A method and system for preoperative risk scoring of extremity, trunk soft tissue sarcomas

Abstract

The application relates to a preoperative risk scoring method for limb and trunk soft tissue sarcoma, which comprises the following steps: obtaining standardized MRI features of a patient, and classifying the standardized MRI features; and inputting the classification results of the standardized MRI features into a scoring model to obtain a preoperative risk score of the patient. The application can accurately evaluate the risk of limb and trunk soft tissue sarcoma non-invasively before operation by converting complex image information into a simple and understandable risk score through unified MRI feature definition and objective weight allocation, so as to provide a precise solution for preoperative planning and postoperative management of the patient.

Description

Technical Field

[0001] This invention relates to the field of medical imaging technology, and in particular to a method and system for preoperative risk scoring of soft tissue sarcomas of the limbs and trunk. Background Technology

[0002] Soft tissue sarcoma (STS) is a heterogeneous group of malignant tumors originating from mesenchymal tissue, accounting for approximately 1% of all malignant tumors, with tens of thousands of new cases worldwide each year. Due to its high heterogeneity, strong local invasiveness, high recurrence rate, and distant metastasis, the clinical treatment and prognostic assessment of STS have always been challenging in the field of oncology. Accurate preoperative risk assessment is crucial for developing individualized treatment plans, selecting the extent of surgery, determining the need for neoadjuvant chemoradiotherapy, and predicting patient prognosis.

[0003] Currently, preoperative risk assessment for soft tissue sarcomas primarily relies on histopathological grading, with the French National Cancer Centers Committee (FNCLCC) grading system and the National Cancer Institute (NCI) grading system being the most commonly used. These systems are mainly based on morphological indicators such as tumor differentiation, mitotic count, and tumor necrosis rate, and are widely regarded as the "gold standard" for assessing the malignancy and metastatic risk of STS. However, traditional grading systems have significant limitations: First, the assessment process is highly subjective, with poor consistency among different pathologists, especially regarding the prognosis of intermediate-grade (G2) tumors, which presents considerable uncertainty; second, grading depends on postoperative specimens and cannot provide preoperative risk assessment; third, the sample size of core biopsy is limited, which may not fully reflect tumor heterogeneity, especially when assessing necrotic and proliferating areas, which can easily lead to bias.

[0004] To address the limitations of traditional morphological grading, recent research has focused on identifying more objective and reproducible biomarkers. Among these, the cell proliferation marker Ki-67 and the hypoxia-associated marker hypoxia-inducible factor-1α (HIF-1α) have been shown to be closely related to the histological grading and prognosis of STS. However, the assessment of Ki-67 and HIF-1α also relies on immunohistochemical testing of tissue specimens, which is an invasive procedure and cannot be obtained dynamically and non-invasively preoperatively. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for preoperative risk assessment of soft tissue sarcomas of the limbs and trunk, which can accurately assess the risk of soft tissue sarcomas of the limbs and trunk without surgery.

[0006] The technical solution adopted by this invention to solve its technical problem is: to provide a preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk, including:

[0007] Acquire standardized MRI features of the patient and classify the standardized MRI features;

[0008] The classification results of the standardized MRI features are input into the scoring model to obtain the patient's preoperative risk score; the scoring model is obtained in the following way:

[0009] Collect historical patient medical record datasets and extract standardized MRI features and postoperative progression of historical patients;

[0010] Using FNCLCC high-grade pathological markers, Ki-67 high-expression pathological markers, and HIF-1α high-expression pathological markers as dependent variables and standardized MRI features as independent variables, multivariate logistic regression analysis was performed using standardized MRI features from historical patients to screen out standardized MRI features that were independently associated with each pathological marker, and the non-standardized regression coefficients of each screened standardized MRI feature in the three logistic regression models were recorded.

[0011] Kendall's rank correlation coefficients were calculated between high-grade FNCLCC, high expression of cell proliferation marker Ki-67, high expression of hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients.

[0012] Based on each selected standardized MRI feature, its unstandardized regression coefficient in the three logistic regression models is used to calculate the fixed weight of each selected standardized MRI feature with the Kendall rank correlation coefficient of the corresponding pathological index.

[0013] The feature-weighted scoring model is constructed based on the fixed weights of each selected standardized MRI feature.

[0014] The standardized MRI features include maximum tumor diameter, tumor location, area of ​​tumor necrosis, peritumoral edema, tumor heterogeneity, invasion of important tissues around the tumor, peritumoral enhancement, and ADC value measurement.

[0015] The maximum diameter of the tumor refers to the maximum diameter length of the tumor measured on at least two orthogonal planes or isotropic three-dimensional images;

[0016] The tumor location is divided into superficial and deep locations. The superficial location refers to a tumor that is completely confined to subcutaneous adipose tissue and has not invaded the deep fascia. The deep location refers to a tumor that originates from or has invaded tissues below the deep fascia.

[0017] The tumor necrosis area refers to the region within the tumor that shows no enhancement on the enhanced T1-weighted image and a high signal on the T2-weighted image.

