Risk Prediction Model for Severe Postoperative Complication in Bariatric Surgery

Development of predictive model for predicting postoperative BMI and optimize bariatric surgery: a single center pilot study

BACKGROUND

The pilot study addresses the challenge of predicting postoperative outcomes, particularly body mass index (BMI) trajectories, following bariatric surgery. The complexity of this task makes preoperative personalized obesity treatment challenging.

OBJECTIVES

To develop and validate sophisticated machine learning (ML) algorithms capable of accurately forecasting BMI reductions up to 5 years following bariatric surgery aiming to enhance planning and postoperative care. The secondary goal involves the creation of an accessible web-based calculator for healthcare professionals. This is the first article that compares these methods in BMI prediction.

SETTING

The study was carried out from January 2012 to December 2021 at GZOAdipositas Surgery Center, Switzerland. Preoperatively, data for 1004 patients were available. Six months postoperatively, data for 1098 patients were available. For the time points 12 months, 18 months, 2 years, 3 years, 4 years, and 5 years the following number of follow-ups were available: 971, 898, 829, 693, 589, and 453.

METHODS

We conducted a comprehensive retrospective review of adult patients who underwent bariatric surgery (Roux-en-Y gastric bypass or sleeve gastrectomy), focusing on individuals with preoperative and postoperative data. Patients with certain preoperative conditions and those lacking complete data sets were excluded. Additional exclusion criteria were patients with incomplete data or follow-up, pregnancy during the follow-up period, or preoperative BMI ≤30 kg/m2.

RESULTS

This study analyzed 1104 patients, with 883 used for model training and 221 for final evaluation, the study achieved reliable predictive capabilities, as measured by root mean square error (RMSE). The RMSE values for three tasks were 2.17 (predicting next BMI value), 1.71 (predicting BMI at any future time point), and 3.49 (predicting the 5-year postoperative BMI curve). These results were showcased through a web application, enhancing clinical accessibility and decision-making.

CONCLUSION

This study highlights the potential of ML to significantly improve bariatric surgical outcomes and overall healthcare efficiency through precise BMI predictions and personalized intervention strategies.

Concomitant Panniculectomy in Abdominal Wall Reconstruction: A Narrative Review Focusing on Obese Patients

The global prevalence of obesity continues to rise, contributing to an increased frequency of abdominal wall reconstruction procedures, particularly ventral hernia repairs, in individuals with elevated body mass indexes. Undertaking these operations in obese patients poses inherent challenges. This review focuses on the current literature in this area, with special attention to the impact of concomitant panniculectomy. Obese individuals undergoing abdominal wall reconstruction face elevated rates of wound healing complications and hernia recurrence. The inclusion of concurrent panniculectomy heightens the risk of surgical site occurrences but does not significantly influence hernia recurrence rates. While this combined approach can be executed in obese patients, caution is warranted, due to the higher risk of complications. Physicians should carefully balance and communicate the potential risks, especially regarding the increased likelihood of wound healing complications. Acknowledging these factors is crucial in shared decision making and ensuring optimal patient outcomes in the context of abdominal wall reconstruction and related procedures in the obese population.

Risk factors for bariatric surgery in patients over 65 years of age-a multicenter retrospective cohort study

PURPOSE

Societies are aging, life expectancy is increasing, and as a result, the percentage of elderly people in the population is constantly increasing. When qualifying patients over 65 years of age for bariatric surgery, the benefits and risks should be carefully assessed. Weighing risk factors against each other to improve the quality of life and better control of obesity-related diseases. The study aimed to determine risk factors for bariatric surgery among patients over 65 years of age.

METHODS

A multicenter, retrospective analysis of patients undergoing laparoscopic bariatric procedures from 2008 to 2022. The patients were divided into two groups: complicated (C) and uncomplicated (UC). Uni- and multivariate logistic regression analysis was performed to obtain significant, independent risk factors.

RESULTS

There were 20 (7.0%) patients in C group and 264 (93.0%) patients in UC group. The most common complication was intraperitoneal bleeding (8, 2.8). There was no postoperative mortality. The mean follow-up was 47.5 months. In a multivariate logistic regression analysis, length of stay and %EWL significantly corresponded to general complications (OR 1.173, OR 1.020). A higher weight loss before surgery lowered the risk for hemorrhagic events after surgery (OR 0.889). A longer length of stay corresponded to leak after surgery (OR 1.175).

CONCLUSIONS

Bariatric and metabolic surgery appears to be a safe method of obesity treatment in patients over 65 years of age. The most common complication was intraperitoneal bleeding. A prolonged hospital stay may increase the risk of leakage, while a higher weight loss before the surgery may lower the risk of bleeding.

The Effects of Preoperative Glycaemic Control (HbA1c) on Bariatric and Metabolic Surgery Outcomes: Data from a Tertiary-Referral Bariatric Centre in the UK

BACKGROUND

Current recommendations advocate the achievement of an optimal glucose control (HbA1c < 69 mmol/mol) prior to elective surgery to reduce risks of peri- and post-operative complications, but the relevance for this glycaemic threshold prior to Bariatric Metabolic Surgery (BMS) following a specialist weight management programme remains unclear.

METHODS

We undertook a retrospective cohort study of patients with type 2 diabetes mellitus (T2DM) who underwent BMS over a 6-year period (2016-2022) at a regional tertiary referral following completion of a specialist multidisciplinary weight management. Post-operative outcomes of interest included 30-day mortality, readmission rates, need for Intensive Care Unit (ICU) care and hospital length of stay (LOS) and were assessed according to HbA1c cut-off values of < 69 (N = 202) and > 69 mmol/mol (N = 67) as well as a continuous variable.

