A multi-dimensional management system based on clinical data of kidney stone surgery

Abstract

The application relates to the technical field of kidney stone operation management, and discloses a multi-dimensional management system based on clinical data of kidney stone operation. A preoperative data integration module of the system receives and integrates image data from CT, ultrasound, X-ray and other devices, and generates a three-dimensional stone distribution model through spatial registration. A surgical risk assessment module extracts parameters such as stone volume, density and position based on the model, combines physiological indicators and historical operation data of the patient, and calculates a surgical difficulty coefficient and a complication probability. An intraoperative real-time monitoring module collects endoscope video, laser parameter and flushing liquid pressure data, and updates the stone crushing state and residual volume in real time. A tissue damage early warning module analyzes the laser energy distribution and the tissue contact time, assesses the thermal damage risk and generates a safe operation boundary. A postoperative recovery prediction module integrates the operation data and the metabolic characteristics of the patient, and predicts the stone clearance rate and the recovery progress of the kidney function.

Description

Technical Field

[0001] This invention relates to the field of kidney stone surgery management technology, specifically a multi-dimensional management system based on clinical data from kidney stone surgery. Background Technology

[0002] Kidney stones are a common disease of the urinary system, and surgical treatment mainly includes minimally invasive procedures such as percutaneous nephrolithotomy (PCNL) and ureteroscopic lithotripsy. In traditional surgical management models, preoperative planning relies heavily on the surgeon's experience in interpreting two-dimensional imaging data, making it difficult to accurately assess key parameters such as the spatial distribution, volume, and density of the stones. This shift from two-dimensional to three-dimensional thinking is subjective and affects the accuracy of surgical path planning. Intraoperative procedures depend on the surgeon's visual feedback and tactile feedback, lacking quantitative monitoring methods for the lithotripsy process, energy usage, and tissue impact. Postoperative evaluation is mostly based on short-term imaging follow-ups and the patient's subjective symptoms, lacking accurate prediction of long-term renal function recovery.

[0003] Existing medical information systems typically store data from different stages independently. For example, image archiving systems store preoperative images, surgical anesthesia systems record intraoperative parameters, and electronic medical records systems collect postoperative follow-up information. These systems lack effective data linkage and in-depth analysis capabilities, creating information silos. Physicians struggle to quickly obtain integrated information about the patient throughout their entire lifecycle for comprehensive judgment. Furthermore, surgical risk assessments are often based on limited clinical experience, lacking quantitative models that incorporate individualized patient characteristics. Intraoperative safety relies primarily on the physician's skill and real-time judgment, lacking objective early warning mechanisms for tissue damage risks. Postoperative recovery predictions often remain at the qualitative descriptive level, lacking personalized predictive models based on multi-dimensional data fusion.

[0004] Clinical management of kidney stone surgery suffers from problems such as fragmented data, subjective assessment, lack of early warning, and crude prediction. There is an urgent need for a management system that can cover the entire process of preoperative, intraoperative, and postoperative care, and achieve multi-dimensional data integration and intelligent analysis to improve surgical accuracy, safety, and prognosis. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-dimensional management system based on clinical data from kidney stone surgery to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a multi-dimensional management system based on clinical data from kidney stone surgery, the system comprising:

[0007] The preoperative data integration module is used to receive and integrate imaging examination data of kidney stone patients from different medical devices, spatially register CT scan results, ultrasound images and X-ray films, and generate a three-dimensional stone distribution model.

[0008] Based on the three-dimensional stone distribution model, the surgical risk assessment module extracts the stone volume, density, and location parameters, and calculates the surgical difficulty coefficient and the probability of potential complications by combining the patient's physiological indicators and historical surgical data.

[0009] The intraoperative real-time monitoring module is used to collect endoscopic video streams, laser lithotripsy parameters, and irrigation fluid pressure data during the operation, and to update the stone fragmentation status and residual volume simultaneously.

[0010] Based on the stone fragmentation state and residual volume, the tissue damage early warning module analyzes the laser energy distribution and contact time with kidney tissue, assesses the thermal damage risk area, and generates a safe operating boundary.

[0011] The postoperative recovery prediction module integrates surgical time, lithotripsy efficiency, and complication occurrence, combined with the patient's individual metabolic characteristics, to predict the stone clearance rate and the progress of kidney function recovery.

[0012] Preferably, the generation process of the three-dimensional stone distribution model is as follows:

[0013] Threshold segmentation is performed on the CT scan results to extract the stone pixel set in different density ranges and establish an initial stone volume framework.

[0014] The echo intensity data in the ultrasound image is mapped to the initial stone volume frame to supplement information on calcified areas and soft tissue boundaries.

[0015] The contour of the supplemented stone volume frame is corrected using two-dimensional projection data from X-ray films to eliminate artifact interference caused by image overlap.

[0016] By fusing spatial coordinate parameters from multimodal images, a three-dimensional stone distribution model including density gradient and anatomical location is reconstructed.

[0017] Preferably, the calculation process for the surgical difficulty coefficient is as follows:

[0018] The maximum transverse diameter of the stones and the branch angle of the aggregate system are extracted from the three-dimensional stone distribution model to calculate the physical obstacle removal index.

[0019] Analyze the peak differences in the stone density distribution curve to determine the compositional heterogeneity score;

[0020] Assess the degree of limitation in instrument operation space by combining data on urinary tract anatomy variations and previous surgical records;

[0021] By combining the physical obstacle removal index, component heterogeneity score, and instrument operation space limitation, a quantitative surgical difficulty coefficient is generated.

[0022] Preferably, the detection process of the intraoperative real-time monitoring module includes:

[0023] Inter-frame differential processing is performed on the endoscopic video stream to identify the movement trajectory of the laser fiber tip and the dynamics of stone fragmentation;

[0024] Extract the pulse frequency and single-shot energy parameters output by the laser generator, and calculate the total amount of broken stone per unit time;

[0025] Monitor the fluctuation curve of the flushing fluid pressure sensor to estimate the trend of intrarenal pelvic pressure change and perfusion efficiency;

[0026] The distribution map of residual stone volume is dynamically updated based on the total amount of fragments and the trend of intrarenal pelvis pressure.

