CN-121978261-A - Intelligent evaluation method and system for medical equipment extract

CN121978261ACN 121978261 ACN121978261 ACN 121978261ACN-121978261-A

Abstract

The invention provides an intelligent evaluation method and system for medical equipment extracts, which are applied to the technical field of chemical analysis, wherein an adaptive standardized kit is used for carrying out standardized extraction on the medical equipment extracts, the standardized kit comprises a standard solvent, a monitoring standard sample and an extraction container, an extract sample is analyzed, mass spectrograms and chromatograms of all components are collected in real time and raw data are input into an intelligent evaluation platform, the intelligent evaluation platform automatically extracts spectral characteristics and compares the spectral characteristics with a preset database to identify known chemical components, toxicity and biocompatibility evaluation is carried out on each chemical component, the change of the monitoring standard sample before and after extraction is analyzed, whether the extraction process is stable or not is evaluated, and a report is automatically generated if the extraction process is stable. The evaluation method and the evaluation system not only improve the accuracy and the reliability of the extraction evaluation process, but also solve the problems of inconsistent extraction, difficult component identification, low data processing efficiency, insufficient toxicity evaluation and the like in the prior art.

Inventors

LIN CHUNXIN
HUANG PING
ZHANG LILI
BAI YANG

Assignees

南京明捷生物医药检测有限公司

Dates

Publication Date: 20260505
Application Date: 20260108

Claims (10)

1. An intelligent evaluation method of medical equipment extract is characterized by comprising the following steps: performing standardized extraction on the medical equipment extract under specific extraction conditions by using an adaptive standardized kit, wherein the standardized kit comprises a gradient polar standard solvent, a monitoring standard sample with chemical stability and a low-adsorption extraction container; the extract sample is sent to a chromatographic mass spectrometry system for analysis, mass spectrograms and chromatograms of all components are collected in real time and stored as original data, and the original data are input to an intelligent evaluation platform; the intelligent evaluation platform pre-processes the original data and automatically extracts spectral features through a deep learning algorithm; Comparing the spectral characteristics of the extract with a preset database to identify known chemical components; Based on the identified chemical components, evaluating toxicity and biocompatibility of each chemical component through a toxicity analysis model, and simultaneously, analyzing and monitoring the change of the standard sample before and after extraction to evaluate whether the extraction process is stable or not; If the extraction process is determined to be stable, automatically generating a report based on the evaluation result, wherein the report comprises chemical components, risk evaluation, toxicity level and optimization suggestion.
2. The intelligent evaluation method of medical equipment extract according to claim 1, wherein the gradient polar standard solvent comprises a low-polarity solvent, a medium-polarity solvent and a high-polarity solvent, and a certain proportion of the low-polarity solvent, the medium-polarity solvent and the high-polarity solvent mixed solvent is set for extraction of a medical equipment extract sample according to the type of the medical equipment extract.
3. The intelligent evaluation method of medical device extract according to claim 1, wherein the deep learning algorithm is a convolutional neural network, and the process of obtaining spectral features through the convolutional neural network comprises the following steps: Preprocessing and normalizing the original data, and converting the normalized original data into a two-dimensional matrix suitable for convolutional neural network input, wherein a first dimension represents a mass-to-charge ratio, and a second dimension represents a sample dimension and/or a time dimension; Inputting the two-dimensional matrix into a convolutional neural network, and performing convolutional operation on the two-dimensional matrix through at least one convolutional layer to generate a feature map; performing a multi-level convolution operation on the feature map to extract multi-level features from low level to high level; performing a pooling operation after the convolution layer to reduce the dimension of the feature map; flattening the pooled features, and adopting a Lasso algorithm to perform feature selection on the flattened features so as to remove redundant features; and inputting the features subjected to feature selection into a full-connection layer, processing the features by an activation function, and outputting feature vectors for subsequent prediction or classification tasks.
4. The method of claim 3, wherein the Lasso algorithm performs feature screening by compressing the weights of uncorrelated features to zero and outputs a subset of features that are correlated with subsequent chemical component classifications or toxicity predictions.
5. The method of claim 3, wherein the feature vector output by the convolutional neural network is a spectral feature comprising at least peak intensity, retention time, mass-to-charge ratio.
6. The intelligent evaluation method of medical device extract according to claim 1, wherein the database stores standard spectra and chemical components corresponding to the standard spectra data, and the similarity between the extract sample and the standard spectra in the database is calculated by a matching algorithm, so as to identify the corresponding known chemical components.
7. The method of claim 6, wherein the matching algorithm comprises one or more of pearson correlation coefficient algorithm, cosine similarity algorithm, and dynamic time warping algorithm.
8. The method of claim 1, wherein the toxicity analysis model comprises a quantitative structure-activity relationship model for predicting biological activity of the chemical component according to the chemical structure and a threshold toxicity concentration model for judging safety of the chemical component according to a set safety threshold.
9. The method for intelligent evaluation of medical device extracts according to any one of claims 1 to 8, wherein in the chemical component identification process, deep learning analysis is performed on incompletely matched or unknown chemical components by using a graph neural network, specifically comprising the following steps: When the matching degree of the spectral characteristics of the extract sample and a preset chemical component database does not meet a preset threshold value or unknown chemical components which are not recognized exist, constructing chemical components to be recognized into graph structure data, inputting the graph structure data into a graph neural network for deep learning analysis, wherein chemical molecules or components are used as graph nodes, chemical bonds or interrelationships among molecules are used as graph edges, and at least one of mass-to-charge ratios, molecular weights or peak intensities is used as node characteristics; And carrying out weighted aggregation on adjacent node characteristics through a graph convolution layer, iteratively updating node representation, inputting the updated node characteristic vector to a full-connection layer, carrying out characteristic fusion and optimization, and outputting at least one of classification results, concentration predictions or risk scores of the unknown chemical components.
10. An intelligent assessment system for medical device extracts, comprising: the standardized kit comprises a gradient polar standard solvent, a monitoring standard sample with chemical stability and a low-adsorption extraction container, and is used for carrying out standardized extraction on medical equipment extracts under specific extraction conditions; the non-target analysis module is used for analyzing the extract sample through a chromatographic mass spectrometry system, collecting mass spectrograms and chromatograms of all components in real time, storing the mass spectrograms and chromatograms as original data, and inputting the original data into the intelligent evaluation platform; The intelligent evaluation platform comprises: the data processing module is used for preprocessing the original data and automatically extracting spectral features through a deep learning algorithm; The data comparison and identification module is used for comparing the spectral characteristics of the extract with a preset database and identifying known chemical components; the risk evaluation module is used for evaluating the toxicity and biocompatibility of each chemical component through a toxicity analysis model, and simultaneously analyzing and monitoring the change of the standard sample before and after extraction to evaluate whether the extraction process is stable or not; And the report generation module is used for automatically generating a report according to the analysis result, wherein the report comprises chemical components, risk assessment, toxicity level and optimization advice.

