Predicting postoperative nausea and vomiting with the application of an artificial neural network

Artificial intelligence and its clinical application in Anesthesiology: a systematic review

PURPOSE

Application of artificial intelligence (AI) in medicine is quickly expanding. Despite the amount of evidence and promising results, a thorough overview of the current state of AI in clinical practice of anesthesiology is needed. Therefore, our study aims to systematically review the application of AI in this context.

METHODS

A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We searched Medline and Web of Science for articles published up to November 2022 using terms related with AI and clinical practice of anesthesiology. Articles that involved animals, editorials, reviews and sample size lower than 10 patients were excluded. Characteristics and accuracy measures from each study were extracted.

RESULTS

A total of 46 articles were included in this review. We have grouped them into 4 categories with regard to their clinical applicability: (1) Depth of Anesthesia Monitoring; (2) Image-guided techniques related to Anesthesia; (3) Prediction of events/risks related to Anesthesia; (4) Drug administration control. Each group was analyzed, and the main findings were summarized. Across all fields, the majority of AI methods tested showed superior performance results compared to traditional methods.

CONCLUSION

AI systems are being integrated into anesthesiology clinical practice, enhancing medical professionals' skills of decision-making, diagnostic accuracy, and therapeutic response.

Artificial neural network application for identifying risk of depression in high school students: a cross-sectional study

BACKGROUND

Identifying important factors contributing to depression is necessary for interrupting risk pathways to minimize adolescent depression. The study aimed to assess the prevalence of depression in high school students and develop a model for identifying risk of depression among adolescents.

METHODS

Cross-sectional study was conducted. A total of 1190 adolescents from two high schools in eastern China participated in the study. Artificial neurol network (ANN) was used to establish the identification model.

RESULTS

The prevalence of depression was 29.9% among the students. The model showed the top five protective and risk factors including perceived stress, life events, optimism, self-compassion and resilience. ANN model accuracy was 81.06%, with sensitivity 65.3%, specificity 88.4%, and area under the receiver operating characteristic (ROC) curves 0.846 in testing dataset.

CONCLUSION

The ANN showed the good performance in identifying risk of depression. Promoting the protective factors and reducing the level of risk factors facilitate preventing and relieving depression.

Machine learning applications to clinical decision support in neurosurgery: an artificial intelligence augmented systematic review

Machine learning (ML) involves algorithms learning patterns in large, complex datasets to predict and classify. Algorithms include neural networks (NN), logistic regression (LR), and support vector machines (SVM). ML may generate substantial improvements in neurosurgery. This systematic review assessed the current state of neurosurgical ML applications and the performance of algorithms applied. Our systematic search strategy yielded 6866 results, 70 of which met inclusion criteria. Performance statistics analyzed included area under the receiver operating characteristics curve (AUC), accuracy, sensitivity, and specificity. Natural language processing (NLP) was used to model topics across the corpus and to identify keywords within surgical subspecialties. ML applications were heterogeneous. The densest cluster of studies focused on preoperative evaluation, planning, and outcome prediction in spine surgery. The main algorithms applied were NN, LR, and SVM. Input and output features varied widely and were listed to facilitate future research. The accuracy (F_(2,19) = 6.56, p < 0.01) and specificity (F_(2,16) = 5.57, p < 0.01) of NN, LR, and SVM differed significantly. NN algorithms demonstrated significantly higher accuracy than LR. SVM demonstrated significantly higher specificity than LR. We found no significant difference between NN, LR, and SVM AUC and sensitivity. NLP topic modeling reached maximum coherence at seven topics, which were defined by modeling approach, surgery type, and pathology themes. Keywords captured research foci within surgical domains. ML technology accurately predicts outcomes and facilitates clinical decision-making in neurosurgery. NNs frequently outperformed other algorithms on supervised learning tasks. This study identified gaps in the literature and opportunities for future neurosurgical ML research.

Prediction of Bispectral Index during Target-controlled Infusion of Propofol and Remifentanil

Background

The discrepancy between predicted effect-site concentration and measured bispectral index is problematic during intravenous anesthesia with target-controlled infusion of propofol and remifentanil. We hypothesized that bispectral index during total intravenous anesthesia would be more accurately predicted by a deep learning approach.

