Search

CN-121983333-A - Continuous medical index and clinical outcome optimal association boundary value determination model and application

CN121983333ACN 121983333 ACN121983333 ACN 121983333ACN-121983333-A

Abstract

The patent relates to the technical field of medical statistics and clinical information processing, in particular to a data driving model for determining an optimal association boundary value between a continuous medical index and a classification clinical outcome and application thereof. The invention is driven by data, avoids bias of subjective experience judgment, and automatically and objectively searches the most obvious statistical critical value associated with clinical outcome. And traversing a large number of candidate critical values, combining with a restrictive cubic spline to perform nonlinear fitting so as to accurately position the lowest point of the p value on the fitting curve, assisting the stacking histogram to intuitively present the ending distribution differences corresponding to different thresholds, and finally determining the optimal association critical value, wherein the precision is obviously higher than that of the traditional grouping method. The method can be applied to the association analysis of other continuous medical indexes (such as CD4 count, viral load, biochemical indexes and the like) and various two-class clinical outcomes (such as survival states, treatment reactions, complications and the like) without depending on specific diseases or indexes. The method can integrate visual chart output, so that the result is more visual, and is convenient for clinicians to understand and apply.

Inventors

  • Mo Minghua
  • LI YUWEI
  • LI WEIMIN
  • WANG YAN

Assignees

  • 首都医科大学附属北京胸科医院

Dates

Publication Date
20260505
Application Date
20260104

Claims (7)

  1. 1. A method for automatically and accurately determining a CD4 + T lymphocyte count threshold for an HIV combined TBM prognosis, the method comprising the steps of: s1, inputting a baseline CD4 count and two kinds of classification ending data of a patient to form an initial data set; s2, preprocessing data, namely checking the distribution of continuity independent variables, namely observing the distribution state of the data, determining a method for removing extreme values, and removing the extreme values; s3, generating candidate critical values, and systematically generating a series of candidate critical values by taking every 0.5 unit of the data processed in the S2 as a step length; s4, traversing and calculating a relevance p value, namely dividing patients into two groups with a CD4 greater than or equal to a threshold value and a CD4 smaller than the threshold value by taking each candidate critical value as a limit, and constructing a 2x2 list linkage table of the group and the ending; s5, carrying out nonlinear fitting on the trend of the p value along with the change of the threshold value based on a restrictive cubic spline (RESTRICTED CUBIC SPLINE, RCS), wherein each p value is represented in a scattered point form in a graph, and an RCS curve is drawn; S6, drawing a stacking histogram, displaying the ending distribution differences corresponding to different thresholds based on the stacking histogram, further determining an optimal association boundary value, and outputting the optimal association value to obtain a result.
  2. 2. The method of claim 1, wherein in step S4, the test is selected from a chi-square test or a Fisher exact test, wherein the Fisher exact test is used when any desired frequency in the list is less than 5, otherwise the chi-square test is used, and the generated series of consecutive thresholds are tested one by one to improve the accuracy of predicting the best associated threshold.
  3. 3. The method of claim 1, wherein in step S6, the optimal threshold is a point on the RCS fitting curve where the p-value is globally minimum.
  4. 4. The method of claim 1, wherein all analyses are performed in R software, the plots are plotted using ggplot packages, the data sets are processed using dplyr and tidyr packages, and the restricted cubic spline fitting is performed using splines packages.
  5. 5. A system for automatically and accurately determining a CD4 + T lymphocyte count threshold for an HIV combined TBM prognosis, the system comprising the following modules: and a data input module: Inputting baseline CD4 counts and the two-class outcome data of the patient to form an initial dataset; the data preprocessing module is used for checking the continuity independent variable distribution, observing the data distribution state, determining a method for removing extreme values, and removing initial data An extreme value; and a data processing module: Systematically generating a series of candidate critical values by taking each 0.5 unit of preprocessed data as a step length; Dividing patients into two groups with CD4 greater than or equal to a threshold value and CD4 less than the threshold value by taking each candidate critical value as a limit, and constructing a 2x2 list of the group and the death outcome of one year; non-linear fitting of the trend of p values with threshold values based on a restrictive cubic spline (RESTRICTED CUBIC SPLINE, RCS) is performed by representing each p value in a scattered form in a graph and drawing an RCS curve thereof; displaying the ending distribution difference corresponding to different thresholds based on the stacking histogram, and further determining an optimal association boundary value; And the data output module is used for: And outputting the optimal association boundary value.
  6. 6. The system of claim 5, wherein the data processing module is configured to perform a test selected from the group consisting of chi-square test and Fisher's exact test, to use Fisher's exact test when any desired frequency in the list is less than 5, to use chi-square test otherwise, to perform a test on the generated series of consecutive thresholds, to improve the accuracy of predicting the optimal correlation threshold.
  7. 7. The system of claim 5, wherein the optimal correlation threshold is a point on the RCS fit curve where the p-value is globally minimum.

