Pulmonary nodule auxiliary detection method and device for paired CT (Computed Tomography) images

Abstract

The invention discloses a pulmonary nodule auxiliary detection method and device for paired CT (Computed Tomography) images. According to the method, a first CT image of the lung of a patient and a two-dimensional CT slice sequence corresponding to the first CT image are obtained, each two-dimensional CT slice is divided into a plurality of sub-block matrixes with the same size and number, and the matrix transformation result of each sub-block matrix is extracted through a preset image algorithm. And calculating nodule progress display factors by comparing matrix transformation results of corresponding sub-block matrixes of a subsequent second CT image of the same patient, obtaining nodule progress detection quantity of each two-dimensional CT slice based on the factors, and finally comprehensively outputting nodule progress detection results. According to the method, the nodule change of the patient is measured by adopting the detection quantity with unique design, so that a doctor can be helped to more accurately observe and analyze the progress of the pulmonary nodule.

Description

Technical Field

[0001] The present invention relates to a lung nodule auxiliary detection method for paired CT images, and also relates to a corresponding lung nodule auxiliary detection device, belonging to the field of medical care informatics. Background Art

[0002] Monitoring of lung nodules is crucial for early detection and treatment of lung cancer. Traditionally, doctors evaluate the size and density changes of lung nodules by analyzing CT images, but this method is highly subjective and prone to errors. To improve accuracy, researchers have developed 3D nodule volume measurement methods based on image segmentation and 3D filtering techniques, which show better results on lung images than 2D methods.

[0003] In the Chinese patent application with application number 202410586882.8, a method for comparing historical examinations of lung nodules combining deep learning algorithms and image processing technology is proposed. This method improves the efficiency of detection and comparison by identifying the bifurcation points of the main bronchus in the lungs, avoiding the tedious steps of manually finding the corresponding positions of the lesions. Nevertheless, this method still relies on volume measurement and requires precise nodule morphology outlining, and its results may be affected by multiple factors.

[0004] Although these technologies have improved detection efficiency, they have not yet fully resolved the issue of statistical significance of detected changes. Future research is needed to further optimize the algorithms, improve the accuracy and robustness of detection, and develop new evaluation methods to determine whether the detected changes are clinically significant. Summary of the invention

[0005] The primary technical problem to be solved by the present invention is to provide a lung nodule auxiliary detection method for paired CT images.

[0006] Another technical problem to be solved by the present invention is to provide a lung nodule auxiliary detection device for paired CT images.

[0007] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0008] According to a first aspect of an embodiment of the present invention, a method for auxiliary detection of pulmonary nodules in paired CT images is provided, comprising the following steps:

[0009] Acquire a first CT image of a patient's lung scan, and acquire a two-dimensional CT slice sequence corresponding to the first CT image;

[0010] For each two-dimensional CT slice in the two-dimensional CT slice sequence, dividing each of the two-dimensional CT slices into a plurality of sub-block matrices of the same size and number;

[0011] For each of the sub-block matrices, extracting a matrix transformation result of each of the sub-block matrices by using a preset image algorithm;

[0012] Reacquire a second CT image of the patient's lung scan, and extract the matrix transformation result of each of the sub-block matrices of the second CT image in the same manner;

[0013] For the sub-block matrix with the same number of slices and the same position, based on the matrix transformation result of the first CT image and the matrix transformation result of the second CT image, obtaining the nodule display factor corresponding to the sub-block matrix at the position;

[0014] Based on the nodule progression display factors of all sub-block matrices of the two-dimensional CT slice under the same number of slices, obtaining the nodule progression detection amount of the two-dimensional CT slice under the number of slices;

[0015] Based on the nodule progression detection amount of each of the two-dimensional CT slices, a final nodule progression detection result is comprehensively output.

[0016] According to a second aspect of an embodiment of the present invention, there is provided a device for auxiliary detection of lung nodules for paired CT images, comprising a processor and a memory; wherein the memory is coupled to the processor and is used to store a computer program, and when the computer program is executed by the processor, the processor implements the above-mentioned method for auxiliary detection of lung nodules.