[0027] Preferably, the tissue damage early warning module is implemented in the following way:

[0028] A laser energy attenuation model in the renal parenchyma was established, and the cumulative value of absorbed heat in tissues at different depths was calculated.

[0029] Based on the rate of tissue color change in the endoscopic field of view, the local temperature rise gradient can be inferred.

[0030] Based on the accumulated heat absorption and the local temperature rise gradient, an isothermal damage risk contour map is drawn.

[0031] Based on the isothermal damage risk contour map, operational recommendations including safe power thresholds and action time are generated.

[0032] Preferably, the postoperative recovery prediction module operates as follows:

[0033] The total laser energy output and effective lithotripsy time during the operation were statistically analyzed, and the energy utilization efficiency parameters were calculated.

[0034] Analyze the proportion of crystal components in immediate postoperative urine to assess the residual microstone burden;

[0035] By combining the patient's 24-hour urinary metabolic indicators, an individualized stone recurrence risk curve was established;

[0036] Based on the energy utilization efficiency parameters, residual microstone load, and stone recurrence risk curve, the degree of renal function recovery within three months is predicted.

[0037] Preferably, the system further includes:

[0038] The data visualization interaction module is used to overlay the three-dimensional stone distribution model with real-time surgical data, supporting multi-planar reconstruction and virtual profile viewing;

[0039] The surgical procedure optimization module provides suggestions on instrument selection and operation path for the current surgery based on a database of similar historical cases.

[0040] The quality control indicator generation module automatically calculates key performance indicators for surgery, including stone clearance rate and lithotripsy volume per unit time.

[0041] Preferably, the working process of the data visualization interaction module is as follows:

[0042] Establish a three-dimensional coordinate system for the surgical scene and perform real-time spatial registration between the endoscopic field of view and the preoperative images;

[0043] Mark the area of ​​residual stones and important vascular and nerve structures in the registered 3D view;

[0044] It provides a virtual laser path simulation function to predict the energy distribution under different incident angles;

[0045] It allows surgeons to customize the sectional view and simultaneously displays the anatomical relationship in both transverse and coronal views.

[0046] Preferably, the operation process of the surgical plan optimization module is as follows:

[0047] Based on the current characteristics of the stones, multi-parameter similarity matching is performed in the case database to screen comparable surgical records;

[0048] Extract laser setting parameters and surgical path data from matched cases to generate a set of suggested operating parameters;

[0049] Analyze the types and management methods of complications in comparable surgical records to develop risk response plans;

[0050] The recommended set of operating parameters is adjusted in a personalized manner based on the current anatomical variations of the patients.

[0051] Preferably, the working process of the quality control index generation module is as follows:

[0052] Compare the preoperative three-dimensional stone distribution model with the postoperative imaging results to calculate the percentage of volume clearance.

[0053] Analyze the time allocation data for each stage of the surgery to evaluate the efficiency of the operational process;

[0054] Record the time points of complication occurrences and treatment measures to generate a safety event timeline;

[0055] The volume clearance percentage, operational efficiency, and safety event timeline are integrated to form a standardized quality control report.

[0056] Compared with the prior art, the beneficial effects of the present invention are:

[0057] This invention integrates multi-source imaging data and generates a three-dimensional stone distribution model, providing intuitive and precise anatomical basis for preoperative planning. Traditional methods relying on doctors to imagine three-dimensional structures from two-dimensional images are inherently uncertain, while the three-dimensional model provided by this system clearly shows the spatial location, shape, and relationship of the stones with surrounding tissues. This helps doctors develop better surgical approaches and lithotripsy strategies, improving the scientific rigor and accuracy of surgical planning.

[0058] The surgical risk assessment module achieves objective and personalized assessment of surgical risks through quantitative analysis of multi-dimensional data such as stone characteristics and patient physiological indicators. This assessment no longer relies solely on the doctor's subjective experience but is based on a data model to calculate the difficulty coefficient and complication probability. This provides crucial decision support for doctor-patient communication, surgical plan selection, and emergency response planning, helping to reduce surgical risks and improve medical safety.

[0059] The intraoperative real-time monitoring module enables quantitative management of the surgical process. The system continuously collects data such as endoscopic video and laser energy parameters, and provides real-time feedback on the degree of stone fragmentation and residual stone status. This allows doctors to dynamically adjust surgical procedures, avoiding excessive stone fragmentation or excessive residue, thus improving the precision and efficiency of the surgery.

[0060] The tissue damage early warning module analyzes laser energy distribution and tissue contact time to promptly detect potential thermal damage risks. The system-generated safe operating boundaries provide doctors with intuitive visual cues, helping to avoid excessive energy concentration that could damage kidney tissue, significantly improving surgical safety and reducing postoperative complications.

[0061] The postoperative recovery prediction module combines surgical data and individual patient characteristics to provide personalized predictions of stone clearance rate and renal function recovery. This data-driven prediction model provides a scientific basis for the development of postoperative follow-up plans and rehabilitation guidance, contributing to more refined postoperative management and improving long-term patient outcomes. This system effectively connects fragmented surgical data, forming a closed-loop management system from preoperative assessment to postoperative prediction, thus promoting the transformation of kidney stone surgery towards a precision medicine model. Attached Figure Description

[0062] Figure 1 This is a timeline diagram of the multi-dimensional management system for clinical data from kidney stone surgery described in this invention.

