CN-122000072-A - Construction method and system of acute obstruction suppurative cholangitis conservation treatment failure risk prediction model

Abstract

The application provides a construction method and a construction system of a conservative treatment failure risk prediction model of acute obstructive suppurative cholangitis. The construction method comprises the steps of S1, collecting clinical and laboratory index data of a plurality of cases of acute obstructive suppurative cholangitis patients meeting nanodischarge standards, preprocessing, dividing the data into a training set and a testing set, S2, balancing the training set data by adopting a ROSE algorithm, then carrying out feature screening by sequentially utilizing Spearman correlation analysis, single factor ROC analysis and LASSO regression, S3, taking whether the patient aggravates as a dependent variable within 24 hours or not, establishing a prediction model by adopting multivariate logistic regression, optimizing model parameters through 10-fold cross validation, and S4, carrying out model performance verification evaluation based on the testing set. The method and the device remarkably improve early prediction precision and realize risk stratification and treatment decision quantification.

Inventors

• LIN RUI
• WANG ZHIYONG
• XIE JI
• LI XIANZHEN
• LI LIANGHUI
• Shen Xiaqing

Assignees

• 上海市同济医院
• 江西省崇义县人民医院

Dates

Publication Date

20260508

Application Date

20251219

Claims (6)

1. 1. The method for constructing the acute obstruction suppurative cholangitis conservation treatment failure risk prediction model is characterized by comprising the following steps: Step S1, collecting clinical and laboratory index data of a plurality of cases of acute obstructive suppurative cholangitis patients meeting the nano-array standard, preprocessing the data, and dividing the data into a training set and a testing set; step S2, carrying out balancing on training set data by adopting a ROSE algorithm, and then carrying out feature screening by sequentially utilizing Spearman correlation analysis, single-factor ROC analysis and LASSO regression, wherein the finally screened features comprise TBIL, LDH, IL-1 beta, IL-18, stone diameter and common bile duct diameter; S3, taking whether the patient aggravates within 24 hours as a dependent variable, taking the characteristics as the independent variable, establishing a prediction model by adopting multivariate logistic regression, and optimizing model parameters through 10-fold cross verification, wherein the model formula is as follows: ; wherein P represents the probability of patient exacerbation within 24 hours; and S4, performing model performance verification evaluation based on the test set.
2. 2. The method for constructing a model for predicting risk of failure in conservative treatment of acute obstructive suppurative cholangitis according to claim 1, wherein in step S3, the model formula is: ; Wherein, the unit of TBIL index is mu mol/L, the unit of LDH index is U/L, the unit of IL-1 beta index is pg/mL, the unit of IL-18 index is pg/mL, the diameter of stone is the maximum diameter of stone, the unit of stone is mm, and the unit of CBD diameter is mm.
3. 3. The method for constructing a model for predicting risk of failed conservative treatment of acute obstructive suppurative cholangitis according to claim 2, wherein in step S3, when P is not less than 0.23, it is determined that the patient has a higher risk of failed conservative treatment for 24 hours.
4. 4. The method for constructing a predictive model of risk of failure in conservative treatment of acute obstructive suppurative cholangitis according to claim 1, further comprising: And S5, constructing a nomogram, and further verifying a prediction conclusion of the acute obstructive suppurative cholangitis conservation treatment failure risk prediction model through the nomogram.
5. 5. The utility model provides a acute obstruction suppurative cholangitis conservative treatment failure risk prediction system which is characterized by comprising: the index value input module is used for obtaining the values of TBIL, LDH, IL-1 beta, IL-18, calculus diameter and CBD diameter of the current patient; A model prediction module, configured to perform probability prediction by using the acute obstructive suppurative cholangitis conservative treatment failure risk prediction model obtained by the construction method according to any one of claims 1 to 3; and the result output module is used for outputting the probability value of the current patient suffering from the exacerbation within 24 hours and the corresponding patient 24 hours conservation treatment failure risk and recommended treatment strategy.
6. 6. The system for predicting risk of failure in conservative treatment of acute obstructive suppurative cholangitis according to claim 5, further comprising a nomogram module for automatically calculating a total score and a corresponding probability value of exacerbation of the current patient within 24 hours based on the values of TBIL, LDH, IL-1β, IL-18, stone diameter, and CBD diameter of the current patient and the nomogram.

