A breast biopsy system

Abstract

The present application relates to the field of breast biopsy, and particularly relates to a breast biopsy rotary cutting system, comprising: a tissue stiffness calculation module, which calculates the tissue stiffness contacted by the needle based on the needle acting force and the needle moving speed; a needle stiffness prediction module, which generates the needle dynamic stiffness by a needle motion model based on a multi-layer perception architecture, by the needle acting force and the tissue stiffness; and a needle stiffness adjustment module, which determines the rotary cutting acting force of the transmission assembly of the needle according to the needle dynamic stiffness. The present application adapts the rotary cutting acting force of the needle to individual differences, effectively avoids the subcutaneous blood vessel injury caused by the excessive penetration of the needle, and ensures the success rate of the needle rotary cutting target tissue.

Description

Technical Field

[0001] This invention relates to the field of breast biopsy, and more particularly to a breast biopsy excision system. Background Technology

[0002] Breast cancer is one of the most common malignant tumors among women worldwide, and early and accurate diagnosis is key to improving treatment outcomes. Breast biopsy excision systems, as a commonly used minimally invasive sampling device in clinical practice, use a 2-3 mm diameter, 150 mm long needle to puncture through fatty tissue until reaching the target glandular tissue and potential breast cancer tissue area. The needle then excises the breast cancer tissue, which is then drawn into the biopsy needle under negative pressure to obtain a breast cancer tissue sample for pathological biopsy analysis.

[0003] However, during the process of doctors manually pushing the biopsy needle tip into the target tissue, the cutting force of existing biopsy needles cannot be controlled according to the significant differences between adipose tissue (stiffness of 0.1-1 kPa), glandular tissue (stiffness of 2-10 kPa), and breast cancer tissue (stiffness of 10-50 kPa). On the one hand, adipose tissue has the lowest stiffness, and if the cutting force is too large, it is easy to cause the needle to penetrate excessively, damaging subcutaneous blood vessels, mammary ducts, and other normal structures. On the other hand, the target tissue has the highest stiffness, and if the cutting force is insufficient, it will lead to insufficient cutting by the cutting head, resulting in problems such as sample breakage and insufficient sample volume.

[0004] Although existing breast biopsy needles have attempted to incorporate force control technology, with some devices integrating force sensors into the needle's transmission components to collect force data when the needle contacts the tissue, they only treat this force data as a static parameter, completely ignoring the dynamic influence of the needle's movement speed. This overlooks the issue of the tissue's instantaneous, artificially high stiffness due to inertia during high-speed puncture, and the fact that tissue deformation is more complete and stiffness measurements are closer to the true value during low-speed puncture.

[0005] Therefore, how to control the needle stiffness differently for different tissues based on the needle movement speed, so as to achieve appropriate force puncture protection for adipose tissue and strong puncture sampling of target tissue, is a technical problem that needs to be solved. Summary of the Invention

[0006] To address this, the present invention provides a breast biopsy rotary cutting system that combines the dual parameters of needle force and needle movement speed to calculate tissue stiffness and uses a multilayer perceptron model to control the needle stiffness of the biopsy needle. This avoids misjudgment of tissue stiffness and adapts the cutting force of the needle to individual differences, effectively avoiding subcutaneous vascular damage caused by excessive needle penetration and ensuring the success rate of needle rotary cutting of the target tissue.

[0007] To achieve the above objectives, the present invention proposes a breast biopsy excision system, comprising:

[0008] The tissue stiffness calculation module is used to obtain the needle force collected by the force sensor set in the transmission assembly of the needle of the breast biopsy needle, and to calculate the tissue stiffness of the needle contact based on the needle force and the needle moving speed.

[0009] The needle stiffness prediction module is used to generate the dynamic stiffness of the needle by using the needle motion model based on a multilayer perceptron architecture to measure the tissue stiffness of the needle force.

[0010] The needle stiffness adjustment module is used to determine the shearing force of the transmission component based on the dynamic stiffness of the needle, so that the shearing force of the needle piercing the adipose tissue is less than the shearing force piercing the target tissue.

[0011] Furthermore, the needle motion model includes an input layer, a multi-hop fully connected layer, and an output layer, and the needle stiffness prediction module includes:

[0012] An input data integration unit is used to determine the force error based on the needle force and the expected force, and at least constructs a comprehensive input vector by combining the time rate of change and historical values ​​of the force error and the tissue stiffness through the input layer.