[0063] Figure 2 Flowchart generated for a three-dimensional stone distribution model;

[0064] Figure 3 This is a flowchart of the intraoperative real-time monitoring module detection process. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Please see Figure 1 This invention provides a multi-dimensional management system for clinical data of kidney stone surgery. The system integrates multiple functional modules to achieve data management throughout the entire kidney stone surgery process. The preoperative data integration module receives imaging data from CT scanners, ultrasound equipment, and X-ray machines. Using image processing algorithms, it performs spatial registration of multimodal images, aligning CT scan results, ultrasound images, and X-ray films to a unified coordinate system, generating a three-dimensional model containing the spatial distribution of the stones. The surgical risk assessment module extracts physical parameters of the stones, such as volume, density, and location information, from the three-dimensional stone distribution model. Combined with the patient's physiological indicators such as heart rate and blood pressure, as well as historical surgical records, it calculates the surgical difficulty coefficient and the probability of complications using a predefined algorithm model. The intraoperative real-time monitoring module connects to the endoscopic video system, laser lithotripsy equipment, and irrigation fluid pressure sensor, acquiring video streams, laser pulse parameters, and pressure data in real time. It dynamically analyzes and updates the stone fragmentation status and residual volume. The tissue damage early warning module analyzes the distribution of laser energy in the kidney tissue based on real-time monitoring data, assesses the risk of thermal damage in conjunction with tissue contact time, and generates safe operating boundary prompts. The postoperative recovery prediction module integrates surgery duration, lithotripsy efficiency indicators, and complication events, and correlates them with individual patient metabolic characteristics such as urine composition to predict stone clearance rate and renal function recovery trends. All modules interact via a data bus to ensure real-time information synchronization. The system employs a distributed architecture deployed on a hospital server cluster, supporting multi-terminal access.

[0067] Example 1: See Figure 2The generation of the three-dimensional stone distribution model begins with the depth processing of CT scan results. CT data is acquired from the image archiving system in DICOM format, with slice thickness typically set to 0.5 to 0.625 mm to ensure sufficient spatial resolution. Thresholding segmentation employs a multi-level adaptive thresholding algorithm. This algorithm does not fix a global threshold but dynamically adjusts the segmentation boundary based on local image features and grayscale histogram distribution, identifying high-density calcified stone regions, whose Hounsfield unit values ​​are typically higher than 800. Then, a lower threshold range is set for medium-density uric acid stones or cystine stones. A region growing method is used to aggregate continuous and similarly attributed pixels to form a preliminary stone point cloud set. This point cloud set constitutes the geometric basis of the initial stone volume framework, which is presented as a three-dimensional mesh model, but at this stage only contains basic morphological and density information extracted from the CT data.

[0068] The process of ultrasound image fusion requires addressing the spatial registration problem between multimodal data. During image acquisition, the spatial position of the ultrasound probe is tracked in real-time by an electromagnetic positioning device, recording the probe's position and orientation relative to the patient's reference coordinate system. Echo intensity data in ultrasound images reflect the acoustic characteristics of tissues; hyperechoic areas typically correspond to calcified portions of stones or tissue interfaces. When mapping these data to the initial stone volume frame, a thin-plate spline interpolation algorithm is used for non-rigid registration. This allows clear soft tissue boundaries in the ultrasound image, such as the renal pelvis wall and calyx contours, to effectively complement the stone frame primarily defined by CT. This is particularly effective in highlighting stone edges with poor contrast on CT images or small stones attached to soft tissue. This step significantly enriches the model's boundary information and histological background.

[0069] The two-dimensional projection data from X-ray images provides orthogonal verification and correction for the model. X-ray images, typically abdominal plain films or intravenous pyelography images, have two-dimensional projection characteristics and are easily affected by intestinal gas or bone overlap. When using this data for contour correction, digital reconstruction radiography is employed. This involves simulating the X-ray projection of the initial three-dimensional stone volume framework to generate a two-dimensional digital image. This digital image is then compared with the actual X-ray image, and the position of the midpoint in the three-dimensional model is adjusted through optimization algorithms to minimize the difference between the simulated and actual projections. This effectively eliminates artifacts caused by image overlap, such as correcting misjudgments of the stone contour caused by rib shadows, making the model's geometry more accurately conform to the actual anatomical structure.

[0070] The fusion process of multimodal image spatial coordinate parameters begins with establishing CT images as the reference for spatial registration. CT images, acquired with a fixed scan bed position and well-defined slice thickness and pixel spacing parameters, allow for the establishment of a stable and accurate three-dimensional Cartesian coordinate system. In this system, each pixel contains not only grayscale information but also its unique three-dimensional spatial coordinates, providing a geometric basis for subsequent registration. Ultrasound and X-ray images need to be mapped to the CT coordinate system using a calculated coordinate transformation matrix. This transformation matrix is ​​typically a 4x4 affine transformation matrix, incorporating rotation, translation, and scaling parameters. Solving this matrix usually relies on identifying a set of corresponding anatomical or external positioning markers in the two sets of images. For example, implanted reference markers or inherent anatomical landmarks on the patient's bones that are clearly identifiable in different modalities can be used. By minimizing the positional error between these corresponding point sets, the optimal transformation matrix parameters can be iteratively calculated.

[0071] The fusion process relies not only on sparse corresponding point matching but also introduces a mutual information maximization strategy based on image voxel grayscale information to optimize registration accuracy. Mutual information is an indicator of the statistical dependence between two images. During iterative optimization, one image is resampled using the currently estimated transformation matrix, and its mutual information value with the other image is calculated. The optimization algorithm continuously adjusts the parameters of the transformation matrix to maximize the mutual information value, indicating that the two images have achieved optimal alignment geometrically, even if their imaging principles and grayscale features differ significantly. The final generated 3D stone distribution model is a structured comprehensive data volume. Each voxel in this data volume not only records its 3D spatial coordinates in this unified coordinate system but also contains the fused density value and the modal identification information from which the voxel data originated. For example, the data of a certain voxel may mainly come from CT, but its edge information may be supplemented and corrected by higher-sensitivity ultrasound data.

[0072] This data organization structure enables the model to support advanced 3D visualization operations. Physicians or the system can rotate and observe the model at any angle on the workstation to examine the spatial relationship between the stone and surrounding tissues from different perspectives. The model supports cutting operations on any plane and can generate virtual cross-sectional views, clearly showing the depth of the stone's embedding within the renal collecting system. The transparency adjustment function allows users to adjust the transparency of different tissues individually, enabling them to visually observe the precise location, shape, and size of the encased stone through the outer tissue, as well as its complex 3D anatomical relationship with the renal pelvis, calyces, and even surrounding blood vessels. This provides intuitive and accurate geometric basis for surgical path planning and risk assessment.

