CN110664524A

CN110664524A - Devices, systems, and media for guiding stent implantation in a vessel

Info

Publication number: CN110664524A
Application number: CN201910950451.4A
Authority: CN
Inventors: 宋麒; 尹游兵; 李育威; 智英轩; 刘潇潇; 祝烨; 马斌
Original assignee: Kunlun Beijing Medical Cloud Technology Co Ltd
Current assignee: Kunlun Beijing Medical Cloud Technology Co Ltd
Priority date: 2018-10-08
Filing date: 2019-10-08
Publication date: 2020-01-10
Anticipated expiration: 2039-10-08
Also published as: CN110664524B

Abstract

The present disclosure relates to a device, system, and medium for guiding stent implantation in a blood vessel. The device comprises: an interface configured to receive a sequence of 2D images of a blood vessel; a processor configured to: reconstructing a first 3D model of a vessel based on the 2D image sequence; acquiring structural parameters and a position to be implanted of a stent to be implanted; modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted; and predicting a second blood flow reserve fraction of the blood vessel by using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model. The device can be used for FFR prediction before stent implantation, can also accurately and conveniently predict the FFR of various stents after being implanted into blood vessels at various positions, has quick and accurate prediction, and provides sufficient and accurate guide for a user on selection and positioning of the stent to be implanted and the prognosis effect of stent implantation.

Description

Devices, systems, and media for guiding stent implantation in a vessel

Cross Reference to Related Applications

This application claims priority from U.S. provisional application No. 62/742,914 filed on 8.10.2018, the entire contents of which are incorporated herein by reference.

Technical Field

The present disclosure relates generally to calculating fractional flow reserve to implant a stent. More particularly, the present disclosure relates to devices, systems, and media for guiding stent implantation in a blood vessel.

Background

Fractional flow reserve is a gold standard for functional assessment of coronary artery stenosis. However, pressure-guidewire-based measurement of Fractional Flow Reserve (FFR) is often percutaneously invasive, expensive, and time consuming, often in conjunction with drug-induced hyperemia. Therefore, some computer modeling techniques have been proposed to calculate Virtual Fractional Flow Reserve (VFFR) without using a pressure guidewire. However, existing models employ simplified fluid dynamics equations that assume various empirical constants and use a weighted average of multi-scale calculations, and the determination and weighting of the empirical constants is subject to error. The choice of scale is also empirical. To calculate the virtual FFR, some intermediate features, such as stenosis degree, etc., also need to be calculated. Such modeling, also known as physical-based models, does not accurately predict the accuracy of fractional flow reserve and is inconvenient to operate.

Deep learning has recently been introduced to enable accurate prediction of FFR, which can be trained with the true FFR values obtained from invasive surgery, and thus with higher accuracy, and can be continuously improved with acquiring more data, e.g. training data for different patient groups making it easier to adapt to the prediction of FFR for that patient group, than physical-based models employing various simplified physical modeling assumptions. However, the deep learning models in the prior art are generally only used for prediction of FFR of diseased vessels and do not provide sufficient guidance for stent implantation in vessels.

The present disclosure is provided to overcome the above disadvantages in the prior art.

Disclosure of Invention

Therefore, there is a need for a device, system and medium for guiding stent implantation in a blood vessel, which can be used not only for FFR before stent implantation (also referred to herein as pre-stent FFR) prediction, but also for FFR after various stents are implanted in blood vessels at various positions (also referred to herein as post-stent FFR) with accuracy and convenience, with rapid and accurate prediction, thereby providing a user with sufficient and accurate guidance in selection and positioning of the stent to be implanted and the prognostic effect of stent implantation.

In a first aspect, the present disclosure provides a device for guiding stent implantation in a blood vessel. The device includes: an interface configured to receive a sequence of 2D images of a blood vessel. The apparatus also includes a processor configured to: reconstructing a first 3D model of a vessel based on the 2D image sequence; acquiring structural parameters and a position to be implanted of a stent to be implanted; modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted; and predicting a second blood flow reserve fraction of the blood vessel by using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model.

In a second aspect, the present disclosure provides a device for guiding stent implantation in a blood vessel. The device includes: a first acquisition unit configured to acquire a 2D image sequence of a blood vessel; a reconstruction unit configured to reconstruct a first 3D model of the vessel based on the sequence of 2D images; a second acquisition unit configured to acquire structural parameters of a stent to be implanted and a position to be implanted; a model modification unit configured to: modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted; a prediction unit configured to: and predicting a second blood flow reserve fraction of the blood vessel by using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model.

In a third aspect, the present disclosure provides a system for guiding stent implantation in a blood vessel. The system comprises an apparatus according to the first aspect; and a display configured to: displaying the first 3D model for the user to perform stenosis analysis of the vessel at least where a pre-operative analysis of the vessel is performed; and displaying the second 3D model and a second fractional flow reserve of the blood vessel in a case where an implantation effect of the stent is simulated.

