Chemical property analysis method for drug quality consistency evaluation

Abstract

The application relates to the technical field of medicine detection, and discloses a chemical property analysis method for medicine quality consistency evaluation, which comprises the following steps: adjusting fluid physicochemical parameters in a controlled flow field to induce a transient dissociation response of medicine, collecting absorbance data sequences and complex impedance signal sequences, mapping first-order differential values of the absorbance data and complex impedance phase angles to a two-dimensional physicochemical parameter distribution domain, thereby extracting a physicochemical response distribution point set, and converting the point set into state nodes according to grid division rules, and then statistically counting migration frequencies and directions between the nodes to construct a physicochemical state transition map. Through topological characterization of a dynamic evolution process of physicochemical properties of the medicine, the application can accurately capture an evolution path of a microstate, effectively solves a discrimination risk caused by feature collapse, and realizes high-fidelity evaluation of quality consistency of complex preparations.

Description

Technical Field

[0001] This invention belongs to the field of pharmaceutical testing technology, and in particular relates to a chemical property analysis method for evaluating the consistency of pharmaceutical quality. Background Technology

[0002] Current drug quality consistency evaluation involves the determination of the physicochemical characteristics of drugs in simulated physiological environments. Conventional analytical methods use timed sampling under constant medium conditions to collect the cumulative drug release concentration at characteristic time points and use statistical parameters such as similarity factors to evaluate the equivalence between different formulations. This approach is applicable when dealing with formulation systems in quasi-static equilibrium and provides a basic evaluation criterion for standardized quality control of drugs. Oral solid dosage forms face changes in the physicochemical environment during gastrointestinal transport, involving changes in pH gradients, fluctuations in ionic strength, and changes in fluid shear forces. When the polymeric excipient system undergoes abrupt changes in the above boundary conditions, it produces microscopic responses such as local supersaturation, metastable crystal transformation, or gel layer erosion, which directly determine the transient effective concentration of the active ingredient at key absorption sites. Existing analytical methods set static buffers as the test environment and have a low measurement frequency, resulting in the feedback physical quantity being limited to the time-series accumulation of the overall concentration. This evaluation model loses the transient characteristics that reflect the true interaction mechanism between excipients and active ingredients.

[0003] The physical structure of detection equipment has limitations, and the characterization logic and analysis methods for complex formulation release processes are insufficient. For example, Chinese invention patent application CN116525137A discloses a method for evaluating the quality consistency of traditional Chinese medicine compound preparations. By establishing a biomimetic drug absorption system combined with a real-time label-free dynamic cell analysis system, the consistency of drug efficacy is evaluated using the cell index evolution curve over time. This scheme achieves overall evaluation from the perspective of biological effects. However, in the underlying engineering logic, cell response is a secondary signal. Due to the physiological inertia and signal hysteresis effect of the biological probe, the sampling resolution cannot map the millisecond-level transient dissociation action of the preparation. The biological evaluation system cannot decouple the interference of complex matrix physical scattering and it is difficult to restore the microscopic evolution trajectory of the preparation in phase space. When evaluating chemical preparations with complex release characteristics, there is a fundamental feature perception defect.

[0004] Therefore, how to capture and analyze the transient physicochemical evolution trajectory under controlled perturbation, and construct a dynamic evaluation mechanism with topological fidelity and retrieval capability, has become the technical problem to be solved by this invention. Summary of the Invention

[0005] This invention provides a chemical property analysis method for drug quality consistency evaluation, comprising the following steps:

[0006] Step S1: In a closed and controlled flow field, the pH value, ionic strength and surfactant concentration of the analytical fluid are adjusted according to a preset time sequence to induce the drug to be tested to produce a transient dissociation response that changes with the fluid environment.

[0007] Step S2: The absorbance data sequence and complex impedance signal sequence characterizing the drug under test in the transient dissociation response are acquired by the detection unit composed of an ultraviolet absorbance sensor and an impedance sensor.

[0008] Step S3: Extract the first-order difference value from the absorbance data sequence, and map the first-order difference value and the complex impedance phase angle in the complex impedance signal sequence as coordinate components to the two-dimensional physicochemical parameter distribution domain, and extract the physicochemical response distribution point set composed of the correlation points of the first-order difference value and the complex impedance phase angle at the same sampling time.