[0018] The peritumoral edema refers to a non-tumor, identifiable, high-signal region around the tumor that can be distinguished from the tumor boundary on T2-weighted fat-suppressed sequences or STIR sequences.

[0019] Tumor heterogeneity refers to the simultaneous presence of heterogeneous regions with low, medium, and high signal intensity within a tumor in T2-weighted images.

[0020] The invasion of important tissues around the tumor refers to the relationship between the tumor and adjacent key structures. A central angle corresponding to the arc length of the contact between the tumor and the nerve or blood vessel exceeding 180° is defined as nerve / blood vessel encirclement. Bone invasion is defined as the interruption of the continuity of the adjacent bone cortex or abnormal bone marrow signal on T2-weighted and enhanced T1-weighted images.

[0021] Peritumoral enhancement refers to the presence of a clearly defined band-like or nodular enhancement region that extends beyond the main boundary of the tumor on enhanced T1-weighted images during the arterial or portal venous phases.

[0022] The ADC value measurement refers to combining T2WI and enhanced T1WI images, avoiding hemorrhage, necrosis and cystic areas, delineating a circular or elliptical region of interest on the DWI image at the point of highest signal in the solid part of the tumor, and obtaining the average ADC value of the region of interest on the corresponding ADC image.

[0023] The classification of the standardized MRI features specifically involves:

[0024] Based on clinical criteria, the maximum diameter of a tumor is classified into three categories.

[0025] Tumors are classified into two categories based on their location: superficial and deep.

[0026] For the area of ​​tumor necrosis, the percentage of the area of ​​the tumor that shows no enhancement on the T1-weighted image after enhancement and is a high signal on the T2-weighted image is calculated, and the area is divided into two categories according to the percentage.

[0027] Peritumoral edema is classified into three categories based on the size of the high signal area;

[0028] Regarding tumor heterogeneity, it is divided into two categories based on the size of the heterogeneous regions;

[0029] In cases of invasion of important tissues surrounding the tumor, it is divided into two categories based on whether there is nerve / vascular encapsulation or bone invasion;

[0030] Regarding peritumoral enhancement, it is divided into two categories based on the presence or absence of band-like or nodular enhancement areas;

[0031] For ADC value measurement, it is divided into two categories based on the preset cutoff value.

[0032] The fixed weights of the standardized MRI features are obtained through The calculation yielded that, For the first Fixed weights for standardized MRI features, , and The first The unstandardized regression coefficients of standardized MRI features in three logistic regression models , and Kendall's rank correlation coefficients were used to correlate high-grade FNCLCC, high expression of the cell proliferation marker Ki-67, high expression of the hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients.

[0033] The technical solution adopted by this invention to solve its technical problem is: to provide a preoperative risk scoring system for soft tissue sarcomas of the limbs and trunk, comprising:

[0034] An acquisition module is used to acquire standardized MRI features of the patient and classify the standardized MRI features;

[0035] The scoring module is used to input the classification results of the standardized MRI features into the scoring model to obtain the patient's preoperative risk score; the scoring model is obtained in the following way:

[0036] Collect historical patient medical record datasets and extract standardized MRI features and postoperative progression of historical patients;

[0037] Using FNCLCC high-grade pathological markers, Ki-67 high-expression pathological markers, and HIF-1α high-expression pathological markers as dependent variables and standardized MRI features as independent variables, multivariate logistic regression analysis was performed using standardized MRI features from historical patients to screen out standardized MRI features that were independently associated with each pathological marker, and the non-standardized regression coefficients of each screened standardized MRI feature in the three logistic regression models were recorded.

[0038] Kendall's rank correlation coefficients were calculated between high-grade FNCLCC, high expression of cell proliferation marker Ki-67, high expression of hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients.

[0039] Based on each selected standardized MRI feature, its unstandardized regression coefficient in the three logistic regression models is used to calculate the fixed weight of each selected standardized MRI feature with the Kendall rank correlation coefficient of the corresponding pathological index.

[0040] The feature-weighted scoring model is constructed based on the fixed weights of each selected standardized MRI feature.

[0041] The standardized MRI features include maximum tumor diameter, tumor location, area of ​​tumor necrosis, peritumoral edema, tumor heterogeneity, invasion of important tissues around the tumor, peritumoral enhancement, and ADC value measurement.

[0042] The maximum diameter of the tumor refers to the maximum diameter length of the tumor measured on at least two orthogonal planes or isotropic three-dimensional images;

[0043] The tumor location is divided into superficial and deep locations. The superficial location refers to a tumor that is completely confined to subcutaneous adipose tissue and has not invaded the deep fascia. The deep location refers to a tumor that originates from or has invaded tissues below the deep fascia.

[0044] The tumor necrosis area refers to the region within the tumor that shows no enhancement on the enhanced T1-weighted image and a high signal on the T2-weighted image.

[0045] The peritumoral edema refers to a non-tumor, identifiable, high-signal region around the tumor that can be distinguished from the tumor boundary on T2-weighted fat-suppressed sequences or STIR sequences.

[0046] Tumor heterogeneity refers to the simultaneous presence of heterogeneous regions with low, medium, and high signal intensity within a tumor in T2-weighted images.

