scholarly journals Integration of Feature Vectors from Raw Laboratory, Medication and Procedure Names Improves the Precision and Recall of Models to Predict Postoperative Mortality and Acute Kidney Injury

Author(s):  
Ira S. Hofer ◽  
Marina Kupina ◽  
Lori Laddaran ◽  
Eran Halperin

Abstract Introduction: Manuscripts that have successfully used machine learning (ML) to predict a variety of perioperative outcomes often use only a limited number of features selected by a clinician. We hypothesized that techniques leveraging a broad set of features for patient laboratory results, medications, and the surgical procedure name would improve performance as compared to a more limited set of features chosen by clinicians. Methods Feature vectors for laboratory results included 702 features total derived from 39 laboratory tests, medications consisted of a binary flag for 126 commonly used medications, procedure name used the Word2Vec package for create a vector of length 100. Nine models were trained: Baseline Features, one for each of the three types of data Baseline+Each data type (, all features, and then all features with feature reduction algorithm. Results Across both outcomes the models that contained all features (model 8) (Mortality ROC-AUC 94.42, PR-AUC 31.0; AKI ROC-AUC 92.47, PR-AUC 76.73) was superior to models with only subsets of features Conclusion Featurization techniques leveraging a broad away of clinical data can improve performance of perioperative prediction models.

2021 ◽  
Vol 10 (1) ◽  
Author(s):  
Yi Xu ◽  
Siying Wang ◽  
Leilei He ◽  
Hong Yu ◽  
Hai Yu

Abstract Background The safety of perioperative intravenous hydroxyethyl starch (HES) products, specifically HES 130/0.4, continues to be the source of much debate. The aim of this meta-analysis was to update the existing evidence and gain further insight into the clinical effects of HES 130/0.4 on postoperative outcomes for volume replacement therapy in surgical patients. Methods MEDLINE, EMBASE, and Cochrane Library databases were searched from inception to March 2020 for relevant randomized controlled trials (RCTs) on perioperative use of HES 130/0.4 in adult surgical patients. The primary outcome was postoperative mortality and secondary outcomes were the incidence of acute kidney injury (AKI) and requirement for renal replacement therapy (RRT). The analysis was performed using the random-effects method and the risk ratio (RR) with a 95% confidence interval (CI). We performed the risk-of-bias assessment of eligible studies and assessed the overall quality of evidence for each outcome. Results Twenty-five RCTs with 4111 participants were finally included. There were no statistical differences between HES 130/0.4 and other fluids in mortality at 30 days (RR 1.28, 95% CI 0.88 to 1.86, p = 0.20), the incidence of AKI (RR 1.23, 95% CI 0.99 to 1.53, p = 0.07), or requirement for RRT (RR 0.75, 95% CI 0.37 to 1.53, p = 0.43). Overall, there was a moderate certainty of evidence for all the outcomes. There was no subgroup difference related to the type of surgery (p = 0.17) in the incidence of AKI. As for the type of comparator fluids, however, there was a trend that was not statistically significant (p = 0.06) towards the increased incidence of AKI in the HES 130/0.4 group (RR 1.22, 95% CI 0.97 to 1.54) compared with the crystalloid group (RR 1.21, 95% CI 0.27 to 3.91). Subgroup analyses according to the type of surgery demonstrated consistent findings. Conclusions This systematic review and meta-analysis suggests that the use of HES 130/0.4 for volume replacement therapy compared with other fluids resulted in no significant difference in postoperative mortality or kidney dysfunction among surgical patients. Given the absent evidence of confirmed benefit and the potential trend of increased kidney injury, we cannot recommend the routine clinical use of HES 130/0.4 for volume replacement therapy in surgical patients from the perspective of benefit/risk profile. However, the results need to be interpreted with caution due to the limited sample size, and further well-powered RCTs are warranted. Trial registration PROSPERO registry reference: CRD42020173058


2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
L Lei ◽  
Y He ◽  
Z Guo ◽  
B Liu ◽  
J Liu ◽  
...  

