CN112998651B

CN112998651B - Application of photoacoustic imaging in breast tumor scoring system and scoring system

Info

Publication number: CN112998651B
Application number: CN202110183802.0A
Authority: CN
Inventors: 杨萌; 姜玉新; 李建初; 张睿; 赵辰阳; 王铭; 赵瑞娜; 王若蛟; 刘思锐
Original assignee: Peking Union Medical College Hospital Chinese Academy of Medical Sciences
Current assignee: Peking Union Medical College Hospital Chinese Academy of Medical Sciences
Priority date: 2021-02-10
Filing date: 2021-02-10
Publication date: 2021-08-27
Anticipated expiration: 2041-02-10
Also published as: CN112998651A

Abstract

The invention relates to the application of photoacoustic imaging in a breast tumor scoring system, which comprises the following steps of (1) photoacoustic/ultrasonic bimodal imaging is used for acquiring image information of a breast tumor in an in-vitro mode; (2) analyzing the collected image information and respectively carrying out quantitative scoring and semi-quantitative scoring; (3) combining the quantitative score and the semi-quantitative score to obtain a comprehensive score and judge whether the breast tumor has a malignant tendency result; if one or all of the quantitative scores or semi-quantitative scores are judged to be malignant tendency, the tumor is considered to be malignant tendency. The specific scoring system is also related. Compared with the scoring system applied in the previous research, the method for distinguishing the malignant tumor from the benign tumor based on the diagnosis model and the nomogram is more convenient, has high repeatability and is more objective in diagnosis.

Description

Application of photoacoustic imaging in breast tumor scoring system and scoring system

Technical Field

The invention relates to the technical field of medical diagnosis, in particular to application of photoacoustic imaging in a breast tumor scoring system and the scoring system.

Background

The breast cancer is the most common malignant tumor of women worldwide, new breast cancer cases of women worldwide reach 208.9 ten thousand in 2018, and 62.7 ten thousand women die of the breast cancer. Early diagnosis and treatment are of great significance in improving the prognosis of breast cancer patients. The current Breast Imaging examination means commonly used in clinic mainly comprise molybdenum target, ultrasound and nuclear magnetic resonance, the classification System of the universal Breast Imaging reporting and Data System (BI-RADS) mainly takes the morphological characteristics of the focus, and the information of blood flow supply and distribution and the like plays a limited role. However, tumor vessels, one of the focuses of many years of breast cancer research, actually contain considerable diagnostic value. Angiogenesis is the basis of occurrence, development, infiltration and metastasis of breast cancer and other solid tumors, and obviously influences the biological behavior, curative effect and prognosis of the tumors. The previous research indicates that when the diameter of the tumor is 1-3mm, no new blood vessels are generated around the tumor, and the tumor grows slowly. When enough new vessels are generated around the tumor, the new vessels are generated by the surrounding vessels in a budding mode and enter the tumor tissue, and the tumor can grow rapidly. Therefore, the imaging means capable of accurately displaying tumor vessels and quantitatively evaluating the tumor vessels has important clinical significance for improving the diagnosis efficiency of breast tumors and improving the prognosis of patients.

The photoacoustic imaging (PAI) is a novel imaging technology, has the characteristics of high optical imaging contrast and strong acoustic imaging penetrating power, and quantitatively analyzes a series of physiological parameter changes of tissues to realize functional imaging while obtaining high-resolution tissue images. Hemoglobin in blood vessels can strongly absorb electromagnetic waves in a visible light band, and a photoacoustic imaging system can perform high-quality imaging on tumor blood vessels by detecting the hemoglobin. In addition, by utilizing dual-wavelength imaging and photoacoustic imaging, the relative content of oxygenated hemoglobin and deoxygenated hemoglobin in the tumor tissue can be quantitatively evaluated, so that information such as the oxygen content of the tissue can be obtained, the growth and metabolism of the tumor can be further evaluated, and the method has great potential in breast cancer diagnosis.

