Double-field scanning system and method for detecting circulating tumor cells
Through the combination of the dual-field scanning system and the MeanShift algorithm, the complexity and false positive false negative problems of circulating tumor cell detection in the prior art are solved, efficient and accurate CTC recognition is achieved, and the integrity and detection efficiency of cells are ensured.
Patent Information
- Application Number
- CN202411575437.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-06
- Publication Date
- 2025-01-07
AI Technical Summary
There are many steps for the detection of circulating tumor cells, making it difficult to effectively enrich and detect all tumor cells, resulting in false negative or false positive results, and traditional methods are harmful to cell activity and morphology.
The dual-field scanning system is combined with the MeanShift algorithm to receive signals through the bright-field reflected light path and the dark-field scattered light path, and CTC enrichment is used to use micro-filtration technology, and the CTC is automatically recognized through the MeanShift algorithm to improve detection efficiency and accuracy.
It realizes CTC detection with high resolution accuracy, simplifies operation steps, reduces detection costs, ensures cell integrity and activity, and significantly improves the accuracy and detection efficiency of CTC recognition.
Smart Images
Figure CN119269452A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of circulating tumor cell detection, and in particular to a dual-field scanning system and method for detecting circulating tumor cells. Background Art
[0002] Due to the rarity, heterogeneity, and diverse morphological types of circulating tumor cells (CTCs), the enrichment and detection of circulating tumor cells from patients' blood samples remains a technical challenge. Before CTC detection, tumor cells need to be separated from the sample, but due to the common heterogeneity of tumor cells, a single antibody cannot be used to label all tumor cells. Therefore, some tumor cells are often lost during enrichment, which can result in false negative or false positive test results.
[0003] At present, common CTC enrichment and detection methods include magnetic cell separation (MACS), microfluidic detection devices, immunofluorescence staining, etc. However, most of these methods have many steps and cannot enrich and detect all tumor cells. For example, magnetic cell separation technology cannot detect CTCs that lack epithelial markers. For commercial detection methods, such as CellSearch, this method uses immunomagnetic beads to enrich CTCs, uses epithelial cell adhesion factor (EpCAM) to distinguish positive CTCs from normal blood cells, and then uses anti-CD45 fluorescent dye, cytokeratin and nuclear fluorescent staining reagent DAPI to jointly identify CTCs. The overall operation steps are numerous, and the detection efficiency is relatively low, which often leads to the loss or destruction of CTCs and changes in cell activity. Due to the heterogeneity of tumor cells, not all tumor cells express epithelial markers, so this method is also affected by epithelial markers.
[0004] With the rapid development of artificial intelligence technology, more and more machine learning algorithms are applied in medical research. The unsupervised learning MeanShift algorithm is an algorithm that can automatically find the extreme points of the probability density function. It can be used for applications such as peak seeking and clustering. By setting the area radius, it can automatically find the maximum point of the probability density function in the sample point.
[0005] Therefore, the object of the present invention is to develop a method for detecting circulating tumor cells with high resolution accuracy. Summary of the invention
[0006] Based on the technical problems existing in the background technology, the present invention proposes a dual-field scanning system and method for detecting circulating tumor cells, which has the advantages of fast scanning speed and high resolution accuracy.
[0007] The dual-field scanning system for detecting circulating tumor cells proposed by the present invention comprises: A sample stage, used to place the sample to be tested; A detection light path, used to provide a detection light source for the sample stage; Bright field reflected light path, used to receive reflected light signals from the sample stage; Dark field scattered light path, used to receive scattered light signals from the sample stage; The data processing module recognizes circulating tumor cells using the MeanShift algorithm based on the signals received by the bright field reflected light path and the dark field scattered light path.
[0008] Preferably, ellipsoidal reflectors are further provided on both sides of the sample stage, and the scattered light of the sample stage is received by a dark field scattered light path after passing through the ellipsoidal reflectors.
[0009] Preferably, the dark field scattering light path comprises a filter, a second aperture and a photomultiplier tube PMT which are arranged in sequence, and the filter is located on a side close to the sample stage.
[0010] Preferably, the bright field reflection light path comprises a focusing lens, a first aperture and a photodetector PDA which are arranged in sequence, and the focusing lens is located on a side close to the sample stage.
[0011] Preferably, the calculation method steps of the MeanShift algorithm are as follows: S11: Assume there are n sample points, i=1, 2, 3, ..., n, then the kernel density estimation function is: in, is the kernel function; is the window radius; is the dimension of the vector; is the normalization constant; S12: Determine the MeanShift vector as: in, ; S13: Using formula By continuously iterating, the algorithm can converge to the density extreme point and realize the identification of circulating tumor cells.
