CN114366037A - Spectral diagnostic system for intraoperative guidance of brain cancer and its working method - Google Patents

Abstract

The spectrum diagnosis system and the working method for guiding in brain cancer operation can give the diagnosis results of normal and pathological changes in real time or near real time, the detection probe can easily reach the operation area to meet the demand of in-operation in-vivo diagnosis, and the brain tissue is not damaged when the detection probe is used in vivo. The control unit controls the laser light source unit and the spectrum acquisition unit to cooperatively work to acquire spectrum data, and the data processing unit and the intraoperative early warning unit are combined to make a tumor grade early warning for an intraoperative surgeon; the data processing unit is connected with the control unit and the intraoperative warning unit, the control unit inputs the spectral data into the data processing unit, and the data processing unit analyzes the data and outputs the result to the intraoperative warning unit.

Description

Spectral diagnostic system for intraoperative guidance of brain cancer and its working method

technical field

The present invention relates to the technical field of photoelectric detection, in particular to a spectral diagnosis system used for intraoperative guidance of brain cancer, and a working method of the spectral diagnosis system used for intraoperative guidance of brain cancer.

Background technique

Glioma is a very common clinical brain tumor, which is characterized by: 1) glioma has no clear tumor boundary; 2) glioma has strong permeability, and malignant glioma will penetrate into the surrounding 3) The growth rate of glioma is very fast, and its survival period is about 12-15 months. According to the World Health Organization's glioma grading standards, gliomas can be divided into four grades, among which glioblastoma (GBM) has high permeability and growth rate and needs to be removed or treated as soon as possible to prolong the patient survival. In intracranial surgery, in order to maximize the resection margin, the surgeon can only decide the extent of the tumor resection margin based on experience and preoperative medical images, combined with the results of intraoperative frozen section. Traditional frozen sections are time-consuming and may have sampling errors. Surgeons urgently need an intraoperative diagnosis method with the following characteristics: 1) The diagnostic equipment needs to realize the rapid detection of brain tissue during the operation, and can give the diagnosis results of normal and lesions in real time or near real time; 2) The diagnostic equipment, especially the equipment The detection probe can be miniaturized, so that the detection probe can easily reach the surgical area and meet the needs of intraoperative in vivo diagnosis; 3) When used in vivo, there is no damage to the brain tissue.

SUMMARY OF THE INVENTION

In order to overcome the defects of the prior art, the technical problem to be solved by the present invention is to provide a spectral diagnosis system for intraoperative guidance of brain cancer, which can provide the diagnosis results of normal and lesions in real time or near real time, and the detection probe can It can easily reach the surgical area to meet the needs of intraoperative in vivo diagnosis, and there is no damage to brain tissue when used in vivo.

The technical scheme of the present invention is: this kind of spectral diagnosis system for brain cancer intraoperative guidance, which comprises:

Excitation light source unit, spectrum acquisition unit, control unit, data processing unit, intraoperative warning unit;

The control unit controls the laser light source unit and the spectral acquisition unit to work together to collect spectral data, and the data processing unit is combined with the intraoperative early warning unit to make tumor grade early warning for the intraoperative surgeon; the data processing unit is connected with the control unit and the intraoperative warning unit , the control unit inputs the spectral data into the data processing unit, and after the data processing unit analyzes the data, the result is output to the intraoperative warning unit;

The data processing unit includes a deep learning development board (5), which calculates three groups of tissue optical parameters with different light source-detector distances from the three groups of diffuse reflectance spectral data respectively through Monte Carlo inversion operation; The parameters and the endogenous fluorescence spectrum data with the same light source-detector distance of the three groups are input into the fluorescence quantitative algorithm to obtain three groups of fluorescence quantitative data;

The three groups of tissue optical parameters and three groups of fluorescence quantitative data are respectively input into the multilayer perceptron model to obtain tumor grade prediction. The multilayer perceptron model can classify the data set into high-grade tumors, low-grade tumors, and normal tissues; The second execution will get the prediction results of 3 sets of tissue optical parameters and 3 sets of fluorescence quantitative data.