[0013] The feature mapping unit is used to perform feature mapping on the comprehensive input vector through a multi-hop fully connected layer to generate tissue stiffness features;

[0014] The output mapping unit is used to output and map the tissue stiffness characteristics through the output layer to generate the needle dynamic stiffness.

[0015] Furthermore, the multi-hop fully connected layer includes a first fully connected layer, a second fully connected layer, a third fully connected layer, and a fourth fully connected layer, and the feature mapping unit includes:

[0016] The feature extraction subunit is used to extract basic features from the comprehensive input vector through the first fully connected layer to generate primary features;

[0017] The feature integration subunit is used to extract the primary features through the second fully connected layer to generate intermediate features, and to fuse the intermediate features with the primary features through a first skip connection to generate identity mapping features.

[0018] The feature refinement subunit is used to refine the identity mapping features through the third fully connected layer to generate high-level features;

[0019] The feature mapping stabilization subunit is used to perform feature mapping on the high-level features through the fourth fully connected layer to generate decision-oriented features, and to perform feature fusion stabilization by connecting the decision-oriented features with the high-level features through a second skip connection to generate the organization stiffness features.

[0020] The output layer includes an output fully connected layer and an activation function layer, and the output mapping unit includes:

[0021] The output mapping subunit is used to pass the tissue stiffness features through the output fully connected layer without activation function to generate the original predicted stiffness value, the original predicted stiffness change value, and the stiffness convergence rate.

[0022] The target steady-state stiffness calculation subunit is used to generate the target steady-state stiffness by passing the original predicted stiffness value through the Sigmoid activation function.

[0023] The predicted stiffness change calculation subunit is used to generate a predicted stiffness change value by passing the original predicted stiffness change value through the Tanh activation function.

[0024] The dynamic stiffness calculation subunit is used to generate the dynamic stiffness of the needle tip by means of a stiffness function based on the target steady-state stiffness, the predicted stiffness change value, and the stiffness convergence rate.

[0025] Furthermore, the transmission assembly of the needle includes a transmission screw connected to the needle drive, and the needle stiffness adjustment module includes:

[0026] The damping calculation unit is used to generate predicted damping by applying the critical damping formula to the dynamic stiffness of the needle tip, and to calculate the screw damping term based on the product of the predicted damping and the error in the needle tip's movement speed.

[0027] The stiffness calculation unit is used to calculate the screw stiffness term based on the product of the needle dynamic stiffness and the needle position error;

[0028] A screw shearing force control unit is used to generate the shearing force based on the difference between the screw damping term and the screw stiffness term, and to cause the drive screw to perform the shearing force.

[0029] Furthermore, the needle force includes the needle force at the current moment and the needle force in the previous sampling period;

[0030] The tissue stiffness calculation module is used to take the ratio of the difference between the needle force at the current moment and the needle force in the previous sampling period to the product of the needle movement speed and the sampling period interval as the tissue stiffness.

[0031] Furthermore, the breast biopsy excision system also includes:

[0032] The prediction model training module is used to construct a data fitting term based on the mean square error between the dynamic stiffness of the needle and the true stiffness of the sample, construct an overshoot penalty term based on the difference between the dynamic stiffness of the needle and the expected steady-state stiffness, construct a total loss function based on the data fitting term and the overshoot penalty term, and perform backpropagation training on the needle motion model through the total loss function.

[0033] In particular, a needle motion model based on multilayer perceptron (MLP) is adopted, which uses a multi-hop fully connected layer structure to achieve high-precision, nonlinear prediction of tissue stiffness. This significantly improves the model's feature extraction capability and gradient transfer efficiency, enabling the system to respond quickly to changes in tissue characteristics, adapt to the differences in different patients and tissue types, and ensure the stability of the puncture process.

[0034] Furthermore, the breast biopsy needle includes:

[0035] The vacuum suction chamber, connected to the suction tube, is used to suck in and contain the target tissue that has been cut by the needle.

[0036] A transmission component is disposed inside the vacuum suction cavity and fixed to the needle;

[0037] A transmission assembly, fixed to the transmission component, is used to convert the applied pressing force into a torque that sequentially drives the transmission component and the needle to rotate.