[0017] Compared with the prior art, the present invention has the following technical effects:

[0018] (1) From the perspective of hypothesis testing, a unique and precise detection method was designed to measure the changes in patients’ nodules in combination with a specific image processing algorithm. Thus, by quantifying the changes in the lesions in the two CT images before and after, it helps doctors observe and analyze the progression of lung nodules more accurately.

[0019] (2) By studying whether there are significant differences in the local nodule images between the two CT scans of the patient, the detection quantity is used to provide an answer to whether the changes in the lung nodules are statistically significant, which makes up for the fact that existing methods cannot answer the key question of whether the detected changes are statistically significant.

[0020] (3) The nodule progression detection value of each two-dimensional CT slice is represented by a pre-set hyperparameter, and a two-dimensional curve between the nodule progression detection value and the hyperparameter is constructed. The influence of the hyperparameter is eliminated by calculating the area under the curve of the two-dimensional curve, thereby improving the accuracy of the final nodule progression detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1A flowchart of a method for auxiliary detection of pulmonary nodules in paired CT images provided by the first embodiment of the present invention;

[0022] Figure 2A This is a schematic diagram of a two-dimensional CT slice of the same physical position of patient A before and after treatment in the first embodiment of the present invention;

[0023] Figure 2B This is a schematic diagram of a two-dimensional curve constructed for patient A in the first embodiment of the present invention;

[0024] Figure 3A This is a schematic diagram of a two-dimensional CT slice of the same physical position of patient B before and after treatment in the first embodiment of the present invention;

[0025] Figure 3B This is a schematic diagram of a two-dimensional curve constructed for patient B in the first embodiment of the present invention;

[0026] Figure 4 This is the confusion matrix of each diameter grouping under the simulation experiment in the first embodiment of the present invention;

[0027] Figure 5 is the evaluation index value under the simulation experiment in the first embodiment of the present invention;

[0028] Figure 6 This is the confusion matrix of each diameter grouping and the whole sample under real clinical data in the first embodiment of the present invention;

[0029] Figure 7 is the evaluation index value under real clinical data in the first embodiment of the present invention;

[0030] Figure 8 A structural diagram of a pulmonary nodule auxiliary detection device for paired CT images provided in a second embodiment of the present invention;

[0031] Fig. 9 A structural diagram of a lung nodule auxiliary detection device for paired CT images provided in the third embodiment of the present invention. DETAILED DESCRIPTION

[0032] The technical content of the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] First embodiment

[0034] The first embodiment of the present invention proposes a method for auxiliary detection of lung nodules for paired CT images. The method is based on the principle of hypothesis testing and focuses on analyzing whether there are significant differences in the images of lung nodules after the patient undergoes two CT scans. This method is intended to help doctors observe and evaluate the progression of lung nodules more accurately. In medical practice, doctors usually judge whether a nodule has progressed based on the changes in the nodule and the surrounding normal tissue. If the nodule has no significant changes compared to the surrounding tissue, the nodule is considered to be stable; if there are significant changes, the nodule is considered to be progressing.

[0035] Based on this clinical experience, the embodiment of the present invention designs a unique and sophisticated statistic - nodule progression detection result (DNP), which is used to quantify the nodule changes of a single patient. DNP is not only constructed based on the clinical experience of doctors, but also can provide objective conclusions with statistical significance, thereby assisting doctors to make more accurate judgments on the progression of patients' lung nodules. This method reduces the subjectivity and errors in traditional methods through quantitative analysis, and improves the accuracy and reliability of lung nodule detection.

[0036] The specific steps of the pulmonary nodule auxiliary detection method are described below:

[0037] S1: Acquire a two-dimensional CT slice sequence.

[0038] First, collect the first CT image of the patient's lung scan before treatment, using a three-dimensional tensor Define the first CT image; wherein, represents the t-th two-dimensional CT slice, T is the total number of CT slices, p represents the dimension of the two-dimensional CT; R represents the real number field; t is a positive integer.

[0039] In actual situations, a set of real lung CT images that are diagnosed as positive often contains a large number of CT slices without nodules. Such CT slices will be excluded in subsequent research and analysis to reduce the interference of irrelevant information of nodules. Therefore, in actual implementation, only the continuous two-dimensional CT slice sequence containing the target lung nodules is selected and recorded. ,in represent In the tth CT slice Logarithmic transformation of the HU value of the pixel at that location.