The invention can combine the diffuse reflection spectrum with the endogenous fluorescence spectrum, detect the tissue optical characteristics and endogenous fluorescence information of the target lesion, and use the machine learning algorithm to grade the tumor of the target lesion, so that the normal or near real time can be given. And the diagnosis results of lesions, the detection probe can easily reach the surgical area to meet the needs of intraoperative in vivo diagnosis, and there is no damage to brain tissue when used in vivo.

Also provided is a working method of a spectroscopic diagnostic system for intraoperative guidance of brain cancer, which includes the following steps:

(1) Data acquisition: Connect the multimode fiber to the spectrometer and the excitation light source, and use the USB controller to connect the spectrometer, the excitation light source and the computer. The multimode fiber probe was aimed at the target lesion, and the data acquisition software based on LabVIEW system was used to collect the diffuse reflectance spectrum and endogenous fluorescence spectrum;

(2) Calculation of tissue optical parameters and fluorescence quantitative data: using the diffuse reflectance spectral data obtained in step (1), the tissue optical parameters of the target lesion are extracted, and the endogenous fluorescence spectral data obtained in step (1) is used to perform fluorescence. Quantitative operation to obtain fluorescence quantitative data;

(3) Machine learning model training: using the tissue optical parameters and fluorescence quantitative data obtained in step (2) as a data set, input the data set into the machine learning model to train the model, determine the optimal threshold of the classification model, and classify the target lesions into The following three grades are: normal tissue, low-grade tumor, and high-grade tumor;

(4) The system makes a warning prompt to the clinical surgeon during the operation according to the grade of the target lesion obtained in step (3).

Description of drawings

FIG. 1 shows a schematic structural diagram of a spectroscopic diagnosis system for intraoperative guidance of brain cancer according to the present invention.

FIG. 2 shows the front view of the excitation light source and the spectrometer of the spectroscopic diagnosis system for intraoperative guidance of brain cancer according to the present invention.

FIG. 3 shows a longitudinal cross-sectional view of the excitation light source and the spectrometer of the spectroscopic diagnosis system for intraoperative guidance of brain cancer according to the present invention.

FIG. 4 shows an example diagram of a multimode fiber optic probe for a brain cancer intraoperative guided spectroscopic diagnosis system according to the present invention.

FIG. 5 shows an example diagram of a control unit interface of a spectroscopic diagnosis system for brain cancer intraoperative guidance according to the present invention.

FIG. 6 shows a schematic diagram of an intraoperative warning interface of the spectral diagnosis system for brain cancer intraoperative guidance according to the present invention.

FIG. 7 shows a flow chart of spectral data analysis for the intraoperative guided spectral diagnosis system for brain cancer according to the present invention.

FIG. 8 shows a schematic diagram of the intraoperative guidance of the spectroscopic diagnosis system for brain cancer intraoperative guidance according to the present invention in guiding brain cancer surgery diagnosis.

Detailed ways

In order to make those skilled in the art better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only Embodiments are part of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

It should be noted that the term "comprising" and any modification in the description and claims of the present invention and the above drawings are intended to cover non-exclusive inclusion, for example, a process, method, device including a series of steps or units , products or devices are not necessarily limited to those steps or units expressly listed, but may include other steps or units not expressly listed or inherent to these processes, methods, products or devices.

As shown in Figures 1-6, this spectroscopic diagnosis system for brain cancer intraoperative guidance includes:

Excitation light source unit, spectrum acquisition unit, control unit, data processing unit, intraoperative warning unit;

The control unit controls the laser light source unit and the spectral acquisition unit to work together to collect spectral data, and the data processing unit is combined with the intraoperative early warning unit to make tumor grade early warning for the intraoperative surgeon; the data processing unit is connected with the control unit and the intraoperative warning unit , the control unit inputs the spectral data into the data processing unit, and after the data processing unit analyzes the data, the result is output to the intraoperative warning unit;

The data processing unit includes a deep learning development board 5, which calculates three groups of tissue optical parameters with different light source-detector distances from three groups of diffuse reflectance spectral data through Monte Carlo inversion operation; Three sets of endogenous fluorescence spectral data with the same light source-detector distance were input into the fluorescence quantitative algorithm to obtain three sets of fluorescence quantitative data; three sets of tissue optical parameters and three sets of fluorescent quantitative data were respectively input into the multilayer perceptron model to obtain tumor grading It is predicted that the multi-layer perceptron model can classify the data set into high-grade tumors, low-grade tumors, and normal tissues; each execution of the data processing unit will obtain the prediction results of 3 sets of tissue optical parameters and 3 sets of fluorescence quantitative data.