[0038] Furthermore, the transmission assembly includes:

[0039] The pressing screw is used to convert the pressing force into torque;

[0040] The drive gear is fixedly sleeved on the pressing screw and rotates synchronously with the pressing screw;

[0041] The driven gear meshes with the driving gear and rotates synchronously with the driving gear;

[0042] The transmission screw is fixed to the driven gear and the transmission component, and is used to adjust the torque of the driven gear to the shearing force before transmitting it to the transmission component.

[0043] Furthermore, the force sensor is sleeved on the transmission screw.

[0044] In particular, the integrated design of the vacuum suction chamber and the rotary cutting needle allows for the simultaneous rotary cutting of the target tissue by the needle, while the excised tissue is rapidly drawn into and contained within the suction chamber using a negative pressure suction mechanism. The needle's transmission component employs a mechanical structure combining gear meshing and a screw, efficiently and smoothly converting external pressure into precise control torque to drive the needle's rotation, ensuring reliable power transmission and precise force control during rotary cutting.

[0045] Compared with the prior art, the beneficial effects of the present invention are that it combines the two parameters of needle force and needle movement speed to calculate tissue stiffness, and controls the needle stiffness of biopsy needles based on a multilayer perceptron model, which avoids misjudgment of tissue stiffness and adapts the needle's rotary cutting force to individual differences, effectively avoiding subcutaneous vascular damage caused by excessive needle penetration, and ensuring the success rate of needle rotary cutting of target tissue.

[0046] In particular, this invention employs a needle motion model based on multilayer perceptron (MLP), which, with the help of a multi-hop fully connected layer structure, achieves high-precision, nonlinear prediction of tissue stiffness. This significantly improves the model's feature extraction capability and gradient transfer efficiency, enabling the system to quickly respond to changes in tissue characteristics, adapt to the differences in different patients and tissue types, and ensure the stability of the puncture process.

[0047] In particular, this invention employs an integrated design of a vacuum suction chamber and a rotary cutting needle, enabling the simultaneous rotary cutting of the target tissue by the needle while simultaneously using a negative pressure suction mechanism to rapidly draw in and contain the excised tissue within the suction chamber. The needle's transmission assembly utilizes a mechanical structure combining gear meshing and a screw, efficiently and smoothly converting external pressure into precise control torque to drive the needle's rotation, ensuring reliable power transmission and precise force control during rotary cutting. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the breast biopsy excision system according to an embodiment of the present invention;

[0049] Figure 2 This is a schematic diagram of the breast biopsy excision system according to an embodiment of the present invention;

[0050] Figure 3 This is a schematic cross-sectional view of the breast biopsy needle in the breast biopsy excision system according to an embodiment of the present invention.

[0051] Figure 4 The breast biopsy excision system of this invention is an embodiment of the present invention. Figure 3 A partial structural diagram of part A in the middle;

[0052] Figure 5 This is a schematic diagram of the overall structure of the breast biopsy needle in the breast biopsy excision system according to an embodiment of the present invention.

[0053] The main components in the diagram are as follows: 1. Vacuum suction chamber; 2. Transmission component; 3. Transmission assembly; 31. Press cap; 32. Press screw; 33. Drive gear; 34. Driven gear; 35. Transmission screw; 4. Suction pipe; 5. Needle; 6. Force sensor. Detailed Implementation

[0054] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0055] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0056] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0057] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0058] like Figures 1 to 5 As shown, this invention provides a breast biopsy rotary cutting system that combines the dual parameters of needle force and needle movement speed to calculate tissue stiffness, and uses a multilayer perceptron model to control the needle stiffness of the biopsy needle. This avoids misjudgment of tissue stiffness and adapts the cutting force of the needle to individual differences, effectively avoiding subcutaneous vascular damage caused by excessive needle penetration, and ensuring the success rate of needle rotary cutting of the target tissue.

[0059] like Figure 1 and 2 As shown, this embodiment proposes a breast biopsy excision system, comprising:

[0060] The tissue stiffness calculation module is used to obtain the needle force collected by the force sensor 6 set in the transmission assembly 3 of the needle 5 of the breast biopsy needle, and to calculate the tissue stiffness of the needle contact based on the needle force and the needle moving speed.

[0061] The needle stiffness prediction module is used to generate the dynamic stiffness of the needle by using the needle motion model based on a multilayer perceptron architecture to measure the tissue stiffness of the needle force.