Abstract Background Patients with congestive heart failure (CHF) are vulnerable to contrast-induced acute kidney injury (CI-AKI), but few prediction models are currently available. Objectives We aimed to establish a simple nomogram for CI-AKI risk assessment for patients with CHF undergoing coronary angiography. Methods A total of 1876 consecutive patients with CHF (defined as New York Heart Association functional class II-IV or Killip class II-IV) were enrolled and randomly (2:1) assigned to a development cohort and a validation cohort. The endpoint was CI-AKI defined as serum creatinine elevation of ≥0.3 mg/dL or 50% from baseline within the first 48–72 hours following the procedure. Predictors for the nomogram were selected by multivariable logistic regression with a stepwise approach. The discriminative power was assessed using the area under the receiver operating characteristic (ROC) curve and was compared with the classic Mehran score in the validation cohort. Calibration was assessed using the Hosmer–Lemeshow test and 1000 bootstrap samples. Results The incidence of CI-AKI was 9.06% (n=170) in the total sample, 8.64% (n=109) in the development cohort and 9.92% (n=61) in the validation cohort (p=0.367). The simple nomogram including four predictors (age, intra-aortic balloon pump, acute myocardial infarction and chronic kidney disease) demonstrated a similar predictive power as the Mehran score (area under the curve: 0.80 vs 0.75, p=0.061), as well as a well-fitted calibration curve. Conclusions The present simple nomogram including four predictors is a simple and reliable tool to identify CHF patients at risk of CI-AKI, whereas further external validations are needed. Figure 1 Funding Acknowledgement Type of funding source: None


2021 ◽  
pp. 000313482199506
Author(s):  
Youngbae Jeon ◽  
Kyoung-Won Han ◽  
Won-Suk Lee ◽  
Jeong-Heum Baek

Purpose This study is aimed to evaluate the clinical outcomes of surgical treatment for nonagenarian patients with colorectal cancer. Methods This retrospective single-center study included patients diagnosed with colorectal cancer at the age of ≥90 years between 2004 and 2018. Patient demographics were compared between the operation and nonoperation groups (NOG). Perioperative outcomes, histopathological outcomes, and postoperative complications were evaluated. Overall survival was analyzed using Kaplan-Meier methods and log-rank test. Results A total of 31 patients were included (16 men and 15 women), and the median age was 91 (range: 90‐96) years. The number of patients who underwent surgery and who received nonoperative management was 20 and 11, respectively. No statistical differences in baseline demographics were observed between both groups. None of these patients were treated with perioperative chemotherapy or radiotherapy. Surgery comprised 18 (90.0%) colectomies and 2 (10.0%) transanal excisions. Short-term (≤30 days) and long-term (31‐90 days) postoperative complications occurred in 7 (35.0%) and 4 (20.0%) patients, respectively. No complications needed reoperation, such as anastomosis leakage or bleeding. No postoperative mortality occurred within 30 days: 90-day postoperative mortality occurred in two patients (10.0%), respectively. The median overall survival of the operation group was 31.6 (95% confidence interval: 26.7‐36.5) and that of NOG was 12.5 months (95% CI: 2.4‐22.6) ( P = 0.012). Conclusion Surgical treatment can be considered in carefully selected nonagenarian patients with colorectal cancer in terms of acceptable postoperative morbidity, with better overall survival than the nonsurgical treatment.


Renal Failure ◽  
2020 ◽  
Vol 42 (1) ◽  
pp. 869-876
Author(s):  
Yang Li ◽  
Xiaohong Chen ◽  
Ziyan Shen ◽  
Yimei Wang ◽  
Jiachang Hu ◽  
...  

2021 ◽  
Vol 36 (Supplement_1) ◽  
Author(s):  
Enrico Favaro ◽  
Roberta Lazzarin ◽  
Daniela Cremasco ◽  
Erika Pierobon ◽  
Marta Guizzo ◽  
...  