[0047] The invasion of important tissues around the tumor refers to the relationship between the tumor and adjacent key structures. A central angle corresponding to the arc length of the contact between the tumor and the nerve or blood vessel exceeding 180° is defined as nerve / blood vessel encirclement. Bone invasion is defined as the interruption of the continuity of the adjacent bone cortex or abnormal bone marrow signal on T2-weighted and enhanced T1-weighted images.

[0048] Peritumoral enhancement refers to the presence of a clearly defined band-like or nodular enhancement region that extends beyond the main boundary of the tumor on enhanced T1-weighted images during the arterial or portal venous phases.

[0049] The ADC value measurement refers to combining T2WI and enhanced T1WI images, avoiding hemorrhage, necrosis and cystic areas, delineating a circular or elliptical region of interest on the DWI image at the point of highest signal in the solid part of the tumor, and obtaining the average ADC value of the region of interest on the corresponding ADC image.

[0050] The classification of the standardized MRI features specifically involves:

[0051] Based on clinical criteria, the maximum diameter of a tumor is classified into three categories.

[0052] Tumors are classified into two categories based on their location: superficial and deep.

[0053] For the area of ​​tumor necrosis, the percentage of the area of ​​the tumor that shows no enhancement on the T1-weighted image after enhancement and is a high signal on the T2-weighted image is calculated, and the area is divided into two categories according to the percentage.

[0054] Peritumoral edema is classified into three categories based on the size of the high signal area;

[0055] Regarding tumor heterogeneity, it is divided into two categories based on the size of the heterogeneous regions;

[0056] In cases of invasion of important tissues surrounding the tumor, it is divided into two categories based on whether there is nerve / vascular encapsulation or bone invasion;

[0057] Regarding peritumoral enhancement, it is divided into two categories based on the presence or absence of band-like or nodular enhancement areas;

[0058] For ADC value measurement, it is divided into two categories based on the preset cutoff value.

[0059] The fixed weights of the standardized MRI features are obtained through The calculation yielded that, For the first Fixed weights for standardized MRI features, , and The first The unstandardized regression coefficients of standardized MRI features in three logistic regression models , and Kendall's rank correlation coefficients were used to correlate high-grade FNCLCC, high expression of the cell proliferation marker Ki-67, high expression of the hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients.

[0060] The technical solution adopted by the present invention to solve its technical problem is: to provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the steps of the above-mentioned preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk.

[0061] The technical solution adopted by the present invention to solve its technical problem is: to provide a computer-readable storage medium on which a computer program is stored, wherein when the computer program is executed by a processor, the steps of the above-mentioned method for preoperative risk scoring of soft tissue sarcomas of the limbs and trunk are implemented.

[0062] Beneficial effects

[0063] By employing the aforementioned technical solution, this invention offers the following advantages and positive effects compared to existing technologies: By integrating the independent correlation strength between imaging features and multiple gold-standard pathological indicators (FNCLCC classification, Ki-67, HIF-1α), this invention calibrates the prognostic weights of each pathological indicator, ensuring that each score in the final evaluation simultaneously contains both pathological discriminative power and clinical prognostic information, thereby guaranteeing the accuracy of the prediction. This invention advances risk assessment from postoperative pathological analysis to preoperative imaging evaluation, enabling clinicians to obtain objective risk quantification before formulating surgical plans or deciding whether to perform neoadjuvant therapy, thus promoting the individualization and precision of treatment strategies. Attached Figure Description

[0064] Figure 1 This is a flowchart of the preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk according to the first embodiment of the present invention;

[0065] Figure 2 This is a schematic diagram illustrating the standardization of tumor maximum diameter measurement in the first embodiment of the present invention;

[0066] Figure 3 This is a schematic diagram of standardized tumor location assessment in the first embodiment of the present invention;

[0067] Figure 4 This is a schematic diagram of the standardized assessment of tumor necrosis area in the first embodiment of the present invention;

[0068] Figure 5 This is a schematic diagram of the standardized assessment of T2-weighted peritumoral edema in the first embodiment of the present invention;

[0069] Figure 6 This is a schematic diagram of the standardization of tumor heterogeneity assessment in the first embodiment of the present invention;

[0070] Figure 7 This is a schematic diagram illustrating the standardized assessment of invasion of important tissues around a tumor in the first embodiment of the present invention;

[0071] Figure 8 This is a schematic diagram of the standardized peritumoral enhancement assessment of tumors in the first embodiment of the present invention;

[0072] Figure 9 This is a schematic diagram illustrating the standardization of tumor ADC value measurement in the first embodiment of the present invention;

[0073] Figure 10 These are MRI images of three patients in an embodiment of the present invention. Detailed Implementation

[0074] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0075] The first embodiment of the present invention relates to a preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk. This method converts multiple qualitative and quantitative MRI features extracted from preoperative MRI images into standardized MRI features for non-invasive prediction of tumor invasiveness (i.e., high-risk pathological features) and the risk of postoperative disease progression in patients.