RESULTS

A total of 269 patients with T2D were included in this study. Patients underwent primary Roux en-Y gastric bypass (RYGB, n = 136), Sleeve Gastrectomy (SG, n = 124), insertion of gastric band (n = 4) or one-anastomosis gastric bypass (OAGB, n = 4). No significant differences in the rates of complications were observed between the two groups of pre-operative HbA1c cut-off values. No HbA1c threshold was observed for glycaemic control that would affect the peri- and post-operative complications following BMS.

CONCLUSIONS

We observed no associations between pre-operative HbA1C values and the risk of peri- and post-operative complications. In the context of a specialist multidisciplinary weight management programme, optimising pre-operative HbA1C to a recommended target value prior to BMS may not translate into reduced risks of peri- and post-operative complications.

Association of Malnutrition Diagnosed Using Global Leadership Initiative on Malnutrition Criteria with Severe Postoperative Complications After Gastrectomy in Patients with Gastric Cancer

Background: The purpose of this study was to investigate the relationship between malnutrition assessed by the Global Leadership Initiative on Malnutrition (GLIM) criteria and the occurrence of severe postoperative complications (SPCs) after gastrectomy in patients with gastric cancer. Methods: A total of 220 patients with gastric cancer were included in this retrospective study. According to the GLIM criteria, the first step was to use the Nutrition Risk Screening Score 2002 to conduct nutritional risk screening for patients and the second step was to diagnose and grade the severity of malnutrition in patients at risk of malnutrition. According to the Clavien-Dindo classification system, SPCs were defined as C-D Grade IIIa or higher. Results: Overall, 66 (30.0%) patients were diagnosed with malnutrition, including 32 (14.5%) with moderate malnutrition and 34 (15.5%) with severe malnutrition. The incidence of SPCs was 14.5%, and the most frequent postoperative event was anastomotic leakage. In the multivariate regression analysis, malnutrition was considered an independent risk factor for SPCs (P < .001). After adjusting for various factors, the grading association remained statistically significant. Compared with patients with normal nutrition, patients with moderate and severe malnutrition have a nearly 15-fold (OR = 15.682, 95% CI: 4.481-54.877, P < .001) and 20-fold (OR = 20.554, 95% CI: 5.771-73.202, P < .001) increased risk of developing SPCs, respectively. Conclusions: Malnutrition assessed by GLIM was an independent risk factor for SPCs in gastric cancer patients. Therefore, early identification of malnourished patients is crucial for timely implementation of nutritional treatment and reducing the occurrence of postoperative complications.

Characteristics of Kidney Transplant Recipients with Prolonged Pre-Transplant Dialysis Duration as Identified by Machine Learning Consensus Clustering: Pathway to Personalized Care

Longer pre-transplant dialysis duration is known to be associated with worse post-transplant outcomes. Our study aimed to cluster kidney transplant recipients with prolonged dialysis duration before transplant using an unsupervised machine learning approach to better assess heterogeneity within this cohort. We performed consensus cluster analysis based on recipient-, donor-, and transplant-related characteristics in 5092 kidney transplant recipients who had been on dialysis ≥ 10 years prior to transplant in the OPTN/UNOS database from 2010 to 2019. We characterized each assigned cluster and compared the posttransplant outcomes. Overall, the majority of patients with ≥10 years of dialysis duration were black (52%) or Hispanic (25%), with only a small number (17.6%) being moderately sensitized. Within this cohort, three clinically distinct clusters were identified. Cluster 1 patients were younger, non-diabetic and non-sensitized, had a lower body mass index (BMI) and received a kidney transplant from younger donors. Cluster 2 recipients were older, unsensitized and had a higher BMI; they received kidney transplant from older donors. Cluster 3 recipients were more likely to be female with a higher PRA. Compared to cluster 1, cluster 2 had lower 5-year death-censored graft (HR 1.40; 95% CI 1.16-1.71) and patient survival (HR 2.98; 95% CI 2.43-3.68). Clusters 1 and 3 had comparable death-censored graft and patient survival. Unsupervised machine learning was used to characterize kidney transplant recipients with prolonged pre-transplant dialysis into three clinically distinct clusters with variable but good post-transplant outcomes. Despite a dialysis duration ≥ 10 years, excellent outcomes were observed in most recipients, including those with moderate sensitization. A disproportionate number of minority recipients were observed within this cohort, suggesting multifactorial delays in accessing kidney transplantation.

Prediction of quality-adjusted life years (QALYs) after bariatric surgery using regularized linear regression models: results from a Swedish nationwide quality register

PURPOSE

To investigate whether the quality-adjusted life years (QALYs) of the patients who underwent bariatric surgery could be predicted using their baseline information.

MATERIALS AND METHODS

All patients who received bariatric surgery in Sweden between January 1, 2011 and March 31, 2019 were obtained from the Scandinavian Obesity Surgery Registry (SOReg). Baseline information included patients' sociodemographic characteristics, details regarding the procedure, and postsurgical conditions. QALYs were assessed by the SF-6D at follow-up years 1 and 2. The general and regularized linear regression models were used to predict postoperative QALYs.

RESULTS

All regression models demonstrated satisfactory and comparable performance in predicting QALYs at follow-up year 1, with R2 and relative root mean squared error (RRMSE) values of about 0.57 and 9.6%, respectively. The performance of the general linear regression model increased with the number of variables; however, the improvement was ignorable when the number of variables was more than 30 and 50 for follow-up years 1 and 2, respectively. Although minor L1 and L2 regularization provided better prediction ability, the improvement was negligible when the number of variables was more than 20. All the models showed poorer performance for predicting QALYs at follow-up year 2.

CONCLUSIONS

Patient characteristics before bariatric surgery including health related quality of life, age, sex, BMI, postoperative complications within six weeks, and smoking status, may be adequate in predicting their postoperative QALYs after one year. Understanding these factors can help identify individuals who require more personalized and intensive support before, during, and after surgery.