Description

Continuous medical index and clinical outcome optimal association boundary value determination model and application Technical Field The patent relates to the technical field of medical statistics and clinical information processing, in particular to a data driving model for determining an optimal association boundary value between a continuous medical index and a classification clinical outcome and application thereof. Background Tuberculosis (tuberculous meningitis, TBM) is a central nervous system tuberculosis （Sanjay K Jain, David M Tobin, Elizabeth W Tucker, et al.Tuberculous eningitis: a roadmap for advancing basic and translational research. Nat Immunol. 2018;19(6):521-525.）, which is mainly caused by the invasion of primary pulmonary mycobacterium tuberculosis into subarachnoid space through blood circulation or direct route to cause hidden infection of meninges, is the most serious manifestation of the mycobacterium tuberculosis infection and is the tuberculosis type with the highest death rate. In 2019, about 164000 adults have TBM in the tuberculosis cases diagnosed in the global 710 million new, wherein about one quarter of cases occur in HIV infected persons, （Dodd PJ, Osman M, Cresswell FV, et al. The global burden of tuberculous meningitis in adults: A modelling study. PLOS Glob Public Health. 2021;1(12):e0000069. Published 2021 Dec 8.）. co-infected persons often face worse outcome （Boonyagars L, Sangketchon C, Pholtawornkulchai K. Presentation, Clinical Characteristics, and Treatment Outcomes among Tuberculous Meningitis Patients with and Without HIV Infection at Vajira Hospital, Thailand: A Retrospective Cohort Study. J Int Assoc Provid AIDS Care. 2021;20:23259582211045551. ;Thwaites G E, Bang N D, Dung N H, et al. Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults[J]. New England Journal of Medicine, 2004, 351(17): 1741-1751.）. although the overall mortality rate of TBM is about 25%, whereas the mortality rate of HIV co-infected persons increases to nearly 50%（Stadelman AM, Ellis J, Samuels THA, et al. Treatment Outcomes in Adult Tuberculous Meningitis: A Systematic Review and Meta-analysis. Open Forum Infect Dis. 2020;7(8):ofaa257.）. severe immunosuppression, particularly depletion of CD4 + T lymphocytes, is a key driver of the sign of HIV infection and tuberculosis susceptibility. A25 times （Ellis PK, Martin WJ, Dodd PJ. CD4 count and tuberculosis risk in HIV-positive adults not on ART: a systematic review and meta-analysis. PeerJ. 2017;5:e4165. Published 2017 Dec 14.）. Vietnam study of 70 HIV-infected persons with combined tubercular meningitis with a CD4 count of less than 200/μl at risk of developing tuberculosis is a study of the 1000 CD4 count/μl population showing that the median CD4 count is only 67/μl (IQR 19-124), which highlights the level of immunodeficiency in this population （Hai HT, Thanh Hoang Nhat L, Tram TTB, et al. Whole blood transcriptional profiles and the pathogenesis of tuberculous meningitis. Elife. 2024;13:RP92344. Published 2024 Oct 30.）. The continuous medical index CD4+ T lymphocyte count is a key contributor to the outcome of a broad spectrum of clinical outcomes (e.g., death/survival, improvement/worsening). Determining an optimal threshold has central value for disease stratification, prognosis, treatment decision-making and guideline formulation. Currently, methods for determining such thresholds are mostly dependent on clinical experience, reference to past literature, or based on statistical distributions. However, these methods have the problems of strong subjectivity, failure to fully utilize the association information of the data itself and the end, failure to accurately find the point of highest statistical association, and the like. The world health organization defines a CD4 count < 50/μl as "advanced HIV disease", a central feature of this stage is immune system breakdown with almost complete loss of defenses, leading to a series of extremely serious consequences （Guidelines for Managing Advanced HIV Disease and Rapid Initiation of Antiretroviral Therapy. Geneva: World Health Organization; 2017.）. with studies indicating that CD4 counts are low, more likely to face poor outcome risk （Loghin II, Vâță A, Miftode EG, et al. Characteristics of Tuberculous Meningitis in HIV-Positive Patients from Northeast Romania. Clin Pract. 2023;13(6):1488-1500. Published 2023 Nov 21.）., yet if the above-mentioned threshold is the one most strongly associated with HIV combined TBM outcome, still requiring greater data volume verification. The prior art lacks an automatic, objective and universal calculation method capable of accurately capturing the optimal correlation critical value of continuous medical indexes and two kinds of clinical outcomes, so that disease risk stratification is inaccurate, and subjective deviation exists in treatment decisions. In view of the foregoing, there is a need for a patent application that automat