The invention can combine the diffuse reflection spectrum with the endogenous fluorescence spectrum, detect the tissue optical characteristics and endogenous fluorescence information of the target lesion, and use the machine learning algorithm to grade the tumor of the target lesion, so that the normal or near real time can be given. And the diagnosis results of lesions, the detection probe can easily reach the surgical area to meet the needs of intraoperative in vivo diagnosis, and there is no damage to brain tissue when used in vivo.

Preferably, the excitation light source unit comprises: 3 white LED light sources 14, 15, 16 with a wavelength range of 300-1000 nm, 3 405 nm purple LED light sources 11, 12, 13, an LED controller 17, and an excitation light source panel 1; The LED controller 17 is connected to the excitation light source panel 1 through a data line, and controls the LED switch, input current, and excitation time.

Preferably, the spectrum acquisition unit includes: a multimode optical fiber 3, an optical fiber probe 38, and a spectrometer 2; the total number of channels of the optical fiber probe 38 is 7, which are arranged in a "one" shape, and the core diameter of the optical fiber is 200 μm. The center-to-center distance between the two adjacent fibers is 240 μm; the first channel 34 is connected to the spectrometer, the second channel 35, 36, 37 is connected to the three white LED light sources 14, 15, 16 in the excitation light source unit, and the third channel 31, 32 , 33 are connected to the three 405nm violet LED light sources 11, 12, and 13 in the excitation light source unit; the diffuse reflection spectrum data and endogenous fluorescence spectrum data of three different light source-detector distances are obtained for each acquisition, and different light source-detector distances are obtained. The distance data has tissue optical information and endogenous fluorescence information at different depths of the target lesion.

Preferably, the control unit includes: a computer 4, a USB controller 41 and a data acquisition unit 42 based on the LabVIEW system; the computer 4 is connected with the LED controller 17 and the spectrometer 2 by using the USB controller 41, and the The data acquisition unit 42 controls the switches and system parameters of the system, and the system parameters include: the excitation sequence of the light source, the excitation time of the light source, and the integration time of the spectrometer.

Preferably, the intraoperative warning unit includes: a display 6 , a display interface 60 , and a warning buzzer 63 . The contents of the display interface include: tumor grade prediction icon 61, warning light 62, diffuse reflection spectrum graph 64, and fluorescence spectrum graph 65; wherein,

When more than one group of prediction results are high-grade tumors, the predicted target lesions are high-grade tumors, and the warning buzzer 63 emits a high-frequency alarm sound to warn the surgeon during the operation. The arrow in the tumor warning icon 61 points to In the high-grade tumor area, the high-grade tumor warning light in the warning light 62 is on;

When more than one prediction result is a low-grade tumor, the predicted target lesion is a low-grade tumor, and the warning buzzer 63 emits a high-frequency alarm sound to warn the surgeon during the operation. The arrow in the tumor warning icon 61 points to In the low-grade tumor area, the low-grade tumor warning light in the warning light 62 is on;

When the predicted target lesion is normal tissue, the arrow in the tumor warning icon 61 points to the normal tissue area, and the normal tissue warning light in the warning light 62 lights up.

As shown in FIG. 7 , a working method of a spectroscopic diagnosis system for intraoperative guidance of brain cancer is also provided, which includes the following steps:

(1) Data acquisition: Connect the multimode fiber to the spectrometer and the excitation light source, and use the USB controller to connect the spectrometer, the excitation light source and the computer. The multimode fiber probe was aimed at the target lesion, and the data acquisition software based on LabVIEW system was used to collect the diffuse reflectance spectrum and endogenous fluorescence spectrum;

(2) Calculation of tissue optical parameters and fluorescence quantitative data: using the diffuse reflectance spectral data obtained in step (1), the tissue optical parameters of the target lesion are extracted, and the endogenous fluorescence spectral data obtained in step (1) is used to perform fluorescence. Quantitative operation to obtain fluorescence quantitative data;

(3) Machine learning model training: using the tissue optical parameters and fluorescence quantitative data obtained in step (2) as a data set, input the data set into the machine learning model to train the model, determine the optimal threshold of the classification model, and classify the target lesions into The following three grades are: normal tissue, low-grade tumor, and high-grade tumor;

(4) The system makes a warning prompt to the clinical surgeon during the operation according to the grade of the target lesion obtained in step (3).