Network prediction of surgical complication clusters: a prospective multicenter cohort study

Complicated relationships exist in both occurrence and progression of surgical complications, which are difficult to account for using a separate quantitative method such as prediction or grading. Data of 51,030 surgical inpatients were collected from four academic/teaching hospitals in a prospective cohort study in China. The relationship between preoperative factors, 22 common complications, and death was analyzed. With input from 54 senior clinicians and following a Bayesian network approach, a complication grading, cluster-visualization, and prediction (GCP) system was designed to model pathways between grades of complication and preoperative risk factor clusters. In the GCP system, there were 11 nodes representing six grades of complication and five preoperative risk factor clusters, and 32 arcs representing a direct association. Several critical targets were pinpointed on the pathway. Malnourished status was a fundamental cause widely associated (7/32 arcs) with other risk factor clusters and complications. American Society of Anesthesiologists (ASA) score ⩾3 was directly dependent on all other risk factor clusters and influenced all severe complications. Grade III complications (mainly pneumonia) were directly dependent on 4/5 risk factor clusters and affected all other grades of complication. Irrespective of grade, complication occurrence was more likely to increase the risk of other grades of complication than risk factor clusters.

Future of bariatric surgery beyond simple weight loss: Metabolic surgery

Anatomical modifications implemented during bariatric surgery not only result in weight loss, but also lead to metabolic corrections that translate into better glycemia stability and improvement in cardiovascular and liver disorders. The logical extension of surgical indications beyond mere reduction of the body mass index (BMI) (i.e. patients with<35kg/m2) is a hot topic today in France and worldwide. Metabolic surgeries make use of multiple modalities (endoscopic, mini-invasive, invasive) that should be carried out by trained physicians and within the same type of multidisciplinary formation as that for bariatric surgery. The aim of this update is to describe the physiological mechanisms that result in the benefits of bariatric surgery, the various procedures currently available and the perspectives for this new field in visceral and digestive surgery.

High preoperative HbA1c does not affect early or late complication rates after bariatric surgery

BACKGROUND

Preoperative HbA1c has been associated with an increased incidence of postoperative morbidity and mortality after abdominal and cardiovascular surgery. The literature on bariatric surgery is inconclusive and guidelines recommend postponement of surgery when HbA1c is above an arbitrary threshold (≥ 8.5%). In this study, we sought to understand the impact of preoperative HbA1c on early and late postoperative complications.

METHODS

We performed a retrospective analysis of prospectively collected data on obese patients with diabetes who underwent laparoscopic bariatric surgery. Patients were categorized into three groups according to their preoperative HbA1c level: < 6.5% (group 1), 6.5-8.4% (group 2) and ≥ 8.5% (group 3). Primary outcomes were early and late postoperative complications (< and > 30 days, respectively) that were differentiated based on severity (major/minor). Secondary outcomes were length of stay (LOS), duration of surgery, and rate of readmission.

RESULTS

In total, 6798 patients underwent laparoscopic bariatric surgery from 2006 to 2016, of which 1021 (15%) patients had Type 2 Diabetes (T2D). Complete data with a median follow-up of 45 months (3-120) were available for 914 patients with HbA1c < 6.5% (n = 227, 24.9%), 6.5-8.4% (n = 532, 58.5%) and ≥ 8.5% (n = 152, 16.6%). Early major surgical complication rate was similar across the groups ranging from 2.6 to 3.3%. No associations between high preoperative HbA1c and late complications-medical as well as surgical-was observed. Groups 2 and 3 had statistically significant more pronounced inflammatory status. LOS (1.8-1.9 days), readmission rates (1.7-2.0%) and surgical time was similar across the three groups.

CONCLUSION

Elevated HbA1c is not associated with more early or late postoperative complications, longer LOS, longer surgical time or higher rates of readmission.

Association of body composition with postoperative complications after laparoscopic sleeve gastrectomy and Roux-en-Y gastric bypass

OBJECTIVE

To evaluate predictive values of body composition parameters measured from preoperative CT/MRIs for postoperative complications after laparoscopic sleeve gastrectomy (LSG) and Roux-en-Y gastric bypass (LRYGB) in patients with obesity.

METHODS

In this retrospective case-control study, patients performing abdominal CT/MRIs within one month before and developing 30-day complications after bariatric procedures were matched for age, sex, and type of surgery with patients without complications (1/3 ratio, respectively). Complications were determined by documentation in the medical record. Two readers blindly segmented the total abdominal muscle area (TAMA) and visceral fat area (VFA) using predetermined thresholds for the Hounsfield unit (HU) on unenhanced CT and the signal intensity (SI) on T1-weighted MRI at the L3 vertebral level. Visceral obesity (VO) was defined as VFA > 136 cm² in males and > 95 cm² in females. These measures, along with perioperative variables, were compared. Multivariate logistic regression analyses were performed.

RESULTS

Of 145 included patients, 36 had postoperative complications. No significant differences between LSG and LRYGB were present regarding complications and VO. Hypertension (p = 0.022), impaired lung function (p = 0.018), American Society of Anesthesiologists (ASA) grade (p = 0.046), VO (p = 0.021), and VFA/TAMA ratio (p < 0.0001) were associated with postoperative complications in the univariate logistic analysis; the VFA/TAMA ratio was the only independent predictor in multivariate analyses (OR 2.01, 95% CI 1.37-2.93, p < 0.001).

CONCLUSION

The VFA/TAMA ratio provides important perioperative information in predicting patients who are likely to develop postoperative complications undergoing bariatric surgery.

Selection of Robotic Bariatric Surgery Candidates: a Nationwide Analysis

INTRODUCTION

This study aims to identify risk factors associated with 30-day major complications, readmission, and delayed discharge for patients undergoing robotic bariatric surgery.