[0012] The dual-field scanning method for detecting circulating tumor cells proposed by the present invention adopts the dual-field scanning system for detecting circulating tumor cells, and the method steps are as follows: S21: emitting laser light through the detection light path to illuminate the sample on the sample stage; S22: by adjusting the height of the sample stage and the bright field reflection light path, the bright field reflection light path converts the received reflection light signal into an electrical signal; S23: by adjusting the height of the sample stage and the dark field scattering light path, the dark field scattering light path converts the received scattered light signal into an electrical signal; S24: Through the data processing module, the MeanShift algorithm is used to identify circulating tumor cells according to the signals received by the bright field reflection light path and the dark field scattered light path.
[0013] Beneficial technical effects of the present invention: The present invention uses microfiltration technology to enrich CTCs, and utilizes the advantages of microfiltration technology, which is simple to operate, low cost, independent of CTC surface antigen expression, and can ensure the complete morphology and activity of cells, to overcome the disadvantages of traditional CTC enrichment, which is high cost and easy to cause cell damage.
[0014] The present invention can greatly improve the accuracy of CTC recognition through dual recognition of bright field and dark field.
[0015] The present invention realizes automatic recognition of CTC through MeanShift algorithm, which can greatly improve the detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 This is a schematic diagram of the structure of the dual-field scanning system for detecting circulating tumor cells proposed by the present invention; Figure 2 This is a schematic diagram of a sample plate suitable for a dual-field confocal scanning device used in the present invention.
[0017] In the figure: 1-laser diode, 2-beam expansion lens, 3-objective lens, 4-ellipsoidal reflector, 5-sample stage, 6-focusing lens, 7-first aperture, 8-PDA, 9-filter, 10-PMT, 11-second aperture. DETAILED DESCRIPTION
[0018] The present invention will be further explained below in conjunction with specific embodiments.
[0019] Example 1 Reference Figure 1 The dual-field scanning system for detecting circulating tumor cells proposed in the present invention includes: a sample stage 5 for placing a sample to be tested; a detection light path for providing a detection light source for the sample stage 5; a bright field reflection light path for receiving a reflection light signal of the sample stage 5; a dark field scattering light path for receiving a scattered light signal of the sample stage; and a data processing module for realizing the identification of circulating tumor cells through a MeanShift algorithm according to the signals received by the bright field reflection light path and the dark field scattering light path.
[0020] In order to realize the reception of the optical signal by the dark field scattered light path, ellipsoidal reflectors 4 are further arranged on both sides of the sample stage 5 , and the scattered light of the sample stage 5 is received by the dark field scattered light path after passing through the ellipsoidal reflectors 4 .
[0021] As a preferred solution of the present invention, the dark field scattering light path includes a filter 9 , a second aperture 11 and a photomultiplier tube PMT10 which are arranged in sequence, and the filter 9 is located on a side close to the sample stage 5 .
[0022] As a preferred solution of the present invention, the bright field reflection light path includes a focusing lens 6 , a first aperture 7 and a photodetector PDA8 which are arranged in sequence, and the focusing lens 6 is located on a side close to the sample stage 5 .
[0023] The calculation method steps of the MeanShift algorithm in the data processing module of the present invention are as follows: S11: Assume there are n sample points, i=1, 2, 3, ..., n, then the kernel density estimation function is: in, is the kernel function; is the window radius; is the dimension of the vector; is the normalization constant; S12: Determine the MeanShift vector as: in, ; S13: Using formula By continuously iterating, the algorithm can converge to the density extreme point and realize the identification of circulating tumor cells.
[0024] Example 2 The dual-field scanning method for detecting circulating tumor cells proposed by the present invention has the following steps: First, prepare the circulating tumor cell sample and filter it with an 8μm pre-fixed polycarbonate membrane to screen CTCs. Red blood cells and platelets will be filtered out due to their small size, while larger white blood cells and CTCs will be left to make a blood smear. Then, the blood smear is stained with silver stain and Wright-Giemsa stain and placed in a sample tray (such as Figure 2 The samples were fixed as shown in Figure 2 for detection.
[0025] The overall experimental detection optical path uses a laser diode as the light source, emitting 405nm laser light through a beam expander lens so that the incident light pupil of the objective lens (focal length is 1.512cm) is completely filled, which makes the light spot on the sample surface as small as possible.
[0026] In the bright field, we use a focusing lens to converge the reflected light from the sample surface, and use a photodetector PDA to collect the light intensity signal after the reflected light passes through a 5-micron pinhole aperture. After the pinhole aperture is precisely coupled to the PDA, it will shield the reflected light not on the sample surface and the noise caused by other impurities in the light path, thus achieving layered scanning. In the dark field, we use an ellipsoidal reflector to collect scattered light. According to the principle of geometric optics, the light emitted from one focus F1 will converge at another focus F2 after being reflected by the ellipsoidal reflector. In the experiment, by precisely adjusting the height of the displacement platform and the position of the photomultiplier tube PMT, the sample surface and the photosensitive surface of the photomultiplier tube are located at the two foci of the ellipsoidal reflector. The scattered light from the sample surface is converged by the ellipsoidal mirror, filtered by the 5-micron pinhole aperture, and collected by the PMT. Through dual recognition of bright and dark fields, the accuracy of CTC recognition can be greatly improved.