[0062] The needle stiffness adjustment module is used to determine the rotary cutting force of the transmission component 3 based on the dynamic stiffness of the needle, so that the rotary cutting force of the needle 5 piercing the adipose tissue is less than the rotary cutting force piercing the target tissue.

[0063] In particular, by employing an algorithm model based on multilayer perceptron (MLP) and integrating current and historical force and stiffness data, nonlinear calculations and predictions are performed to accurately estimate the shearing force of the needle piercing adipose tissue and the shearing force of the target tissue (glandular tissue and breast cancer tissue), thus avoiding the lag and oscillation of traditional control methods.

[0064] Furthermore, the needle motion model includes an input layer, a multi-hop fully connected layer, and an output layer, and the needle stiffness prediction module includes:

[0065] An input data integration unit is used to determine the force error based on the needle force and the expected force, and at least constructs a comprehensive input vector by combining the time rate of change and historical values ​​of the force error and the tissue stiffness through the input layer.

[0066] The feature mapping unit is used to perform feature mapping on the comprehensive input vector through a multi-hop fully connected layer to generate tissue stiffness features;

[0067] The output mapping unit is used to output and map the tissue stiffness characteristics through the output layer to generate the needle dynamic stiffness.

[0068] In particular, by integrating the input vectors, the model not only focuses on the current force error, but also comprehensively analyzes the error's changing trend, historical values, and real-time organizational stiffness.

[0069] Specifically, the synthesized input vector generated by the input layer can be represented as:

[0070] X(t)=[e(t),derivative_e(t),normalized_force(t),needle_velocity(t),e(t-1),e(t-2),K_tissue(t),normalized_position(t)]

[0071] In the formula, e(t) represents the force error determined by the difference between the current needle force and the expected force; derivative_e(t) represents the rate of change of the first derivative of the force error at the current time; normalized_force(t) represents the normalized needle force at the current time; needle_velocity(t) represents the needle velocity; e(t-1) and e(t-2) represent the force errors of the previous sampling period and the period before that, respectively, i.e., the historical values ​​of the force error; K_tissue(t) represents the tissue stiffness at the current time; and normalized_position(t) represents the normalized needle position. Specifically, each element of the above-mentioned integrated input vector is standardized.

[0072] Furthermore, the multi-hop fully connected layer includes a first fully connected layer, a second fully connected layer, a third fully connected layer, and a fourth fully connected layer, and the feature mapping unit includes:

[0073] The feature extraction subunit is used to extract basic features from the comprehensive input vector through the first fully connected layer to generate primary features;

[0074] The feature integration subunit is used to extract the primary features through the second fully connected layer to generate intermediate features, and to fuse the intermediate features with the primary features through a first skip connection to generate identity mapping features.

[0075] The feature refinement subunit is used to refine the identity mapping features through the third fully connected layer to generate high-level features;

[0076] The feature mapping stabilization subunit is used to perform feature mapping on the high-level features through the fourth fully connected layer to generate decision-oriented features, and to perform feature fusion stabilization by connecting the decision-oriented features with the high-level features through a second skip connection to generate the organization stiffness features.

[0077] In particular, by extracting features in layers to adapt to their complexity, using skip connections to preserve full-scale information, and employing gradient stabilization mechanisms to ensure training efficiency, the generated tissue stiffness features can accurately distinguish between adipose tissue (low stiffness) and diseased tissue (high stiffness), while also adapting to individual differences in clinical data.

[0078] Specifically, the process by which a multi-hop fully connected layer generates tissue stiffness features can be represented as:

[0079] H1 = ReLU(BN(W1*X+b1))

[0080] H2 = ReLU(BN(W2*H1+b2)) + H1

[0081] H3 = ReLU(BN(W3*H2+b3))

[0082] H4 = ReLU(BN(W4*H3+b4)) + H3

[0083] In the formula, H1, H2, H3, and H4 represent primary features, identity mapping features, advanced features, and organizational stiffness features, respectively; W1 and b1 represent the weight matrix and bias vector of the first fully connected layer, respectively; W2 and b2 represent the weight matrix and bias vector of the second fully connected layer, respectively; W3 and b3 represent the weight matrix and bias vector of the third fully connected layer, respectively; W4 and b4 represent the weight matrix and bias vector of the fourth fully connected layer, respectively; BN represents the batch normalization operation; ReLU represents the ReLU activation function; +H1 represents the first skip connection; and +H3 represents the second skip connection.