METHODS

From the metabolic and bariatric surgery and accreditation quality improvement program (2015-2018) datasets, adult patients who underwent elective robotic bariatric operations were included. Predictors for 30-day major complications, readmission, and delayed discharge (hospital stay ≥ 3 days) were identified using univariable and multivariable analyses.

RESULTS

Major complications in patients undergoing robotic bariatric surgery were associated with both pre-operative and intraoperative factors including pre-existing cardiac morbidity (OR = 1.41, CI = [1.09-1.82]), gastroesophageal reflux disease [GERD] (OR = 1.23, CI = [1.11-1.38]), pulmonary embolism (OR = 1.51, CI = [1.02-2.22]), prior bariatric surgery (OR = 1.66, CI = [1.43-1.94]), increased operating time (OR = 1.003, CI = [1.002-1.004]), gastric bypass or duodenal switch (OR = 1.58, CI = [1.40-1.79]), and intraoperative drain placement (OR = 1.28, CI = [1.11-1.47]). With regard to 30-day readmission, non-white race (OR = 1.25, CI = [1.14-1.39]), preoperative hyperlipidemia (OR = 1.16, CI = [1.14-1.38]), DVT (OR = 1.48, CI = [1.10-1.99]), therapeutic anticoagulation (OR = 1.48, CI = [1.16-1.89]), limited ambulation (OR = 1.33, CI = [1.01-1.74]), and dialysis (OR = 2.14, CI = [1.13-4.09]) were significantly associated factors. Age ≥ 65 (OR = 1.18, CI = [1.04-1.34]), female gender (OR = 1.21, CI = [1.10-1.32]), hypertension (OR = 1.08, CI = [1.01-1.15]), renal insufficiency (OR = 2.32, CI = [1.69-3.17]), COPD (OR = 1.49, CI = [1.23-1.82]), sleep apnea (OR = 1.10, CI = [1.03-1.18]), oxygen dependence (OR = 1.47, CI = [1.10-2.0]), steroid use (OR = 1.26, CI = [1.02-1.55]), IVC filter (OR = 1.52, CI = [1.15-2.0]), and BMI ≥ 40 (OR = 1.12, CI = [1.04-1.21]) were risk factors associated with delayed discharge.

CONCLUSION

When selecting patients for bariatric surgery, surgeons early in their learning curve for utilizing robotics should avoid individuals with pre-existing cardiac or renal morbidities, venous thromboembolism, and limited functional status. Patients who have had previous bariatric surgery or require technically demanding operations are at higher risk for complications. An evidence-based approach in selecting bariatric candidates may potentially minimize the overall costs associated with adopting the technology.

The development of machine learning in bariatric surgery

Background

Machine learning (ML), is an approach to data analysis that makes the process of analytical model building automatic. The significance of ML stems from its potential to evaluate big data and achieve quicker and more accurate outcomes. ML has recently witnessed increased adoption in the medical domain. Bariatric surgery, otherwise referred to as weight loss surgery, reflects the series of procedures performed on people demonstrating obesity. This systematic scoping review aims to explore the development of ML in bariatric surgery.

Methods

The study used the Preferred Reporting Items for Systematic and Meta-analyses for Scoping Review (PRISMA-ScR). A comprehensive literature search was performed of several databases including PubMed, Cochrane, and IEEE, and search engines namely Google Scholar. Eligible studies included journals published from 2016 to the current date. The PRESS checklist was used to evaluate the consistency demonstrated during the process.

Results

A total of seventeen articles qualified for inclusion in the study. Out of the included studies, sixteen concentrated on the role of ML algorithms in prediction, while one addressed ML's diagnostic capacity. Most articles (n = 15) were journal publications, whereas the rest (n = 2) were papers from conference proceedings. Most included reports were from the United States (n = 6). Most studies addressed neural networks, with convolutional neural networks as the most prevalent. Also, the data type used in most articles (n = 13) was derived from hospital databases, with very few articles (n = 4) collecting original data via observation.

Conclusions

This study indicates that ML has numerous benefits in bariatric surgery, however its current application is limited. The evidence suggests that bariatric surgeons can benefit from ML algorithms since they will facilitate the prediction and evaluation of patient outcomes. Also, ML approaches to enhance work processes by making data categorization and analysis easier. However, further large multicenter studies are required to validate results internally and externally as well as explore and address limitations of ML application in bariatric surgery.

No Weekday Effect in Bariatric Surgery-a Retrospective Cohort Study

PURPOSE

Major abdominal surgery carried out in the later part of the week has been associated with increased complication rates. The aim of this study was to explore whether the weekday of surgery affects the 30-day complication risks after primary Roux-en-Y gastric bypass (RYGB) and sleeve gastrectomy (SG).

MATERIAL AND METHODS

Prospectively collected data, extracted from the Scandinavian Obesity Surgery Registry (SOReg), of all patients who underwent primary laparoscopic RYGB or SG between 2010 and 2017 were included in this retrospective cohort study. Multivariate logistic regression adjusted for differences in case-mix and operating center by weekday of surgery.

RESULTS

In total, 49,349 patients were included in this study. The overall 30-day complication rate was 7.2% (n = 3574), whereof 2.9% (n = 1428) had a severe complication, i.e., requiring intervention in general anesthesia or more. The 30-day mortality rate and readmission rate were 0.02% (n = 12) and 7.6% (n = 3726), respectively. The highest overall complication rate was seen in patients operated on Wednesdays and Thursdays (7.7%), while severe complications were most common on Wednesdays (3.3%). However, a large variation in severe complications was seen between centers, from 0.4 to 8.0%. After adjustment for case-mix and operating center, there was no significant increased risk of overall complications, severe complications, or readmission rates by weekday of surgery, except for a lower readmission rate in patients operated on Tuesdays.