Methods

Long short-term memory and the feed-forward neural network were sequenced to simulate the pharmacokinetic and pharmacodynamic parts of an empirical model, respectively, to predict intraoperative bispectral index during combined use of propofol and remifentanil. Inputs of long short-term memory were infusion histories of propofol and remifentanil, which were retrieved from target-controlled infusion pumps for 1,800 s at 10-s intervals. Inputs of the feed-forward network were the outputs of long short-term memory and demographic data such as age, sex, weight, and height. The final output of the feed-forward network was the bispectral index. The performance of bispectral index prediction was compared between the deep learning model and previously reported response surface model.

Results

The model hyperparameters comprised 8 memory cells in the long short-term memory layer and 16 nodes in the hidden layer of the feed-forward network. The model training and testing were performed with separate data sets of 131 and 100 cases. The concordance correlation coefficient (95% CI) were 0.561 (0.560 to 0.562) in the deep learning model, which was significantly larger than that in the response surface model (0.265 [0.263 to 0.266], P &lt; 0.001).

Conclusions

The deep learning model–predicted bispectral index during target-controlled infusion of propofol and remifentanil more accurately compared to the traditional model. The deep learning approach in anesthetic pharmacology seems promising because of its excellent performance and extensibility.

Applying Naive Bayesian Networks to Disease Prediction: a Systematic Review

INTRODUCTION

Naive Bayesian networks (NBNs) are one of the most effective and simplest Bayesian networks for prediction.

OBJECTIVE

This paper aims to review published evidence about the application of NBNs in predicting disease and it tries to show NBNs as the fundamental algorithm for the best performance in comparison with other algorithms.

METHODS

PubMed was electronically checked for articles published between 2005 and 2015. For characterizing eligible articles, a comprehensive electronic searching method was conducted. Inclusion criteria were determined based on NBN and its effects on disease prediction. A total of 99 articles were found. After excluding the duplicates (n= 5), the titles and abstracts of 94 articles were skimmed according to the inclusion criteria. Finally, 38 articles remained. They were reviewed in full text and 15 articles were excluded. Eventually, 23 articles were selected which met our eligibility criteria and were included in this study.

RESULT

In this article, the use of NBN in predicting diseases was described. Finally, the results were reported in terms of Accuracy, Sensitivity, Specificity and Area under ROC curve (AUC). The last column in Table 2 shows the differences between NBNs and other algorithms.

DISCUSSION

This systematic review (23 studies, 53,725 patients) indicates that predicting diseases based on a NBN had the best performance in most diseases in comparison with the other algorithms. Finally in most cases NBN works better than other algorithms based on the reported accuracy.

CONCLUSION

The method, termed NBNs is proposed and can efficiently construct a prediction model for disease.

Gender-Specific Differences in Low-Dose Haloperidol Response for Prevention of Postoperative Nausea and Vomiting: A Register-Based Cohort Study

Background

Postoperative nausea and vomiting (PONV) is one of the most common and distressing complications after general anesthesia and surgery, with young non-smoking females receiving postoperative opioids being high-risk patients. This register-based study aims to evaluate the effect of low-dose haloperidol (0.5 mg intravenously) directly after induction of general anesthesia to reduce the incidence of PONV in the postoperative anesthesiological care unit (PACU).

Methods

Multivariable regression models were used to investigate the association between low-dose haloperidol and the occurrence of PONV using a patient registry containing 2,617 surgical procedures carried out at an university hospital.

Results

Haloperidol 0.5 mg is associated with a reduced risk of PONV in the total collective (adjusted odds ratio = 0.75, 95% confidence interval: [0.56, 0.99], p = 0.05). The results indicate that there is a reduced risk in male patients (adjusted odds ratio = 0.45, 95% confidence interval: [0.28, 0.73], p = 0.001) if a dose of 0.5 mg haloperidol was administered while there seems to be no effect in females (adjusted odds ratio = 1.02, 95% confidence interval: [0.71, 1.46], p = 0.93). Currently known risk factors for PONV such as female gender, duration of anesthesia and the use of opioids were confirmed in our analysis.

Conclusion

This study suggests that low-dose haloperidol has an antiemetic effect in male patients but has no effect in female patients. A confirmation of the gender-specific effects we have observed in this register-based cohort study might have major implications on clinical daily routine.

Does previous chemotherapy-induced nausea and vomiting predict postoperative nausea and vomiting?

BACKGROUND

Postoperative nausea and vomiting (PONV) remains a problem in the postoperative period. Previous PONV in oncology patients has recently been associated with chemotherapy-induced nausea and vomiting (CINV). We assessed if CINV could improve Apfel's heuristic for predicting PONV.