Specific embodiments of the present invention will be described in detail below.

Example 1: Diagnosis of brain tissue as a high-grade tumor

As shown in Figure 8, the workflow of the present invention is:

(1) Connect the multimode fiber 3 with the excitation light source 1 and the spectrometer 2, use the USB controller 41, connect the computer 4 with the excitation light source 1 and the spectrometer 2, control the excitation light source 1 and the spectrometer 2 to work together, and connect the computer 4 , Deep Learning Development Board 5, Display 6.

(2) sticking the multimode fiber probe 38 to the lesion 7 to be tested;

(3) On the control unit interface (Figure 5), set the excitation sequence of the excitation light source to channel01, channel02, channel03, channel04, channel05, and channel06, the excitation time is set to 200ms, and the integration time of the spectrometer is 100ms, click "Start Acquisition".

(4) In the intraoperative warning unit, the spectral curves of the diffuse reflectance spectrum and the fluorescence spectrum of the brain tissue of this part can be seen. The intraoperative warning unit predicts and classifies the lesion 7 as a high-grade tumor, the tumor grade prediction icon 61 will point to the high-grade tumor area, the high-grade tumor warning light in the warning light 62 will light up, and the warning buzzer 63 will emit a high-grade tumor. beeps frequently to alert the surgeon.

Example 2: Diagnosis of brain tissue as normal tissue

As shown in Figure 8, the workflow of the present invention is:

(1) Connect the multimode fiber 3 with the excitation light source 1 and the spectrometer 2, use the USB controller 41, connect the computer 4 with the excitation light source 1 and the spectrometer 2, control the excitation light source 1 and the spectrometer 2 to work together, and connect the computer 4 , Deep Learning Development Board 5, Display 6.

(2) sticking the multimode fiber probe 38 to the lesion 7 to be tested;

(3) On the control unit interface (Figure 5), set the excitation sequence of the excitation light source to channel01, channel02, channel03, channel04, channel05, and channel06, the excitation time is set to 200ms, and the integration time of the spectrometer is 100ms, click "Start Acquisition".

(4) In the intraoperative warning unit, the spectral curves of the diffuse reflectance spectrum and the fluorescence spectrum of the brain tissue of this part can be seen. The intraoperative warning unit predicts and classifies the lesion 7 as normal tissue, the tumor grade prediction icon 61 will point to the normal tissue area, and the normal tissue warning light in the warning light 62 will be on.

The above are only preferred embodiments of the present invention, and do not limit the present invention in any form. Any simple modifications, equivalent changes and modifications made to the above embodiments according to the technical essence of the present invention still belong to the present invention The protection scope of the technical solution of the invention.

Claims (6)