[0040] In one embodiment of the present invention, the process of obtaining a two-dimensional CT slice sequence is as follows: First, a lung CT image of the patient is directly obtained by tomography. Subsequently, a professionally qualified doctor carefully reviews and annotates these images to identify and determine the specific area covering the range of the lung nodules. According to the doctor's annotations, continuous two-dimensional CT slices containing the target lung nodules are selected to form a sequence. This sequence of two-dimensional CT slices will be used in subsequent lung nodule auxiliary detection methods to facilitate accurate analysis and evaluation of the presence and changes of lung nodules.

[0041] S2: Divide each 2D CT slice into multiple sub-block matrices.

[0042] Specifically, after obtaining a two-dimensional CT slice sequence based on step S1, for each two-dimensional CT slice in the two-dimensional CT slice sequence, the number n to be divided is first determined based on the size of the target lung nodule. Here, n is a positive integer, and its value is mainly related to the size of the nodule. For example, the dimension of a two-dimensional CT slice is 512 512, the sub-blocks are 32 32, that is, the size of each sub-block matrix is ​​m = 32, so a 512 512 2D CT slices can be divided into 16 16 of 32 32 sub-blocks. That is, 16 horizontal cuts and 16 vertical cuts, so n = 16.

[0043] Then, based on the number of slices to be divided n, each 2D CT slice is divided into n n sub-block matrices, namely , where each sub-block matrix is ​​m m matrix, p=n m. Moreover, the size and number of the sub-block matrices divided by each two-dimensional CT slice are equal, so that subsequent comparative analysis can be facilitated.

[0044] It is understandable that for any and , of Position elements can be represented by Given, where , For the convenience of description, define the index set ,therefore can be rewritten as . Then, you can As The tth CT slice sub-block matrices.

[0045] S3: extracting the matrix transformation result of each sub-block matrix through a preset image algorithm.

[0046] For each two-dimensional CT slice, after being divided into a plurality of sub-block matrices, a matrix transformation result needs to be extracted for each sub-block matrix of the two-dimensional CT slice through a preset image algorithm.

[0047] In this embodiment, the preset image algorithm is as follows:

[0048] ⅰ. Perform logarithmic transformation on the pixel HU value of each position element in each sub-block matrix.

[0049] Specifically, for the t-th two-dimensional CT slice of the first CT image, 1 , n 2 ) sub-block matrix, and perform logarithmic transformation on the pixel point HU value of each position element in the sub-block matrix, so as to obtain the (n)th t-th two-dimensional CT slice. 1 , n 2 ) sub-block matrices ; where 1≤n 1 , n 2 ≤n.

[0050] ⅰⅰ. For the sub-block matrix in step i , define its binary characteristic parameter .

[0051] Specifically, when the sub-block matrix When the target nodule center point is included, ,otherwise, .

[0052] iii. Based on the hypothesis test, the matrix transformation results of each sub-block matrix are calculated.

[0053] In this embodiment, according to the binary characteristic parameters of each sub-block matrix in step ii, for each two-dimensional CT slice of the first CT image, it is assumed that the sum of the binary characteristic parameters of all sub-block matrices is , that is, there is only one target nodule in each 2D CT slice. And, assuming that for each 2D CT slice, the center point of the target nodule appears at The sub-block matrix of the position, that is, the sub-block position where the nodule appears It has nothing to do with the CT slice number t.

[0054] Based on this, for any ,definition , 1≤ ≤n; thus calculate , as the matrix transformation result of the (n1, n2)th sub-block matrix in the tth CT slice. Among them, “ "and" ” factors a matrix element by element.

[0055] S4: Calculate the nodule presentation factor corresponding to each sub-block matrix.