[0009] Step S4: Divide the two-dimensional physicochemical parameter distribution domain into multiple discrete physicochemical state regions according to the preset grid division rules, and define the physicochemical state region into which the physicochemical response distribution point set falls as a state node characterizing a specific chemical state.

[0010] Step S5: Statistically analyze the migration direction and frequency between state nodes at adjacent time points, and construct a physicochemical state transition map that characterizes the dynamic evolution topological features of the chemical properties of the drug under test. This allows for in-depth evaluation of drug quality through consistency comparison of the map structure.

[0011] Preferably, the method further includes the following steps: Step S6, according to a preset topology serialization rule, converting the grid coordinate attributes of each state node in the physicochemical state transition map and the migration frequency information between adjacent nodes into a character sequence with a specific encoding, generating a kinetic trajectory feature string to characterize the drug dissolution kinetics; Step S7, using the kinetic trajectory feature string as a retrieval key, matching in a preset standard drug fingerprint database through an inverted index mechanism to retrieve a subset of reference maps corresponding to the drug to be tested; Step S8, calculating the subgraph isomorphic matching degree between the physicochemical state transition map and each reference map in the subset of reference maps, and outputting the chemical property consistency evaluation result between the drug to be tested and the reference preparation based on the subgraph isomorphic matching degree.

[0012] Preferably, in step S1, acidic buffer solution, alkaline buffer solution, and simulated gastrointestinal fluid are alternately injected into the closed controlled flow field at a preset frequency using a high-precision plunger pump group to construct a perturbed flow field with a dynamic concentration gradient.

[0013] Preferably, in step S2, the sampling frequency of the signal acquired by the detection unit is not less than 100Hz; the complex impedance signal sequence includes a complex impedance phase angle sequence and a complex impedance amplitude sequence that change with time.

[0014] Preferably, the physicochemical state transition map includes state switching paths composed of directed arcs and a weight matrix for storing the dwell time of the physicochemical response distribution point set in each discrete physicochemical state region.

[0015] Preferably, in step S6, the topology serialization rule includes: extracting the maximally connected components in the physicochemical state transition map, and using a depth-first traversal algorithm to extract the key evolution paths in the maximally connected components, and mapping the coordinate components of each state node on the key evolution path to feature characters.

[0016] Preferably, in step S7, the inverted indexing mechanism includes: splitting the dynamic trajectory feature string into multiple fixed-length feature segments, retrieving the standard map containing the most feature segments from the standard drug fingerprint database, and using it as a member of the reference map subset.

[0017] Preferably, in step S8, the calculation process of the subgraph isomorphic matching degree includes: extracting the topological feature vector of the physicochemical state transition map, and calculating the Euclidean distance between the topological feature vector and the reference feature vector of the reference map.

[0018] Preferably, in step S8, when outputting the consistency evaluation results, if the subgraph isomorphic matching degree is higher than the preset similarity threshold, it is determined that the in vitro dissociation behavior of the drug under test is consistent with that of the reference preparation.

[0019] Compared with existing technologies, the chemical property analysis method for drug quality consistency evaluation of the present invention has the following advantages:

[0020] 1. In the chemical property analysis of drug quality consistency evaluation, by implementing chemical microenvironment pulse perturbation during the characteristic period of drug release and simultaneously collecting transient absorbance and transient conductivity signals in the detection area, a kinetic response link characterizing the unsteady evolution of the test system is constructed. This mechanism elevates the evaluation dimension of drug physicochemical properties from the traditional static endpoint concentration or smooth cumulative release curve to a nonlinear relaxation trajectory reflecting the microscopic phase transition process. By artificially triggering the instantaneous imbalance of the dissolution system, the interaction mechanism between the active ingredient and excipients is made explicit at the physical response level. This eliminates the microscopic release defects masked by the similarity of overall dissolution behavior at the measurement principle level, and improves the accuracy of in vitro testing in predicting in vivo bioequivalence.

