In-vitro diagnostic kit of embedded fluorescence labeling detection system and method thereof

Abstract

The invention belongs to the technical field of fluorescence labeling detection, and discloses an in-vitro diagnostic kit of an embedded fluorescence labeling detection system and a method of the in-vitro diagnostic kit. The kit comprises a micro-fluidic chip carrier, an optical detection module, a fluid control module and a signal processing module. The optical detection module adopts an excitation light source array with accurately aligned wavelengths and a narrow-band spectrum filtering assembly, and optical crosstalk and background noise are physically suppressed in combination with a specific included angle light path design. The fluid control module achieves high-precision and stable control over micro-upgrading fluid through cooperation of a micro-valve array, a pulsation damper and a high-frequency pressure feedback loop. Through three innovations of hardware light path purification, precise fluid control and software algorithm decoupling, the problems of crosstalk, noise and hydraulic control precision in multi-marker fluorescence detection are fundamentally solved, and the specificity, accuracy and reliability of detection are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of fluorescent label detection technology, and in particular to an in vitro diagnostic kit of an embedded fluorescent label detection system and a method thereof. BACKGROUND

[0002] In vitro diagnostic technology is a key technical field in modern medical systems for disease screening, diagnosis, monitoring and prognosis evaluation, which provides clinical information by detecting specific markers in human samples (such as blood, urine). Fluorescent label detection technology, with its high sensitivity and specificity, has become one of the core detection methods widely used in in vitro diagnostics.

[0003] Embedded detection systems show great potential in improving detection efficiency and convenience due to their integration and automation. However, existing fluorescent label detection schemes based on embedded systems generally face technical bottlenecks such as severe optical signal crosstalk and significant background noise interference when implementing high-throughput and multi-index joint detection.

[0004] The invention patent with publication number CN116500004A discloses a small fluorescent analysis optical system and method for microfluidic chip detection. The system includes a laser, a first pinhole diaphragm, a first filter, a dichroic mirror, a light trap, a microfluidic chip, a mask, a collimating lens, a condenser lens, a second filter, a second pinhole diaphragm, and a photodetector. The invention improves the optical system layout and reduces the overall volume of the optical system while ensuring detection stability and sensitivity through miniaturized optical structure design. The dichroic mirror and light trap are set on the optical path to filter the excitation light twice, effectively reducing stray light and background noise interference during oblique optical system detection. The mask is set on the surface of the microfluidic chip to effectively suppress the autofluorescence phenomenon of the microfluidic chip under excitation light irradiation and reduce the scattering of excitation light on the surface of the microfluidic chip detection area. By scanning the microfluidic chip, quantitative analysis of low concentration samples can be achieved.

[0005] Traditional optical design is difficult to effectively separate the excitation and emission spectra of different fluorescent labels in limited space, resulting in decreased detection specificity and increased false positive risk.

[0006] The fluid control system of existing kits lacks precision in manipulating microliter-level samples, which can cause uneven reactions and poor signal stability, seriously affecting the accuracy and repeatability of quantitative detection. Therefore, there is an urgent need to develop an embedded fluorescent label detection system that can effectively overcome optical interference and achieve precise fluid control to solve the physical separation problem of multi-fluorescent label spectral overlap and fluid pulsation suppression in microfluidic environment. SUMMARY

[0007] The application aims to provide an in-vitro diagnostic kit of an embedded fluorescent marker detection system and a method thereof, and solve the technical problems of physical separation of spectral overlap of multiple fluorescent markers and fluid pulsation suppression in a microfluidic environment.

[0008] To solve the above technical problems, the application adopts the technical scheme of:

[0009] An in-vitro diagnostic kit of an embedded fluorescent marker detection system comprises:

[0010] An optical detection module for exciting fluorescent markers and collecting fluorescent signals, comprising an excitation light source array, a miniature spectral filter assembly and a photoelectric sensor; the excitation light source array adopts four groups of semiconductor lasers with different wavelengths, and the central wavelengths are strictly aligned with the maximum absorption peaks of four commonly used fluorescent markers; the miniature spectral filter assembly is integrated in front of the light path entrance of the photoelectric sensor and is composed of eight layers of alternately deposited silicon dioxide and tantalum pentoxide dielectric films, the bandpass range of which is accurately matched with the emission spectral peak region of the corresponding fluorescent markers, and the half-width is controlled within 12 nanometers; the photoelectric sensor adopts a back-illuminated scientific grade complementary metal oxide semiconductor image sensor with a pixel size of 2.8 μm, and a thermoelectric refrigeration device is built-in to stabilize the working temperature at minus 15 degrees Celsius;