In a fourth aspect, the present disclosure provides a non-transitory computer-readable medium storing instructions. The instructions, when executed by the processor, perform a method for guiding stent implantation in a vessel. The method comprises the following steps: receiving a sequence of 2D images of a blood vessel; reconstructing a first 3D model of a vessel based on the 2D image sequence; acquiring structural parameters and a position to be implanted of a stent to be implanted; modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted; and predicting a second blood flow reserve fraction of the blood vessel by using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model.

The device, the system and the medium for guiding the stent implantation in the blood vessel can be used for predicting the FFR before stent implantation, can also accurately and conveniently predict the FFR of various stents after the stents are implanted in various positions in the blood vessel, and the prediction is rapid and accurate, thereby providing sufficient and accurate guidance for a user on the selection and the positioning of the stent to be implanted and the prognosis effect of the stent implantation. It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.

Drawings

In the drawings, which are not necessarily drawn to scale, like reference numerals may depict like parts in different views. Like numbers with letter suffixes or different letter suffixes may represent different instances of similar components. The drawings illustrate various embodiments, generally by way of example and not by way of limitation, and together with the description and claims, serve to explain the disclosed embodiments. Such embodiments are illustrative and not intended to be exhaustive or exclusive embodiments of the present apparatus, system, or non-transitory computer readable medium having stored thereon instructions for carrying out the methods.

Fig. 1 shows a flow diagram of a method for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure;

fig. 2(a) shows a diagram of an angiographic image of a diseased right coronary artery;

fig. 2(B) shows a 3D model reconstructed using the sequence of angiographic images of the diseased right coronary artery shown in fig. 2(a) and a graphical representation of a virtual FFR (pre-stented virtual FRR) at the various locations of the vessel predicted using the method for guiding stenting in a vessel according to an embodiment of the present disclosure;

FIG. 2(C) shows a graphical representation of the actual FFR (Pre-stented FFR) measured using a pressure guidewire for the diseased right coronary artery shown in FIG. 2 (A);

fig. 2(D) shows a graph simulating a virtual FFR (post-stent virtual FFR) at each location of a blood vessel predicted after implantation of a stent having a size of 3.4mm (deployed diameter) × 24mm (length) in the right coronary artery shown in fig. 2(a) using a method for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure;

fig. 2(E) is a diagram showing a contrast image after a stent having a size of 3.4mm (expanded diameter) × 24mm (length) is actually implanted at the implantation position shown in fig. 2(D) in the right coronary artery shown in fig. 2 (a);

fig. 2(F) shows a graphical representation of the actual FFR (post-stenting FFR) measured using a pressure guidewire after actual implantation of a stent having dimensions 3.4mm (deployed diameter) × 24mm (length) at the implantation location shown in fig. 2(D) in the right coronary artery shown in fig. 2 (a);

fig. 3 shows a schematic view of a device for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure;

fig. 4 shows a schematic view of a device for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure;

fig. 5 shows a block diagram of a system for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure.

Detailed Description

Fig. 1 shows a flow diagram of a method for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure. As shown in fig. 1, the process 100 begins with step 101 of receiving a 2D image sequence of a blood vessel. In some embodiments, the 2D image sequence of the vessel may comprise an angiographic image sequence of the vessel at least two different projection angles. The sequence of angiographic images may be acquired by any available image acquisition device, which may be embedded in the device 100; it may also be independent of the device 100, i.e. the image acquisition device is an external device. The 2D image sequence of the blood vessel may be directly received from the image acquiring device or indirectly received from the image acquiring device, for example, the angiography device acquires angiography image sequences of the blood vessel at two different projection angles for the blood vessel, stores the angiography image sequences in a medical image database (e.g., a medical image database of a hospital), and acquires the angiography image sequence of the corresponding blood vessel by connecting and accessing the medical image database.

At step 102, a first 3D model of a vessel may be reconstructed based on the 2D image sequence. In some embodiments,

steps

101 and 102 may be implemented by using existing blood vessel image analysis and reconstruction software, and

steps

103 and 105 may be integrated with the existing blood vessel image analysis and reconstruction software or may be written as a software module independent of the existing blood vessel image analysis and reconstruction software. The first 3D model almost restores the 3D geometry of the vessel, and by displaying the first 3D model, the physician can roughly notice which segments of the vessel may present a stenosis. The reconstructed first 3D model may be used as a basis to predict a first FFR of the blood vessel (pre-stent vFFR) by using a trained learning model, and specifically, may be used to predict a first FFR of the blood vessel based on a structure-related parameter of the first 3D model (e.g., a sequence of structure-related parameters along the centerline), may be used in combination with a blood flow-related parameter (e.g., a sequence of flow velocities along the centerline) based on the structure-related parameter of the first 3D model, and may be used to predict a first FFR of the blood vessel based on a sequence of image blocks of the first 3D model along the centerline of the blood vessel, such as, but not limited to, a sequence-to-sequence learning network. Note that as the parameters based on are different (e.g. structure-related parameter sequence, fused sequence of structure-related parameters and blood flow-related parameters, image patch sequence), the learning network structure changes accordingly, and the training data set used for training also changes accordingly. For example, the learning network may be selected from a recurrent neural network, a gated recurrent unit, a long-short term memory unit, or a bidirectional variant thereof, and so on, which are not described herein again. The above exemplary description of the learning network also applies to the learning model applied to the second 3D model in step 105, which is not described in detail below.