[0076] To address the issues of ambiguous feature definitions and inconsistent quantification methods in existing technologies, this implementation first defines a set of features that include conventional MRI morphological features and ADC functional parameters. Clear interpretation criteria and quantification grading rules are then established for each feature, resulting in standardized MRI features, as detailed below:

[0077] (1) The maximum diameter of a tumor refers to the maximum diameter length of the tumor measured on at least two orthogonal planes (usually axial, coronal, or sagittal) or isotropic three-dimensional images. When classifying tumors, they can be divided into three categories according to commonly used clinical standards: <5 cm, 5-10 cm, and >10 cm. Figure 2 In the middle section, a is a lesion in the left forearm with a maximum tumor diameter of 4.75 cm; b is a lesion in the left shoulder with a maximum tumor diameter of 7.28 cm; and c is a lesion in the left thigh with a maximum tumor diameter of 11.0 cm.

[0078] (2) Tumor location is divided into superficial location and deep location. Superficial location is defined as a tumor that is completely confined to subcutaneous adipose tissue and has not invaded the deep fascia; deep location is defined as a tumor that originates from or has invaded tissues below the deep fascia (such as muscles and intermuscular spaces). When classifying, it is divided into two categories according to superficial location and deep location. Figure 3 In the middle image, a represents a mass in the right thigh, with the tumor located superficially under the skin; b represents a mass in the left thigh, with the tumor located deep within the soft tissue.

[0079] (3) Tumor necrosis area refers to the area within the tumor that shows no enhancement on the T1-weighted image after enhancement and has a high signal on the T2-weighted image. When classifying, it can be divided into two categories: <50% and ≥50% by calculating the percentage of this area to the total tumor area. Figure 4In the middle section, a is a lesion in the right thigh with a necrotic area ≥50%; b is a lesion beside the right elbow joint with a necrotic area <50%.

[0080] (4) Peritumoral edema refers to a non-tumor, identifiable high-signal area around the tumor that can be distinguished from the tumor boundary on T2-weighted fat-suppressed sequences or STIR sequences. When classifying, it can be divided into three categories: "none", "localized (range < 50% of the maximum diameter of the tumor)" and "extensive (range ≥ 50% of the maximum diameter of the tumor)". Figure 5 a) is a lesion on the right thigh with no edema around the lesion; b) is a lesion next to the left knee joint with localized edema around the lesion; c) is a lesion next to the right elbow joint with widespread edema around the lesion.

[0081] (5) Tumor heterogeneity refers to the simultaneous presence of heterogeneous regions with low, medium, and high signal intensity within a tumor on a T2-weighted image. When classifying tumors, the proportion of heterogeneous regions to the total tumor area can be calculated, classifying them into two categories: <50% and ≥50%. Figure 6 In the middle section, a is a lesion in the right forearm with tumor heterogeneity <50%; b is a lesion in the left thigh with tumor heterogeneity ≥50%.

[0082] (6) Invasion of important tissues around the tumor refers to the relationship between the tumor and adjacent key structures. A central angle exceeding 180° corresponding to the arc length of the contact between the tumor and a nerve or blood vessel is defined as nerve / blood vessel encirclement. Discontinuity of adjacent cortical bone or abnormal bone marrow signal on T2-weighted and enhanced T1-weighted images is defined as bone invasion. In classification, the presence of any of the above signs indicates "presence," otherwise it is considered "non-presence." Figure 7 In the middle image, a is a lesion in the left groin, where the central angle corresponding to the arc length of the contact between the tumor and the adjacent external iliac artery is less than 180°, indicating that the blood vessels have not been invaded; in the middle image, b is a lesion in the left thigh, where the central angle corresponding to the arc length of the contact between the tumor and the adjacent sciatic nerve is greater than 180°, indicating that the nerve has been invaded.

[0083] (7) Peritumoral enhancement refers to the presence of a clearly defined band-like or nodular enhancement area extending beyond the main tumor boundary on enhanced T1-weighted images during the arterial or portal venous phase. It is classified into two categories: "present" and "absent". Figure 8 In the middle image, a is a lesion located beside the left ankle joint, with no peritumoral enhancement; b is a lesion located on the left thigh, with peritumoral enhancement visible.

[0084] (8) ADC value measurement refers to drawing a circular or elliptical region of interest on the solid part of the tumor on the DWI image (b=800 s / mm²) at the point of highest signal in the solid part of the tumor, using a combination of T2WI and enhanced T1WI images, avoiding areas of hemorrhage, necrosis, and cystic degeneration. The average ADC value of this region of interest is then obtained on the corresponding ADC image. When classifying the tumor, the optimal cutoff value (1.2×10⁻³ mm² / s) for the ADC value can be determined based on the patient's postoperative disease progression as the clinical outcome, using the receiver operating characteristic curve. The tumor is then divided into two categories based on this optimal cutoff value. Figure 9 In the middle, a and b represent lesions on the left shoulder, with an ADC value of 0.67 × 10⁻³ mm² / s; c and d represent lesions on the right thigh, with an ADC value of 1.94 × 10⁻³ mm² / s.

[0085] Therefore, this implementation method performs standardized quantitative processing on each image feature, transforming vague qualitative descriptions into clear quantitative or semi-quantitative classifications, thus eliminating subjective arbitrariness from the source.