METHODS

We conducted a retrospective study of 1500 consecutive patients undergoing intermediate or major cancer surgery between April and July 2011. PONV was assessed in the first postoperative day during post-anaesthesia care. The assigned anaesthetist completed an electronic medical record with all of the studied variables. Multiple logistic regression analyses were performed to assess whether any of the variables could add predictive ability to Apfel's tallying heuristic, and receiver operating characteristic (ROC) curves were modelled. The areas under the curve (AUC) were used to compare the model's discriminating ability for predicting patients who vomited from those who did not vomit.

RESULTS

The overall incidence of PONV was 26%. Multiple logistic regressions identified two independent predictors for PONV (odds ratio; 95% CI), Apfel's score (1.78; 1.23-2.63) and previous chemotherapy-induced vomiting (3.15; 1.71-5.9), Hosmer-Lemeshow's P < 0.0001. Previous CINV was the most significant predictor to be added to Apfel's heuristic in this population.

CONCLUSIONS

A history of chemotherapy-induced nausea vomiting was a strong predictor for PONV and should be investigated as an added risk factor for PONV in the preoperative period of oncology surgery in prospective studies.

Description and pilot evaluation of the Metabolic Irregularities Narrowing down Device software: a case analysis of physician programming

Background

There is a gap between the abilities and the everyday applications of Computerized Decision Support Systems (CDSSs). This gap is further exacerbated by the different ‘worlds’ between the software designers and the clinician end-users. Software programmers often lack clinical experience whereas practicing physicians lack skills in design and engineering.

Objective

Our primary objective was to evaluate the performance of Metabolic Irregularities Narrowing down Device (MIND) intelligent medical calculator and differential diagnosis software through end-user surveys and discuss the roles of CDSS in the inpatient setting.

Setting

A tertiary care, teaching community hospital.

Study participants

Thirty-one responders answered the survey. Responders consisted of medical students, 24%; attending physicians, 16%, and residents, 60%.

Results

About 62.5% of the responders reported that MIND has the ability to potentially improve the quality of care, 20.8% were sure that MIND improves the quality of care, and only 4.2% of the responders felt that it does not improve the quality of care. Ninety-six percent of the responders felt that MIND definitely serves or has the potential to serve as a useful tool for medical students, and only 4% of the responders felt otherwise. Thirty-five percent of the responders rated the differential diagnosis list as excellent, 56% as good, 4% as fair, and 4% as poor.

Discussion

MIND is a suggesting, interpreting, alerting, and diagnosing CDSS with good performance and end-user satisfaction. In the era of the electronic medical record, the ongoing development of efficient CDSS platforms should be carefully considered by practicing physicians and institutions.

Predicting postoperative vomiting for orthopedic patients receiving patient-controlled epidural analgesia with the application of an artificial neural network

Patient-controlled epidural analgesia (PCEA) was used in many patients receiving orthopedic surgery to reduce postoperative pain but is accompanied with certain incidence of vomiting. Predictions of the vomiting event, however, were addressed by only a few authors using logistic regression (LR) models. Artificial neural networks (ANN) are pattern-recognition tools that can be used to detect complex patterns within data sets. The purpose of this study was to develop the ANN based predictive model to identify patients with high risk of vomiting during PCEA used. From January to March 2007, the PCEA records of 195 patients receiving PCEA after orthopedic surgery were used to develop the two predicting models. The ANN model had a largest area under curve (AUC) in receiver operating characteristic (ROC) curve. The areas under ROC curves of ANN and LR models were 0.900 and 0.761, respectively. The computer-based predictive model should be useful in increasing vigilance in those patients most at risk for vomiting while PCEA is used, allowing for patient-specific therapeutic intervention, or even in suggesting the use of alternative methods of analgesia.

Self-adjustments may account for the contradictory correlations between HRV and motion-sickness severity

This study investigates the relationship between heart rate variability (HRV) and the level of motion sickness (MS) induced by simulated tunnel driving. The HRV indices, normalized low frequency (NLF, 0.04-0.15 Hz), normalized high frequency (NHF, 0.15-0.4 Hz), and LF/HF ratio were correlated with the subjectively and continuously rated MS levels of 20 participants. The experimental results showed that for 13 of the subjects, the MS levels positively correlated with the NLF and the LF/HF ratio and negatively correlated with the NHF. The remaining seven subjects had negative correlations between the MS levels and the NLF and the LF/HF ratio and a positive correlation between the MS levels and the NHF. To clarify this contradiction, this study also inspected the effects of subjects' self-adjustments on the correlations between the MS levels and the HRV indices and showed that the variations in the relationship might be attributed to the subjects' self-adjustments, which they used to relieve the discomfort of MS.