[0011] A fluid control module for manipulating microliter-level fluid, comprising a microvalve array, a pulsation damper and a pressure feedback loop; the microvalve array adopts 24 pinch valves driven by thermosensitive shape memory alloy, and the travel accuracy of the pinch valves reaches 5 μm, and the response time is less than 10 milliseconds; the pulsation damper is composed of a section of elastic silica gel pipe with a length of 35 mm and an inner diameter of 0.5 mm and a gas chamber with a volume of 80 μl coupled; the pressure feedback loop monitors the pressure fluctuation in the flow path in real time, the sampling frequency is 1000 Hz, and the opening timing of the microvalve array is dynamically adjusted through proportional-integral-derivative algorithm;

[0012] A signal processing module for processing photoelectric signals and calculating concentration values, embedded with a digital signal processor, and the digital signal processor runs an adaptive spectral deconvolution algorithm; the algorithm first estimates the background noise of the original photoelectric signal, then constructs a 4x4 crosstalk matrix based on the preset fluorescent marker emission spectrum database, and finally solves the real concentration signals of each marker through non-negative least squares method.

[0013] In specific implementation, the algorithm first estimates the background noise of the original photoelectric signal, then constructs a 4x4 crosstalk matrix A based on the preset fluorescent marker emission spectrum database, wherein the element Aij represents the signal contribution degree of the jth fluorescent marker to the ith detection channel. Finally, the real concentration signals of each marker are solved through non-negative least squares method Where y is the observed signal vector, x is the true concentration vector, the constraint is x≥0, and the iteration termination condition is the norm of the residuals obtained from two consecutive iterations (i.e., The relative rate of change is less than 0.1% or the number of iterations reaches 100.

[0014] In one embodiment of the present invention, the excitation optical path and the collection optical path of the optical detection module are arranged at a 52° angle; a band-stop filter with a center wavelength of 532 nm is embedded in the excitation optical path, and the optical density value of the filter is 6; a microscope objective with a numerical aperture of 0.65 is installed at the front end of the collection optical path, and its focal plane is precisely aligned with the bottom of the reaction cell of the microfluidic chip.

[0015] In one embodiment of the present invention, the fluid control module is driven by a pneumatic micropump. The pneumatic micropump has a cavity volume of 120 microliters and is equipped with a ceramic piston driven by a stepper motor. The stepper motor has a step angle of 1.8 degrees and works with a ball screw with a lead of 1 mm to achieve fluid propulsion. The outlet of the micropump is connected to a three-way connector. One branch of the three-way connector leads to the reaction tank, and the other branch leads to a waste liquid tank with a volume of 500 microliters.

[0016] In one embodiment of the present invention, the adaptive spectral deconvolution algorithm comprises three consecutive computational stages;

[0017] The first stage involves signal preprocessing, applying a fourth-order Butterworth low-pass digital filter with a cutoff frequency of 50Hz to the raw photoelectric signal, and performing baseline correction.

[0018] In the second stage, crosstalk modeling is performed. Based on the standard emission spectra of each fluorescent label obtained in advance through standard calibration, a crosstalk matrix characterizing the degree of spectral overlap between channels is constructed.

[0019] In the third stage, signal separation is performed. A system of linear equations with crosstalk matrix as coefficients is solved using the non-negative least squares method. The objective function is to minimize the sum of squared residuals between the observed signal and the reconstructed signal. The iteration termination condition is set to the residual change rate being less than 0.1% or the number of iterations reaching 100.

[0020] In one embodiment of the present invention, the reagent kit is integrated inside a sealed microfluidic chip carrier; the microfluidic chip carrier is manufactured by injection molding using a cyclic olefin copolymer material and its surface is treated with oxygen plasma; the chip has 12 independent microfluidic channels etched inside, each network including a sample injection area, a reagent mixing area, a temperature-controlled reaction chamber and an optical detection window.

[0021] In one embodiment of the present invention, a thin-film heater and a platinum resistance temperature sensor are integrated below the temperature-controlled reaction tank to form a closed-loop temperature control system with a temperature control accuracy of ±0.2℃.

[0022] In one embodiment of the present invention, the reagent mixing zone achieves vortex mixing through an asymmetric flow channel structure; the flow channel structure includes three consecutive 90-degree bends, with an inner radius of curvature of 0.8 mm and an outer radius of curvature of 1.6 mm.