[0086] like Figure 1 As shown, the preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk in this embodiment includes:

[0087] Step 1: Obtain the patient's standardized MRI features and classify the standardized MRI features;

[0088] Step 2: Input the classification results of the standardized MRI features into the scoring model to obtain the patient's preoperative risk score.

[0089] The scoring model in this embodiment can be obtained in the following way:

[0090] In this implementation, tumors with any of the adverse pathological markers (FNCLCC grade II / III, Ki-67 labeling index ≥30%, or high expression of HIF-1α) are defined as "high-risk pathological features".

[0091] First, we collected historical patient medical record datasets and extracted standardized MRI features and postoperative progression of historical patients.

[0092] Next, using high-grade FNCLCC (G2 / 3), high Ki-67 expression (LI≥30%), and high HIF-1α expression as dependent variables, and all standardized MRI features as independent variables, multivariate logistic regression analysis was performed using standardized MRI features from historical patients. Standardized MRI features independently correlated with each pathological indicator were selected, and the non-standardized regression coefficients of each selected standardized MRI feature in the three logistic regression models were recorded, denoted as follows: , and .

[0093] Then, Kendall's rank correlation coefficients were calculated between high-grade FNCLCC, high expression of the cell proliferation marker Ki-67, high expression of the hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients, respectively, and denoted as follows: , and The Kendall rank correlation coefficient quantifies the influence of each pathological indicator on adverse clinical outcomes.

[0094] Next, based on each selected standardized MRI feature, its unstandardized regression coefficients in the three logistic regression models were compared with the Kendall's rank correlation coefficients of the corresponding pathological indicators to calculate the fixed weights of each selected standardized MRI feature. The fixed weights of the standardized MRI features can be determined by... The calculation yielded that, For the first Fixed weights for each standardized MRI feature.

[0095] Finally, a feature-weighted scoring model is constructed based on the fixed weights of each selected standardized MRI feature. The final scoring model can be expressed as: ,in, The output of the scoring model, For the first The classification results of each standardized MRI feature are assigned values, and the assignment table is shown in Table 1.

[0096]

[0097] After completing the above scoring, it is necessary to determine the optimal diagnostic threshold for the MRI score, i.e., a score above a certain threshold indicates a high-risk patient, and a score below a certain threshold indicates a low-risk patient. In this embodiment, the statistical analysis used to determine the MRI score threshold is ROC analysis, and the outcome is a high-risk pathological feature (any one of the following: high-grade FNCLCC, high Ki-67 expression, or high HIF-1α expression). The optimal diagnostic score threshold can be obtained through ROC analysis.

[0098] like Figure 10 As shown, it provides three case studies demonstrating that risk assessment can be performed using MRI scoring models.

[0099] Case 1: Female, 55 years old, with a mass in the left lower leg. The MRI features are as follows: (1) Maximum tumor diameter, 8 cm → 1 point; (2) Tumor location, deep → 1 point; (3) Tumor necrosis area < 50% → 0 points; (4) Peritumoral edema, none → 0 points; (5) Tumor heterogeneity, < 50% → 0 points; (6) Invasion of important tissues around the tumor, none → 0 points; (7) Peritumoral enhancement, none → 0 points; (8) ADC, 0.81×10⁻ 3 mm 2 / s → 5 points. The final score output by the scoring model was S_score = 7 points (high risk). Pathological examination confirmed myxoid fibrosarcoma, FNCLCC grade 2, low expression of Ki-67 and HIF-1α. The patient developed lung metastases 4 months after surgery.

[0100] Case 2: Male, 42 years old, with a mass in the left lower leg. The MRI features are as follows: (1) Maximum tumor diameter, 6cm → 1 point; (2) Tumor location, deep → 1 point; (3) Tumor necrosis area < 50% → 0 points; (4) Peritumoral edema, none → 0 points; (5) Tumor heterogeneity, < 50% → 0 points; (6) Invasion of important tissues around the tumor, none → 0 points; (7) Peritumoral enhancement, present → 3 points; (8) ADC, 2.23×10⁻ 3 mm 2 / s→0 points. The final score output by the scoring model was S_score = 5 points (low risk). Pathological examination confirmed it to be myxoid liposarcoma, FNCLCC grade I, with low expression of Ki-67 and HIF-1α. The patient had no metastasis, no recurrence, and no death 5 years after surgery.

[0101] Case 3: Female, 31 years old, with a mass on the left sole of her foot. The MRI features are as follows: (1) Maximum tumor diameter, 12cm → 2 points; (2) Tumor location, superficial → 0 points; (3) Tumor necrosis area < 50% → 0 points; (4) Peritumoral edema, none → 0 points; (5) Tumor heterogeneity, < 50% → 0 points; (6) Invasion of important tissues around the tumor, none → 0 points; (7) Peritumoral enhancement, none → 0 points; (8) ADC, 1.10×10⁻ 3 mm 2 / s →5 points. The final score output by the scoring model was S_score = 7 points (high risk). Pathology confirmed synovial sarcoma, FNCLCC grade III, high expression of Ki-67, low expression of HIF-1α. The patient experienced local recurrence 3 months after surgery and died 10 months later.