[0073] The calculation of surgical difficulty is a multi-parameter comprehensive evaluation process. Its first step is to extract key geometric and physical parameters from the completed three-dimensional stone distribution model. Measuring the maximum transverse diameter of the stone is not simply the maximum diameter itself, but rather by calculating the minimum enclosing ellipsoid of the stone point cloud to obtain its longest axis dimension in three-dimensional space. This better reflects the actual spatial perception of instrument manipulation during surgery. The measurement of the branching angles of the renal calyx focuses on the width of the renal calyx neck and the angle of the ureteropelvic junction where the stone is located. These angles directly affect the accessibility and bending freedom of instruments such as the ureteroscope. These angle values ​​are obtained through geometric calculations on the three-dimensional model, and then the physical clearance obstacle index is derived. This index quantifies the mechanical resistance that may be encountered during stone removal. Specifically, the three-dimensional stone distribution model not only includes the morphology and location information of the stone, but also reconstructs the complete three-dimensional geometric structure of the renal pelvis, calyx, and other renal systems by fusing multimodal images. The system automatically identifies and extracts key anatomical landmarks on this model, such as the junction of the renal pelvis and calyx, the narrowing points of the renal calyx neck, and the ureteropelvic junction. The branching angle of the collecting system is quantitatively obtained by calculating the angle between the central axis of the renal calyces and the main axis of the renal pelvis. This angle directly reflects the spatial degrees of freedom required for the bending and movement of instruments within the kidney; a smaller angle generally indicates more restricted instrument manipulation. The derivation of the physical clearance barrier index is based on the above angle measurements combined with the geometric parameters of the stone's maximum transverse diameter. The maximum transverse diameter is obtained by calculating the longest axis dimension of the stone's point cloud model in three-dimensional space, reflecting the physical scale of the stone itself. The system weights and fuses the branching angle and the maximum transverse diameter; a smaller angle and a larger transverse diameter result in a higher index value, indicating that the stone is located in a narrow location and is relatively large, together constituting higher mechanical clearance resistance. This index quantifies the physical difficulty required to remove the stone in a specific anatomical environment, providing an objective basis for surgical planning.

[0074] Analysis of stone density distribution curves reveals the heterogeneity of stone composition. Density distribution curves generated from CT values ​​may exhibit single-peak, double-peak, or multi-peak patterns. Peak difference analysis is achieved by calculating the statistical characteristics of curve kurtosis. A broad and gentle peak suggests relatively homogeneous stone composition, while multiple sharp and separated peaks indicate a possible mixture of different density components, such as calcium oxalate and calcium phosphate. The compositional heterogeneity score is quantified based on this peak shape characteristic; a higher score indicates a more complex stone composition, potentially requiring adjustments to laser parameters or surgical strategies. The introduction of patient-specific factors makes the difficulty assessment more targeted. Anatomical variation data of the urinary tract is extracted from the 3D model, such as ureteropelvic junction stenosis or horseshoe kidney abnormalities. These variations significantly limit the instrument's operating space. Previous surgical records provide historical information, such as whether there have been multiple failed lithotripsy attempts or postoperative infections. A rule engine comprehensively analyzes these anatomical variations and historical records to assess the degree of instrument operating space limitation within the specific patient's anatomy—a qualitative and quantitative judgment process.

[0075] The final surgical difficulty coefficient was generated using a weighted linear combination model. Different weighting coefficients were assigned to the scores of three dimensions: physical clearance obstacle index, component heterogeneity score, and instrument operation space limitation. These weighting coefficients were determined based on analysis of a large amount of historical surgical data, reflecting the relative importance of each factor influencing the surgical difficulty. The calculated quantitative surgical difficulty coefficient is a continuous numerical value or a graded result. This result can provide objective decision support for surgeons in preoperative planning, such as predicting the need for longer, more delicate procedures or preparing alternative surgical plans. The entire calculation process emphasizes the transformation from concrete data to abstract indicators, striving to objectively reflect the surgical challenges posed by the characteristics of the stone itself and the individual patient's conditions.

[0076] Example 2: See Figure 3The intraoperative real-time monitoring module's detection process is based on the synchronous acquisition and fusion of multi-source data. The endoscopic video stream is acquired from the digital endoscope system at a rate of 25 frames per second, with a video resolution of 1920×1080 pixels, ensuring the clarity of tissue details. Inter-frame differential processing is performed on consecutive video frames. The algorithm calculates the difference in grayscale values ​​of corresponding pixels in adjacent frames and sets a dynamic threshold to distinguish between static background tissue and moving objects, thereby accurately identifying the movement trajectory of the laser fiber tip. The identification of the fiber tip also incorporates a shape template matching algorithm. Because its metallic material exhibits a high-brightness characteristic under illumination, its movement path in the intrarenal space can be reconstructed by tracking its centroid coordinate sequence. The dynamic identification of stone fragmentation relies on the analysis of texture changes within the inter-frame differential region. When the laser pulse hits the stone, the grayscale values ​​of the pixels in the region fluctuate drastically. By monitoring the contour changes and diffusion trends of these fluctuating regions, the scale and direction of stone fragmentation can be determined.

[0077] Laser generator parameters are acquired via a digital interface provided by the device. Pulse frequency and single-shot energy parameters are sampled every millisecond, and these raw parameters are converted into physical energy values ​​using a calibration curve. The calculation of the total lithotripsy volume per unit time requires combining visual analysis results. The three-dimensional volume of the fragmented area identified from the video is estimated using stereoscopic vision principles. This estimation is based on a pre-calibrated correspondence between endoscope pixel size and actual distance. The volume of newly added fragmented areas in each frame is accumulated, and the lithotripsy efficiency per unit time is calculated by combining this with the laser action time window. Monitoring the flushing fluid pressure fluctuations uses a high-frequency pressure sensor implanted in the perfusion tubing. Sensor data is low-pass filtered to remove noise caused by pump pulsation, and the fluctuation curve reflects the real-time changes in pressure within the renal pelvis. By establishing a dynamic balance model of inflow and outflow rates, combined with pressure changes, the actual volume change trend of the renal pelvis can be calculated, thereby assessing whether the perfusion system remains unobstructed and whether there is a risk of excessive pressure within the renal pelvis.