CONCLUSION

The result of the present study supports the notion that bariatric surgery can be performed safely on all weekdays.

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

BACKGROUND

The predictive values of the platelet to lymphocyte ratio (PLR) and red cell distribution width (RDW) have been demonstrated in different types of abdominal surgery. The aim of this study was to investigate the interest of the preoperative PLR and RDW as predictors of 30-day postoperative complications in patients with acute mesenteric ischemia (AMI).

METHODS

Clinical data of 105 AMI patients were retrospectively reviewed. Postoperative complications were evaluated by the Clavien-Dindo classification. The cutoff values for neutrophil to lymphocyte ratio (NLR), PLR, and RDW were determined by receiver operating characteristic curves. Univariate and multivariate analyses evaluating the risk factors for postoperative complications were performed.

RESULTS

In the univariate analyses, advanced age, female, anemia, high white blood cell (WBC), high PLR, high NLR, high RDW, Charlson comorbidity index (CCI) score ≥2, and bowel resection were associated with the postoperative complications. A multivariable analysis revealed that advanced age, high PLR, high RDW, and bowel resection were independent predictors of postoperative complications.

CONCLUSIONS

The PLR and RDW might play important roles in evaluation of the risk of postoperative complications in AMI patients. The preoperative PLR and RDW are simple and useful predictors of postoperative complications in AMI patients.

Probability calibration-based prediction of recurrence rate in patients with diffuse large B-cell lymphoma

Background

Although many patients receive good prognoses with standard therapy, 30–50% of diffuse large B-cell lymphoma (DLBCL) cases may relapse after treatment. Statistical or computational intelligent models are powerful tools for assessing prognoses; however, many cannot generate accurate risk (probability) estimates. Thus, probability calibration-based versions of traditional machine learning algorithms are developed in this paper to predict the risk of relapse in patients with DLBCL.

Methods

Five machine learning algorithms were assessed, namely, naïve Bayes (NB), logistic regression (LR), random forest (RF), support vector machine (SVM) and feedforward neural network (FFNN), and three methods were used to develop probability calibration-based versions of each of the above algorithms, namely, Platt scaling (Platt), isotonic regression (IsoReg) and shape-restricted polynomial regression (RPR). Performance comparisons were based on the average results of the stratified hold-out test, which was repeated 500 times. We used the AUC to evaluate the discrimination ability (i.e., classification ability) of the model and assessed the model calibration (i.e., risk prediction accuracy) using the H-L goodness-of-fit test, ECE, MCE and BS.

Results

Sex, stage, IPI, KPS, GCB, CD10 and rituximab were significant factors predicting the 3-year recurrence rate of patients with DLBCL. For the 5 uncalibrated algorithms, the LR (ECE = 8.517, MCE = 20.100, BS = 0.188) and FFNN (ECE = 8.238, MCE = 20.150, BS = 0.184) models were well-calibrated. The errors of the initial risk estimate of the NB (ECE = 15.711, MCE = 34.350, BS = 0.212), RF (ECE = 12.740, MCE = 27.200, BS = 0.201) and SVM (ECE = 9.872, MCE = 23.800, BS = 0.194) models were large. With probability calibration, the biased NB, RF and SVM models were well-corrected. The calibration errors of the LR and FFNN models were not further improved regardless of the probability calibration method. Among the 3 calibration methods, RPR achieved the best calibration for both the RF and SVM models. The power of IsoReg was not obvious for the NB, RF or SVM models.

Conclusions

Although these algorithms all have good classification ability, several cannot generate accurate risk estimates. Probability calibration is an effective method of improving the accuracy of these poorly calibrated algorithms. Our risk model of DLBCL demonstrates good discrimination and calibration ability and has the potential to help clinicians make optimal therapeutic decisions to achieve precision medicine.

Using the Super Learner algorithm to predict risk of 30-day readmission after bariatric surgery in the United States

BACKGROUND

Risk prediction models that estimate patient probabilities of adverse events are commonly deployed in bariatric surgery. The objective was to validate a machine learning (Super Learner) prediction model of 30-day readmission after bariatric surgery in comparison with a traditional logistic regression.

METHODS

This prognostic study for validation of risk prediction models used data from the Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program. Patients who underwent elective laparoscopic gastric bypass or laparoscopic sleeve gastrectomy between 2015 and 2018 were included. Models used 5-fold cross-validation and were evaluated using the area under the receiver operating characteristic curve, the net reclassification index, and the integrated discrimination improvement.

RESULTS

The 30-day readmission rate among 393,833 patients was 3.9%. Super Learner area under the receiver operating characteristic curve was 0.674 (95% confidence interval 0.670-0.679), compared to 0.650 (95% confidence interval 0.645-0.654) for logistic regression. The net reclassification index was 0.239 (95% confidence interval 0.223-0.254), and 0.252 (95% confidence interval 0.249-0.255) for those who were and were not readmitted within 30 days. The integrated discrimination improvement was 0.0032 (95% confidence interval 0.0030-0.0033).

CONCLUSION

The Super Learner outperformed traditional logistic regression in predicting risk of 30-day readmission after bariatric surgery. Machine learning models may help target high-risk patients more optimally and prevent unnecessary readmissions.

Development of the "OS-SEV90 Score" to Predict Severe Postoperative Complications at 90 Days Following Bariatric Surgery

BACKGROUND

Bariatric surgery may be associated with severe postoperative complications (SPC). Factors associated with the risk of SPC have not been fully investigated.

OBJECTIVES

This study aimed to identify preoperative risk factors of SPC within 90 days and to develop a risk prediction model based on these factors.

