Liver cancer capillary density evaluation system based on CD34 and GPC-3 combined marker

Abstract

The invention provides a liver cancer capillary density evaluation system based on a CD34 and GPC-3 combined marker. The liver cancer microvessel density evaluation system based on the CD34 and GPC-3 combined marker comprises a digital pathology scanning module, a tumor area recognition module, a microvessel feature extraction module, a spatial correlation analysis module, a molecular subtype integration module, a microvessel invasion evaluation module, a prognosis prediction module, a dynamic layering module, a report generation module and a treatment recommendation module. The liver cancer microvessel density evaluation system based on the CD34 and GPC-3 combined marker has the advantages that the limitation of single quantification of microvessels is broken through, multi-dimensional accurate analysis of the liver cancer microenvironment is achieved through blood vessel function typing and spatial topology modeling, and conversion from pathological diagnosis to individualized treatment is promoted.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a liver cancer microvessel density assessment system based on combined labeling of CD34 and GPC-3. Background Art

[0002] Pathological assessment of liver cancer has long been challenged by insufficient resolution of microenvironmental heterogeneity. Traditional methods rely on the single marker CD34 to detect microvessel density (MVD), which cannot distinguish between tumor-specific and stromal vessels and overlooks key features such as vascular maturity and spatial distribution. The commonly used TNM staging system has a 35% miss rate for microvascular invasion (MVI), leading to misjudgment of postoperative recurrence risk in 30% of patients.

[0003] Therefore, it is necessary to provide a liver cancer microvessel density assessment system based on the combined labeling of CD34 and GPC-3 to solve the above technical problems. Summary of the Invention

[0004] The technical problem solved by the present invention is to provide a liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling, which breaks through the limitations of single microvascular quantification, realizes multi-dimensional and accurate analysis of the liver cancer microenvironment through vascular function classification and spatial topological modeling, and promotes the transformation of pathological diagnosis to personalized treatment.

[0005] To solve the above technical problems, the present invention provides a liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling, comprising: a digital pathology scanning module: for obtaining CD34 and GPC-3 double-stained digital images of liver cancer tissue sections; Tumor region identification module: used to divide the tumor core area and the invasion front area based on the GPC-3 expression intensity threshold; Microvascular feature extraction module: used to identify CD34-positive microvessels in the tumor area and calculate microvessel density and vascular maturity index; Spatial correlation analysis module: used to generate topological heat maps of GPC-3 expression intensity and CD34+ blood vessel distribution; Molecular subtype integration module: used to receive additional marker data and classify liver cancer subtypes; Microvascular invasion assessment module: used to detect GPC-3+ cancer cell clusters within CD34+ vessels within a 1 cm area adjacent to the cancer; Prognostic prediction module: used to calculate the invasion risk score based on the formula: RISK = (microvessel density × 0.3) + (vascular maturity index × (-0.5)) + (number of microvascular invasions × 0.8); Dynamic stratification module: used to divide patients into low-risk group, medium-risk group and high-risk group according to RISK value; Report generation module: used to output pathology reports containing topological heat maps and layering recommendations; Treatment recommendation module: used to generate targeted treatment recommendations based on stratified results.

[0006] Preferably, the tumor region recognition module adopts a U-Net deep learning model, and the training data includes labeled samples of GPC-3 strong positive, weak positive and negative.

[0007] Preferably, the microvascular feature extraction module distinguishes three types of vascular morphologies: Type I vessels: diameter >15 μm, complete lumen; Type II vessels: diameter ≤15 μm, branches ≥3; Type III vessels: discontinuous lumen segments.

[0008] Preferably, the vascular maturity index calculation formula is: Vascular maturity index = number of type I vessels / (number of type I vessels + number of type II vessels + number of type III vessels).

[0009] Preferably, when the spatial correlation analysis module generates a two-dimensional topological heat map, a red gradient is used to represent a GPC-3 high expression area, and a blue point cluster is used to represent a CD34+ blood vessel dense area.

[0010] Preferably, the subtypes defined by the molecular subtype integration module include: GPC-3+ / β-catenin nuclear+ type, GPC-3+ / CK19+ type and GPC-3 single positive type.

[0011] Preferably, for GPC-3+ / CK19+ liver cancer, a risk coefficient of 0.5 is automatically added to the prognosis prediction module.