[0078] The process of dynamically updating the residual stone volume distribution map begins with using an initial 3D stone distribution model constructed preoperatively or intraoperatively as a geometric reference. This model defines the original spatial occupancy and density distribution of the stones. During the surgery, the video analysis module continuously processes the real-time video stream transmitted from the endoscope, using computer vision algorithms to identify and track stone fragments broken up by the laser and moved out of the field of view with the irrigation fluid. Based on the pixel size, movement speed, and preset calibration parameters of the particles, the total volume of stones removed is calculated cumulatively. Stone removal is not uniformly peeled off from the entire stone model, but is highly dependent on the specific application location of the laser fiber. The system accurately determines the 3D coordinate region corresponding to each laser ablation on the initial 3D model by integrating the spatial positioning data of the endoscope and image recognition at the working end of the fiber. Based on the actual fragmentation effect observed in the field of view, such as large stones breaking down into small particles or only surface etching, the system performs local volume reduction calculations at the corresponding locations in the 3D model. The amount of volume reduction is based on the stone removal amount estimated by video analysis and is distributed to the vicinity of the application point according to a certain spatial attenuation model.

[0079] The generated residual stone volume distribution map needs to be presented to the surgeon in an intuitive form. This distribution map is overlaid on the real-time endoscopic image or 3D model in the form of a 3D heatmap. The heatmap uses a continuously changing color gradient to encode the thickness or relative density information of the residual stones. Generally, warm colors represent areas with larger residual thickness or higher density, and cool colors represent areas with less residual stone or areas that have been largely removed. The entire monitoring and updating process forms a real-time closed loop. The new residual distribution map immediately becomes the benchmark for the next round of updates. The system continuously acquires the latest laser action position and video analysis data, and performs local volume reduction and heatmap updates again. This iterative process allows the 3D model to dynamically and almost in real-time reflect the real changes in the surgical area, providing the surgeon with continuous and intuitive visual feedback on the amount and spatial distribution of residual stones, assisting them in deciding on the next lithotripsy strategy, such as adjusting laser energy, changing the bombardment position, or determining the surgical endpoint.

[0080] The core of the tissue damage early warning module is to establish a model of laser energy propagation and thermal effects in biological tissues. The laser energy attenuation model considers the tissue penetration characteristics of lasers of different wavelengths, such as the absorption and scattering coefficients of holmium lasers in kidney tissue. The model calculates the radial diffusion of laser energy from the point of application to the surrounding kidney parenchyma, showing an exponential decay of energy with increasing depth. The cumulative heat absorbed by each micro-tissue voxel is calculated through time integration. This calculation needs to consider the thermal conduction effect of the tissue and the cooling effect of blood perfusion, using the basic principles of the biothermal equation for approximate solution to obtain an estimate of the temperature field distribution at different locations within the kidney parenchyma. Tissue color changes in the endoscopic field of view provide real-time feedback for thermal damage assessment. When heated, the tissue color changes from normal pink to white and brownish-red, a change characteristic of the saturation and hue components in the HSV color space. The algorithm analyzes the pixel colors in the area surrounding the laser application point in real time, inferring the local tissue temperature rise gradient by establishing a mapping relationship between color feature values ​​and known temperature calibration data. Combining the temperature field estimation based on the physical model with the color change inference based on vision allows for cross-validation, improving the reliability of temperature monitoring.

[0081] The isothermal injury risk contour map is constructed using spatial interpolation, generating a continuous three-dimensional temperature distribution field from the calculated and inferred discrete point temperature data. Within this temperature field, regions reaching the critical temperature for tissue denaturation are marked; for example, the 60°C isothermal surface is generally considered the boundary of irreversible damage. These boundaries are projected onto a three-dimensional kidney model, forming an isothermal injury risk contour map surrounding the laser application point. Based on this contour map, the system generates operational recommendations including a safe power threshold and application time. The safe power threshold is not a fixed value but is dynamically adjusted based on the real-time distance between the laser application point and important structures such as the renal pelvis wall and blood vessels. When the risk contour line approaches critical structures, the system suggests reducing the power or changing the application point. The application time recommendation provides predictive prompts based on the real-time calculated temperature rise rate, issuing a warning before the temperature approaches the critical value. This entire warning mechanism aims to quantify and visualize the risk of thermal injury, assisting operators in ensuring tissue safety while pursuing lithotripsy efficiency.

[0082] Example 3: The postoperative recovery prediction module begins with the integrated analysis of surgical data and patient biosamples. The total laser energy output is extracted from the laser device's operation log, which records the energy value and emission time of each pulse. The total energy output is obtained by accumulating the energy values ​​of all effective pulses. The determination of the effective lithotripsy time depends on the synchronous analysis of the endoscopic video and the device activation signal. The system identifies the time period during which the stone fragmentation is actually observed in the endoscopic field of view during laser emission, excluding empty shots or aiming adjustment time. A quantitative index is introduced when calculating the energy utilization efficiency parameter. This parameter is defined as the ratio of the fragmented stone volume to the consumed energy, reflecting the degree of laser energy utilization. Its calculation formula is as follows: in: This represents a parameter indicating energy utilization efficiency. This represents the effective lithotripsy volume, obtained through dynamically updated data from the intraoperative real-time monitoring module. This represents the total laser energy output. During the calculation, the effective stone fragment volume is obtained by subtracting the intraoperatively updated residual volume from the initial volume of the three-dimensional stone distribution model, ensuring the accuracy of the data source.

[0083] Postoperative immediate urinalysis focuses on the crystalline components in the urine sample. The urinalysis is performed using infrared spectroscopy or X-ray diffraction, and samples are collected within one hour after surgery to avoid contamination. The proportion of crystalline components is determined by integrating the peak area of ​​the spectra, identifying characteristic peaks of common stone components such as calcium oxalate, calcium phosphate, and uric acid, and calculating the relative percentage of each component. When assessing the residual microstone load, the concentration of crystals in the urine and the total urine volume are considered. The crystal concentration is obtained from the spectral data using a standard curve method, and the total urine volume is measured using the collection container. The microstone load is expressed in milligrams per liter (mg / L), reflecting the early postoperative excretion of residual fragments. A high load suggests the need for increased postoperative fluid intake or medication intervention.