[0023] In one embodiment of the present invention, lyophilized reagent microspheres are pre-placed in the liquid storage chamber upstream of the reagent mixing zone; the microspheres have a diameter of 200 μm and their shells are formed by cross-linking sodium alginate and calcium chloride.

[0024] In addition, this invention also discloses an embedded fluorescent labeling detection method for an in vitro diagnostic kit, comprising the following steps:

[0025] Step 1: Inject 50 μL of the sample to be tested into the sample injection area of ​​the microfluidic chip through the injection port;

[0026] Step 2: Activate the fluid control module. By coordinating the control of the micro-valve array and the pneumatic micro-pump, the sample and the pre-stored lyophilized reagent are delivered to the reagent mixing zone at a flow rate of 15 μL / s for continuous vortex mixing for 10 seconds.

[0027] Step 3: Transfer the mixed reaction solution to a temperature-controlled reaction tank and incubate at 37°C for 8 minutes;

[0028] Step 4: After incubation, activate the four sets of excitation light sources with different wavelengths in the optical detection module, and at the same time, the photoelectric sensor synchronously collects the fluorescence emission signals of the four channels.

[0029] Step 5: The signal processing module performs an adaptive spectral deconvolution algorithm on the acquired raw signal to calculate the concentration value of each fluorescent marker, and outputs the final result through the universal serial bus interface.

[0030] In step 2, vortex mixing is achieved through an asymmetric flow channel structure; the asymmetric flow channel structure induces Dean vortices at low Reynolds numbers.

[0031] Furthermore, the lyophilized reagent in step 2 is released in the form of microspheres; when the sample flows through the reservoir, the microspheres completely dissolve within 30 seconds under the action of fluid shear force.

[0032] Furthermore, the adaptive spectral deconvolution algorithm in step 5 includes: estimating the background noise of the original photoelectric signal; constructing a 4x4 crosstalk matrix based on a preset fluorescent marker emission spectrum database; and iteratively solving for the true concentration signal of each marker using the non-negative least squares method.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] This invention integrates an optical detection module, employs a wavelength-precisely aligned excitation light source array and a miniature spectral filtering component composed of eight dielectric films, and combines a 52° angled optical path design with a band-stop filter to achieve efficient physical separation of the excitation and emission spectra of multiple fluorescent markers. This improves the specificity and accuracy of multi-index joint detection and fundamentally overcomes the technical bottleneck of false positives caused by signal crosstalk.

[0035] The fluid control module designed in this invention achieves sub-microliter precision control of microliter-level fluid transfer through the synergistic effect of a microvalve array driven by a thermosensitive shape memory alloy, a pulsation damper in a coupled gas chamber, and a pressure feedback loop based on high-frequency sampling. This reduces volumetric errors in sample distribution and effectively suppresses fluid pulsation, ensuring uniform mixing and stable reaction of reagents and samples within the reaction system. It provides a highly consistent physical environment for quantitative detection, greatly improving the repeatability and reliability of detection results.

[0036] The signal processing module deployed in this invention and its adaptive spectral deconvolution algorithm, by constructing a crosstalk matrix and applying non-negative least squares for mathematical separation, can accurately reconstruct the true concentration signals of each fluorescent marker from the composite photoelectric signal contaminated by crosstalk and noise. This software-level signal purification capability and the hardware-level optical purification measures effectively complement each other, further improving the detection signal-to-noise ratio of the system by two orders of magnitude, enabling it to maintain excellent detection performance in complex biological sample matrices. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0038] Figure 1 This is a simplified diagram illustrating the structure and function of the microfluidic chip carrier of the present invention.

[0039] Figure 2 This is a simplified diagram illustrating the structure and function of the optical detection module of the present invention.

[0040] Figure 3 This is a simplified diagram of the structure and function of the fluid control module of the present invention.