[0021] 2. Utilizing the physical complementarity between transient conductivity and transient absorbance signals, an intrinsic signal decoupling mechanism is established to address interference from complex excipient matrices. Under conditions where background scattering fluctuations are caused by formulation disintegration, the high-frequency component in the transient conductivity signal accurately maps the spatial occupancy frequency of insoluble particles. By converting the high-frequency component into a dynamic scattering compensation factor and applying it to the transient absorbance signal, in-situ cancellation of optical interference is achieved. This processing path ensures that the measurement focus always converges to the active molecules in a true solution state, thereby improving the fidelity of the determination of the chemical component concentration evolution sequence in complex formulation systems without physically filtering the analytical fluid.

[0022] 3. By mapping the acquired dual-channel time series to a two-dimensional phase space and constructing a state transition map based on gridded clustering logic, a high-fidelity topological representation of the chemical kinetic process is achieved. This method abandons the conventional approach of forcibly reducing nonlinear signals to a single scalar constant, and instead extracts a topological data structure composed of steady-state nodes and state transition edges. This shift from vector comparison to graph isomorphic comparison enables the evaluation system to identify different metastable evolution paths, intercepts the risk of heterogeneous isomorphic discrimination caused by feature collapse at the data structure level, and establishes a data firewall for the rigorous evaluation of drug micro-quality. Attached Figure Description

[0023] Figure 1 This is a flowchart of the chemical property analysis method for drug quality consistency evaluation according to the present invention;

[0024] Figure 2 This is a logic block diagram for constructing the physicochemical state transition map and serializing the dynamic trajectory in this invention. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0026] It should be noted that all directional and positional terms used in this invention, such as: up, down, left, right, front, back, vertical, horizontal, inner, outer, top, bottom, transverse, longitudinal, center, etc., are only used to explain the relative positional relationship and connection between components in a specific state (as shown in the accompanying drawings). They are only for the convenience of describing this invention and do not require that this invention be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention. In addition, the descriptions of "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated.

[0027] In the description of this invention, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections; they can refer to direct connections or indirect connections through an intermediate medium; they can refer to the internal connection of two components. For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances.

[0028] In the description of this specification, references to the terms "an embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example, and the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0029] A chemical property analysis method for drug quality consistency evaluation includes the following steps:

[0030] Step S1: In a closed and controlled flow field, the pH value, ionic strength and surfactant concentration of the analytical fluid are adjusted according to a preset time sequence to induce the drug to be tested to produce a transient dissociation response that changes with the fluid environment.

[0031] Step S2: The absorbance data sequence and complex impedance signal sequence characterizing the drug under test in the transient dissociation response are acquired by the detection unit composed of an ultraviolet absorbance sensor and an impedance sensor.

[0032] Step S3: Extract the first-order difference value from the absorbance data sequence, and map the first-order difference value and the complex impedance phase angle in the complex impedance signal sequence as coordinate components to the two-dimensional physicochemical parameter distribution domain, and extract the physicochemical response distribution point set composed of the correlation points of the first-order difference value and the complex impedance phase angle at the same sampling time.

[0033] Step S4: Divide the two-dimensional physicochemical parameter distribution domain into multiple discrete physicochemical state regions according to the preset grid division rules, and define the physicochemical state region into which the physicochemical response distribution point set falls as a state node characterizing a specific chemical state.

[0034] Step S5: Statistically analyze the migration direction and frequency between state nodes at adjacent time points, and construct a physicochemical state transition map that characterizes the dynamic evolution topological features of the chemical properties of the drug under test. This allows for in-depth evaluation of drug quality through consistency comparison of the map structure.

[0035] Preferably, the method further includes the following steps: Step S6, according to a preset topology serialization rule, converting the grid coordinate attributes of each state node in the physicochemical state transition map and the migration frequency information between adjacent nodes into a character sequence with a specific encoding, generating a kinetic trajectory feature string to characterize the drug dissolution kinetics; Step S7, using the kinetic trajectory feature string as a retrieval key, matching in a preset standard drug fingerprint database through an inverted index mechanism to retrieve a subset of reference maps corresponding to the drug to be tested; Step S8, calculating the subgraph isomorphic matching degree between the physicochemical state transition map and each reference map in the subset of reference maps, and outputting the chemical property consistency evaluation result between the drug to be tested and the reference preparation based on the subgraph isomorphic matching degree.