[0056] In this embodiment, after the matrix transformation results corresponding to each sub-block matrix of each two-dimensional CT slice of the first CT image are calculated through the above steps S1 to S3, a second CT scan image of the same patient after a period of follow-up is obtained. Perform the same operation and calculation to obtain the second CT scan image The matrix transformation result of the (n1, n2)th sub-block matrix of the t-th two-dimensional CT slice is , Based on the above results, the nodule display factor of the (n1, n2)th sub-block matrix of the t-th two-dimensional CT slice is calculated. :

[0057]

[0058] in, is a pre-set hyperparameter; It means that the nodule has a tendency to increase in size on the tth CT slice; It means that the nodule has a tendency to shrink on the tth CT slice; It means that there is no clear change in the nodule on the tth CT slice.

[0059] S5: Obtain the nodule progression detection amount for each two-dimensional CT slice.

[0060] Specifically, since each 2D CT slice consists of n 2 Sub-block matrices, and each sub-block matrix has its corresponding nodule display factor Therefore, for the same two-dimensional CT slice, it is necessary to integrate the nodule progression display factors of n2 sub-block matrices to form a more effective and comprehensive nodule progression detection quantity.

[0061] In this embodiment, the nodule progression factor is integrated by taking the average value. Specifically, the integrated nodule progression detection value is .

[0062] S6: Output the final nodule progression detection result.

[0063] After the nodule progression detection amount of each two-dimensional CT slice is calculated based on step S5, it is easy to verify that for any , However, nodule progression detection There are still drawbacks, namely is a hyperparameter function, and the hyperparameter It is difficult to select a priori. In order to eliminate the hyperparameter In this embodiment, The two-dimensional curve, .

[0064] Figure 2A The figure shows a schematic diagram of a two-dimensional CT slice of the same physical position before and after treatment of patient A in this embodiment. The position marked by the box is the location of the nodule, and the diameter of the nodule of patient A increases by about 50%. Figure 2B The example shown is the one constructed for patient A in this embodiment. Schematic diagram of a two-dimensional curve. Similarly, Figure 3A FIG. 1 is a schematic diagram of a two-dimensional CT slice of the same physical position scanned before and after treatment of patient B in this embodiment. The diameter of the nodule of patient B is reduced by about 50%. Figure 3B Shown is the constructed Schematic diagram of a two-dimensional curve.

[0065] When constructed After the two-dimensional curve is calculated The area under the curve is recorded as the nodule progression detection result DNPt of the t-th two-dimensional CT slice. Then, based on the nodule progression detection results of all two-dimensional CT slices, the final nodule progression detection result is comprehensively output; .

[0066] It can be understood that when the embodiment of the present invention is actually applied, it is only necessary to obtain CT images of the same patient twice before and after scanning, and then according to the center of gravity position of the nodule marked by the doctor, the final value of the nodule progression detection result can be calculated, and then a hypothesis test can be performed to draw a conclusion on whether the nodule has progressed.

[0067] In order to verify the effectiveness of the auxiliary detection method for lung nodules provided in the embodiment of the present invention, the inventors verified the technical solution through large-scale simulation experiments and real clinical data.

[0068] 1. Simulation Experiment Verification

[0069] First, the design of the simulation experiment is to select the CT image of a patient who first discovered a nodule, denoted as Then add the mean value of all pixels in the nodule area to , the variance is Gaussian random perturbation.

[0070] Randomly select two such samples, denoted as and ,Bundle As a baseline CT sample, As a follow-up CT sample. and They are all generated by adding the same random perturbation to the same CT image, so they can be considered as a pair of unchanged nodule samples.

[0071] The above method generates two nodule samples from the perspective of nodule density change. Similarly, we can also generate unchanged samples by changing the diameter of the nodule. Therefore, in order to generate unchanged samples, we are based on , Different combinations of the three parameters of φ and nodule diameter are designed to generate nodule-invariant samples. In this embodiment, The value is , The value is The nodule diameter was considered to increase by 10%, 20% or 50%, and to decrease by 10%, 20% or 50%.

[0072] Based on the above simulation experiment, 100,000 pairs of unchanged nodule samples were generated, and then the Boostrap method was used to calculate the empirical distribution of nodule progression detection results, giving a 95% confidence interval, thereby obtaining the sampling distribution of the test statistic under the null hypothesis (nodules remain unchanged). Based on the same method mentioned above, 20,000 pairs of test samples were generated, in which the ratio of progressive nodules to non-progressive nodules was approximately 1:1, and the generation parameters of the progressive nodule samples in the test samples were a parameter set outside the specified range of the above three parameters.

