our hospital between 2014 and 2019. Patients were divided into two groups based on the presence or absence of postoperative infectious complications, such as pneumonia, anastomotic leakage, surgical site infections, pyothorax, acalculous cholecystitis, and peripheral phlebitis. Neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, and Onodera prognostic nutritional index were used as indicators of systemic inflammation and nutritional status. Skeletal muscle mass was evaluated using the skeletal muscle index (SMI), calculated by evaluating the total cross-sectional area of muscle tissue at the third lumbar level in computed tomography imaging. Univariable and multivariable logistic regression analyses were used to identify predictors of postoperative infectious complications. Postoperative infectious complications occurred in 41 patients (33.3%). A reduction in SMI was observed in 105 patients (87.8%) during NAC. Univariable and multivariable analyses indicated that the reduction in SMI during NAC was an independent predictor of postoperative complications (odds ratio=0.89; 95% confidence interval=0.79-0.99; p=0.048). Skeletal muscle loss during NAC is a useful predictor of postoperative complications in patients with esophageal cancer undergoing esophagectomy.

BackgroundPreoperative nutritional status in patients with gastric cancer after surgery has attracted considerable interest. The nutritional risk index (NRI) has been widely used as a convenient and effective nutritional assessment index, but the relationship between preoperative NRI and postoperative complications in patients with gastric cancer has not been adequately studied. Our study aimed to investigate the effects of preoperative NRI on 30-day postoperative complications in patients with gastric cancer.MethodsThis retrospective analysis investigated 578 patients with gastric cancer. Preoperative NRI calculations were based on serum albumin levels and body weight, and receiver operating characteristic curves were used in analyzing NRI values and establishing optimal cutoff points. Patients were categorized into two groups according to cutoff value: low NRI group (NRI<96.7) and high NRI group (NRI≥96.7). The hazard ratio (HR) for postoperative complications was calculated through Cox regression analysis and adjusted for potential confounders, and the effects of NRI on postoperative complications in patients with gastric cancer were examined. In addition, we conducted subgroup analyses to examine whether there was an interaction between the effect of NRI on the cumulative incidence of postoperative complications and other confounding factors.ResultsOf the 578 patients with gastric cancer who underwent radical surgery, 120 (20.8%) experienced postoperative complications. The optimal NRI threshold of 96.7 was identified using ROC curve analysis. Cox regression analysis demonstrated that preoperative NRI was independently associated with 30-day postoperative complications after adjusting for confounding factors (HR=0.93; 95%CI: 0.90–0.96; P<0.001). Patients in the low NRI group had significantly higher rates of postoperative complications than those in the high NRI group(HR=2.89, 95%CI:1.71–4.88; P<0.001). The cumulative incidence analysis revealed a higher risk of postoperative complications over time in the low NRI group compared with the high NRI group (P<0.001). These associations remained robust in subgroup analyses.ConclusionsNRI is an independent predictor of 30-day postoperative complications in gastric cancer patients and is a convenient and useful nutritional screening tool for identifying patients with gastric cancer at high risk of postoperative complications.

OA01.05 The Impact of Lung Age on Postoperative Complications in Patients with Lung Cancer Combined with Pulmonary Fibrosis and Emphysema

Objective: to study the diagnostic and prognostic values of the Mannheim peritoneal index (MPI) in the development of postoperative local and systemic complications in patients with peritonitis. Materials and methods. The case histories of 92 patients with generalized peritonitis of varying etiology (other than pancreatogenic one) were analyzed. The patients were retrospectively divided into 3 groups according to the outcomes and occurrence of postoperative local complications. The postoperative complications were classified by the procedure developed by A. L. Kostyuchenko et al. as local and systemic ones. When the patients had two signs or more of the systemic inflammatory response syndrome, they were stated to have systemic complications and to be diagnosed as having abdominal sepsis with the pattern of organ dysfunctions being described in accordance with the sepsis classification proposed by R. S. Bone et al. (1992). The number of organ dysfunctions was daily counted in each patient over time in the postoperative period. On the first postoperative day, MPI was calculated in scores for each patient; the mean MPI was estimated for all patient groups. The predictable mortality was calculated using the MPI plot. Results. All the patients with generalized peritonitis in the development of local postoperative complications were observed to have sepsis in the postoperative period, without developing local complication in 84.6% of the patients. A direct correlation was found between the MPI and the quantity of organ dysfunctions (r=0.6; p=0.001). In patients with local postoperative complications being developed, the MPI values were higher (p&lt;0.05) than in those without them. The mortality rates that have been predicted by means of MPI (16.3%) and actual (15.2%) are actually in agreement. Conclusion. There is evidence for the diagnostic and prognostic values of MPI in the development of local and systemic postoperative complications in patients with peritonitis. MPI calculation from postoperative day 1 along with later dynamic estimation of the number of organ dysfunctions makes it to objectively assess not only prognosis, but also the pattern of postoperative peritonitis. The major advantage of MPI is the simplicity of calculation, which permits its use in clinics. Key words: peritonitis, prognosis, complications, sepsis, Mannheim peritoneal index.