[0027] The use of a pinhole aperture can significantly reduce background scattering noise, and the rotating stage rotates at a constant speed (10r / s) to avoid 1 / f noise in static mode. After bright and dark field scanning, the photodetector PDA and the photomultiplier tube PMT convert the collected light signals into electrical signals. Based on the obtained data points, the Archimedean spiral algorithm is used to restore the scanned sample morphology into an image. Then, the MeanShift algorithm is used to set the area radius according to the scanned CTC type, and the sample data of suspected CTCs can be quickly found and identified in bright field and dark field respectively. The scanning system has the characteristics of fast scanning speed and high resolution accuracy.
[0028] In the test, the biological sample is controlled to move along a specified direction at a constant speed while rotating around the central axis at a constant angular velocity. The motion trajectory is the Archimedean spiral: In the formula, The distance between two points of the Archimedean spiral; w is the angular velocity of the rotating stage; v is the moving speed of the translation stage; fs is the sampling frequency; t is the time; The signal is sampled in a custom time series so that the sampling points are evenly distributed on the sampling plane.
[0029] Since the surface of the cell sample is light-transmissive, during the scanning process, there will inevitably be reflected and scattered light from the back of the sample and the rotating stage.
[0030] In our optical system, the size of the light spot irradiated on the sample will affect the final detection accuracy. The spot diameter D can be expressed as: In the formula, represents the wavelength of light, f represents the focal length of the objective lens, d represents the diameter of the pupil of the objective lens, NA represents the numerical aperture of the lens, is the angle between the incident light beam and the sample stage.
[0031] After collecting the data, we use the MeanShift algorithm to realize the automatic identification of CTC. The MeanShift algorithm can automatically identify the number of clusters, and we can set different radius thresholds according to the different types of circulating tumor cells to be identified. Multiple types of circulating tumor cells can be identified. Suppose there are n sample points in total, i=1, 2, 3, ..., n, then the kernel density estimation function is: in, is the kernel function; is the window radius; is the dimension of the vector; is the normalization constant; Determine the MeanShift vector as: in, ; Using formula Through continuous iteration, the algorithm can converge to the density extreme point and realize the identification of circulating tumor cells. It can greatly improve the detection efficiency.
Claims
1. A dual-field scanning system for detecting circulating tumor cells, characterized in that: include: A sample table (5) for placing a sample to be tested; A detection light path, used to provide a detection light source for the sample stage (5); A bright field reflected light path, used for receiving a reflected light signal from the sample stage (5); Dark field scattered light path, used to receive scattered light signals from the sample stage; The data processing module recognizes circulating tumor cells using the MeanShift algorithm based on the signals received by the bright field reflected light path and the dark field scattered light path.
2. The dual-field scanning system for detecting circulating tumor cells according to claim 1, characterized in that: Ellipsoidal reflectors (4) are also provided on both sides of the sample stage (5), and scattered light from the sample stage (5) is received by a dark field scattered light path after passing through the ellipsoidal reflectors (4).
3. The dual-field scanning system for detecting circulating tumor cells according to claim 1 or 2, characterized in that: The dark field scattering light path comprises a filter (9), a second aperture (11) and a photomultiplier tube (PMT) (10) which are arranged in sequence, and the filter (9) is located on a side close to the sample stage (5).
4. The dual-field scanning system for detecting circulating tumor cells according to claim 1, characterized in that: The bright field reflection light path comprises a focusing lens (6), a first aperture (7) and a photodetector PDA (8) which are arranged in sequence, and the focusing lens (6) is located on a side close to the sample stage (5).
5. The dual-field scanning system for detecting circulating tumor cells according to claim 1, characterized in that: The calculation steps of the MeanShift algorithm are as follows: S11: Assume there are n sample points, i=1, 2, 3, ..., n, then the kernel density estimation function is: ； ； in, is the kernel function; is the window radius; is the dimension of the vector; is the normalization constant; S12: Determine the MeanShift vector as: ； in, ; S13: Using formula By continuously iterating, the algorithm can converge to the density extreme point and realize the identification of circulating tumor cells.
6. A dual-field scanning method for detecting circulating tumor cells, using a dual-field scanning system for detecting circulating tumor cells as described in any one of claims 1 to 5, characterized in that: The steps are as follows: S21: emitting laser light through the detection light path to illuminate the sample on the sample stage; S22: by adjusting the height of the sample stage and the bright field reflection light path, the bright field reflection light path converts the received reflection light signal into an electrical signal; S23: by adjusting the height of the sample stage and the dark field scattering light path, the dark field scattering light path converts the received scattered light signal into an electrical signal; S24: Through the data processing module, the MeanShift algorithm is used to identify circulating tumor cells according to the signals received by the bright field reflection light path and the dark field scattered light path.