[0041] Figure 4 This is a simplified diagram illustrating the structure and function of the signal processing module of the present invention. Detailed Implementation

[0042] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0043] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0044] Example 1:

[0045] See Figures 1-4 This embodiment discloses an in vitro diagnostic kit for an embedded fluorescent labeling detection system. The kit uses a microfluidic chip carrier manufactured from a cyclic olefin copolymer material through injection molding. The surface of the microfluidic chip carrier is treated with oxygen plasma to significantly enhance its surface energy, thereby achieving excellent hydrophilicity and ensuring stable filling and delivery of micro-level fluids through capillary forces. The interior of the microfluidic chip is formed into 12 completely independent microchannel networks using precision etching technology. Each independent microchannel network includes a sample injection area for receiving external samples, a reagent mixing area for thorough mixing of samples and reagents, a temperature-controlled reaction chamber with precise temperature control, and an optical detection window dedicated to optical signal reading. These 12 microchannel networks are physically isolated from each other, enabling the kit to process up to 12 independent test samples simultaneously or sequentially, or to perform multiple replicate experiments, greatly improving detection throughput and efficiency.

[0046] The core function of the optical detection module is to excite the fluorescent markers pre-placed in the detection system and collect the emitted fluorescence signals, while suppressing background noise and inter-channel crosstalk to the greatest extent possible.

[0047] The excitation source array consists of four groups of semiconductor lasers with different wavelengths. The center wavelengths of these four lasers have been rigorously selected and calibrated to precisely align with the maximum absorption peaks of four fluorescent labels most commonly used in in vitro diagnostics.

[0048] In practice, these four wavelengths can be 488 nm, 532 nm, 635 nm and 785 nm, respectively, covering the spectral range from blue to near-infrared.

[0049] Each laser group is equipped with an independent drive circuit and temperature control unit to ensure that its output optical power stability is better than 1% and the center wavelength drift is less than 0.5 nanometers. The excitation optical path is designed with a specific geometric configuration.

[0050] After the excitation beam is emitted from the laser, it first passes through a collimating lens group, compressing the divergence angle to within 1 milliradian, forming a nearly parallel beam. Subsequently, the beam passes through a band-stop filter embedded in the optical path. The band-stop filter has a center wavelength of 532 nm and an optical density value as high as 6, capable of attenuating the intensity of light near 532 nm to one millionth of its original value. The main function of the band-stop filter is to strongly attenuate the Rayleigh scattering light generated by the excitation beam in the sample, preventing it from entering the signal collection channel, thereby significantly reducing optical background noise.

[0051] The filtered excitation beam is precisely focused at an incident angle of 52° onto the reaction liquid column within the optical detection window of the microfluidic chip. This 52° angle effectively separates the excitation beam from the fluorescence collection beam, minimizing the possibility of direct reflection or scattering of the excitation light into the collection optics. On the fluorescence collection side, a 0.65-numerical aperture microscope objective is mounted at the very front of the collection beam.

[0052] The focal plane of the microscope objective is precisely adjusted to be aligned with the bottom plane of the microfluidic chip's reaction cell, ensuring efficient collection of fluorescence signals emitted from the reaction region. The collected fluorescence signals then enter a miniature spectral filter assembly. This spectral filter assembly, directly integrated in front of the optical path entrance of the photoelectric sensor, consists of eight alternating layers of silicon dioxide and tantalum pentoxide dielectric films.

[0053] A high-performance bandpass filter was constructed using a multilayer dielectric film structure. The bandpass range was designed to precisely match the emission spectrum peak region of the corresponding fluorescent label. In specific implementation, for fluorophores with an emission peak at 520 nm, the center wavelength of the bandpass filter was also set at 520 nm, and its full width at half maximum (FWHM) was strictly controlled within a narrow range of no more than 12 nm. This narrow-band filtering characteristic ensures that each channel allows only the characteristic emission light of the target fluorescent label to pass through, while effectively blocking stray light from other wavelength ranges, including the emission light of other fluorescent labels and various background fluorescence.

[0054] The photoelectric sensor is the final receiver of optical signals and employs a complementary metal-oxide-semiconductor (CMOS) image sensor. With a pixel size of 2.8 μm, it provides high spatial resolution, enabling localized analysis of fluorescence signals within microchannels. More importantly, the sensor incorporates a thermoelectric cooling system, allowing it to stably maintain its operating temperature at -15°C. This low-temperature operation significantly reduces the sensor's dark current noise and improves the signal-to-noise ratio, a crucial measure for detecting weak fluorescence signals.

[0055] The sensor outputs the signals it reads in digital form with a dynamic range of 16 bits, ensuring that it can simultaneously capture high-intensity and extremely weak light signals without saturation or loss.