Next, in step 103, structural parameters of the stent to be implanted and the location to be implanted may be obtained. In some embodiments, the structural parameters of the stent to be implanted include the length and deployed diameter of the stent to be implanted. In some embodiments, the structural parameters of the stent to be implanted and the position to be implanted may be entered by the physician, for example, the physician, after referring to the first 3D model of the vessel presented, manually decides what model of the implanted stent to implant and presets the position to be implanted, so that the physician may autonomously enter all stent implantation protocols intended for prognostic effect prediction via

subsequent steps

104 and 105. In some embodiments, the structural parameters of the stent to be implanted and the position to be implanted may also be automatically determined by the processor according to the reconstructed stenosis distribution of the first 3D model, for example, the maximum value of the stenosis region of the blood vessel and the diameter (healthy diameter) of the healthy state thereof is automatically determined, the expanded diameter of the stent to be implanted is determined to be at least the maximum value of the healthy diameter and the length is at least the continuous length of the stenosis region, and of course, corresponding margins may be set to take into account the bending of the blood vessel and the tension of the muscles inside the blood vessel wall so as to ensure that the stent to be implanted can cover the stenosis region and correct the stenosis inner diameter thereof to the healthy diameter against the inward tension of the muscles. When there are a plurality of discrete stenotic regions, there may be a plurality of stents to be implanted and corresponding locations to be implanted. In some embodiments, semi-automatic acquisition of the structural parameters of the stent to be implanted and the position to be implanted can also be realized, that is, the recommended structural parameters of the stent to be implanted and the recommended position to be implanted can be automatically determined and confirmed and adjusted by a doctor, so that the efficiency and the accuracy can be considered. Through manual (fully manual or semi-automatic) intervention of the structural parameters of the stent to be implanted and the setting of the position to be implanted by a physician, the physician can benefit from rich experience in blood vessel image analysis, and the available stent implantation schemes can be analyzed in a targeted way to guide the physician to select a preferred scheme or verify the feasibility of the implantation scheme, so that the workload of prognosis analysis is relieved.

In step 104, the first 3D model may be modified to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted, wherein the second 3D model represents the 3D geometry assumed by the vessel after the stent having the acquired preset structural parameters is implanted at the acquired preset position to be implanted. By using step 104, the second 3D model can be obtained conveniently and rapidly based on the first 3D model, compared with the conventional method of obtaining the second 3D model by performing 2D image acquisition and 3D reconstruction on the stent-implanted blood vessel in the prior art, step 104 greatly reduces the computational resources and the workload of physicians, consumes less time, and can obtain 3D models after various stents are implanted at various implantation positions that cannot be obtained by the conventional method (it is impossible to repeatedly implement different stent implantation schemes and angiographic imaging on patients).

Next, in step 105, a second FFR of the blood vessel may be predicted using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model. Step 101-.

Steps

104 and 105 are further described in detail below.

In some embodiments, in step 104, a portion of the first 3D model at the location to be implanted may be expanded to conform to the structural parameters of the stent to be implanted and to maintain the curvature of the portion. Through letting the implantation of support keep the curvature of the vascular part of position department of implanting, can avoid implanting the distortion of the vascular shape that leads to and trend to avoid the unfavorable interference to improper tension of the blood vessel of position department of implanting and blood flow distribution, this is the state that should realize when blood vessel is implanted, the extension of the part of position department of waiting to implant in the first 3D model makes it conform with the length and the expansion diameter of waiting to implant the support, make the extension effect of implant support to the vascular section of position department of implanting can conveniently be simulated to the second 3D model that obtains. In some embodiments, the tension of the muscle of the blood vessel wall can be considered, so that the diameter of the blood vessel section at the position to be implanted in the second 3D model is the expansion diameter of the stent to be implanted multiplied by a tension coefficient less than or equal to 1; in further exemplary embodiments, the diameters of the respective portions of the second 3D model can also be adjusted in consideration of the stenosis distribution of the vessel segment at the location to be implanted, so that the diameters are smaller at the stenosis.

In step 105, the structure-related parameter and the blood flow-related parameter may be used as characteristic parameters of the second 3D model for the prediction of post-stent vFFR. In some embodiments, the structure-related parameter includes at least one of a vessel diameter distribution along the centerline of the second 3D model, a healthy diameter distribution (which may be predicted by regression of a diameter peak distribution of the model along the centerline), a curvature distribution, and an optical path depth distribution. In some embodiments, the blood flow related parameters of the second 3D model may be obtained in various ways. For example, the calculation may be simulated by performing a computational fluid dynamics finite element modeling of the second 3D model, the inlet boundary condition may be an inlet blood flow rate and the outlet boundary condition may be a preset microvascular resistance, it is noted that the microvascular resistance does not change much in both the stent implanted and the stent non-implanted states, and the resulting vFFR is not sensitive to deviations in the inlet blood flow rate, where the inlet blood flow rate and the preset microvascular resistance at the time of angiographic imaging of the vessel without the (virtual) stent implanted may be followed, making the simulation calculation simpler and quicker and the calculation load lower. In some embodiments, the second 3D model may directly follow the blood flow related parameters of the first 3D model, and it is found through simulation experiments that the final vFFR is insensitive to the deviation of the blood flow related parameters of the first 3D model and the second 3D model, and the vFFR having a better goodness of fit with the invasive pressure guidewire pullback measurement value may still be predicted by directly following the blood flow related parameters of the first 3D model.