Currently, the commonly used breast imaging examinations in clinical practice are molybdenum target and Ultrasound (US), both of which are based on morphological features to discover and identify good and malignant nodules of the breast. Molybdenum target examination is radioactive, pressing on the breast during examination is uncomfortable for the patient, and it does not image dense breast tissue well. The breast ultrasound has higher resolution ratio on breast tissues and lesions, can perform real-time and dynamic imaging, but has lower specificity, and the examination result is greatly influenced by the subjectivity of doctors. On the other hand, breast tumors have heterogeneity, and benign and malignant nodules have a certain overlap in morphological expression, so that the properties of breast nodules cannot be comprehensively reflected by a simple morphological imaging means. Molybdenum targets and conventional ultrasound examination have certain limitations in identifying benign and malignant nodules of the breast. Color doppler ultrasound and power doppler ultrasound are currently common imaging means for clinically displaying tissue vascularity, but are relatively insensitive to low velocity blood vessels. The Dynamic enhanced MRI (Dynamic contrast-enhanced MRI) (DCE-MRI) has the capability of quantitatively evaluating blood flow perfusion and tumor vascularity, but has relatively low specificity and high cost, is not suitable for partial patients and has not been clinically popularized yet.

Previous studies have demonstrated the feasibility of PAI in displaying tumor vessels and identifying breast cancer, but a reliable and convenient method for assisted diagnosis of breast cancer based on photoacoustic imaging has not been established. In most studies, sophisticated image processing and analysis methods are used to quantitatively assess PA signals, still at some distance from clinical applications. Therefore, a simple and complete photoacoustic imaging analysis method for breast tumors can have potential clinical application value.

Disclosure of Invention

The invention aims to provide a photoacoustic imaging scoring system for breast tumors, which can be used for obtaining PA/US images in a standardized manner, processing the PA/US images in a standardized manner, establishing a diagnosis model, distinguishing malignant tumors and benign tumors according to data obtained by calculation of the diagnosis model, and has the advantages of simple method, high accuracy and high feasibility.

In one aspect of the present invention, there is provided the use of photoacoustic imaging in a breast tumor scoring system, comprising the steps of,

(1) photoacoustic/ultrasonic bimodal imaging carries out image information acquisition on breast tumors in an in vitro mode;

(2) analyzing the collected image information and respectively carrying out quantitative scoring and semi-quantitative scoring;

(3) combining the quantitative score and the semi-quantitative score to obtain a comprehensive score and judge whether the breast tumor has a malignant tendency result; if one or all of the quantitative scores or semi-quantitative scores are judged to be malignant tendency, the tumor is considered to be malignant tendency.

In the application, the quantitative score is obtained by quantitatively calculating the PA signal space density of the tumor and the peripheral area thereof through the acquired image information and then calculating through a quantitative score analysis formula; and the semi-quantitative score is obtained by quantitatively calculating the PA signal condition of the tumor and the peripheral area thereof through the acquired image information and then calculating through a semi-quantitative score analysis formula. The tendency to malignancy was evaluated as a tendency to have a quantitative score of more than 0.841 and a semi-quantitative score of more than 0.597.

The use as described above, wherein the tumour and its surrounding area comprises the region S, i.e. the region outside the tumour, within 5mm of its periphery; region T, the intratumoral region, comprising region P and region S; region P, the tumor margin region, the interior of the tumor, an annular region along the tumor outer margin 1/8 tumor minor axis width; region C, i.e. the central region of the tumor, the interior of the tumor, the central region of the tumor excluding region P.

The above application, wherein the PA signal spatial density is a quotient of the number of pixels having PA signals in a region and the total number of pixels in the region.

The use as described above, wherein the PA signal profile is a score of PA signal for each region in each image, with no score-0, a small score of-1, a medium score of-2, or a significant score of-3.

The above application, wherein the specific criteria for scoring the PA signal are: the detection result includes that the signals are divided into 0 min and no PA signal, 1 min and 1-2 point-shaped or thin line-shaped PA signals with the diameter less than 0.1cm, 2 min and 3-4 point-shaped or one longer strip-shaped penetration signal, and 3 min and more than or equal to 5 or more PA signals.