METHODS

We conducted a retrospective single-center cohort study based on a prospectively maintained database of obese patients undergoing laparoscopic bariatric surgery from October 2005 to May 2019. All SPC occurring up to the 90th postoperative day were recorded according to the Dindo-Clavien classification. Associations between potential risk factors and SPC were analyzed using a logistic regression model, and the risk prediction ("OS-SEV90 score") was computed. Based on the OS-SEV90 score, the patients were grouped into 3 categories of risk: low, intermediate, and high.

RESULTS

Among 1963 consecutive patients, no patient died and 82 (4.2%) experienced SPC within 90 days. History of gastric or esophageal surgery (adjusted odds ratio (aOR) 3.040, 95% confidence interval; CI 1.78-5.20, p< 0.0001), past of thromboembolic event aOR 2.26, 95%; CI 1.12-4.55, p = 0.0225), and surgery performed by a junior surgeon (aOR 1.99, 95%; CI 1.26-3.13, p = 0.003) were all independently associated with risk for SPC, adjusting for ASA physical status system (ASA) score ≥ 3, severe OSA, psychiatric disease, asthma, a history of abdominal surgery, alcohol, cardiac disease, and dyslipidemia. "the OS-SEV90 score" based on these factors was constructed to classify patients into 3 risk groups: low (≤2), intermediate (3-4), and high (≥5). According to "the OS-SEV90 score," SPC increased significantly from 2.9% in the low-risk group, 7.7% in the intermediate-risk group, and 23.3% in the high-risk group.

CONCLUSIONS

A predictive model of SPC within 90 days "the OS-SEV90 score" has been developed using 9 baseline risk factors. The use of the OS-SEV90 score may help the multidisciplinary team to identify the specific risk of each patient and inform them about and optimize the comorbidities before the surgery. Further studies are warranted to validate this score in a new independent cohort before using it in clinical practice.

Preoperative chronic opioid use and its impact on early complications in bariatric surgery: a Swedish nationwide cohort study of 56,183 patients

BACKGROUND

The association between severe obesity and chronic pain makes opioid use common among bariatric patients. Preoperative opioid use has been identified as a risk factor in other surgical procedures.

OBJECTIVES

To examine the impact of preoperative opioid use on complications after primary bariatric surgery.

SETTING

Sweden.

METHODS

All primary laparoscopic Roux-en-Y gastric bypass (LRYGB) and laparoscopic sleeve gastrectomy (LSG) patients from 2007-2017 were identified in the Scandinavian Obesity Surgery Register. Prescriptions for opioids within 90 days prior to surgery were retrieved from the Swedish Prescribed Drug Register and converted into oral morphine equivalents (OMEs). Patients with ≥2 prescription of opioids within 90 days prior to surgery were defined as chronic opioid users. Generalized linear regression was used to adjust for age, sex, body mass index, procedure type, year of operation, and co-morbidities.

RESULTS

Of the 56,183 patients who had undergone primary LRYGB (n = 49,615) or LSG (n = 6568), 17.5% (n = 9825) had at least 1 prescription of opioids prior to surgery, of which 4.3% (n = 2390) were defined as chronic opioid users. Chronic opioid use was associated with a higher risk of severe complications (Clavien Dindo grade ≥ 3b; odds ratio [OR], 1.67; 95% confidence interval [CI], 1.37-2.04), increased lengths of stay (relative risk, 1.11; 95% CI, 1.08-1.14), and higher rates of readmission (OR, 1.70; 95% CI, 1.49-1.94) and reoperation (OR, 1.87; 95% CI, 1.53-2.27; all P values < .001). Furthermore, higher OME exposure was associated with stepwise higher risks.

CONCLUSION

Preoperative opioid use was an independent risk factor for severe complications, as well as prolonged lengths of stay, readmission, and reoperation after primary bariatric surgery.

Applying probability calibration to ensemble methods to predict 2-year mortality in patients with DLBCL

BACKGROUND

Under the influences of chemotherapy regimens, clinical staging, immunologic expressions and other factors, the survival rates of patients with diffuse large B-cell lymphoma (DLBCL) are different. The accurate prediction of mortality hazards is key to precision medicine, which can help clinicians make optimal therapeutic decisions to extend the survival times of individual patients with DLBCL. Thus, we have developed a predictive model to predict the mortality hazard of DLBCL patients within 2 years of treatment.

METHODS

We evaluated 406 patients with DLBCL and collected 17 variables from each patient. The predictive variables were selected by the Cox model, the logistic model and the random forest algorithm. Five classifiers were chosen as the base models for ensemble learning: the naïve Bayes, logistic regression, random forest, support vector machine and feedforward neural network models. We first calibrated the biased outputs from the five base models by using probability calibration methods (including shape-restricted polynomial regression, Platt scaling and isotonic regression). Then, we aggregated the outputs from the various base models to predict the 2-year mortality of DLBCL patients by using three strategies (stacking, simple averaging and weighted averaging). Finally, we assessed model performance over 300 hold-out tests.

RESULTS

Gender, stage, IPI, KPS and rituximab were significant factors for predicting the deaths of DLBCL patients within 2 years of treatment. The stacking model that first calibrated the base model by shape-restricted polynomial regression performed best (AUC = 0.820, ECE = 8.983, MCE = 21.265) in all methods. In contrast, the performance of the stacking model without undergoing probability calibration is inferior (AUC = 0.806, ECE = 9.866, MCE = 24.850). In the simple averaging model and weighted averaging model, the prediction error of the ensemble model also decreased with probability calibration.

CONCLUSIONS

Among all the methods compared, the proposed model has the lowest prediction error when predicting the 2-year mortality of DLBCL patients. These promising results may indicate that our modeling strategy of applying probability calibration to ensemble learning is successful.