[0084] The integration of 24-hour urinary metabolic parameters provides a basis for individualized risk assessment. Urine collection is conducted over a full 24-hour period, and the measured parameters include urinary calcium, urinary uric acid, urinary oxalate, urinary citrate concentrations, and urinary pH. An individualized stone recurrence risk curve was established using a proportional hazards regression model. The model input variables included metabolic parameters, patient age, sex, body mass index, and surgery-related parameters such as energy utilization efficiency. The risk curve was generated based on training with historical cohort data, showing the probability of stone recurrence over time. High-risk areas correspond to patients with significant metabolic abnormalities; for example, high urinary calcium or low urinary citrate levels significantly increase the curve steepness. Model validation was performed through internal cross-validation to ensure predictive stability.

[0085] This system predicts the degree of renal function recovery within three months by integrating multiple parameters. Renal function assessment is based on an estimated glomerular filtration rate (eGFR) calculated from serum creatinine levels using the CKD-EPI formula. The prediction process employs machine learning algorithms, such as random forests or gradient boosting trees. Input features include energy utilization efficiency parameters, residual microstone burden, integral values ​​of the stone recurrence risk curve, and the patient's underlying conditions such as hypertension or diabetes. The algorithm outputs the predicted eGFR value and confidence interval at three months. The prediction results are presented in the form of continuous values ​​and trend graphs to assist physicians in developing personalized follow-up plans. The entire module runs on a hospital server, with data automatically retrieved from the laboratory information system and electronic medical records, reducing human intervention errors.

[0086] Example 4: The implementation of the data visualization interaction module is based on the precise construction of a three-dimensional coordinate system for the surgical scene. This coordinate system uses the patient's coordinate system obtained from the preoperative CT scan as a reference, and establishes its relative relationship with the operating table through positioning markers attached to the patient's body surface. During the operation, the optical positioning system continuously tracks the reflective sphere array on the tip of the ureteroscope, acquiring the six-degree-of-freedom position and attitude data of the endoscope in space in real time. This data is mapped to the preoperative coordinate system through a coordinate transformation matrix, achieving real-time spatial registration between the endoscopic field of view and the preoperative image. The registration algorithm uses an iterative nearest-point method to continuously optimize the matching degree between the renal calyx contour feature points extracted from the real-time endoscopic video frames and the corresponding feature points on the surface of the three-dimensional model, controlling the registration error to the sub-millimeter level. In the registered 3D view, the system highlights the areas of residual stones that have not yet been removed with semi-transparent red, based on the preoperative segmentation results and angiography data. At the same time, it outlines important vascular structures such as renal artery branches with blue fine lines and marks the anatomical boundaries of the urine collection system, such as the renal pelvis and calyces, with yellow. These annotation layers are superimposed on the real organ model, and the surgeon can freely switch the display and hiding of each layer through a touch screen or foot switch.

[0087] The virtual laser path simulation function relies on a light projection algorithm and a physical model of laser energy propagation in tissue. When the surgeon clicks on a preset target point for the stone in the 3D model interface, the system emits a beam of light from the virtual position of the current fiber tip, simulating the laser's propagation path. This path is not only displayed as a visual dotted line, but also calculates the energy distribution based on the laser settings and the type of tissue along the path. For example, when the path approaches a blood vessel marked in the model, the energy distribution map shows the potential energy attenuation or reflection in that area. The surgeon can adjust the virtual incident angle and laser power settings, and observe the changes in the simulated energy distribution map in real time, thereby evaluating the possible effects of different operating strategies before the actual laser emission. The function of supporting custom cross-sectional views is achieved through a planar cutting algorithm. The surgeon can draw cutting lines arbitrarily on the 3D model, and the system instantly generates and displays a transverse view of the cut surface, while automatically generating a coronal view orthogonal to it. The two cross-sectional views are updated synchronously, helping the surgeon understand complex 3D anatomical relationships, such as determining the exact distance between a lower calyx stone and the renal cortex.

[0088] The workflow of the surgical plan optimization module begins with a multi-dimensional analysis of the current case's characteristics. The system extracts the morphological features of the patient's stones from the electronic medical record, including but not limited to stone volume, maximum transverse diameter, location in the renal calyx, CT value range, and the surface area to volume ratio calculated from the 3D model. Simultaneously, it records the patient's individual anatomical parameters, such as the angle of the ureteropelvic junction and the width of the renal calyx neck. These parameters constitute a feature vector, used for similarity matching in a historical case database. The database stores a complete dataset of surgical cases from previous years. The matching algorithm uses weighted Euclidean distance to calculate the difference in feature vectors between the current case and historical cases, assigning high weights to key features such as stone location and size, and selecting several comparable surgical records with the highest similarity.

[0089] From successfully matched comparable cases, the module extracts detailed operational parameters recorded during the surgery, forming a set of operational parameter suggestions. These parameters include laser equipment type, commonly used power and frequency setting ranges, fiber optic model selection, and surgical approach. The system analyzes the parameter settling trends in these successful cases; for example, for similarly sized inferior calyx stones, most successful cases may have used a specific range of low power and high frequency settings. The module also retrieves recorded complication events and their management methods from comparable cases. For example, if multiple similar cases report perforation of the inferior calyx neck, the system compiles operational precautions and emergency treatment procedures for this risk, forming a risk response plan. In the final personalized adjustment stage, the system compares the anatomical differences between the current patient and the matched cases. For example, if the current patient's renal calyx neck is significantly narrower, the system will fine-tune the suggested laser energy parameters downwards and indicate the need for more precise lens manipulation. The entire optimization suggestion is presented to the surgeon in a structured checklist format as a reference for preoperative planning and intraoperative decision-making. Refer to Table 1, which shows a sample of real-time registration error data recorded by the system during a surgery for inferior calyx stones.