Comparison of survival predictions for rats with hemorrhagic shocks using an artificial neural network and support vector machine

Hemorrhagic shock is the cause of one third of deaths resulting from injury in the world. Early diagnosis of hemorrhagic shock makes it possible for physicians to treat patients successfully. The objective of this study was to select an optimal survival prediction model using physiological parameters from rats during our hemorrhagic experiment. These physiological parameters were used for the training and testing of survival prediction models using an artificial neural network (ANN) and support vector machine (SVM). To avoid over-fitting, we chose the optimal survival prediction model according to performance measured by a 5-fold cross validation method. We selected an ANN with three hidden neurons and one hidden layer and an SVM with Gaussian kernel function as a trained survival prediction model. For the ANN model, the sensitivity, specificity, and accuracy of survival prediction were 97.8 ± 3.3 %, 96.3 ± 2.7 %, and 96.8 ± 1.7 %, respectively. For the SVM model, the sensitivity, specificity, and accuracy were 97.5 ± 2.9 %, 99.3 ± 1.1 %, and 98.5 ± 1.2 %, respectively. SVM was preferable to ANN for the survival prediction.

Effect of respiratory modulation on relationship between heart rate variability and motion sickness

This study investigates the interplay among heart rate variability (HRV), respiration, and the severity of motion sickness (MS) in a realistic passive driving task. Although HRV is a commonly used metrically in physiological research or even believed to be a direct measure of sympathovagal activities, the results of MS-effected HRV remain mixed across studies. The goal of this study is to find the source of these contradicting results of HRV associated with MS. Experimental results of this study showed that the group trend of the low-frequency (LF) component and the LF/HF ratio increased and high-frequency (HF) component decreased significantly as self-reported MS level increased (p<0.001), consistent with a perception-driven autonomic response of the cardiovascular system. However, in one of the subjects, the relationship was reversed when individuals intentionally adjust themselves (deep breathing) to relieve the discomfort of MS during the experiments. It appears that the correlations between HRV and MS level were higher when individuals made fewer adjustments (the number of deep breathing) during the passive driving experiments.

Predicting adverse outcomes of cardiac surgery with the application of artificial neural networks

Risk-stratification models based on pre-operative patient and disease characteristics are useful for providing individual patients with an insight into the potential risk of complications and mortality, for aiding the clinical decision for surgery vs non-surgical therapy, and for comparing the quality of care between different surgeons or hospitals. Our study aimed to apply artificial neural networks (ANN) models to predict mortality and morbidity after cardiac surgery, and also to compare the efficacy of this model to that of the logistic regression model and Parsonnet score. The accuracy of the ANN, logistic regression and Parsonnet score in predicting mortality was 83.8%, 87.9% and 78.4%. The accuracy of the ANN, logistic regression and Parsonnet score in predicting major morbidity was 79.0%, 74.3% and 68.6%. The area under the receiver operating characteristic curves (AUC) of the ANN, logistic regression and Parsonnet score in predicting in-hospital mortality were 0.873, 0.852 and 0.829. The AUCs of the ANN, logistic regression and Parsonnet score in predicting major morbidity were 0.852, 0.789 and 0.727. The results showed the ANN models have the best discriminating power in predicting in-hospital mortality and morbidity among these models.

Predicting technique survival in peritoneal dialysis patients: comparing artificial neural networks and logistic regression

Background. Early technique failure has been a major limitation on the wider adoption of peritoneal dialysis (PD). The objectives of this study were to use data from a large, multi-centre, prospective database, the United Kingdom Renal Registry (UKRR), in order to determine the ability of an artificial neural network (ANN) model to predict early PD technique failure and to compare its performance with a logistic regression (LR)-based approach.

Methods. The analysis included all incident PD patients enrolled in the UKRR from 1999 to 2004. The event of interest was technique failure. For both the ANN and LR analyses a bootstrap approach was used: the data were divided into 20 random training (75%) and validation (25%) sets. Models were derived on the latter and then used to make predictions on the former. Predictive accuracy was assessed by area under the ROC curve (AUROC). The 20 AUROC values and their standard errors were then averaged.

