Medical Implant Inspection Using Imaging

Through imaging technology and machine learning model analysis, the quality inspection of automated medical implant devices has been solved, and the problems of low manual inspection efficiency and inconsistent results in the prior art are achieved, achieving efficient and accurate fault determination and device removal processes.

CN112955733BActive Publication Date: 2024-11-01EDWARDS LIFESCIENCES CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201980072436.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2018-09-19
Filing Date
2019-09-17
Publication Date
2024-11-01
Estimated Expiration
2039-09-17

AI Technical Summary

Technical Problem

In the prior art, the quality inspection of medical implant devices relies on manual processing by humans, resulting in low efficiency, high cost and risk of inconsistent results, and manual inspection may damage the device.

Method used

Using imaging technology, the imaging data of the target device is generated through three-dimensional computed tomography and other means, the image features are analyzed using control circuits and machine learning models, the inspection results are automatically determined, and the target device is decided based on the fault prediction model.

Benefits of technology

Automatic quality inspection of medical implant device is realized, which improves efficiency and accuracy, reduces the risk of human intervention, reduces inspection costs, and can determine failures without damaging the device.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112955733B_ABST
    Figure CN112955733B_ABST
Patent Text Reader

Abstract

An apparatus for inspecting a target device includes a data communication interface and a control circuit configured to receive imaging data of the target device using the data communication interface, determine image features associated with the target device based at least in part on the imaging data, and determine an inspection result based at least in part on the image features.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Related Applications

[0002] This application claims the benefit of U.S. Patent Application No. 62 / 733,365, filed on September 19, 2018, and entitled “Medical Implant Inspection Using Imaging,” the disclosure of which is incorporated herein by reference in its entirety. Technical Field

[0003] The present disclosure relates generally to the field of implant device manufacturing and inspection and associated systems, devices, and methods. Background Art

[0004] Manufacturing prosthetic heart valves and other human prosthetic implant devices can involve quality inspection to identify defects in manufactured products and / or make product rejection determinations. The use of manual processing and inspection by humans for quality inspection and product rejection determinations can be random, expensive, and time consuming. Summary of the invention

[0005] In some embodiments, the present disclosure relates to an apparatus for inspecting a target device. The apparatus includes a data communication interface and a control circuit, the control circuit being configured to receive imaging data of the target device using the data communication interface, determine image features associated with the target device based at least in part on the imaging data, and determine an inspection result based at least in part on the image features.

[0006] The imaging data may include three-dimensional computed tomography imaging data. In some embodiments, the target device is a prosthetic heart valve device. The image features may include features related to one or more of the presence, size, color, and shape of contaminants associated with one or more parts of the target device. In some embodiments, the control circuit is further configured to determine the inspection results based at least in part on the fault prediction model. For example, the control circuit can be further configured to obtain a set of image feature data associated with a plurality of reference devices, obtain inspection decision data corresponding to the plurality of reference devices, the inspection decision data including fault decision data, determine a cross-correlation between the fault decision data and a set of imaging feature data, and determine a fault prediction model based on the cross-correlation.

[0007] In some embodiments, the inspection result is a failure determination, and the control circuit is further configured to send a command to a device remover device to remove the target device from the manufacturing process flow based on the failure determination.

[0008] In some embodiments, the present disclosure relates to a method for generating an inspection prediction model. The method includes obtaining imaging data for a plurality of reference devices, determining image features associated with the imaging data, obtaining inspection decision data associated with the plurality of reference devices, determining cross-correlations between the image features and the inspection decision data, and generating an inspection prediction model based at least in part on the cross-correlations.

[0009] The inspection decision data may indicate a human inspection result for each of the plurality of reference devices. In some embodiments, the inspection prediction model is a polynomial model having one or more variables associated with the image features. The inspection prediction model may be based at least in part on product specification data associated with the plurality of reference devices. The method may further include generating imaging data for the plurality of reference devices.

[0010] In some embodiments, the present disclosure relates to a method for determining an inspection decision for a target device. The method includes obtaining imaging data for the target device, determining image feature data of the imaging data, evaluating the image feature data using an inspection prediction model, and determining an inspection result associated with the target device based on the evaluation.

[0011] The imaging data may include computed tomography imaging data. In some embodiments, the target device is a prosthetic medical implant device. The image feature data may include one or more of contaminant presence data, size data, color data, and shape data. In some embodiments, determining the inspection result is based at least in part on a fault prediction model. For example, the method may further include obtaining a set of image feature data associated with a plurality of reference devices, obtaining inspection decision data corresponding to the plurality of reference devices, the inspection decision data including fault decision data, and generating a fault prediction model based on a cross-correlation between the fault decision data and a set of imaging feature data. In some embodiments, the method further includes removing the target device from the process flow based on the inspection result.