[0012] Preferably, the microvascular invasion assessment module adopts a sliding window scanning method for the adjacent cancerous area, and the window size is 1 mm×1 mm.

[0013] Preferably, the recommendation strategy of the treatment recommendation module is: Low-risk group: surgical resection is recommended; Intermediate-risk group: surgery + anti-angiogenic therapy is recommended; High-risk group: Neoadjuvant therapy + liver transplantation evaluation is recommended.

[0014] Preferably, the pathology report output by the report generation module includes a three-dimensional reconstructed blood vessel-tumor interaction model, showing the spatial penetration trajectory of CD34+ blood vessels in the GPC-3+ tumor area.

[0015] Compared with related technologies, the liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling provided by the present invention has the following beneficial effects: The present invention provides a liver cancer microvessel density assessment system based on the combined labeling of CD34 and GPC-3. The system realizes tumor-specific microvessel identification through CD34 / GPC-3 dual-labeling spatial decoupling technology, overcoming the limitation of traditional single-labeling methods that cannot distinguish between tumor blood vessels and stromal blood vessels; establishes a vascular function classification system, and combines the dynamic maturation index (VMI) to quantify the stability and functional status of vascular structure; analyzes the spatial coupling pattern between GPC-3 high-expression areas and CD34+ blood vessels based on spatial topological heat maps, which can reveal the angiogenesis pattern of the core area of ​​tumor proliferation; integrates molecular subtype characteristics through a multimodal risk warning model, which can achieve three-level precise stratification of recurrence risk; and finally generates a visual report containing three-dimensional blood vessel-tumor infiltration trajectories to guide clinicians in formulating personalized treatment decision chains, promoting the leap of liver cancer pathology assessment from a single quantitative indicator to microenvironment system functional analysis, and significantly improving the accuracy of diagnosis and treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of the liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling provided by the present invention; Figure 2 The immunohistochemical image of GPC-3 in hepatocellular carcinoma in the liver cancer microvessel density assessment system based on combined labeling of CD34 and GPC-3 provided by the present invention; Figure 3 This is the CD34 immunohistochemistry image of hepatocellular carcinoma in the liver cancer microvessel density assessment system based on the combined labeling of CD34 and GPC-3 provided by the present invention; Figure 4 The immunohistochemical image of β-catenin in hepatocellular carcinoma in the liver cancer microvessel density assessment system based on combined CD34 and GPC-3 labeling provided by the present invention; Figure 5 This is the immunohistochemistry image of CK19 in hepatocellular carcinoma in the liver cancer microvessel density assessment system based on the combined labeling of CD34 and GPC-3 provided by the present invention. DETAILED DESCRIPTION

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] Please refer to Figure 1-Figure 5 ,in, Figure 1 This is a block diagram of the principle of the liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling provided by the present invention; Figure 2 The immunohistochemical image of GPC-3 in hepatocellular carcinoma in the liver cancer microvessel density assessment system based on combined labeling of CD34 and GPC-3 provided by the present invention; Figure 3This is the CD34 immunohistochemistry image of hepatocellular carcinoma in the liver cancer microvessel density assessment system based on the combined labeling of CD34 and GPC-3 provided by the present invention; Figure 4 The immunohistochemical image of β-catenin in hepatocellular carcinoma in the liver cancer microvessel density assessment system based on combined CD34 and GPC-3 labeling provided by the present invention; Figure 5 This is a CK19 immunohistochemical image of hepatocellular carcinoma in the liver cancer microvessel density assessment system based on the combined labeling of CD34 and GPC-3. The liver cancer microvessel density assessment system based on the combined labeling of CD34 and GPC-3 includes: a digital pathology scanning module: used to obtain CD34 and GPC-3 double-stained digital images of liver cancer tissue sections; Tumor region identification module: Used to divide the tumor core area and the invasion front area based on the GPC-3 expression intensity threshold. The tumor region identification module uses the U-Net deep learning model. The training data includes labeled samples of GPC-3 strong positive (H-score>150), weak positive (50≤H-score≤150), and negative (H-score<50).