Description

Construction method and system of acute obstruction suppurative cholangitis conservation treatment failure risk prediction model Technical Field The invention relates to the technical field of model construction, in particular to a construction method and a construction system of a risk prediction model for conservative treatment failure of acute obstructive suppurative cholangitis. Background Acute obstructive suppurative cholangitis (acute obstructive suppurative cholangitis, AOSC) is a serious acute abdomen caused by biliary tract obstruction and bacterial infection, and if drainage or operation treatment is not performed in time, sepsis is often rapidly progressed to multiple organ failure, and the death rate is extremely high. Continuous updating of Tokyo Guidelines (TG 07, TG13, TG 18) provides international standards for diagnosis and classification of AOSC, but Guidelines lack specific standards on "quantitative determination of initial treatment failure in moderate (Grade II) patients," only propose "if no response to initial treatment should be followed early on for drainage or surgery". This ambiguous definition leads to practical difficulties for the clinician to accurately determine patient condition trends at the time of admission. In large three-level hospitals, teams with ERCP or Laparoscopic Common Bile Duct Exploration (LCBDE) experience can conduct minimally invasive biliary tract drainage or lithotomy at any time, but in most primary hospitals, the number of liver and gall surgeons who can master minimally invasive techniques is limited due to equipment conditions and personnel training level, and in emergency treatment, only anti-infection, fluid infusion and other conservative treatments can be performed in advance. How to determine whether the patient is suffering from an exacerbation within 24 hours from the initial stage of the patient admission becomes critical in determining whether an immediate referral or initiation of an emergency intervention is required. If the patient with successful conservative treatment can be identified, not only can the expert with enough time schedule be strived for carrying out the period selection minimally invasive treatment, but also the satisfaction degree of the patient can be obviously improved, the hospitalization time can be shortened and the treatment cost can be reduced, otherwise, if the patient with high risk can be identified as early as possible, the sepsis progress caused by delay can be avoided. In conclusion, actual medical resources are uneven, and emergency minimally invasive treatment capability of primary hospitals is limited. Common means currently used to determine AOSC early stage disease changes include: The traditional clinical symptoms, such as Charcot's triple symptoms (fever, jaundice, right upper abdominal pain), are only about 50-70% sensitive, and insufficient to identify early-stage high-risk patients 。 The tokyo guideline (Tokyo Guideline 2018, TG 18) hierarchy is a hierarchy that relies on body temperature, white blood cells, bilirubin, etc. for stratification, but where "initial treatment no response" lacks quantification criteria, making rapid decisions difficult at admission. A single laboratory index, such as PCT, CRP, ALT, reflects inflammation, but has limited predictive efficiency and does not form a composite discriminating model. The existing model has insufficient precision, and the retrospective analysis of AOSC patients in the same hospital for the past 10 years shows that the ROC AUC of TG18 for predicting the exacerbation of the illness within 24 hours is only about 0.74, and the clinical distinction is limited. Furthermore, most of the previous studies focused on the treatment strategy for severe (Grade III) patients, whereas the prediction of early course of disease for patients with intermediate AOSC lacks a systematic model. Particularly in basic medical institutions, accurate treatment is difficult to achieve if blindly delayed drainage may lead to rapid exacerbation of the condition, and if too early surgery adds unnecessary trauma and costs. 1.Wada K, Takada T, Kawarada Y, et al. Diagnostic criteria and severity assessment of acute cholangitis: Tokyo Guidelines. J Hepatobiliary Pancreat Surg. 2007;14(1):52-8. doi:10.1007/s00534-006-1156-7. 2.Kiriyama S, Takada T, Strasberg SM, et al. TG13 guidelines for diagnosis and severity grading of acute cholangitis (with videos). J Hepatobiliary Pancreat Sci. Jan 2013;20(1):24-34. doi:10.1007/s00534-012-0561-3. 3.Kiriyama S, Kozaka K, Takada T, et al. Tokyo Guidelines 2018: diagnostic criteria and severity grading of acute cholangitis (with videos). J Hepatobiliary Pancreat Sci. Jan 2018;25(1):17-30. doi:10.1002/jhbp.512. 4.Yokoe M, Takada T, Mayumi T, et al. Accuracy of the Tokyo Guidelines for the diagnosis of acute cholangitis and cholecystitis taking into consideration the clinical practice pattern in Japan. J Hepatobiliary Pancreat Sci. Mar 2011;18(2):