[0012] In some embodiments, the present disclosure relates to a device inspection system, the system comprising an imaging device and one or more non-volatile data storage devices, the one or more non-volatile data storage devices storing image feature data associated with multiple reference devices and inspection decision data associated with multiple reference devices. The device inspection system further includes a control circuit configured to access the image feature data and the inspection decision data, generate a fault prediction model based on a cross-correlation between features indicated by the image feature data and fault determinations indicated by the inspection decision data, receive imaging data for a target device from the imaging device, and determine an inspection result associated with the target device based at least in part on the fault prediction model and one or more features indicated by the imaging data for the target device. The system may further include a conveyor device configured to advance the target device through a scanning area of ​​the imaging device. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Various embodiments are depicted in the accompanying drawings for the purpose of illustration and should in no way be interpreted as limiting the scope of the present invention. In addition, various features of different disclosed embodiments may be combined to form additional embodiments, which are a part of the present disclosure. In all drawings, reference numerals may be reused to indicate correspondence between reference elements.

[0014] Figure 1 is a perspective view of a transcatheter heart valve according to one or more embodiments.

[0015] Figure 2 is a perspective view of a surgical heart valve according to one or more embodiments.

[0016] Figure 3 is a top view of an annuloplasty ring according to one or more embodiments.

[0017] Figure 4 is an exploded perspective view of a prosthetic heart valve according to one or more embodiments.

[0018] Figure 5 An imaging-based inspection system is shown in accordance with one or more embodiments.

[0019] Figure 6 is a flow chart illustrating a process for generating an inspection prediction model in accordance with one or more embodiments.

[0020] Figure 7 is a flow chart illustrating a process for inspecting a device according to one or more embodiments. DETAILED DESCRIPTION

[0021] The headings provided herein are for convenience only and do not necessarily affect the scope or meaning of the claimed invention.

[0022] Although certain preferred embodiments and examples are disclosed below, the subject matter of the invention extends beyond the specifically disclosed embodiments to other alternative embodiments and / or uses and modifications and equivalents thereof. Therefore, the scope of the claims that may arise therefrom is not limited by any specific embodiment described below. For example, in any method or process disclosed herein, the actions or operations of the method or process may be performed in any suitable order and are not necessarily limited to any specific disclosed order. Various operations may be described in sequence as multiple discrete operations in a manner that helps to understand certain embodiments; however, the order of description should not be interpreted as implying that these operations are sequentially related. In addition, the structures, systems and / or devices described herein may be implemented as integrated components or as separate components. For the purpose of comparing various embodiments, certain aspects and advantages of these embodiments are described. It is not necessary to achieve all of these aspects or advantages by any particular embodiment. Therefore, for example, various embodiments may be performed in a manner that achieves or optimizes one advantage or a group of advantages as taught herein, without having to achieve other aspects or advantages as taught or suggested herein.

[0023] Overview

[0024] Prosthetic heart valve implants, as well as many other types of prosthetic implant devices and non-implant devices, may include various welded, connected, sutured, positioned and / or attached components and / or portions. The manufacture and assembly of such devices may result in the appearance or introduction of defects or contaminants associated with one or more components of (one or more) devices, which in some cases may result in the device failing quality inspection. That is, certain defects or contaminants in a medical device may result in the device being unsuitable for use in its current state, or such defects or contaminants may indicate that additional corrective measures should be taken for a defective device. In some embodiments, manual human inspections of medical implant devices or other types of devices are performed on a one-to-one basis, where each individual product being inspected may need to be physically handled and inspected by an inspector. It should be explained that the description of device inspections in the present disclosure may be performed at least in part by an operator inspector or a device assembler or both. Therefore, the term "inspector" is used herein in its broad and general sense, and includes an operator and / or assembler or other inspector individual or entity.

[0025] The present disclosure relates to systems, devices and methods for inspecting medical implant devices and / or other types of equipment or devices and providing fault decisions related thereto. The term "failure" is used here and in its broad and / or general sense, and may refer to any status, state, situation or the like of a device (such as a medical device), wherein such status, state, situation or the like has a nature that makes the device currently considered unsatisfactory for production or progress in a manufacturing and / or evaluation process. For example, a fault decision according to the present disclosure may involve visual defects such as stains or other visual defects that may or may not substantially hinder or affect the function of the device under test. The fault decision may further involve assembly errors or any other type of error or situation of the device or its parts that may affect the suitability of the device for use and / or production. The terms "failure determination", "failure assessment", "failure decision", "failure determination", "negative pass / fail determination", "scrap determination", "scrap assessment" and the like are used herein to describe a negative inspection result or determination, which may result in the disposal of the device under test, the removal of the device from a process (e.g., a manufacturing process) flow, remedial actions to at least partially correct the defective situation, or other response actions. In addition, such terms may refer to a determination process for determining whether a target device has passed or failed a quality inspection. The terms "pass / fail determination" and "pass / fail decision" may refer to a determination as to whether a target device meets certain quality inspection criteria. A rejection or failure determination may result from the identification of a defective condition in the subject device, such as the presence of (one or more) contaminants, or a defective connection or configuration of one or more of its components. The terms "subject device," "target device," and "product" are used herein in their broad and ordinary sense, and may refer to a device that is in quality inspection or otherwise subjected to quality inspection or evaluation, such as a heart valve device or other prosthetic implant device.