[0102] It is easy to see that this invention integrates the independent correlation strength between imaging features and multiple gold standard pathological indicators (FNCLCC classification, Ki-67, HIF-1α), calibrating the prognostic weight of each pathological indicator. This ensures that each score in the final evaluation simultaneously contains both pathological discriminative power and clinical prognostic information, thereby guaranteeing the accuracy of the prediction. This invention advances risk assessment from postoperative pathological analysis to preoperative imaging evaluation, enabling clinicians to obtain objective quantitative risk data before formulating surgical plans or deciding whether to perform neoadjuvant therapy, thus promoting the individualization and precision of treatment strategies.

[0103] The second embodiment of the present invention relates to a preoperative risk scoring system for soft tissue sarcomas of the limbs and trunk, comprising:

[0104] An acquisition module is used to acquire standardized MRI features of the patient and classify the standardized MRI features;

[0105] The scoring module is used to input the classification results of the standardized MRI features into the scoring model to obtain the patient's preoperative risk score; the scoring model is obtained in the following way:

[0106] Collect historical patient medical record datasets and extract standardized MRI features and postoperative progression of historical patients;

[0107] Using FNCLCC high-grade pathological markers, Ki-67 high-expression pathological markers, and HIF-1α high-expression pathological markers as dependent variables and standardized MRI features as independent variables, multivariate logistic regression analysis was performed using standardized MRI features from historical patients to screen out standardized MRI features that were independently associated with each pathological marker, and the non-standardized regression coefficients of each screened standardized MRI feature in the three logistic regression models were recorded.

[0108] Kendall's rank correlation coefficients were calculated between high-grade FNCLCC, high expression of cell proliferation marker Ki-67, high expression of hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients.

[0109] Based on each selected standardized MRI feature, its unstandardized regression coefficient in the three logistic regression models is used to calculate the fixed weight of each selected standardized MRI feature with the Kendall rank correlation coefficient of the corresponding pathological index.

[0110] The feature-weighted scoring model is constructed based on the fixed weights of each selected standardized MRI feature.

[0111] The standardized MRI features include maximum tumor diameter, tumor location, area of ​​tumor necrosis, peritumoral edema, tumor heterogeneity, invasion of important tissues around the tumor, peritumoral enhancement, and ADC value measurement.

[0112] The maximum diameter of the tumor refers to the maximum diameter length of the tumor measured on at least two orthogonal planes or isotropic three-dimensional images;

[0113] The tumor location is divided into superficial and deep locations. The superficial location refers to a tumor that is completely confined to subcutaneous adipose tissue and has not invaded the deep fascia. The deep location refers to a tumor that originates from or has invaded tissues below the deep fascia.

[0114] The tumor necrosis area refers to the region within the tumor that shows no enhancement on the enhanced T1-weighted image and a high signal on the T2-weighted image.

[0115] The peritumoral edema refers to a non-tumor, identifiable, high-signal region around the tumor that can be distinguished from the tumor boundary on T2-weighted fat-suppressed sequences or STIR sequences.

[0116] Tumor heterogeneity refers to the simultaneous presence of heterogeneous regions with low, medium, and high signal intensity within a tumor in T2-weighted images.

[0117] The invasion of important tissues around the tumor refers to the relationship between the tumor and adjacent key structures. A central angle corresponding to the arc length of the contact between the tumor and the nerve or blood vessel exceeding 180° is defined as nerve / blood vessel encirclement. Bone invasion is defined as the interruption of the continuity of the adjacent bone cortex or abnormal bone marrow signal on T2-weighted and enhanced T1-weighted images.

[0118] Peritumoral enhancement refers to the presence of a clearly defined band-like or nodular enhancement region that extends beyond the main boundary of the tumor on enhanced T1-weighted images during the arterial or portal venous phases.

[0119] The ADC value measurement refers to combining T2WI and enhanced T1WI images, avoiding hemorrhage, necrosis and cystic areas, delineating a circular or elliptical region of interest on the DWI image at the point of highest signal in the solid part of the tumor, and obtaining the average ADC value of the region of interest on the corresponding ADC image.

[0120] The classification of the standardized MRI features specifically involves:

[0121] Based on clinical criteria, the maximum diameter of a tumor is classified into three categories.

[0122] Tumors are classified into two categories based on their location: superficial and deep.

[0123] For the area of ​​tumor necrosis, the percentage of the area of ​​the tumor that shows no enhancement on the T1-weighted image after enhancement and is a high signal on the T2-weighted image is calculated, and the area is divided into two categories according to the percentage.

[0124] Peritumoral edema is classified into three categories based on the size of the high signal area;

[0125] Regarding tumor heterogeneity, it is divided into two categories based on the size of the heterogeneous regions;

[0126] In cases of invasion of important tissues surrounding the tumor, it is divided into two categories based on whether there is nerve / vascular encapsulation or bone invasion;

[0127] Regarding peritumoral enhancement, it is divided into two categories based on the presence or absence of band-like or nodular enhancement areas;

[0128] For ADC value measurement, it is divided into two categories based on the preset cutoff value.