[0090] Table 1: Statistical Table of Intraoperative Real-Time Registration Error

[0091]

[0092] Considering a complex renal calculi surgery, the patient's preoperative CT scan showed scattered stones in the renal pelvis and upper and middle calyces. During the surgery, the data visualization interaction module registered the flexible endoscope view with the 3D model. The surgeon noticed a small stone in the middle calyx with blurred location in the real-time video. By activating the virtual laser path simulation function and simulating laser irradiation from different angles, it was found that entering the laser from the fornix of the renal calyx could avoid lateral wall vessels while effectively covering the stone. The surgeon then used the custom profile function to generate a coronal view passing through the stone and the suspected vessel, clearly showing a safe distance between them, thus confidently performing laser lithotripsy. Before the surgery, the surgical plan optimization module, based on the stone distribution characteristics and the patient's long renal pelvis, matched three similar surgeries from the historical database. It recommended a staged lithotripsy strategy for this case, treating the main renal pelvis stones first, then the calyx stones, and recommended an extended laser fiber model suitable for a long renal pelvis. During the procedure, the three-dimensional navigation view provided by the system clearly marked the areas of broken and unbroken stones, avoiding any omissions. The quality control report showed that the surgery ultimately achieved complete stone removal.

[0093] Example 5: The implementation of the quality control index generation module begins in the postoperative data acquisition phase. This module automatically retrieves the preoperative 3D stone distribution model and postoperative follow-up imaging results from the hospital's image archiving system. Postoperative imaging is typically completed within 24 to 48 hours after surgery, using the same CT scanning protocol as preoperative imaging to ensure comparability. The comparison process is achieved through an image registration algorithm, spatially aligning the postoperative CT data with the preoperative 3D model. The registration algorithm, based on the principle of maximizing mutual information, identifies stable anatomical landmarks such as the kidney contour and spine as reference points. Initial alignment is achieved through rigid body transformation, followed by fine-tuning using thin-plate spline non-rigid registration to compensate for minor deformations caused by postoperative tissue edema or positional differences. Based on accurate registration, the module calculates the percentage of volume clearance. Its core is to segment the residual stone area in the postoperative image. The segmentation algorithm uses the same threshold range as the preoperative one. It calculates the difference between the total preoperative stone volume and the postoperative residual stone volume using the voxel comparison method, and then outputs the ratio of the preoperative volume to the preoperative volume as a percentage. For example, in a case of renal calyx stone surgery, the preoperative model showed a stone volume of 650 cubic millimeters. After postoperative image segmentation, the residual stone volume was 30 cubic millimeters. The clearance percentage is calculated as (650-30) / 650×100%. However, the module only outputs the calculation logic and does not display the specific value to avoid misleading the reader. The results are integrated into the report in the form of a progress bar.

[0094] Time allocation data for each stage of the surgery is automatically collected through an integrated operating room timing system. This system is synchronized with surgical equipment such as laser generators and endoscopic camera systems, recording discrete events such as laser activation time, instrument entry / exit time, and irrigation pause time. The statistical process defines standard surgical stages, such as the access establishment phase, laser lithotripsy phase, fragment removal phase, and examination phase. The start and end times of each stage are triggered by equipment status signals; for example, the laser lithotripsy phase begins with the first effective laser pulse and ends with the last pulse. The efficiency of the operational procedure is evaluated using a multi-index weighted method. The efficiency value considers not only the total surgical time but also derivative indicators such as the amount of lithotripsy fragments per unit time and instrument switching frequency. The amount of lithotripsy fragments per unit time is calculated by dividing the total volume of fragments provided by the intraoperative real-time monitoring module by the duration of the laser lithotripsy phase. The instrument switching frequency is calculated as the ratio of the number of times tools such as the stone basket and laser fiber are used to the total time. The evaluation results are presented in radar chart form, showing the deviation of the time proportion of each stage from the ideal value.

[0095] The recording of complication occurrence times and treatment measures relies on multi-source data fusion. Complication events are marked by the surgeon during the operation via voice input or a touchscreen interface, and the system automatically records the timestamps and verifies them against vital sign monitor data. The safety event timeline is generated using event sequence modeling. Each complication event, such as bleeding or perforation, is abstracted into a data object containing event type, occurrence time, duration, and treatment measures. The timeline is visualized as a Gantt chart, with the horizontal axis representing the surgical timeline and the vertical axis representing the event severity level. Events are connected by causal chains; for example, a high-temperature alarm precedes tissue discoloration. The timeline data also integrates keyframe screenshots from endoscopic videos, providing visual evidence for each event.

[0096] The process of integrating volume clearance percentage, procedural efficiency, and safety event timelines to create a standardized quality control report follows medical quality management guidelines. The report structure uses a predefined template and includes header fields such as patient anonymity identifier, surgery date, and surgeon information. The volume clearance percentage is displayed numerically and visually, with pre- and post-operative images shown side-by-side, and areas of difference highlighted. The procedural efficiency section includes pie charts of time distribution for each stage and efficiency indicator tables, highlighting steps that deviate from the standard procedure. The safety event timeline is embedded as an interactive chart, allowing users to click on events to view detailed information. Once generated, the report is automatically uploaded to the hospital's quality control database and pushed to relevant surgical teams for post-operative review and continuous quality improvement.

[0097] Consider a case of renal pelvic stone surgery using ureteroscopic laser lithotripsy. Preoperative 3D model showed the stone located in the middle of the renal pelvis, with a volume of approximately 480 cubic millimeters and uniform density. Postoperative CT scan was performed on the second day after surgery. During automatic image registration, a slight shift in kidney position due to respiratory movement was observed. After adjustment using a non-rigid registration algorithm, the residual stone volume of approximately 25 cubic millimeters was accurately segmented, resulting in a clearance percentage of 94.8%. However, the report only described it as "highly efficient clearance." Time allocation data showed the laser lithotripsy phase lasted 18 minutes, but frequent instrument changes indicated that the fragment clearance phase accounted for too high a proportion of the time, due to the dispersed small stone particles requiring multiple uses of the stone basket. The complication timeline recorded a brief bleeding event that occurred when the laser power was too high and hit a small blood vessel. The surgeon immediately reduced the power and flushed the area. The event lasted 20 seconds and was marked as a minor event on the timeline. The integrated quality control report highlighted the good clearance rate but recommended optimizing the fragment clearance strategy to reduce instrument changes. The report was used for the department's monthly quality control meeting.