In some embodiments, taking the example of obtaining a 2D image sequence using an angiographic apparatus, the blood flow related parameters of the first 3D model may be obtained using a medical imaging modality other than an angiographic apparatus, such as, but not limited to, ultrasound doppler, or the like. In some embodiments, the flow velocity distribution may also be determined as a blood flow related parameter of the first 3D model based on the 2D angiographic image sequence of at least two different projection angles of the blood vessel received in step 101. Various methods for determining the flow velocity distribution based on two 2D angiography image sequences with different projection angles are developed, such as, but not limited to, the TIMI frame counting method, the method described in chinese patent application publication No. CN 108550388 a, with the title of 2018101897296, the device and system for calculating parameters of blood vessel blood flow based on angiography, and the like, which are not described herein again.

That is, the blood flow related parameters of the second 3D model need not be recalculated based on the second 3D model after the stent is implanted, nor need a medical imaging modality other than the angiographic apparatus be introduced, but may be obtained only from the 2D image sequence obtained in step 101, and it is verified by a contrast test that the vFFR with a high degree of matching with the FFR value actually measured by the invasive pressure guidewire can still be obtained. Therefore, the operation in the boot flow can be simplified, the boot flow is accelerated, the hardware cost is reduced, and the calculation load is effectively reduced.

In some embodiments, the structure-related parameter of the second 3D model may include at least one of a vessel diameter distribution, a healthy diameter distribution, a curvature distribution, and an optical path depth distribution of the 3D model along the centerline.

In some embodiments, the healthy diameter distribution of the vessel along the centerline may be predicted by performing a regression from the first 3D model or the second 3D peak distribution of diameters along the centerline of the vessel. For example, the present invention may be applied to various methods for predicting a healthy radius described in chinese patent application publication No. CN109979593A, entitled "method for predicting a healthy radius of a blood vessel route, method for predicting a candidate stenosis of a blood vessel route, and device for predicting a degree of stenosis" filed under the patent application No. 2019102628380. Further, the stenosis region may be determined based on a diameter valley distribution along the centerline of the first 3D model of the blood vessel and a healthy diameter distribution. Various methods may be employed to determine the stenotic region. For example, a ratio of a diameter valley to a healthy diameter throughout the blood vessel path may be determined, and when the ratio is less than a first predetermined threshold, it is determined that the site belongs to a stenotic region. For another example, the difference between the healthy diameter and the diameter trough and the ratio of the healthy diameter may be determined throughout the vascular pathway, and when the ratio is greater than a second predetermined threshold, it is determined that the site belongs to a stenotic region.

The determined stenosis region may be presented to a physician to visually guide the physician in step 103 to confirm the appropriate stent structure parameters and implantation location. In some embodiments, similar to step 105, the first FFR of the blood vessel may also be predicted by using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the first 3D model of the blood vessel, whereby the pre-stent-implantation vFFR may be presented to the physician so that the physician may comprehensively grasp the actual condition of the diseased blood vessel in both an intuitive and quantitative manner by looking at the 3D model of the pre-stent-implantation blood vessel (optionally, the stenotic region) and the pre-stent-implantation vFFR, thereby making it easier to make an appropriate selection on the structure parameters of the stent to be implanted and the location to be implanted. Furthermore, based on the pre-stented vFFR of the vessel, one or more stenosis regions in question can be quickly identified, and critical stenosis that requires (or most requires intervention) can be automatically determined based on preset conditions (e.g., vFFR less than 0.75).

In some embodiments, a region of interest may also be determined in the first 3D model based on the determined stenosis region; modifying the attention area of the first 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted, wherein the modified attention area 3D model is used as the second 3D model. The region of interest may be arranged to cover the stenosis region and a peripheral region thereof at a distance. Here, the deep learning prediction of vFFR is limited to the region of interest only, which greatly speeds up the prediction speed and avoids consuming computational resources and time on healthy vessel segments where no stenosis exists at all (or where the probability of stenosis existence is very low).

Using a guidance method according to various embodiments of the present disclosure, a percutaneous coronary artery (PCI) stent implantation of the right coronary artery was guided and simulated. Fig. 2(a) shows a graphical representation of an angiographic image of the right coronary artery of the lesion, with the stenosis noted.

Next, a 3D model was reconstructed using the sequence of angiographic images of the diseased right coronary artery shown in fig. 2(a), and the vFFR (pre-stented virtual FRR) across the vessel was predicted using the method for guiding stenting in a vessel according to an embodiment of the present disclosure, as shown in fig. 2(B), the predicted overall vFFR for this right coronary artery was 0.78, and fig. 2(C) shows that the actual FFR (pre-stented FFR) measured using a pressure guide wire for the diseased right coronary artery shown in fig. 2(a) was 0.75, both of which have a high degree of agreement.