[0019] Microvascular feature extraction module: used to identify CD34-positive microvessels within the tumor region and calculate: Microvessel density (MVD) = number of CD34+ vessels / mm²; Vascular maturity index (VMI) = (number of vessels with a diameter > 10 μm) / total number of vessels; The microvascular feature extraction module distinguishes three types of vascular morphologies: Type I blood vessels (mature type): diameter >15 μm, complete lumen; Type II vessels (immature): diameter ≤ 15 μm, branches ≥ 3; Type III vessels (fragmented): discontinuous lumen fragments; The vascular maturity index (VMI) calculation formula is: VMI=number of type I vessels / (number of type I vessels + number of type II vessels + number of type III vessels).

[0020] Spatial association analysis module: used to generate a topological heat map of GPC-3 expression intensity and CD34+ blood vessel distribution. When the spatial association analysis module generates a two-dimensional topological heat map, a red gradient represents a high GPC-3 expression area, and a blue point cluster represents a CD34+ blood vessel dense area.

[0021] Molecular subtype integration module: used to receive additional marker data such as β-catenin and CK19 to classify liver cancer subtypes. The subtypes defined by the molecular subtype integration module include: GPC-3+ / β-catenin nuclear+ type, GPC-3+ / CK19+ type and GPC-3 single positive type.

[0022] For GPC-3+ / CK19+ liver cancer, a risk factor of 0.5 is automatically added to the prognosis prediction module.

[0023] Microvascular invasion assessment module: used to detect GPC-3+ cancer cell clusters in CD34+ vessels within a 1 cm area adjacent to the cancer. The microvascular invasion assessment module uses a sliding window scanning method for the adjacent cancer area with a window size of 1 mm × 1 mm.

[0024] Prognostic prediction module: used to calculate the invasion risk score based on the formula: RISK = (MVD × 0.3) + (VMI × (-0.5)) + (number of microvascular invasions × 0.8) Dynamic stratification module: used to divide according to RISK value: low risk group (RISK≤2.0), medium risk group (2.0<RISK≤4.0)和高危组(RISK> 4.0).

[0025] Report generation module: used to output a pathology report containing a topological heat map and stratification suggestions. The pathology report output by the report generation module includes a three-dimensional reconstructed vascular-tumor interaction model, showing the spatial penetration trajectory of CD34+ blood vessels in the GPC-3+ tumor area.

[0026] Treatment recommendation module: used to generate targeted treatment recommendations based on stratification results. The recommendation strategy of the treatment recommendation module is: Low-risk group: surgical resection is recommended; Intermediate-risk group: surgery + anti-angiogenic therapy is recommended; High-risk group: Neoadjuvant therapy + liver transplantation evaluation is recommended.

[0027] The present invention also provides a method for evaluating liver cancer microvessel density based on combined CD34 and GPC-3 labeling, comprising the following steps: S1: Use the digital pathology scanning module to obtain double-stained digital images: Serial sections of liver cancer tissue were prepared for double immunohistochemical staining of CD34 and GPC-3. Digital images were acquired at 0.5 μm / pixel using a high-resolution whole-slide scanner. Background correction and illumination homogenization were performed on the images, and color deconvolution was used to precisely separate the CD34 (brown) and GPC-3 (red) staining channels.

[0028] S2: Use the tumor region recognition module to accurately segment the tumor region: In the GPC-3 staining channel, expression intensity was assessed using the H-score quantitative system: H-score = (3 × strong positive proportion) + (2 × moderate positive proportion) + (1 × weak positive proportion). Tumor regions were categorized into three levels based on the H-score: core zone (H-score ≥ 200, high expression in the central tumor), transition zone (100 ≤ H-score < 200, moderate expression at the tumor margin), and negative zone (H-score < 100, GPC-3 non-expressing). A U-Net deep learning model was used to refine tumor boundaries at the subcellular level and identify isolated clusters of GPC-3-positive cancer cells.

[0029] S3: Quantify microvascular features using the microvascular feature extraction module: In the CD34 staining channel, a multiscale LoG filter was used to detect vascular lumen structure, identifying closed or semi-closed ring structures with a diameter of 5-50 μm. Detected microvessels were categorized and quantified: those with a diameter greater than 15 μm and a lumen integrity of ≥90% were classified as type I (mature vessels); those with a diameter ≤15 μm and ≥3 branches were classified as type II (immature vessels); and the remainder were classified as type III (fragmented vessels). Microvessel density (MVD) was calculated by dividing the total number of type I, II, and III vessels by the effective tissue area (unit: cells / mm2). The vascular maturation index (VMI) was calculated by dividing the number of type I vessels by the total number of vessels.