[0036] Preferably, in step S1, acidic buffer solution, alkaline buffer solution, and simulated gastrointestinal fluid are alternately injected into the closed controlled flow field at a preset frequency using a high-precision plunger pump group to construct a perturbed flow field with a dynamic concentration gradient.

[0037] Preferably, in step S2, the sampling frequency of the signal acquired by the detection unit is not less than 100Hz; the complex impedance signal sequence includes a complex impedance phase angle sequence and a complex impedance amplitude sequence that change with time.

[0038] Preferably, in step S4, the region index value of the state node to which the first-order difference value and the complex impedance phase angle belong at sampling time t is calculated using the following formula. : ,in, Let be the magnitude of the first-order difference at time t. The phase angle of the complex impedance at time t amplitude, and These are the minimum values ​​of the first-order difference value and the complex impedance phase angle within the preset monitoring window, respectively. L is the grid side length of the physicochemical state region, and K is the total number of partitions in the two-dimensional physicochemical parameter distribution domain along the horizontal axis.

[0039] Preferably, the physicochemical state transition map includes state switching paths composed of directed arcs and a weight matrix for storing the dwell time of the physicochemical response distribution point set in each discrete physicochemical state region.

[0040] Preferably, in step S6, the topology serialization rule includes: extracting the maximally connected components in the physicochemical state transition map, and using a depth-first traversal algorithm to extract the key evolution paths in the maximally connected components, and mapping the coordinate components of each state node on the key evolution path to feature characters.

[0041] Preferably, in step S7, the inverted indexing mechanism includes: splitting the dynamic trajectory feature string into multiple fixed-length feature segments, retrieving the standard map containing the most feature segments from the standard drug fingerprint database, and using it as a member of the reference map subset.

[0042] Preferably, in step S8, the calculation process of the subgraph isomorphic matching degree includes: extracting the topological feature vector of the physicochemical state transition map, and calculating the Euclidean distance between the topological feature vector and the reference feature vector of the reference map.

[0043] Preferably, in step S8, when outputting the consistency evaluation results, if the subgraph isomorphic matching degree is higher than the preset similarity threshold, it is determined that the in vitro dissociation behavior of the drug under test is consistent with that of the reference preparation.

[0044] Example 1: In the in vitro bioequivalence testing scenario for generic drugs containing acid-base sensitive gel matrix, the system faces the challenge of local supersaturation and gel layer erosion responses caused by the polymeric excipient system experiencing abrupt changes in acid-base gradients and ionic strength at the simulated gastrointestinal transport interface. Traditional constant-medium timed sampling and measurement methods set the test environment to a static buffer solution. The overall concentration-time cumulative curve obtained reduces the nonlinear spatiotemporal evolution trajectory to a smooth numerical value, failing to distinguish formulations with similar overall dissolution behavior but different local phase transition processes. To address this situation, the system uses a closed-loop controlled... In the flow field, the pH value, ionic strength, and surfactant concentration of the analytical fluid are adjusted according to a preset time sequence to induce the analyte to produce a transient dissociation response that changes with the fluid environment, thus exposing the internal component interaction mechanism of the test system. Based on this, the system acquires the absorbance data sequence and complex impedance signal sequence characterizing the analyte in the transient dissociation response through a detection unit composed of an ultraviolet absorbance sensor and an impedance sensor. The first-order difference value is extracted from the absorbance data sequence, and it is mapped to the two-dimensional physicochemical parameter distribution domain as coordinate components along with the complex impedance phase angle in the complex impedance signal sequence.