The above application, wherein the quantitative score analysis formula is: logit (y) ═ 1.52691+2.15350 × P_S-T-0.91390*P_S-C+2.04078*P_P-C；

The semi-quantitative score analysis formula is: logit (y) ═ 4.37267+ 2.40903S_S-T-0.34354*S_S-C+5.24288*S_P-C；

Wherein P is_S,P_T,P_PAnd P_CAverage PA signal spatial densities representing S, T, P, and C regions, respectively; p_S-C＝(P_S-P_C)/P_C、P_P-C＝(P_P-P_C)/P_cAnd P_S-T＝(P_S-P_T)/P_TRespectively representing the relative difference of PA signal space density of the S area and the C area, the P area and the C area, and the S area and the T area;

S_S,S_T,S_Pand S_CRespectively representing the average values of PA signal condition scores of the S area, the T area, the P area and the C area; s_S-C(S_S-C＝S_S-S_C)、S_P-C(S_P-C＝S_P-S_C) And S_S-T(S_S-T＝S_S-S_T) Respectively representing the score difference between the S area and the C area, between the P area and the C area, and between the S area and the T area;

the logit (y) value of the quantitative analysis formula is the quantitative score value, and the logit (y) value of the semi-quantitative analysis formula is the semi-quantitative score value.

The invention also provides a mammary gland tumor scoring system based on the photoacoustic imaging technology, which comprises an information acquisition module, an information analysis module, a calculation output module and a judgment module, wherein the information acquisition module is connected with the photoacoustic imaging equipment to acquire image information characteristic parameters of the mammary gland tumor and the peripheral area thereof; the information analysis module analyzes and calculates a quantitative score and a semi-quantitative score according to the obtained image information characteristic parameters; the calculation output module is used for calculating quantitative scores and semi-quantitative scores respectively; and the judging module judges the tumor property according to the quantitative score and the semi-quantitative score.

The breast tumor scoring system is characterized in that the quantitative scoring is obtained by quantitatively calculating PA signal space density of the tumor and the peripheral area thereof through the acquired image information and then calculating through a quantitative scoring analysis formula; and the semi-quantitative score is obtained by quantitatively calculating the PA signal condition of the tumor and the peripheral area thereof through the acquired image information and then calculating through a semi-quantitative score analysis formula.

The tendency to malignancy was evaluated as a tendency to have a quantitative score of more than 0.841 and a semi-quantitative score of more than 0.597.

The breast tumor scoring system described above wherein the tumor and its surrounding area comprise the region S, i.e. the region outside the tumor, within 5mm of its periphery; region T, the intratumoral region, comprising region P and region S; region P, the tumor margin region, the interior of the tumor, an annular region along the tumor outer margin 1/8 tumor minor axis width; region C, i.e. the central region of the tumor, the interior of the tumor, the central region of the tumor excluding region P.

The system for scoring a breast neoplasm as described above, wherein the PA signal spatial density is a quotient of the number of pixels having PA signals in a region and the total number of pixels in the region.

The system for scoring a breast neoplasm as described above, wherein the PA signal profile is a score of PA signal for each region in each image, no score of-0, a small score of-1, a medium score of-2, or a significant score of-3.

The breast tumor scoring system described above, wherein the specific criteria for scoring the PA signal are: the detection result includes that the signals are divided into 0 min and no PA signal, 1 min and 1-2 point-shaped or thin line-shaped PA signals with the diameter less than 0.1cm, 2 min and 3-4 point-shaped or one longer strip-shaped penetration signal, and 3 min and more than or equal to 5 or more PA signals.

The breast tumor scoring system of the above, wherein

The quantitative score analysis formula is: logit (y) ═ 1.52691+2.15350 × P_S-T-0.91390*P_S-C+2.04078*P_P-C；

the logit (y) value of the quantitative score analysis formula is the quantitative score value, and the logit (y) value of the semi-quantitative score analysis formula is the semi-quantitative score value.