[0030] S4: Use the spatial correlation analysis module to perform spatial co-localization analysis: Generate a two-dimensional spatial heat map: the X / Y axes represent the spatial coordinates of the tissue section, the red gradient indicates GPC-3 expression intensity (dark red corresponds to strongly positive areas), and the blue dot matrix represents the distribution of CD34-positive blood vessels (dot density reflects local microvascular density). Calculate the spatial colocalization index (CI): the number of CD34-positive blood vessels within the GPC-3-positive area divided by the total number of CD34-positive blood vessels in the entire section.

[0031] S5: Use the molecular subtype integration module to create a molecular subtype integration: Auxiliary molecular markers were introduced: nuclear β-catenin positivity indicates Wnt pathway activation, and cytoplasmic CK19 positivity indicates stem cell characteristics. Subtype classification rules were established: ① Proliferative subtype requires simultaneous GPC-3 positivity and nuclear β-catenin positivity with a CI > 0.6; ② Invasive subtype requires GPC-3 positivity, CK19 positivity, and a proportion of type III vessels > 40%; ③ Common subtype is GPC-3 positivity alone with no other specific markers.

[0032] S6: Microvascular invasion detection using the Microvascular Invasion Assessment Module: Scanning was performed using a 10×10 sliding window within a 1 cm radius outside the cancer tissue boundary. Positive results were determined by the presence of ≥3 GPC-3-positive cancer cell clusters within CD34-positive vessels, with direct contact between the cancer cell clusters and the vascular endothelium. The density of invasion foci (unit: cells / cm²) and the maximum invasion depth were reported.

[0033] S7: Dynamic prognostic stratification through the prognostic prediction module and the dynamic stratification module: The comprehensive risk score (RISK) was calculated as follows: RISK = (MVD×0.3) + (VMI×(-0.5)) + (microvascular invasion density×0.8) + subtype coefficient (proliferative type +0.7, invasive type +1.2). Three-level risk stratification was performed: low-risk group (RISK≤2.0 and type I vascular ratio>60%); intermediate-risk group (2.0<RISK≤4.0或Ⅱ型血管占比> 50%); high-risk group (RISK>4.0 or type III vessel proportion>30% or invasion density>5 / cm2).

[0034] S8: Generate clinical decisions through report generation module and treatment recommendation module: The generated visualization report includes: a dual-labeled spatial heat map (showing the spatial infiltration trajectory of blood vessels and tumors); a 3D reconstruction of the microvascular network (labeling different vessel types); and a risk radar chart (displaying four dimensions: MVD, VMI, invasion density, and subtype coefficient). Treatment strategy mapping: Surgical resection is recommended for the low-risk group (predicted R0 resection rate >90%); surgical resection combined with postoperative adjuvant lenvatinib therapy (recommended duration: 6 months) is recommended for the intermediate-risk group; and neoadjuvant therapy (TACE combined with bevacizumab) followed by evaluation for liver transplantation is recommended for the high-risk group.

[0035] S9: Quality Control and Iterative Optimization: Credibility verification was implemented: vascular typing results in 10% of high-risk areas were manually reviewed and compared with intraoperative indocyanine green fluorescence imaging results. A model update mechanism was established: when the pathological staging error rate of newly received cases exceeded 15%, a retraining process was triggered. Incremental learning technology was used to integrate new data and continuously optimize the tumor segmentation boundary accuracy of the U-Net model.