[0056] The fluid control module is responsible for manipulating all micro-level fluids, and is comprised of multiple functional subunits working collaboratively. The microvalve array acts as the on / off controller for the fluid path, consisting of 24 independently controlled thermosensitive shape memory alloy-driven pinch valves. Each pinch valve is driven by a shape memory alloy actuator that powers a miniature valve core, achieving a stroke accuracy of 5μm. This allows for precise and repeatable flow regulation from fully closed to fully open. When a specific flow path needs to be opened or closed, the control circuit applies a precisely controlled current pulse to the corresponding shape memory alloy, causing a phase change and displacement, thereby actuating the valve core. The timing of all 24 valve actions is coordinated by the main controller to achieve complex fluid delivery sequences.

[0057] The fluid is driven by a pneumatic micropump. This micropump has a 120-microliter chamber containing a ceramic piston precisely driven by a stepper motor. The stepper motor has a step angle of 1.8 degrees, which converts the rotational motion into linear motion of the piston via a high-precision ball screw with a lead of 1 mm. This allows for sub-μm displacement resolution of the piston, enabling nanoliter-level fluid delivery.

[0058] The outlet of the micropump is connected to a tee connector. One branch of the tee connector leads to the reaction tank in the microfluidic network for delivering the reaction liquid; the other branch connects to a 500-microliter waste tank for collecting waste liquid during system cleaning or emptying.

[0059] To suppress fluid pulsation caused by piston movement, a pulsation damper is integrated. The damper consists of a 35 mm long, 0.5 mm inner diameter elastic silicone tube coupled to a sealed 80 μL chamber. When the fluid pressure increases, the silicone tube elastically expands, compressing the gas inside the chamber and absorbing excess energy; when the pressure decreases, the silicone tube rebounds, the chamber expands, and energy is released to maintain pressure stability. This reduces fluid pressure fluctuations by more than 80%.

[0060] The pressure feedback loop includes a high-frequency response pressure sensor with a sampling frequency of 1000Hz, capable of capturing instantaneous pressure changes in the flow path in real time. The acquired pressure data is fed into a proportional-integral-derivative (PID) controller. The required correction is dynamically calculated, and the opening timing of the microvalve array and the stepper motor speed of the pneumatic micropump are adjusted in real time, forming a closed-loop control system. In practical implementation, when an abnormal pressure increase is detected, potentially indicating flow path blockage, the controller will temporarily slow the pump speed or briefly reverse its movement to clear the blockage, thereby ensuring smooth process operation.

[0061] The signal processing module is responsible for performing complex mathematical processing on the raw photoelectric signals acquired by the photoelectric sensor to extract the true concentration information of each fluorescent label. The core hardware of the signal processing module is a high-performance digital signal processor (DSP). The DSP runs a core adaptive spectral deconvolution algorithm. The algorithm aims to solve the inherent spectral crosstalk problem in multi-fluorescent label detection, where the signal detected in one channel may contain contributions from other fluorescent labels.

[0062] The algorithm's execution flow consists of three consecutive and logically rigorous computational stages.

[0063] The first stage is signal preprocessing. In this stage, the system processes the raw digital signal acquired from the photoelectric sensor. First, a fourth-order Butterworth low-pass digital filter with a cutoff frequency of 50Hz is applied. The purpose of the filter is to remove high-frequency electronic noise and interference that may originate from ambient light sources. The filtered signal then undergoes baseline correction. The purpose of baseline correction is to eliminate DC offset or slow drift in the signal, which are typically caused by the sensor's dark current or background fluorescence. The correction method involves calculating the mean of the signal in a stable region before the reaction begins and subtracting this mean from the entire signal sequence.

[0064] The second stage is crosstalk modeling. This stage is based on a pre-built database of fluorescent label emission spectra. The database is a standard emission spectrum obtained by measuring pure standard solutions of each fluorescent label under the same detection conditions. The calibration method includes: injecting each fluorescent label into the system individually, measuring its emission spectrum under default parameters, and storing the normalized spectrum as a standard spectrum in the database.

[0065] Based on these standard spectra, a 4x4 crosstalk matrix is ​​constructed. Each element of the crosstalk matrix represents the signal contribution of the j-th fluorescent label to the i-th detection channel. In practice, the elements on the diagonal of the matrix represent the contribution of the target label to its own channel, which is usually close to 1; while the off-diagonal elements represent the crosstalk coefficients of other labels to that channel.