Development and Validation of a Predictive Model of Success in Bariatric Surgery

PURPOSE

There are no criteria to establish priority for bariatric surgery candidates in the public health system in several countries. The aim of this study is to identify preoperative characteristics that allow predicting the success after bariatric surgery.

MATERIALS AND METHODS

Four hundred and sixty-one patients submitted to Roux-en-Y gastric bypass were included. Success of the surgery was defined as the sum of five outcome variables, assessed at baseline and 12 months after the surgery: excess weight loss, use of continuous positive airway pressure (CPAP) or bilevel positive airway pressure (BiPAP) as a treatment for obstructive sleep apnea (OSA), daily number of antidiabetics, daily number of antihypertensive drugs, and all-cause mortality. Partial least squares (PLS) regression and multiple linear regression were performed to identify preoperative predictors. We performed a 90/10 split of the dataset in train and test sets and ran a leave-one-out cross-validation on the train set and the best PLS model was chosen based on goodness-of-fit criteria.

RESULTS

The preoperative predictors of success after bariatric surgery included lower age, presence of non-alcoholic fatty liver disease and OSA, more years of CPAP/BiPAP use, negative history of cardiovascular disease, and lower number of antihypertensive drugs. The PLS model displayed a mean absolute percent error of 0.1121 in the test portion of the dataset, leading to accurate predictions of postoperative outcomes.

CONCLUSION

This success index allows prioritizing patients with the best indication for the procedure and could be incorporated in the public health system as a support tool in the decision-making process.

Limited Effect of Beta-blockade on Postoperative Outcome After Laparoscopic Gastric Bypass Surgery

BACKGROUND

The benefit of beta-blockade on postoperative outcome remains controversial, though recent studies have suggested a role during major non-cardiac surgery. The benefit of beta-blockade during minimally invasive gastric bypass surgery remains unclear. The aim of the present study was to evaluate the possible association between preoperative beta-blocker therapy and postoperative outcome after laparoscopic gastric bypass surgery.

METHODS

Patients operated with primary laparoscopic gastric bypass surgery in Sweden between 2007 and 2017 were identified through the Scandinavian Obesity Surgery Registry. The dataset was linked to the Swedish National Patient Registry, the Swedish Prescribed Drug Registry, and Statistics Sweden. The main outcome was serious postoperative complication within 30 days of surgery; with postoperative complication, 90-day and 1-year mortality, and weight loss at 2 years after surgery as secondary endpoints. The Poisson regression model was used to evaluate primary and secondary categorical outcomes. A general mixed model was performed to evaluate 2-year weight loss.

RESULTS

In all, 50281 patients were included in the study. No difference was seen between patients on beta-blockade and the control group regarding postoperative complications (adjusted incidence rate ratio 1.04 (95%CI 0.93-1.15), p = 0.506), serious postoperative complication (adjusted IRR 1.06 95%CI 0.89-1.27), p = 0.515), 90-day mortality (adjusted IRR 0.71 (95%CI 0.24-2.10), p = 0.537), and 1-year mortality (adjusted IRR 1.26 (95%CI 0.67-2.36), p = 0.467). Weight loss 2 years after surgery was slightly greater in patients on beta-blockade (adjusted coefficient 0.53 (95%CI 0.19-0.87), p = 0.002).

CONCLUSIONS

Beta-blockade has limited impact on postoperative outcome after laparoscopic gastric bypass surgery.

Using Bayesian Networks to Predict Long-Term Health-Related Quality of Life and Comorbidity after Bariatric Surgery: A Study Based on the Scandinavian Obesity Surgery Registry

Previously published literature has identified a few predictors of health-related quality of life (HRQoL) after bariatric surgery. However, performance of the predictive models was not evaluated rigorously using real world data. To find better methods for predicting prognosis in patients after bariatric surgery, we examined performance of the Bayesian networks (BN) method in predicting long-term postoperative HRQoL and compared it with the convolution neural network (CNN) and multivariable logistic regression (MLR). The patients registered in the Scandinavian Obesity Surgery Registry (SOReg) were used for the current study. In total, 6542 patients registered in the SOReg between 2008 and 2012 with complete demographic and preoperative comorbidity information, and preoperative and postoperative 5-year HROoL scores and comorbidities were included in the study. HRQoL was measured using the RAND-SF-36 and the obesity-related problems scale. Thirty-five variables were used for analyses, including 19 predictors and 16 outcome variables. The Gaussian BN (GBN), CNN, and a traditional linear regression model were used for predicting 5-year HRQoL scores, and multinomial discrete BN (DBN) and MLR were used for 5-year comorbidities. Eighty percent of the patients were randomly selected as a training dataset and 20% as a validation dataset. The GBN presented a better performance than the CNN and the linear regression model; it had smaller mean squared errors (MSEs) than those from the CNN and the linear regression model. The MSE of the summary physical scale was only 0.0196 for GBN compared to the 0.0333 seen in the CNN. The DBN showed excellent predictive ability for 5-year type 2 diabetes and dyslipidemia (area under curve (AUC) = 0.942 and 0.917, respectively), good ability for 5-year hypertension and sleep apnea syndrome (AUC = 0.891 and 0.834, respectively), and fair ability for 5-year depression (AUC = 0.750). Bayesian networks provide useful tools for predicting long-term HRQoL and comorbidities in patients after bariatric surgery. The hybrid network that may involve variables from different probability distribution families deserves investigation in the future.