The invention has the following beneficial effects:

the breast tumor scoring system has the advantages that malignant tumors and benign tumors can be distinguished by quantitative scoring and semi-quantitative scoring, and the diagnosis models of quantitative analysis and semi-quantitative analysis are constructed by utilizing image processing software to carry out pixel analysis on photoacoustic images. In addition, a semi-quantitative PA scoring method is provided by referring to the current Doppler imaging scoring system, and a simple, convenient and feasible method is provided for a clinician to interpret a PA image. In addition, differentiation between malignant and benign tumors based on diagnostic models and nomograms is more convenient, reproducible, and objective than scoring systems used in previous studies.

Drawings

FIG. 1 is a flow chart of the PA image acquisition and data processing standardization of the present invention;

FIG. 2 ROC plot of diagnostic performance for the quantitative score and semi-quantitative score diagnostic models of the present invention (model 1: quantitative analysis, area under curve: 0.824, model 2: semi-quantitative score: area under curve: 0.865);

FIG. 3 is a PA/US image of scoring application example 1 of the present invention;

FIG. 4 nomogram results of the quantitative scoring model and nomogram results of the semi-quantitative scoring model of scoring application example 1 of the present invention;

FIG. 5 is a PA/US visualization of scoring application example 2 of the present invention;

fig. 6 nomogram results of the quantitative scoring model and nomogram results of the semi-quantitative scoring model of the scoring application example 2 of the present invention.

Detailed Description

The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention. Unless otherwise specified, the technical means used in the examples are conventional means well known to those skilled in the art. The invention is described in detail below with reference to the figures and examples.

In the present embodiment, "PA" means photoacoustic, and "US" means ultrasound.

PA/US bimodal imaging system

The PA/US dual-modality system used in this study is based on a high-end clinical ultrasound machine (Resona 7, Mindray Bio-Medical Electronics co., Ltd.) capable of performing and acquiring the data required for PA imaging. And (5) reconstructing the PA imaging result on line by using a delay and sum algorithm. A clinical linear probe (L9-3U, Mindray Bio-Medical Electronics Co., Ltd.) had 192 elements each having a size of 0.2mm and a center frequency of 5.8 MHz. The laser source was an OPO tunable laser (Spitlight 600-OPO, Innolas laser GmbH) that generated 700-850nm laser pulses at 10 Hz. In our study, 750nm and 830nm were used for PA functional imaging.

Patient's health

From 11 months in 2017 to 1 month in 2018, 39 breast focal lesion patients were recruited continuously from the department of breast surgery in the coordination hospital of Beijing, for a total of 40 nodules. Inclusion criteria for breast lesions were: 1) the maximum diameter of the focus is less than 3 cm; 2) the distance between the front edge of the focus and the surface of the skin is more than or equal to 5 mm; 3) due to the limited penetration depth of the laser in the tissue, the trailing edge of the lesion is defined to be < 3cm from the skin surface. All patients were first diagnosed by an experienced imaging physician with ultrasound, X-ray breast molybdenum targets, and/or MRI. 3 imaging physicians completed routine ultrasound examinations on all patients, who had more than 10 years of diagnostic experience in ultrasound diagnosis of breast disease. After routine ultrasound examination, 40 cases of nodule row PA/US were bimodal imaged by 1 imaging physician who had more than 20 years of breast disease ultrasound diagnostic experience and more than 3 years of photoacoustic imaging diagnostic experience. After the imaging examination, the patient received ultrasound guided punch biopsy and/or surgery within a week. All lesions were confirmed by pathological results.

Example 1 construction of a Breast tumor Scoring System

1. Image acquisition

First, a conventional ultrasound examination is performed on each patient to determine the anatomy of each lesion. Subsequently, the whole tumor and the area around the tumor are scanned continuously and uniformly along the short axis direction of the tumor by using a handheld dual-mode PA/US imaging probe. An image of each frame is extracted in the obtained PA/US video.

2. Image partitioning

For each case (nodule), images containing the entire tumor and its surrounding 5mm tissue were taken for analysis in all images, one frame per 1 mm. For each frame of Image selected, we use the Image processing software Image J to subdivide it into different regions according to its position in the tumor and its surrounding regions, as follows.