Fig. 2(D) shows a graph simulating the predicted vFFR (post-stenting vFFR) throughout the vessel after implantation of a stent having a size of 3.4mm (deployed diameter) × 24mm (length) in the right coronary artery shown in fig. 2(a) (the implantation position is as shown in fig. 2 (D)) using a method for guiding stent implantation in a vessel according to an embodiment of the present disclosure, the post-stenting vFFR for the population of the right coronary artery being predicted to be 0.97. Fig. 2(E) is a diagram showing a contrast image after a stent having a size of 3.4mm (expanded diameter) × 24mm (length) is actually implanted at the implantation position shown in fig. 2(D) in the right coronary artery shown in fig. 2 (a). Fig. 2(F) shows a graphical representation of the actual FFR (post-stenting FFR) measured using a pressure guidewire after actual stenting of 3.4mm (deployed diameter) × 24mm (length) dimensions at the implantation site shown in fig. 2(D) in the right coronary artery shown in fig. 2(a), the post-stenting FFR for this right coronary artery population being 0.99, and it can be seen that the post-stenting vFFR (0.97) goodness of fit predicted using the guidance method of the disclosed embodiment is very high.

Fig. 3 shows a device for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure. As shown in fig. 3, the apparatus 300 includes a processor 301 and a communication interface 303. Processor 301 may be a processing device such as a microprocessor, Central Processing Unit (CPU), Graphics Processing Unit (GPU), etc., including one or more general-purpose processing devices. More specifically, the processor may be a Complex Instruction Set Computing (CISC) microprocessor, a Reduced Instruction Set Computing (RISC) microprocessor, a Very Long Instruction Word (VLIW) microprocessor, a processor running other instruction sets, or a processor running a combination of instruction sets. The processor may also be one or more special-purpose processing devices such as application specific integrated circuits (asic), Field Programmable Gate Arrays (FPGAs), digital signal processors (dsp), system on chips (SoCs), or the like.

The communication interface 303 may be configured to perform step 101 shown in fig. 1, i.e. to receive a sequence of 2D images of the blood vessel. The 2D image sequence of the vessel may be used for 3D reconstruction of the vessel, for example, it may be a sequence of contrast images of the vessel acquired at least two different projection angles, a sequence of vessel slice images obtained by tomography along a centerline of the vessel, or the like. In some embodiments, the communication interface 303 may be configured to communicatively connect to an image imaging device, such as, but not limited to, an angiographic device, to receive a sequence of angiographic images of a blood vessel acquired via at least two different projection angles acquired by the angiographic device; it may also be configured to be communicatively connected to a medical image database which may store sequences of angiographic images of vessels acquired at least two different projection angles of the vessel to be stent implanted, such that the apparatus 300 may access and acquire the sequences of angiographic images of vessels acquired at least two different projection angles of the vessel stored on the medical image database via the communication interface 303.

The processor 301 performs at least functions or methods implemented by code or instructions in a program stored in the storage 305.

The processor 301 may be configured to: reconstructing a first 3D model of the vessel based on the received sequence of 2D images; acquiring structural parameters and a position to be implanted of a stent to be implanted; modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted; predicting a second fractional flow reserve of the blood vessel using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model. With the above configuration of the apparatus 300, 3D arterial branch reconstruction and blood flow estimation can be completed in a few minutes, and online pre-stent and post-stent vFFR prediction can be completed in a few seconds, both within clinically acceptable time; and the predicted vFFR after stent implantation has better fit with the FFR measured by the actual invasive pressure guide wire. In this manner, the device 300 can quickly and accurately assess the prognostic function of a blood vessel after implantation of a stent having the acquired preset structural parameters at the preset implantation position, thereby providing a physician with sufficient, accurate and timely guidance in planning a stent implantation plan. Please supplement the beneficial technical effects.

In some embodiments, the processor 301 may be configured to perform various implementations of

steps

102 and 105 and various implementations of associated steps according to various embodiments of the present disclosure.

In some embodiments, the processor 301 is configured to expand a portion of the first 3D model at the location to be implanted to conform to the structural parameters of the stent to be implanted and to maintain the curvature of the portion; wherein the structural parameters of the stent to be implanted include the length and the deployed diameter of the stent to be implanted. By allowing the implantation of the stent to maintain the curvature of the vessel portion at the implantation site, distortions in the shape and orientation of the vessel caused by the implantation can be avoided, thereby avoiding undue strain on the vessel at the implantation site and adverse interference with blood flow distribution, a condition that should be achieved when the vessel is implanted. The expansion of the part of the first 3D model at the position to be implanted is made to conform to the length and the expansion diameter of the stent to be implanted, so that the obtained second 3D model can conveniently simulate the expansion effect of the implanted stent on the vascular section at the position to be implanted.