Description

Intelligent evaluation method and system for medical equipment extract Technical Field The invention belongs to the technical field of chemical analysis, and particularly relates to an intelligent evaluation method and system for medical equipment extracts. Background With advances in medical technology and widespread use of medical devices, more and more medical device materials (e.g., polymers, metals, coatings, etc.) enter or come into contact with the human body. Thus, ensuring the safety and biocompatibility of these materials is a critical issue, especially if their extracts constitute a potential risk to human health. The extract refers to components which may be dissolved or released from the medical device material after the medical device material is contacted with a solvent under specific conditions, and the components may have adverse effects on human bodies, such as allergic reactions, toxic effects, immune reactions, and the like. Currently, evaluating the safety of medical device extracts generally relies on standardized extraction procedures and traditional experimental methods, such as chemical component analysis, cell culture toxicity testing, and the like. However, these conventional methods have certain limitations, mainly in terms of the following: inconsistencies in the extraction process-operating differences in the extraction process may lead to inconsistencies in the composition of the extract, affecting the accuracy and reproducibility of the analysis results, under different laboratories or under different experimental conditions. The complexity of analysis means is that the existing analysis method generally requires complex sample pretreatment and long-time analysis process, and is dependent on manual operation and is easily affected by human errors. Detection of non-target components is difficult-most existing analytical methods favor target analysis, detect for known chemical components, and present a significant challenge for multi-component analysis of unknown components or complex samples. Lack of risk assessment although component analysis is possible, there is also a lack of uniform and standardized technical solutions for toxicity and biocompatibility assessment of extracts, especially comprehensive toxicity assessment based on chemical structure and concentration. Therefore, a more standardized, efficient and intelligent evaluation method is urgently needed in the current technical scheme, and the problems can be solved, so that the safety and biocompatibility of medical equipment materials are guaranteed at a higher level. Disclosure of Invention In view of the problems in the prior art, the invention aims to provide an intelligent evaluation method of medical equipment extracts, which not only improves the accuracy and reliability of the extraction evaluation process, but also solves the problems of inconsistent extraction, difficult component identification, low data processing efficiency, insufficient toxicity evaluation and the like in the prior art by innovative means such as standardized extraction, intelligent analysis, deep learning, automatic report generation and the like. An intelligent evaluation method of medical equipment extracts, comprising the following steps: performing standardized extraction on the medical equipment extract under specific extraction conditions by using an adaptive standardized kit, wherein the standardized kit comprises a gradient polar standard solvent, a monitoring standard sample with chemical stability and a low-adsorption extraction container; the extract sample is sent to a chromatographic mass spectrometry system for analysis, mass spectrograms and chromatograms of all components are collected in real time and stored as original data, and the original data are input to an intelligent evaluation platform; the intelligent evaluation platform pre-processes the original data and automatically extracts spectral features through a deep learning algorithm; Comparing the spectral characteristics of the extract with a preset database to identify known chemical components; Based on the identified chemical components, evaluating toxicity and biocompatibility of each chemical component through a toxicity analysis model, and simultaneously, analyzing and monitoring the change of the standard sample before and after extraction to evaluate whether the extraction process is stable or not; If the extraction process is determined to be stable, automatically generating a report based on the evaluation result, wherein the report comprises chemical components, risk evaluation, toxicity level and optimization suggestion. Preferably, the gradient polar standard solvent comprises low polar solvent, medium polar solvent and high polar solvent, and a certain proportion of low polar solvent, medium polar solvent and high polar solvent mixed solvent is set for extracting the medical device extract sample according to the type of the medical device extra