The impact of preoperative use of infliximab on postoperative complications in patients with Crohn's disease

Role of Platelet to Lymphocyte Ratio and Red Cell Distribution Width in Predicting Postoperative Complications in Patients with Acute Mesenteric Ischemia

Objective To investigate the association between body mass index and postoperative complications in patients with colorectal cancer. Methods 120 patients with colorectal cancer were prospectively studied.According to the body mass index, all patients were assigned to study group(BMI<18.5kg/m2) or control group(BMI≥18.5kg/m2), 60 cases in each group.All patients received elective colorectal cancer resection.The operation situation, postoperative complications, postoperative recovery and postoperative stress reaction of the two groups were observed. Results There were no significant differences in the operation time, the number of lymph node dissection and the amount of bleeding between the two groups during the operation(all P>0.05). Compared with the control group, the total infection rate of the study group was significantly higher(11.67% vs.1.67%, χ2=4.324, P=0.038); the incidence of postoperative complications was significantly increased(25.00% vs.5.00%, χ2=7.908, P=0.005). There were no significant differences in other complications such as wound infection, anastomotic leakage, anastomotic obstruction, adverse cardiovascular events and stress ulcer(all P>0.05). Compared with the control group, the postoperative anal exhaust time of the study group was significantly prolonged[(51.48±9.47)h vs.(43.73±8.63)h, t=5.382, P=0.000]; and the length of hospital duration was significantly prolonged[(20.35±3.21) d vs.(16.28±3.48)d, t=6.462, P=0.000]. There were no significant differences in recurrence rate, mortality rate and progression free survival between the two groups after 1 year (all P>0.05). There were no significant differences in preoperative IL-6 and CRP levels between the two groups(all P>0.05). Compared with the control group, the IL-6 level of the study group at 7-day after operation was significantly higher[(165.56±45.39) ng/L vs.(122.58±29.58)ng/L, t=6.459, P=0.000]; CRP level was significantly higher[(23.49±8.48) mg/L vs.(15.37±6.53)mg/L, t=4.285, P=0.002]. Conclusion Body mass index less than 18.5 kg/m2 is a risk factor for postoperative infection and complications in patients with colorectal cancer, and is associated with increased stress response and prolonged hospital stay. Key words: Body mass index; Colorectal cancer; Emaciation; Postoperative complications

PurposeTo assess the incidence and factors predictive of early postoperative complications in Korean patients who undergo surgery for Crohn's disease (CD).MethodsWe retrospectively assessed 350 patients (246 males, 104 females; mean age, 30 ± 9 years) who underwent surgery for primary or recurrent CD at Asan Medical Center between January 1991 and May 2010. The incidence and predictive factors of early postoperative complications were analyzed by both univariate and multivariate analyses.ResultsOf the 350 patients, 81 patients (23.1%) developed postoperative complications, the most common being septic complications (54 patients), including 19 cases of wound infection. Thirty patients (8.6%) required re-operations, and only one patient died. Multivariate analysis showed that four factors were independently associated with a high risk of early postoperative complications; preoperative moderate to severe anemia (hematocrit concentration <30%; odds ratio [OR], 3.1; 95% confidence interval [CI], 1.6 to 5.9), hypoalbuminemia (serum albumin level <3.0 g/dL; OR, 2.6; 95% CI, 1.4 to 4.7), emergency surgery (OR, 4.0; 95% CI, 1.5 to 10.6), and covering stoma (OR, 2.6; 95% CI, 1.3 to 5.4). Correction of preoperative moderate to severe anemia and hypoalbuminemia decreased the incidence of postoperative complications. Mean hospital stay was significantly longer in patients with than without postoperative complications (31.3 ± 27.2 days vs. 10.3 ± 3.8 days, P < 0.001).ConclusionPreoperative anemia, low albumin level, emergency surgery, and covering stoma significantly increased the risk of early postoperative complications in patients with CD. Correcting preoperatively deficient nutritional factors may reduce postoperative morbidities.