[0084] Specifically, the feature dimensions of the first fully connected layer, the second fully connected layer, the third fully connected layer, and the fourth fully connected layer are set to 128, 128, 64, and 64, respectively.

[0085] Understandably, the first fully connected layer is used to detect preliminary and complex nonlinear combination features from the standardized raw input. Some neurons are specifically activated to detect the impending overshoot state of high tissue stiffness and negative force change rate (reduced force), while others are responsible for detecting the state of safe and rapid propagation in adipose tissue with low stiffness and high positive velocity. The primary features generated by the first fully connected layer are still relatively basic combinations. The second fully connected layer synthesizes more complex and abstract patterns based on the first fully connected layer. For example, it combines the features of impending overshoot with the features of the needle tip being too deep. The identity mapping features generated by this layer have stronger theoretical representation capabilities than the primary features, but they encounter gradient and information loss problems. Therefore, the first skip connection ensures that all valuable primary features extracted by the first layer are not lost or destroyed during the transformation process of the second layer during forward propagation of the model. During backward propagation, the gradient can be directly transmitted back to the first fully connected layer without decay, greatly alleviating the gradient vanishing problem and making the network easier to train. The third fully connected layer, fourth fully connected layer, and second skip connection further map the features to the output tissue stiffness features and needle dynamic stiffness. The third fully connected layer evaluates and weighs the information of the identity mapping features, focusing on solving the current needle dynamic stiffness problem. The fourth fully connected layer maps the features to an abstract concept very close to the final output decision. The second skip connection ensures that the final decision considers both high-level abstract concepts and filtered details. Therefore, the aforementioned multi-skip fully connected layer outperforms the hidden layers of a traditional multilayer perceptron and traditional fully connected layers in extracting tissue stiffness features.

[0086] Furthermore, the output layer includes an output fully connected layer and an activation function layer, and the output mapping unit includes:

[0087] The output mapping subunit is used to pass the tissue stiffness features through the output fully connected layer without activation function to generate the original predicted stiffness value, the original predicted stiffness change value, and the stiffness convergence rate.

[0088] The target steady-state stiffness calculation subunit is used to generate the target steady-state stiffness by passing the original predicted stiffness value through the Sigmoid activation function.

[0089] The predicted stiffness change calculation subunit is used to generate a predicted stiffness change value by passing the original predicted stiffness change value through the Tanh activation function.

[0090] The dynamic stiffness calculation subunit is used to generate the dynamic stiffness of the needle tip by means of a stiffness function based on the target steady-state stiffness, the predicted stiffness change value, and the stiffness convergence rate.

[0091] In particular, the output layer utilizes the powerful fitting ability of neural networks while implementing necessary safety constraints through activation functions.

[0092] Specifically, the process of generating the dynamic stiffness of the needle in the output layer can be represented as:

[0093] Output = [K] ∞ _raw,ΔK_raw,λ]=W out *H4+b out

[0094] K ∞ =Sigmoid(K) ∞ _raw)(K_max-K_min)+K_min

[0095] ΔK = Tanh(ΔK_raw) * ΔK_max

[0096] K(τ)=K ∞ +ΔKexp(-λ*τ)

[0097] In the formula, Output represents the original predicted stiffness value K. ∞ The output vector of the fully connected layer without activation function, consisting of the original predicted stiffness change value ΔK_raw and the stiffness convergence rate λ. out b out These represent the weight matrix and bias vector of the output fully connected layer, respectively; H4 represents the organizational stiffness feature; K ∞denoted by , Sigmoid represents the Sigmoid activation function, K_max and K_min represent the maximum and minimum allowable values ​​of the set steady-state stiffness, respectively, ΔK represents the predicted stiffness change value, Tanh represents the Tanh activation function, ΔK_max represents the maximum allowable rate of change of the steady-state stiffness, K(τ) represents the dynamic stiffness of the needle, exp represents the exponential function, and τ represents the future duration of the timer starting from the current moment.

[0098] Furthermore, the transmission assembly 3 includes a transmission screw 35 that is connected to the needle 5, and the needle stiffness adjustment module includes:

[0099] The damping calculation unit is used to generate predicted damping by applying the critical damping formula to the dynamic stiffness of the needle tip, and to calculate the screw damping term based on the product of the predicted damping and the error in the needle tip's movement speed.