Abstract Background and Aims The modern development of the black box approach in clinical nephrology is inconceivable without a logical theory of renal function and a comprehension of anatomical architecture of the kidney, in health and disease: this is the undisputed contribution offered by Malpighi, Oliver and Trueta starting from the seventeenth century. The machine learning model for the prediction of acute kidney injury, progression of renal failure and tubulointerstitial nephritis is a good example of how different knowledge about kidney are an indispensable tool for the interpretation of model itself. Method Historical data were collected from literature, textbooks, encyclopedias, scientific periodicals and laboratory experimental data concerning these three authors. Results The Italian Marcello Malpighi (1628-1694), born in Crevalcore near Bologna, was Professor of anatomy at Bologna, Pisa and Messina. The historic description of the pulmonary capillaries was made in his second epistle to Borelli published in 1661 and intitled De pulmonibus, by means of the frog as “the microscope of nature” (Fig. 1). It is the first description of capillaries in any circulation. William Harvey in De motu cordis in 1628 (year of publication the same of date of birth of Italian anatomist!) could not see the capillary vessels. This thriumphant discovery will serve for the next reconnaissance of characteristic renal rete mirabile.in the corpuscle of Malpighi, lying within the capsule of Bowman. Jean Redman Oliver (1889-1976), a pathologist born and raised in Northern California, was able to bridge the gap between the nephron and collecting system through meticulous dissections, hand drawn illustrations and experiments which underpin our current understanding of renal anatomy and physiology. In the skillful lecture “When is the kidney not a kidney?” (1949) Oliver summarizes his far-sighted vision on renal physiology and disease in the following sentence: the Kidney in health, if you will, but the Nephrons in disease. Because, the “nephron” like the “kidney” is an abstraction that must be qualified in terms of its various parts, its cellular components and the molecular mechanisms involved in each discrete activity (Fig. 2). The Catalan surgeon Josep Trueta I Raspall (1897-1977) was born in the Poblenou neighborhood of Barcelona. His impact of pioneering and visionary contribution to the changes in renal circulation for the pathogenesis of acute kidney injury was pivotal for history of renal physiology. “The kidney has two potential circulatory circulations. Blood may pass either almost exclusively through one or other of two pathways, or to a varying degree through both”. (Studies of the Renal Circulation, published in 1947). Now this diversion of blood from cortex to the less resistant medullary circulation is known with the eponym Trueta shunt. Conclusion The black box approach to the kidney diseases should be considered by practitioners as a further tool to help to inform model update in many clinical setting. The number of machine learning clinical prediction models being published is rising, as new fields of application are being explored in medicine (Fig. 3). A challenge in the clinical nephrology is to explore the “kidney machine” during each therapeutic diagnostic procedure. Always, the intriguing relationship between the set of nephrological syndromes and kidney diseases cannot disregard the precious notions the specific organization of kidney microcirculation, fruit of many scientific contributions of the work by Malpighi, Oliver and Trueta (Fig. 3).


2020 ◽  
Author(s):  
Sabe Mwape ◽  
Victor Daka ◽  
Scott Matafwali ◽  
Kapambwe Mwape ◽  
Jay Sikalima ◽  
...  

Background Medical laboratory diagnosis is a critical component of patient management in the healthcare setup. Despite the availability of laboratory tests, clinicians may not utilise them to make clinical decisions. We investigated utilsation of laboratory tests for patient management among clinicians at Ndola Teaching Hospital (NTH) and Arthur Davison Childrens Hospital (ADCH), two large referral hospitals in the Copperbelt Province, Ndola, Zambia. Method We conducted a descriptive cross-sectional study among clinicians. The study deployed self-administered questionnaires to evaluate clinician utilisation, querying and confidence in laboratory results. Additional data on demographics and possible laboratory improvements were also obtained. Data were entered in Microsoft excel and exported to SPSS version 16 for statistical analysis. Results Of the 80 clinicians interviewed, 96.2% (77) reported using laboratory tests and their results in patient management. 77.5% (62) of the clinicians indicated they always used laboratory results to influence their patient management decisions. Of the selected laboratory tests, clinicians were more confident in using haemoglobin test results (91.2%). There was no statistically significant association between the clinicians gender or qualification and use of test results in patient management. Conclusion Our findings show that despite the majority querying laboratory results, most of the clinicians use laboratory results for patient management. There is need for interactions between the laboratory and clinical area to assure clinician confidence in laboratory results. Key words: utilisation, clinicians, laboratory tests, Ndola Teaching Hospital, Arthur Davison Childrens Hospital


2020 ◽  
Author(s):  
Vimaladhasan Senthamizhan ◽  
Balaraman Ravindran ◽  
Karthik Raman

AbstractEssential gene prediction models built so far are heavily reliant on sequence-based features and the scope of network-based features has been narrow. Previous work from our group demonstrated the importance of using network-based features for predicting essential genes with high accuracy. Here, we applied our approach for the prediction of essential genes to organisms from the STRING database and hosted the results in a standalone website. Our database, NetGenes, contains essential gene predictions for 2700+ bacteria predicted using features derived from STRING protein-protein functional association networks. Housing a total of 3.5M+ genes, NetGenes offers various features like essentiality scores, annotations and feature vectors for each gene. NetGenes is available at https://rbc-dsai.iitm.github.io/NetGenes/


Sign in / Sign up

Export Citation Format

Share Document