[0045] In response to the physical obstruction and optical scattering drift caused by formulation disintegration and shedding, the complex impedance phase angle is sensitive to the ionic environment and macromolecular network hindrance effects of the system. This dual-channel imaging mechanism extracts a set of physicochemical response distribution points composed of the correlation points of the first-order difference value and the complex impedance phase angle at the same sampling time. It orthogonally fuses the optical data characterizing concentration changes with the electrical data mapping the state of the matrix skeleton, thereby transforming the problem of removing the physical interference of insoluble particles into a coordinate decoupling action in the physicochemical parameter distribution space. Based on this data, the system divides the two-dimensional physicochemical parameter distribution domain into multiple discrete physicochemical state regions according to the preset grid division rules, and defines the physicochemical state region into which the physicochemical response distribution point set falls as a state node characterizing a specific chemical state. By statistically analyzing the migration direction and migration frequency between state nodes at adjacent time points, a physicochemical state transition map characterizing the dynamic evolution topological features of the chemical properties of the drug under test is constructed.

[0046] The extraction of the physicochemical response distribution point set provides an intrinsic signal coordinate source for gridded clustering that filters out physical background scattering noise. The construction of the physicochemical state transition map transforms the discrete data point stream into a structured representation containing metastable phase transition migration paths, enabling the system to output evaluation results for drug quality through consistency comparison of the map structure. This map construction and retrieval mechanism reflects the transient relaxation evolution trajectory of the dissolution system from imbalance to the reconstruction of a new equilibrium, shifting the granularity of consistency evaluation from the comparison of terminal release concentrations to the isomorphic determination of evolutionary state sequences. It utilizes topological features with spatiotemporal evolution correlation to block the link of misjudgment in quality evaluation caused by the collapse of multiple metastable data.

[0047] Example 2: When the system faces the situation where the optical scattering baseline drifts due to insoluble microparticles generated by the disintegration of a generic drug with an acid-base sensitive gel matrix during in vitro equivalence evaluation, this experiment utilizes a closed-loop controlled flow field testing platform for physical verification. This testing platform includes a fluid injection pump, a UV absorbance sensor, and an impedance sensor. The flow rate control precision of the fluid injection pump is set to 0.1 mL / min. The UV absorbance sensor covers a detection band of 200 nm to 400 nm and has a sampling frequency set to 10 Hz. The impedance sensor outputs an excitation frequency of 1 kHz to 1 MHz and has... With a phase angle measurement resolution of 0.1°, the raw data sequences obtained in the experiment are all directly derived from the real-time measurements of the aforementioned physical sensor array. The detection unit is configured as a coaxial integrated flow cell physical structure. The miniature ring electrode of the impedance sensor is embedded in the inner wall of the quartz window of the ultraviolet absorbance sensor, so that the high-frequency AC electric field and the ultraviolet beam penetrate the same cross section of the analytical fluid orthogonally in physical space. This eliminates the inherent spatial transmission delay and physical phase difference of the test fluid flowing through the discrete sensor components in sequence. The output absorbance data sequence and the complex impedance signal sequence correspond to the same test fluid micro-particle entity parameter at the same single sampling time.

[0048] In this test platform, the initial step is to determine the synchronous sampling frequency of optical and electrical signals. The core technical consideration in selecting the synchronous sampling frequency is to balance the completeness of capturing high-frequency transient phase transition characteristics with the system's data processing load. To achieve this balance, the system sets a sampling frequency optimization rule to monitor the rate of dissolution front advancement of the test drug in the flow field. When the rate of dissolution front advancement increases and is accompanied by high-frequency signal perturbations, to avoid the aliasing and loss of phase transition characteristics under the Nyquist sampling theorem, the synchronous sampling frequency tends to the physical upper limit allowed by the test platform. Based on this rule and combined with the typical fluid displacement rate in the gastrointestinal transport simulation scenario, the system selects 10Hz as the standard synchronous sampling frequency for this experiment. This setting allows the system to obtain a time resolution of 100ms to capture the dissolution front. To verify the physicochemical relaxation response of the gel skeleton in the pH abrupt change range, and to study the synergistic characteristics of the feature extraction boundary and channels of the orthogonal fusion mechanism under real interference conditions, a control system containing multiple perturbation gradients was constructed. Commercially available sustained-release gel tablets with clearly defined components were selected as the reference sample. By actively injecting non-reactive silica micropowder with mass fractions of 5%, 10%, and 15% into the test fluid, optical scattering noise sources of formulation erosion and shedding material of different intensities were simulated, forming the first, second, and third experimental groups with gradient distribution of problem intensity. At the same time, a partially missing control group was set up, which blocked the complex impedance phase angle input and relied only on the first-order difference value of UV absorbance to construct the state spectrum, and an out-of-range control group was set up, which changed the grid division scale of the two-dimensional physicochemical parameter distribution domain.