[0100] The stiffness calculation unit is used to calculate the screw stiffness term based on the product of the needle dynamic stiffness and the needle position error;

[0101] The screw shearing force control unit generates the shearing force based on the difference between the screw damping term and the screw stiffness term, and causes the transmission screw 35 to perform the shearing force.

[0102] In particular, motion stability control is achieved through a damping calculation unit, position accuracy control is achieved through a stiffness calculation unit, and finally, the screw force is output through differential logic to adapt to the organizational characteristics.

[0103] Specifically, the process of generating the screw shearing force can be represented as:

[0104]

[0105] F desired (τ)=K(τ)(x desired -x actual )+B(τ)(v desired -v actual )

[0106] In the formula, F desired (τ) represents the screw shearing force over a future time τ, and K(τ) and B(τ) represent the weight matrix of the screw stiffness term and the screw damping term over a future time τ, respectively. This represents the critical damping formula, (x desired -x actual ) represents the needle position error at the current moment, where x desired x actual These represent the target position and actual position of the needle at the current moment, respectively. desired -vactual ) represents the error in the needle's movement speed at the current moment, where v desired v actual These represent the target speed of the needle and the moving speed of the needle at the current moment, respectively.

[0107] Furthermore, the needle force includes the needle force at the current moment and the needle force in the previous sampling period;

[0108] The tissue stiffness calculation module is used to take the ratio of the difference between the needle force at the current moment and the needle force in the previous sampling period to the product of the needle movement speed and the sampling period interval as the tissue stiffness.

[0109] In particular, tissue stiffness can be used to capture stiffness abrupt changes at the junctions where the needle 5 enters the lesion tissue from the adipose tissue.

[0110] Specifically, the process of generating organizational stiffness can be represented as:

[0111]

[0112] In the formula, K tissue (t) represents the organizational stiffness at the current moment, F est (t), F est (t-1) represents the needle force at the current moment and the needle force in the previous sampling period, respectively; v(t) represents the needle movement speed at the current moment; and Δt represents the sampling period interval.

[0113] Furthermore, the breast biopsy excision system also includes:

[0114] The prediction model training module is used to construct a data fitting term based on the mean square error between the dynamic stiffness of the needle and the true stiffness of the sample, construct an overshoot penalty term based on the difference between the dynamic stiffness of the needle and the expected steady-state stiffness, construct a total loss function based on the data fitting term and the overshoot penalty term, and perform backpropagation training on the needle motion model through the total loss function.

[0115] Specifically, the total loss function can be expressed as:

[0116] overshoot=max(0,e(t)*K(τ)-F desired )

[0117] L total =L data +γ*(overshoot) 2

[0118] In the formula, (overshoot) 2 L represents the overshoot penalty. dataL represents the data fitting term. total Let F represent the total loss function, γ represent the weighting coefficient, preferably 0.001, and F desired The target steady-state stiffness is defined, K(τ) represents the needle dynamic stiffness, and e(t) represents the difference between the needle dynamic stiffness and the target steady-state stiffness.

[0119] In particular, a needle motion model based on multilayer perceptron (MLP) is adopted, which uses a multi-hop fully connected layer structure to achieve high-precision, nonlinear prediction of tissue stiffness. This significantly improves the model's feature extraction capability and gradient transfer efficiency, enabling the system to respond quickly to changes in tissue characteristics, adapt to the differences in different patients and tissue types, and ensure the stability of the puncture process.

[0120] like Figure 3 , 4 As shown in Figure 5, further, the breast biopsy needle comprises:

[0121] The vacuum suction chamber 1 is connected to the suction pipe 4 and is used to suck in and contain the target tissue that the needle 5 is cutting.

[0122] The transmission component 2 is disposed inside the vacuum suction cavity 1 and fixed to the needle 5;

[0123] The transmission component 3 is fixed to the transmission member 2 and is used to convert the applied pressure into a torque that sequentially drives the transmission member 2 and the needle 5 to rotate.

[0124] In particular, it enables doctors to drive the transmission component 3 to output rotational torque simply by applying pressure through the handle, without having to manually rotate the needle 5.