In some embodiments, the 2D image sequence of the blood vessel comprises an angiographic image sequence of the blood vessel at least two different projection angles, and the processor 301 is configured to determine the flow velocity distribution as the blood flow related parameter of the second 3D model based on the angiographic image sequence of the blood vessel at least two different projection angles. That is to say, the blood flow related parameters of the second 3D model are not necessarily obtained by recalculation based on the second 3D model after the stent is implanted, and are also not necessarily obtained by introducing a medical imaging modality other than an angiography device, but can be obtained only according to the 2D image sequence, and a contrast test proves that the vFFR with higher matching degree with the actually measured FFR value of the invasive pressure guide wire can still be obtained. Therefore, the operation in the boot flow can be simplified, the boot flow is accelerated, the hardware cost is reduced, and the calculation load is effectively reduced.

In some embodiments, the processor 301 is configured to predict a first fractional flow reserve of the blood vessel using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the first 3D model of the blood vessel. Thus, the pre-stented vFFR can be presented to a physician so that the physician can comprehensively grasp the actual condition of the diseased vessel in both an intuitive and quantitative manner by looking at the 3D model of the pre-stented vessel and the pre-stented vFFR, thereby making it easier to make appropriate selections on the structural parameters of the stent to be implanted and the location to be implanted. In addition, based on the vessel before stent implantation vFFR, one or more stenosis regions with problems can be quickly identified, and key stenosis needing (or needing to be interfered with most) can be automatically judged based on preset conditions.

In some embodiments, the structure-related parameters of each of the first 3D model and the second 3D model include at least one of a vessel diameter distribution, a healthy diameter distribution, a curvature distribution, and an optical path depth distribution of the 3D model along the centerline.

In some embodiments, the processor 301 is further configured to perform a regression on the diameter peak distribution of the first 3D model of the vessel along the centerline to predict a healthy diameter distribution of the vessel along the centerline; the stenosis region is determined based on a diameter trough distribution along the centerline of the first 3D model of the blood vessel and a healthy diameter distribution. The determined stenosis region may be presented to a physician to visually guide the identification of appropriate stent structure parameters and implantation locations.

In some embodiments, the processor 301 is further configured to determine a region of interest in the first 3D model based on the determined stenosis region; and modifying the attention area of the first 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted, wherein the modified attention area 3D model is used as a second 3D model. Here, the deep learning prediction of vFFR is limited to the region of interest only, which greatly speeds up the prediction speed and avoids consuming computational resources and time on healthy vessel segments where no stenosis exists at all (or where the probability of stenosis existence is very low).

Further, the apparatus 300 may also include an input/output interface 302, a memory 304, and a storage 305. The various components of the apparatus 300 may be configured in a centralized fashion, such as being interconnected via a bus 106; the configuration can also be performed in a distributed manner (for example, distributed in the cloud), and the communication between the components is realized through a communication line (wired or wireless communication manner). The apparatus 300 at least implements functions or methods in various embodiments described in the present disclosure through interaction of the processor 301, the memory 304, the storage 305, the input/output interface 302, and the communication interface 303.

The memory 304 temporarily stores the program loaded from the storage 305 and provides a work area for the processor 301. Various pieces of data generated when the processor 301 executes the program are also temporarily stored in the memory 304. For example, memory 304 may include Random Access Memory (RAM) and Read Only Memory (ROM). For example, the memory 305 may store programs executed by the processor 301. For example, the storage 305 includes a hard disk (HDD), a Solid State Drive (SSD), and a flash memory.

The input/output interface 302 includes an input device that inputs various operations to the apparatus 300, and an output device that outputs a result of processing by the apparatus 300.

The communication interface 303 transmits and receives various data through a network. The communication may be performed by cable or wireless, and any communication protocol may be used as long as they can communicate with each other. The communication interface 303 has a function of performing communication with the apparatus 300 and transmitting various pieces of data.

The program for operating the device 300 according to the embodiment may be provided in a state where the program is stored in a computer-readable storage medium. The storage medium may store the program in a "non-transitory tangible medium". The program may also include a software program or a computer program.

Further, at least some of the processes in apparatus 300 may be implemented by cloud computing, which configures one or more computers. Further, some processes in system 300 may be performed at least by a standby device. Further, at least some of the processes of each functional unit implemented by processor 301 may be performed by a standby device.

Fig. 4 shows a device for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure. As shown in fig. 4, the apparatus 400 may include a first obtaining unit 401, a second obtaining unit 402, a reconstruction unit 403, a model modification unit 404, a prediction unit 405, and optionally may further include a stenosis region determining unit 406.

The first acquisition unit 401 is configured to acquire a 2D image sequence of a blood vessel. The second acquiring unit 402 is configured to acquire structural parameters of the stent to be implanted and the position to be implanted. The reconstruction unit 403 is configured to reconstruct a first 3D model of the vessel based on the sequence of 2D images. The model modification unit 404 is configured to modify the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted; the prediction unit 405 is configured to predict a second fractional flow reserve of the blood vessel using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model. The device 400 shown in fig. 4 can be used not only for pre-stent FFR (also referred to herein as pre-stent FFR) prediction, but also for accurately and conveniently predicting FFR after various stents are implanted into a blood vessel at various locations (also referred to herein as post-stent FFR), with rapid and accurate prediction, thereby providing a user with sufficient and accurate guidance in the selection and positioning of the stent to be implanted and the prognostic effect of stent implantation. Please supplement the beneficial technical effects.