Results. There were 3269 patients included in the analysis with a mean age of 59.9 years and a mean observation time of 430 days. Of the patients, 38.3% were female and 90.8% were Caucasian. 1458 patients (44.6%) suffered technique failure. The AUROC for the ANN model was 0.760 ± 0.0167 and the LR model was 0.709 and 0.0208. (P = 0.0164)

Conclusions. Using UKRR data, both ANN and LR models predicted early PD technique failure with moderate accuracy. In this study, an ANN outperformed an LR-based approach. As the scope and the completeness of the UKRR increases, the question of whether more sophisticated ANN models will perform even better remains for further study.

Recent advances, trends and economic considerations in the risk assessment, prevention and treatment of postoperative nausea and vomiting

During the last two decades there have been considerable achievements regarding the management of postoperative nausea and vomiting (PONV). Due to the importance of these symptoms in the aim to streamline clinical processes and to improve patient satisfaction, the debate on the best strategies and also research that focuses on PONV continues. This review summarises the recent developments with respect to the management of PONV. Following a brief review on what is already known on the risk assessment, prevention and treatment of PONV, newer trends in the pharmacological prevention (dexamethasone, neurokinin-1 antagonists, multimodal prevention) will be discussed as well as new insights regarding the value of algorithms for the prevention of PONV. Further, pharmacogenetically based algorithms (according to the metaboliser status) as well as new treatment strategies (dexamethasone, multimodal treatment) will be covered. No drug so far can achieve a reduction of PONV of more than one third. Furthermore, all clinical studies consistently demonstrated that a combination treatment has a simple additive effect without any relevant interaction between different drugs or classes of drugs. The relative reduction of approximately 30% can also be expected from dexamethasone and it is likely that the substances presently in development and in an early clinical use (e.g., neurokinin-1 antagonists) will not represent the new panacea. However, they will probably replenish the existing antiemetic portfolio to better cope with high risk patients. Stratified prevention using pharmacogenetic knowledge is still in the early stages. Algorithms need to be customized to the local settings in order to prove efficient. Treatment remains a most important pillar and there is evidence that the principles of combining antiemetics to prolong effects and improve protection can be similarly applied to treatment. Recent developments in the area of PONV are more related to implementing the already existing evidence than based on the introduction of new molecules. New molecules replenish the pharmacological antiemetic portfolio, which is needed due to the limited efficacy of any single agent available so far. The new neurokinin-1 receptor antagonist, aprepitant, and the long lasting 5-HT(3) receptor antagonist palonosetron are the latest developments in this context. Treatment is most important and can also be regarded as a secondary prevention. Due to limited efficacy of single treatment interventions, combination therapy may gain more widespread use in the future.

Prevention and control of postoperative nausea and vomiting in post-craniotomy patients

Postoperative nausea and vomiting (PONV) are the most frequent side-effects in the postoperative period, impairing subjective well-being and having economic impact due to delayed discharge. However, emetic symptoms can also cause major medical complications, and post-craniotomy patients may be at an increased risk. A review and critical appraisal of the existing literature on PONV in post-craniotomy patients, and a comparison of these findings with the current knowledge on PONV in the general surgical population, leads to the following conclusions: (1) Despite the lack of a documented case of harm caused by retching or vomiting in a post-craniotomy patient, the potential risk caused by arterial hypertension and high intra-abdominal/intra-thoracic pressure leading to high intracranial pressure, forces to avoid PONV in these patients. (2) There is unclarity about a specifically increased (or decreased) risk for PONV in post-craniotomy patients compared with other surgical procedures. (3) The decision whether or not to administer an antiemetic should not be based primarily on risk scores for PONV but on the likelihood for potential catastrophic consequences of PONV. If such a risk cannot be ruled out, a multimodal antiemetic approach should be considered regardless of the individual risk. (4) Randomized controlled trials with antiemetics in post-craniotomy patients are limited with respect to sample size and methodological quality. This also impacts upon the meaning of meta-analyses performed with trials that showed marked heterogeneity and inconclusive results. (5) No studies on the treatment of established PONV are available. This highlights the need to transfer knowledge about PONV treatment from other surgical procedures. (6) Despite the possibility that PONV in post-craniotomy patients can be triggered by specific conditions (e.g. surgery near the area postrema at the floor of the fourth ventricle with the vomiting centre located nearby), recommendations based on trials in post-craniotomy patients may be flawed. Thus, general knowledge on prevention and treatment of PONV must adopted for craniotomy settings.