[0125] like Figure 3 , 4 As shown in Figure 5, the transmission assembly 3 further includes:

[0126] Press screw 32 is used to convert the pressing force into torque;

[0127] The drive gear 33 is fixedly sleeved on the pressing screw 32 and rotates synchronously with the pressing screw 32;

[0128] Driven gear 34 meshes with driving gear 33 and rotates synchronously with driving gear 33;

[0129] The transmission screw 35 is fixed to the driven gear 34 and the transmission component 2, and is used to adjust the torque of the driven gear 34 to the shearing force before transmitting it to the transmission component 2.

[0130] In particular, the transmission component 3 constructs a transmission method adapted to the breast biopsy scenario by converting the force form of the pressing screw 32, optimizing the direction and torque of gear meshing, and precisely executing the transmission screw 35.

[0131] like Figure 3 , 4 As shown in Figure 5, the force sensor 6 is further mounted on the transmission screw 35.

[0132] Specifically, the force sensor 6 is a miniature force sensor, and the tissue stiffness calculation module, needle stiffness prediction module, needle stiffness adjustment module and prediction model training module are all mounted in the controller that controls the air intake duct 4. The controller is not located on the breast biopsy needle, but on the air intake device that is connected to the air intake duct 4.

[0133] Specifically, a pressing cap 31 is provided on the top of the pressing screw 32. The pressing cap 31 is used to convert the doctor's pressing action into the pressing force received by the pressing screw 32.

[0134] In particular, the integrated design of the vacuum suction chamber and the rotary cutting needle allows for the simultaneous rotary cutting of the target tissue by the needle, while the excised tissue is rapidly drawn into and contained within the suction chamber using a negative pressure suction mechanism. The needle's transmission component employs a mechanical structure combining gear meshing and a screw, efficiently and smoothly converting external pressure into precise control torque to drive the needle's rotation, ensuring reliable power transmission and precise force control during rotary cutting.

[0135] In this embodiment, tissue stiffness is calculated using a dual-parameter approach combining needle force and needle movement speed. Needle stiffness control is then performed based on a multilayer perceptron (MLP) model, avoiding misjudgment of tissue stiffness and adapting the needle's rotary cutting force to individual differences. This effectively prevents subcutaneous vascular damage caused by excessive needle penetration, ensuring a high success rate in needle rotary cutting of the target tissue. A needle motion model based on MLP is employed, utilizing a multi-hop fully connected layer structure to achieve high-precision, non-linear prediction of tissue stiffness. This significantly improves the model's feature extraction capability and gradient transfer efficiency, enabling the system to quickly respond to changes in tissue characteristics, adapt to differences in different patients and tissue types, and ensure the stability of the puncture process. An integrated design of the vacuum suction chamber and rotary cutting needle allows for the simultaneous rotary cutting of the target tissue while a negative pressure suction mechanism rapidly draws in the excised tissue and contains it within the suction chamber. The needle's transmission assembly adopts a mechanical structure combining gear meshing and a screw, which efficiently and smoothly converts external pressure into precise control torque to drive the needle's rotation, ensuring reliable power transmission and precise force control of the needle's rotary cutting.

[0136] Those skilled in the art will recognize that the modules and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0137] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0138] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A breast biopsy excision system, characterized in that, include: The tissue stiffness calculation module is used to obtain the needle force collected by the force sensor (6) set in the transmission assembly (3) of the needle (5) of the breast biopsy needle, and to calculate the tissue stiffness of the needle (5) in contact with the needle based on the needle force and the needle moving speed. The needle stiffness prediction module is used to generate the dynamic stiffness of the needle by using the needle force and the tissue stiffness through a needle motion model based on a multilayer perceptron architecture. The needle stiffness adjustment module is used to determine the rotary cutting force of the transmission component (3) based on the dynamic stiffness of the needle, so that the rotary cutting force of the needle (5) piercing the fat tissue is less than the rotary cutting force piercing the target tissue. The needle motion model includes an input layer, a multi-hop fully connected layer, and an output layer. The needle stiffness prediction module includes: An input data integration unit is used to determine the force error based on the needle force and the expected force, and at least constructs a comprehensive input vector by combining the time rate of change and historical values ​​of the force error and the tissue stiffness through the input layer. The feature mapping unit is used to perform feature mapping on the comprehensive input vector through a multi-hop fully connected layer to generate tissue stiffness features; The output mapping unit is used to output and map the tissue stiffness characteristics through the output layer to generate the needle dynamic stiffness.