1. A spectral diagnosis system for intraoperative guidance of brain cancer, characterized in that: it comprises: Excitation light source unit, spectrum acquisition unit, control unit, data processing unit, intraoperative warning unit; The control unit controls the laser light source unit and the spectral acquisition unit to work together to collect spectral data, and the data processing unit is combined with the intraoperative early warning unit to make tumor grade early warning for the intraoperative surgeon; the data processing unit is connected with the control unit and the intraoperative warning unit , the control unit inputs the spectral data into the data processing unit, and after the data processing unit analyzes the data, the result is output to the intraoperative warning unit; The data processing unit includes a deep learning development board (5), which calculates three groups of tissue optical parameters with different light source-detector distances from the three groups of diffuse reflectance spectral data through Monte Carlo inversion operation respectively; The parameters and the three sets of endogenous fluorescence spectrum data with the same light source-detector distance are input into the fluorescence quantitative algorithm to obtain three sets of fluorescence quantitative data; three sets of tissue optical parameters and three sets of fluorescence quantitative data are respectively input into the multilayer perceptron model to obtain For tumor grade prediction, the multi-layer perceptron model can classify the data set into high-grade tumors, low-grade tumors, and normal tissues; each execution of the data processing unit will obtain 3 sets of tissue optical parameters and 3 sets of fluorescence quantitative data prediction results. 2 . The spectral diagnosis system for intraoperative guidance of brain cancer according to claim 1 , wherein the excitation light source unit comprises: 3 white LED light sources (14, 15, 16) with a wavelength range of 300-1000 nm. 3 . ), three 405nm violet LED light sources (11, 12, 13), an LED controller (17), and an excitation light source panel (1); the LED controller (17) is connected to the excitation light source panel (1) through a data cable to control the LED’s switch, input current, excitation time. 3. The spectral diagnosis system for intraoperative guidance of brain cancer according to claim 2, wherein the spectral acquisition unit comprises: a multimode optical fiber (3), an optical fiber probe (38), and a spectrometer (2); The total number of channels of the optical fiber probe (38) is 7, which are arranged in a line, the core diameter of the optical fiber is 200 μm, and the center distance of two adjacent optical fibers is both 240 μm; the first channel (34) is connected to the spectrometer, and the first channel (34) is connected to the spectrometer. The second channel (35, 36, 37) is connected to the three white LED light sources (14, 15, 16) in the excitation light source unit, and the third channel (31, 32, 33) is connected to the three 405nm violet LED light sources in the excitation light source unit (11, 12, 13); Diffuse reflectance spectrum data and endogenous fluorescence spectrum data of 3 different light source-detector distances are obtained for each acquisition, and the data of different light source-detector distances have tissue optical information of different depths of the target lesion and endogenous fluorescence information. 4. The spectroscopic diagnosis system for brain cancer intraoperative guidance according to claim 3, wherein the control unit comprises: a computer (4), a USB controller (41) and a data acquisition unit based on a LabVIEW system (42); use the USB controller (41) to connect the computer (4) with the LED controller (17) and the spectrometer (2), and use the data acquisition unit (42) based on the LabVIEW system to control the switches and system parameters of the system, System parameters include: the excitation sequence of the light source, the excitation time of the light source, and the integration time of the spectrometer. 5. The spectroscopic diagnosis system for brain cancer intraoperative guidance according to claim 4, wherein the intraoperative warning unit comprises: a display (6), a display interface (60), a warning buzzer ( 63); the content of the display interface includes: a tumor grade prediction icon (61), a warning light (62), a diffuse reflection spectrum graph (64), and a fluorescence spectrum graph (65); wherein, When more than one group of prediction results are high-grade tumors, the predicted target lesions are high-grade tumors, and the warning buzzer (63) emits a high-frequency alarm sound to alert the surgeon during the operation, and the tumor warning icon (61) The arrow in the arrow points to the high-grade tumor area, and the high-grade tumor warning light in the warning light (62) lights up; When more than one prediction result is a low-grade tumor, the target lesion is predicted to be a low-grade tumor, and the warning buzzer (63) emits a high-frequency alarm sound to alert the surgeon during the operation, and the tumor warning icon (61) The arrow in the arrow points to the low-grade tumor area, and the low-grade tumor warning light in the warning light (62) lights up; When the predicted target lesion is normal tissue, the arrow in the tumor warning icon (61) points to the normal tissue area, and the normal tissue warning light in the warning light (62) lights up. 6. The working method of a spectroscopic diagnosis system for intraoperative guidance of brain cancer, characterized in that it comprises the following steps: (1) Data acquisition: connect the multimode fiber to the spectrometer and the excitation light source, and use the USB controller to connect the spectrometer and the excitation light source to the computer; point the multimode fiber probe to the target lesion, and use the data acquisition software based on LabVIEW system Collect diffuse reflectance spectra and endogenous fluorescence spectra; (2) Calculation of tissue optical parameters and fluorescence quantitative data: using the diffuse reflectance spectral data obtained in step (1), the tissue optical parameters of the target lesion are extracted, and the endogenous fluorescence spectral data obtained in step (1) is used to perform fluorescence. Quantitative operation to obtain fluorescence quantitative data; (3) Machine learning model training: using the tissue optical parameters and fluorescence quantitative data obtained in step (2) as a data set, input the data set into the machine learning model to train the model, determine the optimal threshold of the classification model, and classify the target lesions into The following three grades are: normal tissue, low-grade tumor, and high-grade tumor; (4) The system makes a warning prompt to the clinical surgeon during the operation according to the grade of the target lesion obtained in step (3).

Images