[0073] For the convenience of subsequent analysis, the above nodule sample pairs were further divided into four groups of subsamples according to the original nodule diameter, namely less than or equal to 5 mm, 5-10 mm, 10-15 mm, and 15-30 mm. , } are the 95% quantile values ​​of DNP obtained for the whole sample and the four subsamples, which are 0.087, 0.002, 0.009, 0.033, and 0.134 respectively. The DNP of the whole sample and each group of test nodule subsamples is compared with the corresponding For comparison, If DNP> , reject the original hypothesis and believe that the nodule has progressed; otherwise, predict that the nodule has no progress. Since the conclusion of this study is to judge whether the nodule has progressed, which is a binary classification problem, the subsequent experimental results will be evaluated using the evaluation indicators of the classification model.

[0074] Figure 4 The confusion matrix for each diameter group is shown. Figure 5 As for the evaluation index values, based on the confusion matrix and evaluation indexes, the sensitivity of the smallest diameter group is the lowest, and only 75.1% of the actual progress test samples are successfully detected. This value increases with the increase of the diameter of the test samples, and the sensitivity of the largest diameter group reaches 99.9%.

[0075] Based on the setting of the simulation experiment, the final results show that for nodules with a diameter greater than 15 mm in the test nodule sample, if they progress, the hypothesis testing method proposed by the present invention can give a correct prediction with a probability of 99.9% without measuring the nodule diameter and volume. The above results are consistent with the actual situation, that is, it is difficult to accurately judge the progress of small nodules based on two CT scans before and after, while the progress of large nodules is more obvious and easy to judge.

[0076] In addition, from the perspective of the full sample test nodules, the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were 85.8%, 93.4%, 92.5% and 87.4% respectively. When the PPV was 92.5%, the sensitivity of the proposed group hypothesis test method could reach 85.8%, which means that the proposed method made correct predictions for 85.8% of the samples with progression at the cost of making 7.5% wrong predictions of progression.

[0077] 2. Verification of real clinical data

[0078] In order to verify the effectiveness of the auxiliary detection method for lung nodules provided by the embodiment of the present invention on a real clinical data set, 436 nodule pair samples were collected from the Department of Thoracic Surgery of a domestic tertiary hospital. These 436 nodule pair samples came from 364 CT pairs of 342 patients. Furthermore, the progress information of the 436 nodule pair samples was extracted from the CT reports of these 342 patients, and was reviewed by three experienced thoracic surgery and imaging experts, and finally the objective and reliable nodule pair progress labels, the nodule pair center coordinates X, Y, and the Z-axis range covered by the nodule (based on the first CT scan) were determined.

[0079] Among them, there are 185 progressive nodule pairs and 251 non-progressive nodule pairs. Among these 436 nodule pair samples, 100 have original nodule diameters less than 5 mm, 153 are in the range of (5,10] mm, 87 are in the range of (10,15] mm, and 96 are in the range of (15,30] mm. The experimental process of the real data set is exactly the same as that of the simulation experiment. After calculating the DNP of the real nodule pair samples of each group, it is compared with the 95% quantile of the unchanged nodule samples of the corresponding group. If the DNP of the real nodule pair sample is greater than the 95% quantile, it is predicted that the sample has progressed, otherwise it is predicted that the sample has not progressed.

[0080] Figure 6 The confusion moments of each group and the entire sample are shown. Figure 7 The PPV of each group was 65.85%, 73.33%, 88.00% and 93.88% respectively, and the accuracy of each group was 83.00%, 86.27%, 89.66% and 90.63% respectively. The PPV and accuracy showed an overall increasing trend with the increase of diameter. The above results are consistent with the actual situation.