Region S is the region outside the tumor within 5mm of its periphery.

Region T is the region within the tumor, including region P and region S.

Region P is the tumor margin region, the interior of the tumor, and the annular region along the tumor outer margin 1/8 tumor minor axis width.

Region C, tumor center region, tumor interior, tumor center region except region P.

3. Quantitative and semi-quantitative assessment

3.1 quantitative evaluation

The PA signal spatial density for each region was calculated using the software MATLAB (Mathworks, inc., USA). The spatial density of the PA signal is defined as the quotient of the number of pixels in a region having the PA signal and the total number of pixels in the region. In this application, we refer to the average PA signal spatial densities of the S-region, T-region, P-region and C-region as P, respectively_S,P_T,P_PAnd P_C. In addition, regions S and C, and regions P and C were also analyzedThe relative difference of PA signal space density of the region, S region and T region is defined as P_S-C＝(P_S-P_C)/P_C,P_P-C＝(P_P-P_C)/P_cAnd P_S-T＝(P_S-P_T)/P_T,。

3.2 semi-quantitative evaluation

In addition to calculating the spatial density of the PA signal, semi-quantitative analysis is also performed by scoring the photoacoustic imaging results for each region. With reference to doppler scoring method of Adler, the PA signals of each region in each image are scored as none (0 point), few (1 point), medium (2 points) or significant (3 points), with the following specific criteria:

0 minute: no PA signal is detected;

1 minute: detecting 1-2 point-shaped or thin-line-shaped PA signals with the diameter less than 0.1 cm;

and 2, dividing: detecting 3-4 point-shaped signals or a longer strip-shaped penetrating signal;

and 3, dividing: 5 or more PA signals are detected.

For each case, the average value of the photoacoustic signal score of each region (S region, T region, P region and C region) in all the images was counted and recorded as: s_S,S_T,S_PAnd S_C. Further calculating and analyzing the difference of scores of the S region and the C region, the P region and the C region and the S region and the T region, and the difference of the P region and the C region PA scores is respectively defined as S_S-C(S_S-C＝S_S-S_C)、S_P-C(S_P-C＝S_P-S_C)、S_S-T(S_S-T＝S_S-S_T)。

Establishing a quantitative analysis variable (P) by a backward stepwise method (backward stepwise) adopting logistic regression_S-C、P_P-CAnd P_S-T) And semi-quantitative method (S)_S-C、S_P-CAnd S_S-T) The diagnostic model of (1). The two model equations are as follows:

quantitative analysis: logit (y) ═ 1.52691+2.15350 × P_S-T-0.91390*P_S-C+2.04078*P_P-C；

Semi-quantitative analysis:Logit(Y)＝-4.37267+2.40903*S_S-T-0.34354*S_S-C+5.24288*S_P-C.

2. statistical analysis

Statistical software SPSS for Windows 22.0(SPSS Inc, Chicago, IL) and EmpowerStats software (X) were used&Y Solutions) for statistical data analysis. Continuous data are presented as mean ± standard deviation. Student's t-test was used to evaluate the difference in benign and malignant tumor signal intensity and scores. P value<0.05 was considered statistically significant. To investigate the diagnostic efficacy of both the PA signal quantitative analysis and semi-quantitative scoring methods, we used a backward stepwise logistic regression method, establishing quantitative analysis variables (P) respectively_S-C、P_P-CAnd P_S-T) And semi-quantitative method (S)_S-C、P_P-CAnd P_S-T) The diagnostic model of (2) is expressed in the form of an equation and a nomogram. The diagnostic effect of the two diagnostic models was evaluated by calculating the sensitivity, specificity, Positive Likelihood Ratio (PLR), Negative Likelihood Ratio (NLR), predictive value (PPV), Negative Predictive Value (NPV) of the diagnostic test. The Receiver Operating Curves (ROCs) for both diagnostic models were plotted and the area under the curve (AUC) was calculated. Sensitivity and positivity of the two models were compared using a 2 x 2 tabulation, chi-square test, and PPV and NPV of the two were compared using a generalized estimation equation.