[0129] The fixed weights of the standardized MRI features are obtained through The calculation yielded that, For the first Fixed weights for standardized MRI features, , and The first The unstandardized regression coefficients of standardized MRI features in three logistic regression models , and Kendall's rank correlation coefficients were used to correlate high-grade FNCLCC, high expression of the cell proliferation marker Ki-67, high expression of the hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients.

[0130] The third embodiment of the present invention relates to an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk of the first embodiment.

[0131] The fourth embodiment of the present invention relates to a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk of the first embodiment.

[0132] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0133] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0134] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction methods implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0135] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0136] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk, characterized in that, include: Acquire standardized MRI features of the patient and classify the standardized MRI features; The classification results of the standardized MRI features are input into the scoring model to obtain the patient's preoperative risk score; the scoring model is obtained in the following way: Collect historical patient medical record datasets and extract standardized MRI features and postoperative progression of historical patients; Using FNCLCC high-grade pathological markers, Ki-67 high-expression pathological markers, and HIF-1α high-expression pathological markers as dependent variables and standardized MRI features as independent variables, multivariate logistic regression analysis was performed using standardized MRI features from historical patients to screen out standardized MRI features that were independently associated with each pathological marker, and the non-standardized regression coefficients of each screened standardized MRI feature in the three logistic regression models were recorded. Kendall's rank correlation coefficients were calculated between high-grade FNCLCC, high expression of cell proliferation marker Ki-67, high expression of hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients. Based on each selected standardized MRI feature, its unstandardized regression coefficient in the three logistic regression models is used to calculate the fixed weight of each selected standardized MRI feature with the Kendall rank correlation coefficient of the corresponding pathological index. The feature-weighted scoring model is constructed based on the fixed weights of each selected standardized MRI feature.

2. The preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk according to claim 1, characterized in that, The standardized MRI features include maximum tumor diameter, tumor location, area of ​​tumor necrosis, peritumoral edema, tumor heterogeneity, invasion of important tissues around the tumor, peritumoral enhancement, and ADC value measurement. The maximum diameter of the tumor refers to the maximum diameter length of the tumor measured on at least two orthogonal planes or isotropic three-dimensional images; The tumor location is divided into superficial and deep locations. The superficial location refers to a tumor that is completely confined to subcutaneous adipose tissue and has not invaded the deep fascia. The deep location refers to a tumor that originates from or has invaded tissues below the deep fascia. The tumor necrosis area refers to the region within the tumor that shows no enhancement on the enhanced T1-weighted image and a high signal on the T2-weighted image. The peritumoral edema refers to a non-tumor, identifiable, high-signal region around the tumor that can be distinguished from the tumor boundary on T2-weighted fat-suppressed sequences or STIR sequences. Tumor heterogeneity refers to the simultaneous presence of heterogeneous regions with low, medium, and high signal intensity within a tumor in T2-weighted images. The invasion of important tissues around the tumor refers to the relationship between the tumor and adjacent key structures. A central angle corresponding to the arc length of the contact between the tumor and the nerve or blood vessel exceeding 180° is defined as nerve / blood vessel encirclement. Bone invasion is defined as the interruption of the continuity of the adjacent bone cortex or abnormal bone marrow signal on T2-weighted and enhanced T1-weighted images. Peritumoral enhancement refers to the presence of a clearly defined band-like or nodular enhancement region that extends beyond the main boundary of the tumor on enhanced T1-weighted images during the arterial or portal venous phases. The ADC value measurement refers to combining T2WI and enhanced T1WI images, avoiding hemorrhage, necrosis and cystic areas, delineating a circular or elliptical region of interest on the DWI image at the point of highest signal in the solid part of the tumor, and obtaining the average ADC value of the region of interest on the corresponding ADC image.

3. The preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk according to claim 2, characterized in that, The classification of the standardized MRI features specifically involves: Based on clinical criteria, the maximum diameter of a tumor is classified into three categories. Tumors are classified into two categories based on their location: superficial and deep. For the area of ​​tumor necrosis, the percentage of the area of ​​the tumor that shows no enhancement on the T1-weighted image after enhancement and is a high signal on the T2-weighted image is calculated, and the area is divided into two categories according to the percentage. Peritumoral edema is classified into three categories based on the size of the high signal area; Regarding tumor heterogeneity, it is divided into two categories based on the size of the heterogeneous regions; In cases of invasion of important tissues surrounding the tumor, it is divided into two categories based on whether there is nerve / vascular encapsulation or bone invasion; Regarding peritumoral enhancement, it is divided into two categories based on the presence or absence of band-like or nodular enhancement areas; For ADC value measurement, it is divided into two categories based on the preset cutoff value.

4. The preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk according to claim 1, characterized in that, The fixed weights of the standardized MRI features are obtained through The calculation yielded that, For the first Fixed weights for a standardized MRI feature. , and The first The unstandardized regression coefficients of standardized MRI features in three logistic regression models , and Kendall's rank correlation coefficients were used to correlate high-grade FNCLCC, high expression of the cell proliferation marker Ki-67, high expression of the hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients.