[0066] The third stage is signal separation, which is the core of the algorithm. In this stage, the system combines the preprocessed observation signal with the constructed crosstalk matrix and uses mathematical methods to solve for the true concentration of each marker. The observed 4-channel signal vector is represented as the product of the crosstalk matrix and the true concentration vector of each marker. To solve for the true concentration vector from the observation signal, a non-negative least squares method is used for iterative solution. The non-negativity constraint is based on the physical fact that fluorescence concentration cannot be negative. The objective function of the algorithm is to minimize the sum of squared residuals between the observed signal and the signal reconstructed from the crosstalk matrix and the estimated concentration. The iterative process continues until a preset termination condition is met. Two termination conditions are set: one is that the rate of change of the residual is less than 0.1%, indicating that the solution has converged; the other is that the number of iterations reaches 100 to prevent infinite loops. Finally, the algorithm outputs a set of 4 values, representing the true relative or absolute concentration values ​​of the 4 fluorescent markers after crosstalk correction and noise suppression.

[0067] The reagent kit's workflow is a highly automated sequence. The user first injects 50 μL of the sample to be tested into the designated sample injection area through the inlet on the chip. Subsequently, the system's main controller activates the fluid control module. Following a preset program, the module coordinates the microvalve array and pneumatic micropump to deliver the sample and pre-stored lyophilized reagents released from the reservoir at a constant flow rate of 15 μL / s to the reagent mixing zone. The design of the reagent mixing zone is crucial for ensuring reaction homogeneity. An asymmetric flow channel structure is employed internally to achieve efficient vortex mixing. This asymmetric flow channel structure contains three consecutive 90-degree bends, with an inner radius of curvature of 0.8 mm and an outer radius of curvature of 1.6 mm. When fluid flows through this asymmetric bend at a low Reynolds number, it induces Dean vortices in the direction perpendicular to the main flow. This secondary flow effect effectively exchanges fluid between the center and edges of the channel, achieving thorough mixing of the sample and reagents in a very short distance and time, with mixing efficiency far exceeding that of straight channels that rely solely on diffusion.

[0068] The lyophilized reagent, pre-stored in the upstream reservoir of the reagent mixing zone, is in the form of microspheres with a diameter of approximately 200 μm. The outer shell of the microspheres is formed by a cross-linking reaction of sodium alginate and calcium chloride. When the sample flows through the reservoir, the shear force generated by the fluid acts on the surface of the microspheres, causing the outer shell to gradually dissolve and release the active reagent components inside. The entire dissolution process is completed within 30 seconds, ensuring that the reagent can participate in the subsequent reaction rapidly and quantitatively.

[0069] The homogenized reaction solution was then transferred to a temperature-controlled reaction chamber for incubation. A thin-film heater and a high-precision platinum resistance temperature sensor were integrated beneath the temperature-controlled reaction chamber, forming a closed-loop temperature control system. The platinum resistance temperature sensor monitored the temperature of the reaction chamber in real time and fed the data back to the temperature controller. The controller adjusted the power output of the thin-film heater using pulse width modulation to precisely stabilize the temperature of the reaction chamber at 37°C, with a temperature control accuracy of ±0.2°C. Under these constant temperature conditions, the reaction solution was incubated for 8 minutes to ensure that biochemical reactions, such as antigen-antibody binding or enzymatic reactions, could proceed fully.

[0070] After incubation, the reaction solution is pushed to the optical detection window. The optical detection module then initiates the detection sequence. The system sequentially activates four sets of excitation light sources with different wavelengths, while the photoelectric sensor simultaneously acquires the fluorescence emission signals of the corresponding four channels. This time-division excitation mode can further reduce potential interference between different light sources. The acquired raw signals from the four channels are immediately sent to the signal processing module. The digital signal processor executes the aforementioned adaptive spectral deconvolution algorithm to estimate the background noise of the raw signals, construct the crosstalk matrix, and iteratively solve it using the non-negative least squares method to finally calculate the accurate concentration values ​​of each fluorescent label. The final calculation results, including the concentration values ​​of each indicator and related quality control information, are output to an external computer or display device through a universal serial bus interface for user reading and recording. The entire process, from sample addition to result output, is fully automated, minimizing human error and ensuring high repeatability and reliability of the detection results.

[0071] Example 2:

[0072] This embodiment is basically the same as Embodiment 1, except that an alternative implementation scheme for an in vitro diagnostic kit of an embedded fluorescent labeling detection system is provided in this embodiment. The excitation light source and signal processing strategy in the optical detection module are optimized to adapt to the fluorescent labeling system that is more sensitive to photobleaching effect and to further improve the detection sensitivity under weak signals.