EVALUATION OF POSTEXODONTIC BLEEDING IN PATIENTS RECEIVING ANTICOAGULANT THERAPY WITH LOW-MOLECULAR-WEIGHT HEPARIN: PILOT STUDY

The preoperative use of anti-tumor necrosis factor-alpha (anti-TNF) in inflammatory bowel disease (IBD) patients undergoing surgery has been controversial due to concern for increased risks of postoperative complications. We aimed to determine the effect of preoperative anti-TNF therapy on postoperative complications in IBD patients undergoing abdominal surgery. A literature search of Google Scholar, PubMed, The Cochrane Library, EMBASE, and CINAHL was performed through October 2019. Studies reporting postoperative complication rates of Crohn's disease (CD), ulcerative colitis (UC), and IBD-unspecified patients with preoperative anti-TNF treatment undergoing abdominal surgery compared to controls without preoperative anti-TNF treatment were included. The main outcomes measured were overall, infectious, and noninfectious postoperative complications. Forty-one studies totaling 20 274 patients were included. There was a significant increase in overall complications in all patients treated with anti-TNF vs. controls [odds ratio (OR) = 1.13, 95% confidence interval (CI), 1.01-1.25, P = 0.03, I2 = 6%] with an absolute risk increase (ARI) of 5.5% and a number needed to harm (NNH) of 18. There was also a significant increase in infectious complications in CD patients (OR = 1.44; 95% CI 1.02-2.03, P = 0.04, I2 = 49%, ARI = 5.5%, NNH = 20) only. Contrastingly, there was a significant increase in noninfectious complications in all patients (OR = 1.44, 95% CI 1.13-1.85, P = 0.003, I2 = 8%, ARI = 6.4%, NNH = 16) and UC patients (OR = 1.57, 95% CI 1.15-2.14, P = 0.005, I2 = 25%, ARI = 8.5%, NNH = 12) only. Preoperative use of anti-TNF agents in IBD patients undergoing abdominal surgery is associated with increases in overall postoperative complications in all patients, infectious postoperative complications in CD patients, and noninfectious postoperative complications in UC patients.

IntroductionThe postoperative mortality rate following a femoral neck fracture remains high. The Surgical Apgar Score (SAS), based on intraoperative blood loss, the lowest mean arterial pressure, and the lowest heart rate, was created to predict 30-day postoperative major complications. Here, we evaluated the relationship between the SAS and postoperative complications in patients who underwent femoral neck surgeries.MethodsWe retrospectively collected data from patients with femoral neck surgeries performed in 2012–2017 at Kumamoto Central Hospital. The variables required for the SAS and the factors presumably associated with postoperative complications including the patients’ characteristics were collected from the medical charts. Intergroup differences were assessed with the χ2 test with Yates’ correlation for continuity in category variables. The Mann-Whitney U test was used to test for differences in continuous variables. We assessed the power of the SAS value to distinguish patients who died ≤ 90 days post-surgery from those who did not, by calculating the area under the receiver operating characteristic curve (AUC).ResultsWe retrospectively examined the cases of 506 patients (94 men, 412 women) aged 87 ± 6 (range 70–102) years old. The 90-day mortality rate was 3.4% (n = 17 non-survivors). There were significant differences between the non-survivors and survivors in body mass index (BMI), the presence of moderate to severe valvular heart disease, albumin concentration, the American Society of Anesthesiologists (ASA) classification, and the SAS. The 90-day mortality rate in the SAS ≤ 6 group (n = 97) was 10.3%, which was significantly higher than that in the SAS ≥ 7 group (n = 409), 1.7%. The AUC value to predict the 90-day mortality was 0.70 for ASA ≥ 3 only, 0.71 for SAS ≤ 6 only, 0.81 for SAS ≤ 6 combined with ASA ≥ 3, and 0.85 for SAS ≤ 6 combined with albumin concentration < 3.5 g/dl, BMI ≤ 20, and the presence of moderate to severe valvular heart disease.ConclusionsOur results suggest that the SAS is useful to evaluate postoperative complications in patients who have undergone a femoral neck surgery. The ability to predict postoperative complications will be improved when the SAS is used in combination with the patient’s preoperative physical status.