[0081] In addition, from the perspective of the full sample test nodules, the sensitivity, specificity, PPV and NPV were 92.97%, 82.87%, 80.00% and 94.12%, respectively. When the PPV was 80.00%, the sensitivity of the proposed group hypothesis test method could reach 92.97%, which means that the proposed method made a correct prediction for 92.97% of the samples with progression at the cost of 20% wrong predictions of progression. When the NPV was 94.12%, the specificity of the proposed group hypothesis test method could reach 82.87%, which means that the proposed method made a correct prediction for 82.87% of the samples without progression at the cost of 5.88% wrong predictions of no progression.

[0082] Second embodiment

[0083] like Figure 8 As shown, based on the above-mentioned first embodiment, the second embodiment of the present invention further provides a lung nodule auxiliary detection device for paired CT images, including an acquisition module 1, a segmentation module 2, a matrix transformation module 3, a calculation module 4 and an output module 5.

[0084] Specifically, the acquisition module 1 is an image acquisition unit, which is used to acquire a first CT image of a patient's lung scan and acquire a two-dimensional CT slice sequence corresponding to the first CT image.

[0085] The segmentation module 2 is connected to the acquisition module 1 to divide each two-dimensional CT slice in the two-dimensional CT slice sequence into a plurality of sub-block matrices.

[0086] The matrix transformation module 3 is connected to the segmentation module 2 and has a preset image algorithm built in, so as to extract the matrix transformation result of each sub-block matrix through the preset image algorithm for each sub-block matrix.

[0087] The calculation module 4 is connected to the matrix transformation module 3 so as to obtain, for the sub-block matrix with the same number of slices and the same position, the nodule progression revelation factor corresponding to the sub-block matrix at the position based on the matrix transformation result of the first CT image and the matrix transformation result of the second CT image; and, based on the nodule progression revelation factors of all sub-block matrices of the two-dimensional CT slice with the same number of slices, obtain the nodule progression detection amount of the two-dimensional CT slice with the same number of slices.

[0088] The output module 5 is connected to the calculation module 4 to comprehensively output the final nodule progression detection result based on the nodule progression detection amount of each two-dimensional CT slice.

[0089] It can be understood that the above-mentioned acquisition module 1, segmentation module 2, matrix transformation module 3, calculation module 4 and output module 5 are module structures corresponding to the above-mentioned steps S1 to S6, so as to implement the above-mentioned operations of S1 to S6, wherein the specific composition of the modules is not elaborated here, and it is only necessary to satisfy the above-mentioned functions.

[0090] Third embodiment

[0091] like Fig. 9 As shown, based on the above-mentioned first embodiment, the third embodiment of the present invention further provides a pulmonary nodule auxiliary detection device for paired CT images. The pulmonary nodule auxiliary detection device includes one or more processors 21 and a memory 22. The memory 22 is coupled to the processor 21 and is used to store one or more programs. When the one or more programs are executed by the one or more processors 21, the one or more processors 21 implement the pulmonary nodule auxiliary detection method for paired CT images as in the above-mentioned embodiment.

[0092] Among them, the processor 21 is used to control the overall operation of the pulmonary nodule auxiliary detection device to complete all or part of the steps of the above-mentioned pulmonary nodule auxiliary detection method for paired CT images. The processor 21 can be a central processing unit (CPU), a graphics processing unit (GPU), a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC), a digital signal processing (DSP) chip, etc. The memory 22 is used to store various types of data to support the operation of the pulmonary nodule auxiliary detection device, and these data may include, for example, instructions for any application or method for operating on the pulmonary nodule auxiliary detection device, and application-related data. The memory 22 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, etc.

[0093] In an exemplary embodiment, the pulmonary nodule auxiliary detection device can be specifically implemented by a computer chip or entity, or by a product with a certain function, for executing the above-mentioned pulmonary nodule auxiliary detection method for paired CT images, and achieving the same technical effect as the above-mentioned method. A typical embodiment is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a vehicle-mounted human-computer interaction device, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0094] In another exemplary embodiment, the present invention further provides a computer-readable storage medium including program instructions, which, when executed by a processor, implements the steps of the method for auxiliary detection of pulmonary nodules in paired CT images in any of the above embodiments. For example, the computer-readable storage medium may be the above-mentioned memory including program instructions, and the above-mentioned program instructions may be executed by a processor of a pulmonary nodule auxiliary detection device to complete the above-mentioned method for auxiliary detection of pulmonary nodules in paired CT images, and achieve the technical effect consistent with the above-mentioned method.