[0073] In this embodiment, the excitation light source array of the optical detection module adopts a different technical approach than that in Embodiment 1.

[0074] Instead of a continuously output semiconductor laser, the excitation source is now a four-group pulse-modulated light-emitting diode (LED) source. The center wavelengths of these four LEDs are precisely matched to the maximum absorption peaks of the four target fluorescent markers. Compared to lasers, while LEDs are slightly inferior in monochromaticity and collimation, they offer advantages such as lower cost, longer lifespan, and easier implementation of high-frequency pulse modulation. Each LED group is controlled by an independent constant-current drive circuit capable of generating a rectangular pulse current with a frequency of 1 kHz and a duty cycle of 10%.

[0075] This allows the LED to remain off most of the time, emitting light only for very short periods at high instantaneous power. This operating method offers two core advantages:

[0076] First, it significantly reduces the total light energy received by the fluorescent label during detection, thereby greatly alleviating the photobleaching effect caused by continuous illumination. This is crucial for maintaining the intensity of the fluorescence signal, especially for applications requiring long-term observation or multiple reads.

[0077] Second, pulse excitation provides a time reference for synchronous phase-locked amplification technology in signal detection.

[0078] Accordingly, the operating mode of the photoelectric sensor is also adjusted to be synchronized with pulse excitation. Within each pulse cycle, a synchronization trigger signal is generated by the digital signal processor. When the LED turns on, the image sensor synchronously opens a very short acquisition window based on this trigger signal, acquiring the total signal containing both fluorescence and background noise. Immediately afterwards, after the LED turns off, the sensor opens another acquisition window of the same duration, acquiring the signal containing only background noise. By subtracting the background noise signal from the total signal in real time, the pure fluorescence signal can be obtained. This time-gated detection technique effectively suppresses continuous background noise that is out of sync with the excitation light, and in practical implementation, it also suppresses ambient stray light and the sensor's dark current, thereby directly improving the signal-to-noise ratio at the hardware level.

[0079] In terms of signal processing, the adaptive spectral deconvolution algorithm executed by the signal processing module adds a pre-signal enhancement step to the algorithm in Example 1. During the signal preprocessing stage, in addition to conventional filtering and baseline correction, the pure fluorescence signal extracted using time-gated technology is accumulated and averaged. Since the excitation is periodic, the system can align and accumulate corresponding signals acquired within multiple consecutive pulse cycles. According to signal processing theory, the amplitude of a periodic signal increases linearly during accumulation, while the amplitude of random noise increases only by its root mean square value. Assuming N accumulations, the signal-to-noise ratio (SNR) can be improved by approximately the square root of N. In practice, performing 100 accumulations theoretically improves the SNR by 10 times. This signal accumulation strategy is particularly suitable for detecting extremely weak fluorescence signals, and is particularly useful in applications such as single-molecule detection or detection of extremely low-abundance targets.

[0080] Furthermore, a real-time optical power monitoring loop was introduced to address potential instantaneous power fluctuations caused by pulsed light sources. A small photodiode is placed in the optical path bypass of each light-emitting diode (LED) to monitor the actual output light pulse intensity. The monitored light intensity data is fed back to the signal processing module. In subsequent signal processing, the fluorescence signals acquired by each channel are normalized and corrected according to the actual pulse intensity of their corresponding excitation source, thereby eliminating detection errors introduced by light source fluctuations and further ensuring the accuracy of the quantitative results.

[0081] Other components of this embodiment, including the microfluidic chip structure, fluid control module, and reagent mixing and temperature-controlled incubation processes, are consistent with those in Embodiment 1.

[0082] This embodiment employs pulse-modulated LED light source, time-gated detection, and signal accumulation technology. While maintaining high specificity and low crosstalk, it particularly enhances the system's ability to resist photobleaching and detect extremely weak signals, thus broadening the application range of the reagent kit.

[0083] At the optical level, this invention does not simply select lasers and filters, but strictly follows the absorption / emission spectral characteristics of fluorescent markers. Through the physical matching of the wavelength-precisely aligned excitation light source array and the narrow-band dielectric film filter component, and combined with the geometric and spectral design of the 52° angled optical path and the high optical density bandstop filter, the physical separation of signal and noise is achieved from the light source origin and the optical path propagation path, breaking through the physical limitation of optical crosstalk in the limited space of the embedded system.