And (3) drawing an ROC curve: analyzing the measurement results of the disease group and the reference group, determining the upper and lower limits, the group distance and the cut-off point (cut-off point) of the measurement value, listing a cumulative frequency distribution table according to the selected group distance interval, and respectively calculating the sensitivity, the specificity and the false positive rate (1-specificity) of all the cut-off points. Sensitivity is taken as an ordinate to represent true positive rate, and (1-specificity) is taken as an abscissa to represent false positive rate, and an ROC curve is drawn by plotting.

The area under the ROC curve is between 1.0 and 0.5. In the case of AUC > 0.5, the closer the AUC is to 1, the better the diagnostic effect. AUC has lower accuracy when being 0.5-0.7, AUC has certain accuracy when being 0.7-0.9, and AUC has higher accuracy when being more than 0.9. When AUC is 0.5, the diagnostic method is completely ineffective and is not valuable. AUC < 0.5 does not correspond to the real case and is rarely found in practice.

Sensitivity: reflecting the ability to diagnose the patient; specificity: reflecting the ability to judge a person who is actually free of disease; positive Predictive Value (PPV): reflecting the possibility of the target disease of the person with positive screening test result. That is, the number of patients who are actually ill among those who are diagnosed as being ill by a certain diagnostic test; negative Predictive Value (NPV): reflecting the likelihood that the test is negative in a subject who is truly not diseased. That is, the number of patients who are diagnosed as being disease-free by a certain diagnostic test is actually disease-free. Likelihood Ratio (LR): combines the advantages of sensitivity, specificity, positive predictive value and negative predictive value, is not influenced by the prevalence rate, and is a relatively stable comprehensive index. The greater the + LR, the greater the probability of true positive if the test result is positive. The smaller the LR, the greater the likelihood of being a true negative if the test result is negative.

3. Results

TABLE 1 diagnostic efficacy of the quantitative analysis score and semi-quantitative score diagnostic models

The specific ROC curve is shown in FIG. 2, and it can be seen that the ROC curve of model 1 (quantitative analysis score) and model 2 (semi-quantitative analysis score) is obviously positioned at the upper left corner of the reference line, which indicates that the diagnosis value is higher. AUC is greater than 0.7, which shows some accuracy. The optimal diagnostic threshold for model 1 was 0.841, corresponding diagnostic sensitivity was 0.769, specificity was 0.786; the optimal diagnostic threshold for model 2 was 0.597, corresponding to a diagnostic sensitivity of 0.808 and specificity of 0.857.

Example 2 Scoring application example

Application example 1

In 2018, in 4 months, 1 female patient with the maximum diameter of breast tumor less than 3cm was recruited in breast surgery department, and the conventional breast ultrasound BI-RADS score is 1 female patient with the age of 40 years of 4 proposed surgical resection, and the image of the patient is shown in FIG. 3.

FIG. 3 shows that there is less PA signal in the central area of the tumor and abundant PA signal in the peripheral and peripheral areas (S)_C:1，S_S:3，S_P:3)。In fig. 4 c, e are two nomograms used to calculate the risk of malignancy. PA parameter (P) of lesion_S:0.10，P_T:0.03，P_C:0.02，P_P:0.06，P_S-T:2.3，P_S-C:4.0，P_P-C:2.0；S_S:3，S_T:3，S_C:1，S_P:3，S_S-T:0，S_S-C:2，S_P-CAnd 2) putting the nomogram into a nomogram, wherein the predicted values of the two models are both more than 0.90. From the results of the nomogram, the tumor was judged to be possibly malignant. The patient was indeed malignant by pathological and clinical diagnosis.

Application example 2

In 2018, in 4 months, 1 female patient with the maximum diameter of breast tumor smaller than 3cm was recruited in breast surgery department, and the conventional breast ultrasound BI-RADS score is 3 proposed for surgical resection, and the image of the patient is shown in FIG. 5.