[0036] Compared with related technologies, the liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling provided by the present invention has the following beneficial effects: The present invention provides a liver cancer microvessel density assessment system based on the combined labeling of CD34 and GPC-3. The system realizes tumor-specific microvessel identification through CD34 / GPC-3 dual-labeling spatial decoupling technology, overcoming the limitation of traditional single-labeling methods that cannot distinguish between tumor blood vessels and stromal blood vessels; establishes a vascular function classification system, and combines the dynamic maturation index (VMI) to quantify the stability and functional status of vascular structure; analyzes the spatial coupling pattern between GPC-3 high-expression areas and CD34+ blood vessels based on spatial topological heat maps, which can reveal the angiogenesis pattern of the core area of ​​tumor proliferation; integrates molecular subtype characteristics through a multimodal risk warning model, which can achieve three-level precise stratification of recurrence risk; and finally generates a visual report containing three-dimensional blood vessel-tumor infiltration trajectories to guide clinicians in formulating personalized treatment decision chains, promoting the leap of liver cancer pathology assessment from a single quantitative indicator to microenvironment system functional analysis, and significantly improving the accuracy of diagnosis and treatment.

[0037] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A liver cancer microvessel density assessment system based on combined CD34 and GPC-3 labeling, characterized in that: include: Digital pathology scanning module: used to obtain CD34 and GPC-3 double-stained digital images of liver cancer tissue sections; Tumor region identification module: used to divide the tumor core area and the invasion front area based on the GPC-3 expression intensity threshold; Microvascular feature extraction module: used to identify CD34-positive microvessels in the tumor area and calculate microvessel density and vascular maturity index; Spatial correlation analysis module: used to generate topological heat maps of GPC-3 expression intensity and CD34+ blood vessel distribution; Molecular subtype integration module: used to receive additional marker data and classify liver cancer subtypes; Microvascular invasion assessment module: used to detect GPC-3+ cancer cell clusters within CD34+ vessels within a 1 cm area adjacent to the cancer; Prognostic prediction module: used to calculate the invasion risk score based on the formula: RISK = (microvessel density × 0.3) + (vascular maturity index × (-0.5)) + (number of microvascular invasions × 0.8); Dynamic stratification module: used to divide patients into low-risk group, medium-risk group and high-risk group according to RISK value; Report generation module: used to output pathology reports containing topological heat maps and layering recommendations; Treatment recommendation module: used to generate targeted treatment recommendations based on stratified results.

2. The liver cancer microvessel density assessment system based on combined CD34 and GPC-3 labeling according to claim 1, characterized in that: The tumor region recognition module adopts the U-Net deep learning model, and the training data includes GPC-3 strongly positive, weakly positive and negative labeled samples.

3. The liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling according to claim 1, characterized in that: The microvascular feature extraction module distinguishes three types of vascular morphologies: Type I vessels: diameter >15 μm, complete lumen; Type II vessels: diameter ≤15 μm, branches ≥3; Type III vessels: discontinuous lumen segments.

4. The liver cancer microvessel density assessment system based on combined CD34 and GPC-3 labeling according to claim 3, characterized in that: The vascular maturity index calculation formula is: Vascular maturity index = number of type I vessels / (number of type I vessels + number of type II vessels + number of type III vessels).

5. The liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling according to claim 1, characterized in that: When the spatial correlation analysis module generates a two-dimensional topological heat map, a red gradient is used to represent a GPC-3 high expression area, and a blue point cluster is used to represent a CD34+ blood vessel dense area.

6. The liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling according to claim 1, characterized in that: The subtypes defined by the molecular subtype integration module include: GPC-3+ / β-catenin nuclear+ type, GPC-3+ / CK19+ type and GPC-3 single positive type.

7. The liver cancer microvessel density assessment system based on combined CD34 and GPC-3 labeling according to claim 6, characterized in that: For GPC-3+ / CK19+ liver cancer, a risk factor of 0.5 is automatically added to the prognosis prediction module.

8. The liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling according to claim 1, characterized in that: The microvascular invasion assessment module adopts a sliding window scanning method for the adjacent cancerous area, and the window size is 1 mm×1 mm.

9. The liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling according to claim 1, characterized in that: The recommendation strategy of the treatment recommendation module is: Low-risk group: surgical resection is recommended; Intermediate-risk group: surgery + anti-angiogenic therapy is recommended; High-risk group: Neoadjuvant therapy + liver transplantation evaluation is recommended.

10. The liver cancer microvessel density assessment system based on CD34 and GPC-3 combined labeling according to claim 1, characterized in that: The pathology report output by the report generation module includes a three-dimensional reconstructed blood vessel-tumor interaction model, showing the spatial penetration trajectory of CD34+ blood vessels in the GPC-3+ tumor area.

Images