[0084] At the fluid control level, this invention directly addresses the physical reality that microfluids are dominated by surface tension and viscosity, constructing a closed loop of perception, decision-making, and execution consisting of a high-frequency (≥1000Hz) pressure feedback loop, a shape memory alloy microvalve array, and an elastic tube-gas chamber coupling damper. A dynamic response model for momentum and pressure changes within the microfluidic system is established, enabling active control of fluid motion at nanoliter precision and ensuring the physical consistency of the reaction system at the microscale.

[0085] At the signal processing level, this invention models the observed signal as a linear transformation between the real concentration signal and the crosstalk matrix, and solves it using the non-negative least squares method, a convex optimization tool. Mathematically, this ensures the optimal estimation of the real signal under physical constraints (non-negative concentration), thus achieving the removal of crosstalk from the information level.

[0086] This invention forms a hardware-software collaborative underlying solution through a triple design of hardware optical path purification, precision fluid control and software algorithm decoupling, achieving a leapfrog improvement in detection specificity, accuracy and reliability on an embedded platform.

[0087] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0088] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An in vitro diagnostic reagent kit for an embedded fluorescent labeling detection system, characterized in that, include: A microfluidic chip carrier, the microfluidic chip carrier having at least one microchannel network etched inside, the microchannel network including a sample injection area, a reagent mixing area, a reaction cell and an optical detection window; The optical detection module includes a multi-wavelength excitation source array, a spectral filtering component, and a photoelectric sensor. The multi-wavelength excitation source array is used to excite different fluorescent markers, the spectral filtering component is used to selectively transmit the target fluorescence emission signal, and the photoelectric sensor is used to collect the fluorescence emission signal. The fluid control module includes a microvalve array, a pulsation damper, and a pressure feedback control unit. The fluid control module is used to precisely control the transmission of micro-level fluid within a microchannel network. The signal processing module, which has an embedded processor and runs a spectral deconvolution algorithm, is used to process the signals collected by the photoelectric sensor to calculate the concentration of each fluorescent label.

2. The in vitro diagnostic reagent kit for an embedded fluorescent labeling detection system according to claim 1, characterized in that: The multi-wavelength excitation source array includes multiple semiconductor lasers, and the center wavelengths of the multiple semiconductor lasers correspond to the absorption peaks of different fluorescent labels.

3. The in vitro diagnostic reagent kit for an embedded fluorescent labeling detection system according to claim 1, characterized in that: The spectral filtering component is a bandpass filter, and the bandpass range of the bandpass filter matches the emission spectrum peak region of the target fluorescent marker.

4. The in vitro diagnostic reagent kit for an embedded fluorescent labeling detection system according to claim 1, characterized in that: The pressure feedback control unit monitors the flow path pressure at a sampling frequency of no less than 1000Hz, and dynamically adjusts the operation of the microvalve array based on the proportional-integral-derivative algorithm.

5. The in vitro diagnostic reagent kit for an embedded fluorescent labeling detection system according to claim 1, characterized in that: A pulsation damper consists of an elastic tube coupled with an air chamber and is used to attenuate fluid pulsation.

6. The in vitro diagnostic reagent kit for an embedded fluorescent labeling detection system according to claim 1, characterized in that: The spectral deconvolution algorithm constructs a crosstalk matrix based on the standard emission spectra of each fluorescent marker. The algorithm separates the overlapping signals in each channel by solving a system of linear equations with the crosstalk matrix as coefficients.

7. The in vitro diagnostic reagent kit for an embedded fluorescent labeling detection system according to claim 1, characterized in that: The reaction tank is equipped with a closed-loop temperature control system.

8. The in vitro diagnostic reagent kit for an embedded fluorescent labeling detection system according to claim 1, characterized in that: The reagent mixing zone contains an asymmetric flow channel structure, which is used to induce eddies as the fluid passes through to enhance mixing.

9. An embedded fluorescent labeling detection method for an in vitro diagnostic reagent kit, characterized in that: In vitro diagnostic kits that utilize the embedded fluorescent labeling detection system according to any one of claims 1-8; Specifically, the following steps are included: Inject the sample to be tested into the sample injection area; The sample and reagents are transported to the reagent mixing area for mixing via a fluid control module, and then the mixed reaction solution is transported to the reaction tank for incubation. After incubation, the reaction solution in the reaction cell is excited by multiple wavelengths and fluorescence signals are collected simultaneously through an optical detection module; The signal processing module runs a spectral deconvolution algorithm to calculate the concentration of each fluorescent marker and output the results.

Images