In some embodiments, the model modification unit 404 is further configured to expand a portion of the first 3D model at the location to be implanted to conform to the structural parameters of the stent to be implanted and to maintain the curvature of the portion. By allowing the implantation of the stent to maintain the curvature of the vessel portion at the implantation site, distortions in the shape and orientation of the vessel caused by the implantation can be avoided, thereby avoiding undue strain on the vessel at the implantation site and adverse interference with blood flow distribution, a condition that should be achieved when the vessel is implanted. The expansion of the part of the first 3D model at the position to be implanted is made to conform to the length and the expansion diameter of the stent to be implanted, so that the obtained second 3D model can conveniently simulate the expansion effect of the implanted stent on the vascular section at the position to be implanted.

In some embodiments, the prediction unit 405 may be further configured to: and predicting a first blood flow reserve fraction of the blood vessel by using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the first 3D model of the blood vessel. Thus, the pre-stented vFFR can be presented to a physician so that the physician can comprehensively grasp the actual condition of the diseased vessel in both an intuitive and quantitative manner by looking at the 3D model of the pre-stented vessel and the pre-stented vFFR, thereby making it easier to make appropriate selections on the structural parameters of the stent to be implanted and the location to be implanted. In addition, based on the vessel before stent implantation vFFR, one or more stenosis regions with problems can be quickly identified, and key stenosis needing (or needing to be interfered with most) can be automatically judged based on preset conditions.

In some embodiments, the apparatus 400 optionally includes a stenosis region determination unit 406. The determination unit 406 is configured to: regression of the diameter peak distribution along the centerline of the first 3D model of the vessel to predict a healthy diameter distribution of the vessel along the centerline; the stenosis region is determined based on a diameter trough distribution along the centerline of the first 3D model of the blood vessel and a healthy diameter distribution. The determined stenosis region may be presented to a physician to visually guide the identification of appropriate stent structure parameters and implantation locations.

Fig. 5 illustrates a system for guiding stent implantation in a blood vessel according to an embodiment of the present disclosure. The system 500 comprises components according to analogy with components of the apparatus 300, such as a processor 501, an input/output interface/502, a communication interface 503, a memory 504 and a storage 505, and a display 507.

The display 507 may be located at the terminal device used by the physician or at a hospital server. The components of system 500, including display 507, may be configured in a centralized manner, interconnected via bus 106, or may be configured in a distributed manner (e.g., distributed in the cloud), with communication between the components being accomplished via communication lines (wired or wireless). The interaction of the processor 501, the memory 504, the storage 506, the input/output interface 502, and the communication interface 503 of the apparatus 500 enable at least the functions or methods described in the various embodiments of the present disclosure.

In some embodiments, the display 507 is configured to display the first 3D model for the user to perform stenosis analysis of the vessel at least if a pre-operative analysis of the vessel is performed; and displaying the second 3D model and a second fractional flow reserve of the blood vessel in a case where an implantation effect of the stent is simulated. Therefore, the physician can know the FFR prediction before the stent is implanted and can also know the FFR prediction after various stents are implanted into the blood vessel at various positions, thereby providing sufficient and accurate guidance for the selection and the positioning of the stent to be implanted and the prognostic effect of the stent implantation.

In some embodiments, the display 507 is configured to also display the healthy diameter distribution of the stenotic region and its periphery in the first 3D model in case of performing a pre-operative analysis of the blood vessel, thereby helping the physician to be able to intuitively guide the confirmation of the appropriate stent structure parameters and implantation position.

In some embodiments, the display 507 is configured to display a user interaction unit configured to set a model of the stent to be implanted and to input a position to be implanted by a user. Wherein the user interaction unit further comprises an indication unit configured to: dragging by the user in a peripheral region of the stenosis region in the first 3D model to preset the position to be implanted. In this way, the manual (fully manual or semi-automatic) intervention of the model of the stent to be implanted and the input of the setting of the position to be implanted by the physician can benefit from the rich experience of the physician in the analysis of the blood vessel images, and the targeted analysis of the available stent implantation schemes to guide the physician to select a preferred scheme or to verify the feasibility of the setting scheme, and to relieve the workload of the prognostic analysis.

Moreover, although illustrative embodiments have been described herein, the scope includes any and all embodiments having equivalent elements, modifications, omissions, combinations (e.g., of aspects across various embodiments), adaptations or alterations based on the present disclosure. The elements in the claims are to be interpreted broadly based on the language employed in the claims and not limited to examples described in the specification or during the life of the application. Further, the steps of the disclosed methods may be modified in any manner, including by reordering steps or inserting or deleting steps. It is intended, therefore, that the description be regarded as examples only, with a true scope being indicated by the following claims and their full scope of equivalents.

The above description is intended to be illustrative and not restrictive. For example, the above-described examples (or one or more aspects thereof) may be used in combination with each other. Other embodiments may be utilized, for example, by one of ordinary skill in the art, upon reading the above description. Also, in the foregoing detailed description, various features may be combined together to simplify the present disclosure. This should not be interpreted as intending that an unclaimed disclosed feature is essential to any claim. Thus, the following claims are hereby incorporated into the detailed description as examples or embodiments, with each claim standing on its own as a separate embodiment, and it is contemplated that these embodiments may be combined with each other in various combinations or permutations. The scope of the invention should be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled.