5. A preoperative risk scoring system for soft tissue sarcomas of the limbs and trunk, characterized in that, include: An acquisition module is used to acquire standardized MRI features of the patient and classify the standardized MRI features; The scoring module is used to input the classification results of the standardized MRI features into the scoring model to obtain the patient's preoperative risk score; the scoring model is obtained in the following way: Collect historical patient medical record datasets and extract standardized MRI features and postoperative progression of historical patients; Using FNCLCC high-grade pathological markers, Ki-67 high-expression pathological markers, and HIF-1α high-expression pathological markers as dependent variables and standardized MRI features as independent variables, multivariate logistic regression analysis was performed using standardized MRI features from historical patients to screen out standardized MRI features that were independently associated with each pathological marker, and the non-standardized regression coefficients of each screened standardized MRI feature in the three logistic regression models were recorded. Kendall's rank correlation coefficients were calculated between high-grade FNCLCC, high expression of cell proliferation marker Ki-67, high expression of hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients. Based on each selected standardized MRI feature, its unstandardized regression coefficient in the three logistic regression models is used to calculate the fixed weight of each selected standardized MRI feature with the Kendall rank correlation coefficient of the corresponding pathological index. The feature-weighted scoring model is constructed based on the fixed weights of each selected standardized MRI feature.

6. The preoperative risk scoring system for soft tissue sarcomas of the limbs and trunk according to claim 5, characterized in that, The standardized MRI features include maximum tumor diameter, tumor location, area of ​​tumor necrosis, peritumoral edema, tumor heterogeneity, invasion of important tissues around the tumor, peritumoral enhancement, and ADC value measurement. The maximum diameter of the tumor refers to the maximum diameter length of the tumor measured on at least two orthogonal planes or isotropic three-dimensional images; The tumor location is divided into superficial and deep locations. The superficial location refers to a tumor that is completely confined to subcutaneous adipose tissue and has not invaded the deep fascia. The deep location refers to a tumor that originates from or has invaded tissues below the deep fascia. The tumor necrosis area refers to the region within the tumor that shows no enhancement on the enhanced T1-weighted image and a high signal on the T2-weighted image. The peritumoral edema refers to a non-tumor, identifiable, high-signal region around the tumor that can be distinguished from the tumor boundary on T2-weighted fat-suppressed sequences or STIR sequences. Tumor heterogeneity refers to the simultaneous presence of heterogeneous regions with low, medium, and high signal intensity within a tumor in T2-weighted images. The invasion of important tissues around the tumor refers to the relationship between the tumor and adjacent key structures. A central angle corresponding to the arc length of the contact between the tumor and the nerve or blood vessel exceeding 180° is defined as nerve / blood vessel encirclement. Bone invasion is defined as the interruption of the continuity of the adjacent bone cortex or abnormal bone marrow signal on T2-weighted and enhanced T1-weighted images. Peritumoral enhancement refers to the presence of a clearly defined band-like or nodular enhancement region that extends beyond the main boundary of the tumor on enhanced T1-weighted images during the arterial or portal venous phases. The ADC value measurement refers to combining T2WI and enhanced T1WI images, avoiding hemorrhage, necrosis and cystic areas, delineating a circular or elliptical region of interest on the DWI image at the point of highest signal in the solid part of the tumor, and obtaining the average ADC value of the region of interest on the corresponding ADC image.

7. The preoperative risk scoring system for soft tissue sarcomas of the limbs and trunk according to claim 5, characterized in that, The classification of the standardized MRI features specifically involves: Based on clinical criteria, the maximum diameter of a tumor is classified into three categories. Tumors are classified into two categories based on their location: superficial and deep. For the area of ​​tumor necrosis, the percentage of the area of ​​the tumor that shows no enhancement on the T1-weighted image after enhancement and is a high signal on the T2-weighted image is calculated, and the area is divided into two categories according to the percentage. Peritumoral edema is classified into three categories based on the size of the high signal area; Regarding tumor heterogeneity, it is divided into two categories based on the size of the heterogeneous regions; In cases of invasion of important tissues surrounding the tumor, it is divided into two categories based on whether there is nerve / vascular encapsulation or bone invasion; Regarding peritumoral enhancement, it is divided into two categories based on the presence or absence of band-like or nodular enhancement areas; For ADC value measurement, it is divided into two categories based on the preset cutoff value.

8. The preoperative risk scoring system for soft tissue sarcomas of the limbs and trunk according to claim 5, characterized in that, The fixed weights of the standardized MRI features are obtained through The calculation yielded that, For the first Fixed weights for a standardized MRI feature. , and The first The unstandardized regression coefficients of standardized MRI features in three logistic regression models , and Kendall's rank correlation coefficients were used to correlate high-grade FNCLCC, high expression of the cell proliferation marker Ki-67, high expression of the hypoxia-related marker HIF-1α, and postoperative disease progression in historical patients.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk as described in any one of claims 1-4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the preoperative risk scoring method for soft tissue sarcomas of the limbs and trunk as described in any one of claims 1-4.

Images