Shown in FIG. 5, in all the regions (S)_C:3、S_S:3、S_P3) all detected rich PA signals. In fig. 6 d, f are two nomograms used to calculate the risk of malignancy. PA parameter (P) of lesion_S:0.15，P_T:0.06，P_C:0.05，P_P:0.06，P_S-T:1.5，P_S-C:2.0，P_P-C:0.2，S_S:3，S_T:3，S_C:3，S_P:3，S_S-T:0，S_S-C:0，S_P-C0) inputting a nomogram, wherein the predicted value of the model 1 is less than 0.50, and the predicted value of the model 2 is less than 0.10. Based on the results of the nomograms, a potentially benign diagnosis of the tumor was made. The patient was indeed benign as a tumor by both pathological and clinical diagnosis.

Although the invention has been described in detail hereinabove with respect to a general description and specific embodiments thereof, it will be apparent to those skilled in the art that modifications or improvements may be made thereto based on the invention. Accordingly, such modifications and improvements are intended to be within the scope of the invention as claimed.

Claims

1. A mammary gland tumor scoring system based on photoacoustic imaging technology is characterized by comprising an information acquisition module, an information analysis module, a calculation output module and a judgment module, wherein the information acquisition module is connected with photoacoustic imaging equipment to acquire image information characteristic parameters of a mammary gland tumor and a peripheral region thereof; the information analysis module analyzes and calculates a quantitative score and a semi-quantitative score according to the obtained image information characteristic parameters; the calculation output module is used for calculating quantitative scores and semi-quantitative scores respectively; the judging module judges the tumor property according to the quantitative score and the semi-quantitative score;

the quantitative score is obtained by quantitatively calculating PA signal space density of the tumor and the peripheral area thereof through the acquired image information and then calculating through a quantitative score analysis formula; the semi-quantitative score is obtained by quantitatively calculating the PA signal condition of the tumor and the peripheral area thereof through the acquired image information and then calculating through a semi-quantitative score analysis formula;

the malignant tendency is taken as an evaluation standard, wherein the quantitative score value is more than 0.841 and the semi-quantitative score value is more than 0.597;

the tumour and its surrounding area comprise the region S, i.e. the region outside the tumour, within 5mm of its periphery; region T, the intratumoral region, comprising region P and region C; region P, the tumor margin region, the interior of the tumor, an annular region along the tumor outer margin 1/8 tumor minor axis width; region C, i.e. the central region of the tumor, the interior of the tumor, the central region of the tumor except region P;

the PA signal space density is the quotient of the number of pixels with PA signals in the area and the total number of pixels in the area;

the PA signal condition is the evaluation of the PA signals of all areas in each image, and no score is-0, a small amount of score is-1, a medium score is-2 or a significant score is-3;

specific criteria for scoring the PA signal are: detecting PA signals at 0 min, detecting point-like or thin-line-like PA signals with the diameter less than 0.1cm at 1 min, detecting point-like or long strip-like penetration signals at 3-4 min or detecting more than or equal to 5 or more PA signals at 3 min;

the quantitative scoring analysis formula is as follows: logit (y) = -1.52691+ 2.15350P_S-T -0.91390* P_S-C +2.04078* P_P-C;

The semi-quantitative score analysis formula is: logit (y) = -4.37267+ 2.40903S_S-T -0.34354* S_S-C +5.24288* S_P-C；

Wherein P is_S, P_T, P_PAnd P_CAverage PA signal spatial densities representing S, T, P, and C regions, respectively; p_S-C = (P_S-P_C)/P_C 、P_P-C= (P_P -P_C)/P_cAnd P_S-T= (P_S -P_T)/P_TRespectively representing the relative difference of PA signal space density of the S area and the C area, the P area and the C area, and the S area and the T area;

S_S, S_T, S_Pand S_CRespectively representing the average values of PA signal condition scores of the S area, the T area, the P area and the C area; s_S-C（S_S-C= S_S -S_C）、S_P-C (S_P-C= S_P -S_C) And S_S-T（S_S-T= S_S -S_T) Respectively representing the score difference between the S area and the C area, between the P area and the C area, and between the S area and the T area;