Claims

1. A device for guiding stent implantation in a blood vessel, the device comprising:

an interface configured to receive a sequence of 2D images of a blood vessel;

a processor configured to:

reconstructing a first 3D model of the vessel based on the sequence of 2D images;

acquiring structural parameters and a position to be implanted of a stent to be implanted;

modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted;

predicting a second fractional flow reserve of the blood vessel using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model.

2. The apparatus according to claim 1, wherein modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted comprises:

and expanding the part of the first 3D model at the position to be implanted to conform to the structural parameters of the stent to be implanted and maintain the curvature of the part.

3. The apparatus according to claim 2, wherein the structural parameters of the stent to be implanted include a length and a deployed diameter of the stent to be implanted.

4. The apparatus of claim 1, wherein the 2D image sequence of the vessel comprises an angiographic image sequence of the vessel at least two different projection angles, the processor further configured to: determining a flow velocity distribution as a blood flow related parameter of the second 3D model based on the angiographic image sequence of at least two different projection angles of the blood vessel.

5. The apparatus of claim 4, wherein the processor is further configured to: predicting a first fractional flow reserve of the blood vessel using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the first 3D model of the blood vessel.

6. The apparatus of claim 1 or 4, wherein the structure-related parameters of each of the first 3D model and the second 3D model comprise at least one of a vessel diameter distribution, a healthy diameter distribution, a curvature distribution, and an optical path depth distribution of the 3D model along the centerline.

7. The apparatus of claim 4, wherein the processor is further configured to: regression of the peak distribution of diameters along the centerline of the first 3D model of the vessel to predict a healthy diameter distribution of the vessel along the centerline; determining a stenosis region based on a diameter trough distribution along a centerline and a healthy diameter distribution of the first 3D model of the blood vessel.

8. The apparatus of claim 7, wherein the processor is further configured to: determining a region of interest in the first 3D model based on the determined stenosis region; modifying the attention area of the first 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted, wherein the modified attention area 3D model is used as the second 3D model.

9. A device for guiding stent implantation in a blood vessel, the device comprising:

a first acquisition unit configured to acquire a 2D image sequence of a blood vessel;

a reconstruction unit configured to reconstruct a first 3D model of the vessel based on a sequence of 2D images;

a second acquisition unit configured to acquire structural parameters of a stent to be implanted and a position to be implanted;

a model modification unit configured to: modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the position to be implanted;

a prediction unit configured to: predicting a second fractional flow reserve of the blood vessel using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the second 3D model.

10. The apparatus of claim 9, wherein the model modification unit is further configured to: and expanding the part of the first 3D model at the position to be implanted to conform to the structural parameters of the stent to be implanted and maintain the curvature of the part.

11. The apparatus of claim 9, wherein the prediction unit is further configured to: predicting a first fractional flow reserve of the blood vessel using the trained learning model based on the structure-related parameters and the blood flow-related parameters of the first 3D model of the blood vessel.

12. The apparatus of claim 9, further comprising a stenosis region determination unit configured to: regression of the peak distribution of diameters along the centerline of the first 3D model of the vessel to predict a healthy diameter distribution of the vessel along the centerline; determining a stenosis region based on a diameter trough distribution along a centerline and a healthy diameter distribution of the first 3D model of the blood vessel.

13. A system for guiding stent implantation in a blood vessel, the system comprising:

the apparatus of any one of claims 1-8; and

a display configured to:

displaying the first 3D model for a user to perform a stenosis analysis of the vessel at least where a pre-operative analysis of the vessel is performed; and

displaying the second 3D model and a second fractional flow reserve of the vessel while simulating an implantation effect of a stent.

14. The system of claim 13, wherein the display is further configured to: in the case of performing a preoperative analysis of the blood vessel, a healthy diameter distribution of the stenosis region and its periphery in the first 3D model is also displayed.

15. The system of claim 13, wherein the display is further configured to display a user interaction unit configured to set a model of a stent to be implanted and to input a position to be implanted by a user.

16. The system of claim 15, wherein the user interaction unit further comprises an indication unit configured to: dragging by a user in a peripheral region of a stenosis region in the first 3D model to preset a location to be implanted.

17. A non-transitory computer readable medium having stored thereon instructions that, when executed by a processor, perform a method for guiding stent implantation in a vessel, the method comprising:

receiving a sequence of 2D images of a blood vessel;

18. The non-transitory computer-readable medium of claim 17, wherein modifying the first 3D model to obtain a second 3D model based on the acquired structural parameters of the stent to be implanted and the location to be implanted comprises:

19. The non-transitory computer-readable medium of claim 17, wherein the 2D image sequence of the vessel comprises an angiographic image sequence of the vessel at least two different projection angles, the processor further configured to: determining a flow velocity distribution as a blood flow related parameter of the second 3D model based on the angiographic image sequence of at least two different projection angles of the blood vessel.