TY - JOUR
T1 - Predicting outcomes in kidney stone endoscopic surgery by rotation forest algorithm
AU - Pooyesh, Shima
AU - Foshati, Saghar
AU - Sabeti, Malihe
AU - Parvin, Hamid
AU - Aminsharifi, Alireza
N1 - Publisher Copyright:
© 2022 Informa UK Limited, trading as Taylor & Francis Group.
PY - 2023
Y1 - 2023
N2 - Artificial intelligence (AI) has developed in medicine significantly during recent decades. Medical decision systems prevent probable errors stemming from clinical experts’ fatigue or inexperience. In addition, researchers can analyse medical data sets with the least time and the most details by the above-mentioned system. Nowadays, health care centres face mass data while extracting knowledge which is the most critical global challenges since diagnosis, cure, and prevention of diseases are considered extremely hard. Physicians should check patient test results and decisions made in the past for similar patients in order to decide properly due to the complexity of predicting the outcome of surgery. However, an automatic instrument such as data mining should be used for exploring patients due to their numbers. The present study utilised k-fold in front of the t-Test method to determine the best algorithm changing the parameters related to classification algorithms for increasing accuracy and found the best amount. Input variables included age, gender, Body Mass Index (BMI), previous surgery in target kidney, history of dialysis, one kidney, a kidney transplant, side of the kidney, stone side, size, and volume, postoperative stone volume, stone location, renal and skeletal anomaly, degree of hydronephrosis, stone lucency, history of high blood pressure and open nephrolithotomy, preoperative and postoperative haemoglobin and creatinine, Hx of hypertension and diabetes, heart disease, access point, radiation exposure, during access, total radiation, and operative time, while output variables included stone-free status and need for further surgeries such as Extracorporeal Shock Wave Lithotripsy (ESWL), Transurethral Lithotripsy (TUL), and Double J (DJ). Based on the results, the rotation forest algorithm and Area under the ROC Curve (AUC = 0.960) had the highest accuracy (85.87%) in predicting the outcomes of kidney stone endoscopic surgery.
AB - Artificial intelligence (AI) has developed in medicine significantly during recent decades. Medical decision systems prevent probable errors stemming from clinical experts’ fatigue or inexperience. In addition, researchers can analyse medical data sets with the least time and the most details by the above-mentioned system. Nowadays, health care centres face mass data while extracting knowledge which is the most critical global challenges since diagnosis, cure, and prevention of diseases are considered extremely hard. Physicians should check patient test results and decisions made in the past for similar patients in order to decide properly due to the complexity of predicting the outcome of surgery. However, an automatic instrument such as data mining should be used for exploring patients due to their numbers. The present study utilised k-fold in front of the t-Test method to determine the best algorithm changing the parameters related to classification algorithms for increasing accuracy and found the best amount. Input variables included age, gender, Body Mass Index (BMI), previous surgery in target kidney, history of dialysis, one kidney, a kidney transplant, side of the kidney, stone side, size, and volume, postoperative stone volume, stone location, renal and skeletal anomaly, degree of hydronephrosis, stone lucency, history of high blood pressure and open nephrolithotomy, preoperative and postoperative haemoglobin and creatinine, Hx of hypertension and diabetes, heart disease, access point, radiation exposure, during access, total radiation, and operative time, while output variables included stone-free status and need for further surgeries such as Extracorporeal Shock Wave Lithotripsy (ESWL), Transurethral Lithotripsy (TUL), and Double J (DJ). Based on the results, the rotation forest algorithm and Area under the ROC Curve (AUC = 0.960) had the highest accuracy (85.87%) in predicting the outcomes of kidney stone endoscopic surgery.
UR - http://www.scopus.com/inward/record.url?scp=85139740273&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85139740273&partnerID=8YFLogxK
U2 - 10.1080/21681163.2022.2131629
DO - 10.1080/21681163.2022.2131629
M3 - Article
AN - SCOPUS:85139740273
SN - 2168-1163
VL - 11
SP - 1397
EP - 1407
JO - Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization
JF - Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization
IS - 4
ER -