[0098] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0099] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-dimensional management system based on clinical data from kidney stone surgery, characterized in that, The system includes: The preoperative data integration module is used to receive and integrate imaging examination data of kidney stone patients from different medical devices, spatially register CT scan results, ultrasound images and X-ray films, and generate a three-dimensional stone distribution model. Based on the three-dimensional stone distribution model, the surgical risk assessment module extracts the stone volume, density, and location parameters, and calculates the surgical difficulty coefficient and the probability of potential complications by combining the patient's physiological indicators and historical surgical data. The intraoperative real-time monitoring module is used to collect endoscopic video streams, laser lithotripsy parameters, and irrigation fluid pressure data during the operation, and to update the stone fragmentation status and residual volume simultaneously. Based on the stone fragmentation state and residual volume, the tissue damage early warning module analyzes the laser energy distribution and contact time with kidney tissue, assesses the thermal damage risk area, and generates a safe operating boundary. The postoperative recovery prediction module integrates surgical time, lithotripsy efficiency, and complication occurrence, and combines these with the patient's individual metabolic characteristics to predict the stone clearance rate and the progress of kidney function recovery. The detection process of the intraoperative real-time monitoring module includes: Inter-frame differential processing is performed on the endoscopic video stream to identify the movement trajectory of the laser fiber tip and the dynamics of stone fragmentation; the pulse frequency and single-shot energy parameters output by the laser generator are extracted to calculate the total amount of stone fragments per unit time; the fluctuation curve of the irrigation fluid pressure sensor is monitored to estimate the trend of intrarenal pelvis pressure change and irrigation efficiency; and the stone residual volume distribution map is dynamically updated based on the total amount of stone fragments and the trend of intrarenal pelvis pressure change. The tissue damage early warning module is implemented in the following ways: A laser energy attenuation model in the renal parenchyma was established to calculate the cumulative heat absorption value of tissues at different depths. Based on the rate of tissue color change in the endoscopic field of view, the local temperature rise gradient was inferred. Combining the cumulative heat absorption value and the local temperature rise gradient, an isothermal injury risk contour map was drawn. Based on the isothermal injury risk contour map, operational recommendations including safe power thresholds and treatment time were generated.

2. The multi-dimensional management system based on clinical data from kidney stone surgery according to claim 1, characterized in that, The generation process of the three-dimensional stone distribution model is as follows: Threshold segmentation is performed on the CT scan results to extract the stone pixel set in different density ranges and establish an initial stone volume framework. The echo intensity data in the ultrasound image is mapped to the initial stone volume frame to supplement information on calcified areas and soft tissue boundaries. The contour of the supplemented stone volume frame is corrected by using two-dimensional projection data from X-ray films to eliminate artifact interference caused by image overlap. By fusing spatial coordinate parameters from multimodal images, a three-dimensional stone distribution model including density gradient and anatomical location is reconstructed.

3. The multi-dimensional management system based on clinical data from kidney stone surgery according to claim 2, characterized in that, The calculation process for the surgical difficulty coefficient is as follows: The maximum transverse diameter of the stones and the branch angle of the aggregate system are extracted from the three-dimensional stone distribution model to calculate the physical obstacle removal index. Analyze the peak differences in the stone density distribution curve to determine the compositional heterogeneity score; Assess the degree of limitation in instrument operation space by combining data on urinary tract anatomy variations and previous surgical records; By combining the physical obstacle removal index, component heterogeneity score, and instrument operation space limitation, a quantitative surgical difficulty coefficient is generated.

4. The multi-dimensional management system based on clinical data from kidney stone surgery according to claim 3, characterized in that, The working process of the postoperative recovery prediction module is as follows: The total laser energy output and effective lithotripsy time during the operation were statistically analyzed, and the energy utilization efficiency parameters were calculated. Analyze the proportion of crystal components in immediate postoperative urine to assess the residual microstone burden; By combining the patient's 24-hour urinary metabolic indicators, an individualized stone recurrence risk curve was established; Based on the energy utilization efficiency parameters, residual microstone load, and stone recurrence risk curve, the degree of renal function recovery within three months is predicted.

5. The multi-dimensional management system based on clinical data from kidney stone surgery according to claim 4, characterized in that, Also includes: The data visualization interaction module is used to overlay the three-dimensional stone distribution model with real-time surgical data, supporting multi-planar reconstruction and virtual profile viewing; The surgical procedure optimization module provides suggestions on instrument selection and operation path for the current surgery based on a database of similar historical cases. The quality control indicator generation module automatically calculates key performance indicators for surgery, including stone clearance rate and lithotripsy volume per unit time.

6. The multi-dimensional management system based on clinical data from kidney stone surgery according to claim 5, characterized in that, The specific working process of the data visualization interaction module is as follows: Establish a three-dimensional coordinate system for the surgical scene and perform real-time spatial registration between the endoscopic field of view and the preoperative images; Mark the area of ​​residual stones and important vascular and nerve structures in the registered 3D view; It provides a virtual laser path simulation function to predict the energy distribution under different incident angles; It allows surgeons to customize the sectional view and simultaneously displays the anatomical relationship in both transverse and coronal views.

7. The multi-dimensional management system based on clinical data from kidney stone surgery according to claim 6, characterized in that, The working process of the surgical plan optimization module is as follows: Based on the current characteristics of the stones, multi-parameter similarity matching is performed in the case database to screen comparable surgical records; Extract laser setting parameters and surgical path data from matched cases to generate a set of suggested operating parameters; Analyze the types and management methods of complications in comparable surgical records to develop risk response plans; The recommended set of operating parameters is adjusted in a personalized manner based on the current anatomical variations of the patients.

8. The multi-dimensional management system based on clinical data from kidney stone surgery according to claim 7, characterized in that, The working process of the quality control index generation module is as follows: Compare the preoperative three-dimensional stone distribution model with the postoperative imaging results to calculate the percentage of volume clearance. Analyze the time allocation data for each stage of the surgery to evaluate the efficiency of the operational process; Record the time points of complication occurrences and treatment measures to generate a safety event timeline; The volume clearance percentage, operational efficiency, and safety event timeline are integrated to form a standardized quality control report.

Images