Predicting Acute Kidney Injury after Cardiac Surgery by Machine Learning Approaches

Cardiac surgery-associated AKI (CSA-AKI) is common after cardiac surgery and has an adverse impact on short- and long-term mortality. Early identification of patients at high risk of CSA-AKI by applying risk prediction models allows clinicians to closely monitor these patients and initiate effective preventive and therapeutic approaches to lessen the incidence of AKI. Several risk prediction models and risk assessment scores have been developed for CSA-AKI. However, the definition of AKI and the variables utilized in these risk scores differ, making general utility complex. Recently, the utility of artificial intelligence coupled with machine learning, has generated much interest and many studies in clinical medicine, including CSA-AKI. In this article, we discussed the evolution of models established by machine learning approaches to predict CSA-AKI.

Deep Learning Neural Networks to Predict Serious Complications After Bariatric Surgery: Analysis of Scandinavian Obesity Surgery Registry Data

Background

Obesity is one of today’s most visible public health problems worldwide. Although modern bariatric surgery is ostensibly considered safe, serious complications and mortality still occur in some patients.

Objective

This study aimed to explore whether serious postoperative complications of bariatric surgery recorded in a national quality registry can be predicted preoperatively using deep learning methods.

Methods

Patients who were registered in the Scandinavian Obesity Surgery Registry (SOReg) between 2010 and 2015 were included in this study. The patients who underwent a bariatric procedure between 2010 and 2014 were used as training data, and those who underwent a bariatric procedure in 2015 were used as test data. Postoperative complications were graded according to the Clavien-Dindo classification, and complications requiring intervention under general anesthesia or resulting in organ failure or death were considered serious. Three supervised deep learning neural networks were applied and compared in our study: multilayer perceptron (MLP), convolutional neural network (CNN), and recurrent neural network (RNN). The synthetic minority oversampling technique (SMOTE) was used to artificially augment the patients with serious complications. The performances of the neural networks were evaluated using accuracy, sensitivity, specificity, Matthews correlation coefficient, and area under the receiver operating characteristic curve.

Results

In total, 37,811 and 6250 patients were used as the training data and test data, with incidence rates of serious complication of 3.2% (1220/37,811) and 3.0% (188/6250), respectively. When trained using the SMOTE data, the MLP appeared to have a desirable performance, with an area under curve (AUC) of 0.84 (95% CI 0.83-0.85). However, its performance was low for the test data, with an AUC of 0.54 (95% CI 0.53-0.55). The performance of CNN was similar to that of MLP. It generated AUCs of 0.79 (95% CI 0.78-0.80) and 0.57 (95% CI 0.59-0.61) for the SMOTE data and test data, respectively. Compared with the MLP and CNN, the RNN showed worse performance, with AUCs of 0.65 (95% CI 0.64-0.66) and 0.55 (95% CI 0.53-0.57) for the SMOTE data and test data, respectively.

Conclusions

MLP and CNN showed improved, but limited, ability for predicting the postoperative serious complications after bariatric surgery in the Scandinavian Obesity Surgery Registry data. However, the overfitting issue is still apparent and needs to be overcome by incorporating intra- and perioperative information.

A Comparative Study of Machine Learning Algorithms in Predicting Severe Complications after Bariatric Surgery

BACKGROUND

Severe obesity is a global public health threat of growing proportions. Accurate models to predict severe postoperative complications could be of value in the preoperative assessment of potential candidates for bariatric surgery. So far, traditional statistical methods have failed to produce high accuracy. We aimed to find a useful machine learning (ML) algorithm to predict the risk for severe complication after bariatric surgery.

METHODS

We trained and compared 29 supervised ML algorithms using information from 37,811 patients that operated with a bariatric surgical procedure between 2010 and 2014 in Sweden. The algorithms were then tested on 6250 patients operated in 2015. We performed the synthetic minority oversampling technique tackling the issue that only 3% of patients experienced severe complications.

RESULTS

Most of the ML algorithms showed high accuracy (>90%) and specificity (>90%) in both the training and test data. However, none of the algorithms achieved an acceptable sensitivity in the test data. We also tried to tune the hyperparameters of the algorithms to maximize sensitivity, but did not yet identify one with a high enough sensitivity that can be used in clinical praxis in bariatric surgery. However, a minor, but perceptible, improvement in deep neural network (NN) ML was found.

CONCLUSION

In predicting the severe postoperative complication among the bariatric surgery patients, ensemble algorithms outperform base algorithms. When compared to other ML algorithms, deep NN has the potential to improve the accuracy and it deserves further investigation. The oversampling technique should be considered in the context of imbalanced data where the number of the interested outcome is relatively small.

Impact of mesenteric defect closure technique on complications after gastric bypass

Background

Closure of mesenteric defects during laparoscopic gastric bypass surgery markedly reduces the risk for small bowel obstruction due to internal hernia. However, this procedure is associated with an increased risk for early small bowel obstruction and pulmonary complication. The purpose of the present study was to evaluate whether the learning curve and subsequent adaptions made to the technique have had an effect on the risk for complications.

Methods

The results of patients operated with a primary laparoscopic gastric bypass procedure, including closure of the mesenteric defects with sutures, during a period soon after introduction (January 1, 2010–December 31, 2011) were compared to those of patients operated recently (January 1, 2014–June 30, 2017). Data were retrieved from the Scandinavian Obesity Surgery Registry (SOReg). The main outcome was reoperation for small bowel obstruction within 30 days after surgery.

Results

A total of 5444 patients were included in the first group (period 1), and 1908 in the second group (period 2). Thirty-day follow-up rates were 97.1 and 97.5% respectively. The risk for early (within 30 days) small bowel obstruction was lower in period 2 than in period 1 (13/1860, 0.7% vs. 67/5285, 1.3%, OR 0.55 (0.30–0.99), p = 0.045). The risk for pulmonary complication was also reduced (5/1860, 0.3%, vs. 41/5285, 0.8%, OR 0.34 (0.14–0.87), p = 0.019).

Conclusion

Closure of mesenteric defects during laparoscopic gastric bypass surgery can be performed safely and should be viewed as a routine part of that operation.