NCT04410848

Symbolic Regression Model To Predict Choledocholithiasis

Symbolic Regression Model To Predict Choledocholithiasis: Prospective Validation

• Status: Unknown
• Phase: —
• Study type: Observational
• Enrollment: 200 (estimated)
• Sponsor: Hospital Universitario Dr. Jose E. Gonzalez · Academic / Other
• Sex: All
• Age: 18 Years
• Healthy volunteers: Not accepted

Summary

Choledocholithiasis refers to the presence of gallstones within the common bile duct. It is proposed to look for markers that help in the diagnosis and in differentiating between retained and migrated gallstones. The selection of patients is a very important aspect, due to the economic aspects and possible complications. Taking advantage of the development of technology, the improvement in computer systems, the use of artificial intelligence and a symbolic regression model that works to predict the presence of choledocholithiasis and provide evidence that clarifies the treatment of patients with this pathology, especially in this group where there is a bigger controversy.

Detailed description

Choledocholithiasis refers to the presence of gallstones within the common bile duct. It is proposed to look for markers that help in the diagnosis and in differentiating between retained and migrated gallstones. The selection of patients to perform endoscopic retrograde cholangiopancreatography (ERCP) is a very important aspect, due to the economic aspects and possible complications. By making a proper patient selection for additional studies or procedure, then the costs, complications and days of stay would be reduced. Avoiding the unnecessary use of ERCP would avoid its complications. Taking advantage of the development of technology, the improvement in computer systems, the use of artificial intelligence and a Symbolic Regression Model that works to predict the presence of choledocholithiasis and provide evidence that clarifies the treatment of patients with this pathology, especially in this group where there is a bigger controversy. Having the historical database of the University Hospital (HU), regarding clinical, laboratory and image variables of patients with suspected choledocholithiasis, using a symbolic regression method, several randomly formed equations are generated. Each equation deducts its coefficient of linear correlation (Pearson's correlation). For the following study we admitted to the emergency department of adults at the University Hospital all patients with clinical suspicion of choledocholithiasis, who meet the inclusion criterion. The study which is realized is a normal one based on the method of clinical predictors, obtaining laboratory studies, image studies, and patient management will be carried out based on the method of clinical predictors. The calculation is made with the equation obtained, and the patient is monitored until discharge. The calculation obtained from the equation will not be taken into account for the decisions in the management of the patient. The variables studied as white blood cells, total bilirubin values, direct bilirubin, indirect bilirubin, Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST), alkaline phosphatase (AP), gamma-glutamyl transpeptidase (GGT) at admission will be taken from the clinical record. Transabdominal ultrasonography will be performed upon admission by the diagnostic radiology department of the HU and the size of the bile duct in mm, presence of gallbladder gallstones and bile duct stones will be taken from the report. ERCP, magnetic resonance cholangiopancreatography (MRCP) or intraoperative cholangiography will be performed and one will be taken as a confirmation of choledocolithiasis, and its absence would rule it out.

Conditions

• Common Bile Duct Calculi

Interventions

Type	Name	Description
OTHER	To determine the diagnostic of Choledocholithiasis with symbolic regression model	To determine the diagnostic of Choledocholithiasis with symbolic regression model

Timeline

• Start date: 2019-06-07
• Primary completion: 2020-05-30
• Completion: 2020-07-31
• First posted: 2020-06-01
• Last updated: 2020-06-01

Locations

1 site across 1 country: Mexico