[0049] The fluid testing process was initiated. At the 30-minute mark, the control system abruptly changed the pH of the flow field environment from 1.2 to 6.8. The detection unit simultaneously output the raw data sequence. The raw optical data, without orthogonal mapping processing, showed that during the pH abrupt change, accompanied by the suspension of silica microparticles, the UV absorbance was affected by the physical obstruction of solid particles, resulting in a baseline step noise with a peak value of 0.52 Au. The complex impedance phase angle in the complex impedance signal sequence was unaffected by the direct physical obstruction caused by the suspended state of solid particles, and its value accurately reflected the transition of the polymer network from a protonated contracted state to an ionized swollen state. The phase angle shifted deterministically from -82.5° to -65.1° due to the change. The system extracts the first-order difference value of the ultraviolet absorbance data sequence and maps the optical rate of change parameter that eliminates absolute baseline drift and the complex impedance phase angle to the two-dimensional physicochemical parameter distribution domain. The calculated physicochemical response distribution point set representing the transient dissociation response is clustered in a specific grid coordinate interval with a first-order difference value of 0.012Au / s and a complex impedance phase angle of -70.5°. The system compensates for the lack of data extraction by the optical channel under the condition of sudden turbidity by introducing an electrical channel to sense the state of the ion hydration layer.

[0050] The system transforms the physicochemical response distribution point set into state nodes according to preset grid division rules, extracts the migration frequency of adjacent time points to generate a physicochemical state transition spectrum, calculates the subgraph isomorphic matching degree between the test spectrum and the reference spectrum subset, and outputs a quantitative judgment index. Output data under different perturbation gradients are extracted. The subgraph isomorphic matching degrees of the first, second, and third experimental groups under 5%, 10%, and 15% microparticle scattering noise interference remain at 98.5%, 97.8%, and 96.2%, respectively. Stable topological feature structures are output under the interference gradient of increasing particle concentration. Based on the output data of the partially missing control group, due to the lack of electrical dimension state constraints, its isomorphic matching degree drops sharply from 88.5% to 62.1% under the same particle concentration gradient. The data confirms that the optical channel and the electrical channel constitute cross-dimensional mutual constraints, presenting a physical interference suppression structure that cannot be achieved by isolated calculation of a single channel. The results showed that the data trajectory of the out-of-range control group was extracted, and the working range of the grid division scale was set to 0.01 to 0.5. When the scale increased to the upper limit of 0.52, the subtle phase transition features were lost due to the excessive coarsening of the state space, and the isomorphic matching degree showed an artificially high value of 89.5%, and the resolution to distinguish the subtle fluctuations of different formulations was lost. When the scale shrank to the lower limit of 0.008, the system incorrectly divided the tiny sensor thermal noise into independent state nodes, resulting in the fragmentation of the topological map structure. The isomorphic matching degree dropped to 45.2%, showing a nonlinear deterioration inflection point that deviated from the working range. The experimental data confirmed the engineering rationality of the grid division scale setting, indicating that the analysis method transferred the evaluation dimension of the in vitro drug release process to the phase transition dynamics domain through physical property decoupling and state space gridding topological representation, and output definite evaluation data in the test scenario containing physical scattering noise.

[0051] Example 3: When the system faces a high-dimensional topology search computational load in a large-scale standard drug fingerprint database, the full-table subgraph isomorphic comparison method generates computational load and fails to meet the data processing requirements of continuous online testing. For this situation, the system applies a preset topology serialization rule to extract the two-dimensional physicochemical parameter distribution domain grid coordinates of each state node in the physicochemical state transition graph. Combined with a pre-established coordinate interval mapping table, the first-order difference value axis coordinates and complex impedance phase angle axis coordinates are converted into letter combination codes to generate node identifiers. The transition frequencies between adjacent state nodes are extracted and appended as numerical suffixes to the end of the corresponding node identifiers to generate local edge sequences. The system connects all local edge sequences according to the traversal path from small to large first-order difference value coordinates and outputs a kinetic trajectory feature string characterizing the dissolution kinetics of the drug under test.