[0095] In summary, the embodiment of the present invention provides a method and device for auxiliary detection of pulmonary nodules in paired CT images, which has the following beneficial effects:

[0096] (1) From the perspective of hypothesis testing, a unique and precise detection method was designed to measure the changes in patients’ nodules in combination with a specific image processing algorithm. Thus, by quantifying the changes in the lesions in the two CT images before and after, it helps doctors observe and analyze the progression of lung nodules more accurately.

[0097] (2) By studying whether there are significant differences in the local nodule images between the two CT scans of the patient, the detection quantity is used to provide an answer to whether the changes in the lung nodules are statistically significant, which makes up for the fact that existing methods cannot answer the key question of whether the detected changes are statistically significant.

[0098] (3) The nodule progression detection value of each two-dimensional CT slice is represented by a pre-set hyperparameter, and a two-dimensional curve between the nodule progression detection value and the hyperparameter is constructed. The influence of the hyperparameter is eliminated by calculating the area under the curve of the two-dimensional curve, thereby improving the accuracy of the final nodule progression detection result.

[0099] It should be noted that the above embodiments are only examples, and the technical solutions of the various embodiments can be combined, all within the protection scope of the present invention.

[0100] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0101] The above is a detailed description of the auxiliary detection method and device for pulmonary nodules in paired CT images provided by the present invention. For a person skilled in the art, any obvious changes made to the invention without departing from the essence of the invention will constitute an infringement of the patent right of the present invention and the person shall bear the corresponding legal liability.

Claims

1. A method for assisting detection of pulmonary nodules in paired CT images, characterized in that The steps include: Acquire a first CT image of a patient's lung scan, and acquire a two-dimensional CT slice sequence corresponding to the first CT image; For each two-dimensional CT slice in the two-dimensional CT slice sequence, dividing each of the two-dimensional CT slices into a plurality of sub-block matrices of the same size and number; For each of the sub-block matrices, extracting a matrix transformation result of each of the sub-block matrices by using a preset image algorithm; Reacquire a second CT image of the patient's lung scan, and extract the matrix transformation result of each of the sub-block matrices of the second CT image in the same manner; For the sub-block matrix with the same number of slices and the same position, based on the matrix transformation result of the first CT image and the matrix transformation result of the second CT image, obtaining the nodule display factor corresponding to the sub-block matrix at the position; Based on the nodule progression display factors of all sub-block matrices of the two-dimensional CT slice under the same number of slices, obtaining the nodule progression detection amount of the two-dimensional CT slice under the number of slices; Based on the nodule progression detection amount of each of the two-dimensional CT slices, a final nodule progression detection result is comprehensively output.

2. The pulmonary nodule auxiliary detection method according to claim 1, characterized in that: The two-dimensional CT slice sequence is a continuous CT slice sequence that only contains the target lung nodules; wherein, the two-dimensional CT slice sequence is obtained through the following steps: first, a lung CT image of the patient is obtained through tomography technology; then, a doctor reviews and annotates the lung CT image, and selects continuous two-dimensional CT slices containing the target lung nodules to form a sequence.

3. The pulmonary nodule auxiliary detection method according to claim 1, characterized in that Dividing the two-dimensional CT slice into a plurality of sub-block matrices comprises the following sub-steps: Based on the size of the target lung nodule, determine the number n to be divided; where n is a positive integer and its value is related to the size of the nodule; Based on the number n to be divided, the two-dimensional CT slice is divided into n n sub-block matrices; Each of the sub-block matrices is m m matrix, where m is the size of the sub-block matrix.

4. The pulmonary nodule auxiliary detection method according to claim 3, characterized in that The preset image algorithm includes the following sub-steps: Performing a logarithmic transformation on the pixel point HU values ​​of the elements at each position in the (n1, n2)th sub-block matrix of the t-th two-dimensional CT slice of the first CT image to obtain the (n1, n2)th sub-block matrix of the t-th two-dimensional CT slice ; Wherein, 1≤n1, n2≤n, t is a positive integer; For the (n1, n2)th sub-block matrix of the t-th two-dimensional CT slice , define its binary characteristic parameter , where when the sub-block matrix When the target nodule center point is included, ,otherwise, ; For the t-th 2D CT slice of the first CT image, assume that the sum of the binary characteristic parameters of all sub-block matrices is 1, and the center point of the target nodule appears at In the sub-block matrix of the position, define , 1≤ ≤n; calculate , as the matrix transformation result of the (n1, n2)th sub-block matrix in the tth CT slice.