2. The breast biopsy excision system according to claim 1, characterized in that, The multi-hop fully connected layer includes a first fully connected layer, a second fully connected layer, a third fully connected layer, and a fourth fully connected layer, and the feature mapping unit includes: The feature extraction subunit is used to extract basic features from the comprehensive input vector through the first fully connected layer to generate primary features; The feature integration subunit is used to extract the primary features through the second fully connected layer to generate intermediate features, and to fuse the intermediate features with the primary features through a first skip connection to generate identity mapping features. The feature refinement subunit is used to refine the identity mapping features through the third fully connected layer to generate high-level features; The feature mapping stabilization subunit is used to perform feature mapping on the high-level features through the fourth fully connected layer to generate decision-oriented features, and to perform feature fusion stabilization by connecting the decision-oriented features with the high-level features through a second skip connection to generate the organization stiffness features.

3. The breast biopsy excision system according to claim 1, characterized in that, The output layer includes an output fully connected layer and an activation function layer, and the output mapping unit includes: The output mapping subunit is used to pass the tissue stiffness features through the output fully connected layer without activation function to generate the original predicted stiffness value, the original predicted stiffness change value, and the stiffness convergence rate. The target steady-state stiffness calculation subunit is used to generate the target steady-state stiffness by passing the original predicted stiffness value through the Sigmoid activation function. The predicted stiffness change calculation subunit is used to generate a predicted stiffness change value by passing the original predicted stiffness change value through the Tanh activation function. The dynamic stiffness calculation subunit is used to generate the dynamic stiffness of the needle tip by means of a stiffness function based on the target steady-state stiffness, the predicted stiffness change value, and the stiffness convergence rate.

4. The breast biopsy excision system according to claim 1, characterized in that, The transmission assembly (3) includes a transmission screw (35) that is connected to the needle (5) for transmission, and the needle stiffness adjustment module includes: The damping calculation unit is used to generate predicted damping by applying the critical damping formula to the dynamic stiffness of the needle tip, and to calculate the screw damping term based on the product of the predicted damping and the error in the needle tip's movement speed. The stiffness calculation unit is used to calculate the screw stiffness term based on the product of the needle dynamic stiffness and the needle position error; The screw shearing force control unit generates the shearing force based on the difference between the screw damping term and the screw stiffness term, and causes the drive screw (35) to perform the shearing force.

5. The breast biopsy excision system according to claim 1, characterized in that, The needle force includes the needle force at the current moment and the needle force in the previous sampling period; The tissue stiffness calculation module is used to take the ratio of the difference between the needle force at the current moment and the needle force in the previous sampling period to the product of the needle movement speed and the sampling period interval as the tissue stiffness.

6. The breast biopsy excision system according to claim 1, characterized in that, Also includes: The prediction model training module is used to construct a data fitting term based on the mean square error between the dynamic stiffness of the needle and the true stiffness of the sample, construct an overshoot penalty term based on the difference between the dynamic stiffness of the needle and the expected steady-state stiffness, construct a total loss function based on the data fitting term and the overshoot penalty term, and perform backpropagation training on the needle motion model through the total loss function.

7. The breast biopsy excision system according to any one of claims 1 to 6, characterized in that, The breast biopsy needle includes: The vacuum suction chamber (1) is connected to the suction pipe (4) and is used to suck in and contain the target tissue cut by the needle (5); The transmission component (2) is disposed inside the vacuum suction cavity (1) and fixed to the needle (5); The transmission assembly (3), fixed to the transmission member (2), is used to convert the applied pressure into a torque that sequentially drives the transmission member (2) and the needle (5) to rotate.

8. The breast biopsy excision system according to claim 7, characterized in that, The transmission assembly (3) includes: The pressing screw (32) is used to convert the pressing force into torque; The drive gear (33) is fixedly sleeved on the pressing screw (32) and rotates synchronously with the pressing screw (32); The driven gear (34) meshes with the driving gear (33) and rotates synchronously with the driving gear (33); The transmission screw (35) is fixed to the driven gear (34) and the transmission member (2) and is used to adjust the torque of the driven gear (34) to the shearing force and then transmit it to the transmission member (2).

9. The breast biopsy excision system according to claim 8, characterized in that, The force sensor (6) is sleeved on the transmission screw (35).

Images