[0052] Based on this dimensionality-reduced data structure, the system uses node identifiers in the dynamic trajectory feature strings as retrieval keys. It invokes an inverted index mechanism in a pre-defined standard drug fingerprint database to locate a set of reference graph identifiers containing the same node identifiers. The system selects the graph data pointed to by this set of identifiers to form a subset of the reference graph. It extracts the migration frequencies of each local edge in the physicochemical state transition graph and the reference migration frequencies at the corresponding edge positions in the reference graph. It then sums the minimum values ​​of both on all matching edges and calculates the maximum values ​​of both on all matching edges. The ratio of the minimum to the maximum sum is output as the subgraph isomorphic matching degree. This topology serialization and inverted index mechanism transforms the high-dimensional graph structure traversal process into key-value addressing and local frequency comparison steps, reducing the complexity of graph comparison. The computational power consumption provides a definite data flow path for drug consistency evaluation retrieval. When extracting the maximally connected components in the physicochemical state transition graph according to the preset topological serialization rules, the migration frequency values ​​between adjacent state nodes are extracted as the weights of unidirectional edges in the directed graph structure. The depth-first traversal algorithm uses the sum of the cumulative edge weights on the state transition path as the search metric to traverse and retrieve. The unidirectional connected routes from the starting node to the terminal node that reach the maximum value of the sum of the cumulative edge weights are marked as key evolution paths. The coordinates of the state nodes and the associated migration frequencies contained in the key evolution paths are extracted according to the time series to form a one-dimensional topological feature vector. By calculating the square root of the sum of the squares of the differences between the corresponding dimensional components of the topological feature vector and the benchmark feature vector, the Euclidean distance in the continuous space is output and used as the basis for determining the isomorphic matching degree of the quantified subgraph.

[0053] Example 4: When the system faces unknown matrix formulation kinetic testing and unknown hardware background noise, the reference calibration procedure of the two-dimensional physicochemical parameter distribution domain is triggered before the drug to be tested is connected. The fluid injection pump is controlled to inject standard buffer fluid into the controlled flow field. The detection unit is used to obtain the empty absorbance data sequence and the empty complex impedance signal sequence within a specific time period. The first time series variance of the empty first-order difference value sequence extracted from the empty absorbance data sequence and the second time series variance of the empty complex impedance phase angle in the empty complex impedance signal sequence are calculated. The first time series variance and the second time series variance are used as the quantification benchmark of the grid division rule. The scale span of a single physicochemical state region on the first-order difference value coordinate axis is set to 3 times the first time series variance. The scale span of the physicochemical state region on the complex impedance phase angle coordinate axis is set to 3 times the second time series variance. The initial grid boundary matching the test flow field is constructed based on the sensor physical drift parameters.

[0054] The system locks the initial grid boundary and extracts the physicochemical response distribution point set containing the drug under test. It maps each data point to a two-dimensional physicochemical parameter distribution domain. For free data points that fall into the boundary of adjacent physicochemical state regions, it calculates the Euclidean distance from the free data point to the geometric center of all adjacent physicochemical state regions, extracts the minimum parameter from the Euclidean distance value, and assigns the free data point to the physicochemical state region corresponding to the minimum parameter to define state nodes. It constructs a physicochemical state transition map by combining the migration direction and migration frequency between state nodes at adjacent time points. The noise floor measurement and boundary distance measurement offset the sensing baseline offset caused by component aging and fluid medium switching, and solidifies the evolution characteristics of the drug's chemical properties in an objective topological matrix that excludes measurement uncertainties.