5. The pulmonary nodule auxiliary detection method according to claim 4, characterized in that The (n1, n2)th sub-block matrix of the t-th 2D CT slice , generated by: Using a three-dimensional tensor The first CT image is defined, wherein represents the t-th two-dimensional CT slice, T is the total number of CT slices, p represents the dimension of the two-dimensional CT; R represents the real number field; The tth CT slice of the first CT image middle The logarithmic transformation of the HU value of the pixel position is recorded as ; The tth CT slice Divide into n n sub-block matrices, recorded as ,in is m m matrix, p=n m; Defining an index collection , to re- Recorded as .

6. The pulmonary nodule auxiliary detection method according to claim 4, characterized in that The nodule presentation factor corresponding to the sub-block matrix is ​​calculated in the following way: For the (n1, n2)th sub-block matrix in the tth CT slice, obtain the matrix transformation results of the first CT image respectively And the matrix transformation result of the second CT image ; Calculate the nodule display factor of the (n1, n2)th sub-block matrix in the tth CT slice : in, is a pre-set hyperparameter and 0< <1; It means that the nodule has a tendency to increase in size on the tth CT slice; It means that the nodule has a tendency to shrink on the tth CT slice; It means that there is no clear change in the nodule on the tth CT slice.

7. The pulmonary nodule auxiliary detection method according to claim 6, characterized in that The nodule progression detection amount of the two-dimensional CT slice is calculated by the following method: Obtain the nodule presentation factor of each sub-block matrix in the t-th two-dimensional CT slice; Based on the nodule display factor of all sub-block matrices in the t-th two-dimensional CT slice, the average value is calculated. as the nodule progression detection amount of the t-th 2D CT slice; in, , for any , .

8. The pulmonary nodule auxiliary detection method according to claim 7, characterized in that The final nodule progression detection result is calculated as follows: Nodule progression detection based on the t-th 2D CT slice And hyperparameters , build Two-dimensional curve of Calculate the The area under the curve is recorded as the nodule progression detection result of the t-th 2D CT slice, and is recorded as DNP t ; Based on the nodule progression detection results of all two-dimensional CT slices, the final nodule progression detection result is output, recorded as DNP; where, .

9. A pulmonary nodule auxiliary detection device for paired CT images, characterized in that include: An acquisition module, used to acquire a first CT image of a patient's lung scan, and acquire a two-dimensional CT slice sequence corresponding to the first CT image; a segmentation module connected to the acquisition module to divide the two-dimensional CT slice into a plurality of sub-block matrices for each two-dimensional CT slice in the two-dimensional CT slice sequence; A matrix transformation module, connected to the segmentation module and having a preset image algorithm built therein, so as to extract the matrix transformation result of each sub-block matrix through the preset image algorithm for each sub-block matrix; a calculation module connected to the matrix transformation module, so as to obtain, for the sub-block matrix of the same number of slices and the same position, a nodule progression revelation factor corresponding to the sub-block matrix at the position based on the matrix transformation result of the first CT image and the matrix transformation result of the second CT image; and, based on the nodule progression revelation factors of all sub-block matrices of the two-dimensional CT slice under the same number of slices, obtain the nodule progression detection amount of the two-dimensional CT slice under the same number of slices; The output module is connected to the calculation module to comprehensively output the final nodule progression detection result based on the nodule progression detection amount of each two-dimensional CT slice.

10. A pulmonary nodule auxiliary detection device for paired CT images, characterized in that It comprises a processor and a memory; wherein the memory is coupled to the processor and is used to store a computer program, and when the computer program is executed by the processor, the processor implements the pulmonary nodule auxiliary detection method according to any one of claims 1 to 8.

Images