[0055] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1. A method for analyzing chemical properties for quality consistency evaluation of a pharmaceutical product, characterized by, Includes the following steps: Step S1: In a closed and controlled flow field, the pH value, ionic strength and surfactant concentration of the analytical fluid are adjusted according to a preset time sequence to induce the drug to be tested to produce a transient dissociation response that changes with the fluid environment. Step S2: The absorbance data sequence and complex impedance signal sequence characterizing the drug under test in the transient dissociation response are acquired by the detection unit composed of an ultraviolet absorbance sensor and an impedance sensor. Step S3: Extract the first-order difference value from the absorbance data sequence, and map the first-order difference value and the complex impedance phase angle in the complex impedance signal sequence as coordinate components to the two-dimensional physicochemical parameter distribution domain, and extract the physicochemical response distribution point set composed of the correlation points of the first-order difference value and the complex impedance phase angle at the same sampling time. Step S4: Divide the two-dimensional physicochemical parameter distribution domain into multiple discrete physicochemical state regions according to the preset grid division rules, and define the physicochemical state region into which the physicochemical response distribution point set falls as a state node characterizing a specific chemical state. Step S5: Statistically analyze the migration direction and frequency between state nodes at adjacent time points, and construct a physicochemical state transition map that characterizes the dynamic evolution topological features of the chemical properties of the drug under test. This allows for in-depth evaluation of drug quality through consistency comparison of the map structure.

2. The chemical property analysis method for drug quality consistency evaluation according to claim 1, characterized in that, It also includes the following steps: Step S6: According to the preset topology serialization rules, the grid coordinate attributes of each state node in the physicochemical state transition map and the migration frequency information between adjacent nodes are converted into a character sequence with a specific encoding to generate a kinetic trajectory feature string for characterizing the drug dissolution kinetics. Step S7: Using the kinetic trajectory feature string as the retrieval key, the reference map subset corresponding to the drug to be tested is retrieved by matching in the preset standard drug fingerprint database through the inverted index mechanism. Step S8: Calculate the subgraph isomorphic matching degree between the physicochemical state transition spectrum and each reference spectrum in the reference spectrum subset, and output the chemical property consistency evaluation results of the test drug and the reference preparation based on the subgraph isomorphic matching degree.

3. The chemical property analysis method for drug quality consistency evaluation according to claim 1, characterized in that, In step S1, acidic buffer solution, alkaline buffer solution, and simulated gastrointestinal fluid are alternately injected into the closed controlled flow field at a preset frequency using a high-precision plunger pump group to construct a perturbed flow field with a dynamic concentration gradient.

4. The chemical property analysis method for drug quality consistency evaluation according to claim 1, characterized in that, In step S2, the sampling frequency of the signal acquired by the detection unit is not less than 100Hz; the complex impedance signal sequence includes a complex impedance phase angle sequence and a complex impedance amplitude sequence that change with time.

5. The chemical property analysis method for drug quality consistency evaluation according to claim 1, characterized in that, The physicochemical state transition map includes state switching paths composed of directed arcs, and a weight matrix used to store the dwell time of the physicochemical response distribution point set in each discrete physicochemical state region.

6. The chemical property analysis method for drug quality consistency evaluation according to claim 2, characterized in that, In step S6, the topology serialization rule includes: extracting the maximally connected components in the physicochemical state transition map, and using a depth-first traversal algorithm to extract the key evolution paths in the maximally connected components, and mapping the coordinate components of each state node on the key evolution path to feature characters.

7. The chemical property analysis method for drug quality consistency evaluation according to claim 2, characterized in that, In step S7, the inverted indexing mechanism includes: splitting the dynamic trajectory feature string into multiple fixed-length feature segments, retrieving the standard map containing the most feature segments from the standard drug fingerprint database, and using it as a member of the reference map subset.

8. The chemical property analysis method for drug quality consistency evaluation according to claim 2, characterized in that, In step S8, the calculation process of the subgraph isomorphic matching degree includes: extracting the topological feature vector of the physicochemical state transition map and calculating the Euclidean distance between the topological feature vector and the reference feature vector of the reference map.

9. The chemical property analysis method for drug quality consistency evaluation according to claim 2, characterized in that, In step S8, when outputting the consistency evaluation results, if the subgraph isomorphic matching degree is higher than the preset similarity threshold, it is determined that the in vitro dissociation behavior of the test drug is consistent with the reference preparation.

Images