[0026] Devices inspected according to the principles disclosed herein may include any type of medical implant device, such as, for example, a prosthetic heart valve device. For example, a heart valve device that may be inspected according to the present disclosure may include a transcatheter heart valve (THV), a surgical heart valve (e.g., a replacement aortic valve or mitral valve device), an annuloplasty ring (e.g., a mitral valve or tricuspid valve annuloplasty device), or other devices. During the manufacturing process, such a device may exhibit a fault or defect that does not allow the device to pass a quality inspection based on certain relevant inspection parameters or standards. Certain exemplary medical implant devices are described below as examples of devices that may undergo quality inspection and / or fault determination according to the present disclosure. For convenience and to provide background for certain aspects of the present disclosure, such a device is described as an example device, and it should be understood that the principles disclosed herein may be applicable to other types of implant devices or other devices or devices that can be evaluated or inspected to identify defects or characteristics associated therewith.

[0027] Example Examination of Prosthetic Implant Devices

[0028] Devices that may be subjects of quality inspection include prosthetic heart valve implant devices. Typically, the human heart includes four valves for assisting blood circulation therein, including the tricuspid valve that separates the right atrium from the right ventricle; the pulmonary valve that separates the right ventricle from the pulmonary artery; the mitral valve that separates the left atrium from the left ventricle; and the aortic valve that separates the left ventricle from the aorta. Heart valves can typically include a relatively dense fibrous ring referred to herein as an annulus, and a plurality of leaflets attached to the annulus. Typically, the size of the leaflets can be such that when the heart contracts, the resulting elevated blood pressure generated in the corresponding heart chamber forces the leaflets to at least partially open to allow outflow from the heart chamber. As the heart chamber pressure decreases, the pressure in the chamber or blood vessel may then dominate and press backward on the leaflets. As a result, the leaflets are juxtaposed to each other, thereby closing the flow channel. The tricuspid valve includes three leaflets, and can typically be closed during ventricular contraction (e.g., systole) and opened during ventricular expansion (e.g., diastole). The pulmonary valve includes three leaflets and is normally configured to open during systole so that blood can be pumped toward the lungs, and to close during diastole to prevent blood from leaking back into the heart from the pulmonary artery. The mitral valve includes two leaflets and is normally configured to open during diastole so that blood in the left atrium can flow into the left ventricle, and to close during diastole to prevent blood from leaking back into the left atrium. The aortic valve includes three leaflets and is normally configured to open during systole to allow blood leaving the left ventricle to enter the aorta, and to close during diastole to prevent blood from leaking back into the left ventricle.

[0029] As described above, inspection and / or fault determination according to embodiments of the present disclosure may be implemented in conjunction with a transcatheter heart valve device. Figure 1 1 is a perspective view of a transcatheter heart valve 110 according to one or more embodiments. The features of the valve 110 described herein can be applied to other valves, including other valves described elsewhere herein. The valve 110 can be, for example, a self-expanding heart valve, a balloon-expandable heart valve, and / or a mechanically expandable heart valve. The valve 110 in the illustrated embodiment can generally include a frame or stent 112, a leaflet structure 193 supported by the frame 112, and a sealing member or skirt 116 fixed to at least a portion of the frame 112 and / or a portion of the leaflet structure 193. In certain embodiments, the valve 110 can be configured to be implanted in the annulus of a natural heart valve of a human being, such as an aortic valve. However, the valve 110 can be additionally or alternatively adapted to be implanted in other natural valves of the heart, or in various other vascular systems, conduits, or orifices of the body, or in a graft, docking stent, docking station, ring, or the like implanted in the body. According to the orientation shown, the lower end 180 of the valve 110 can represent the inflow end, while according to the orientation shown, the upper end 182 of the valve 110 can represent the outflow end.

[0030] The valve 110 and the frame 112 can be configured to be radially contractible to a collapsed or curled state / configuration for introduction into the body using a delivery catheter, and can be further configured to be radially expandable to an expanded state / configuration for implanting the valve into a desired position in the body (e.g., a native aortic valve). In some embodiments, the frame 112 can include plastics, polymers, shape memory materials, or mechanically expandable materials that allow the valve 110 to be crimped to a smaller profile for delivery and expansion of the valve. In some embodiments, an expansion device (such as a balloon of a balloon catheter) or a tool for mechanical expansion can be used to expand the valve directly or indirectly. In some embodiments, the valve 110 can be a self-expanding valve, wherein the frame is made of a self-expanding material such as a shape memory material or a metal (e.g., nitinol). The self-expanding valve can be crimped to a smaller profile and maintained in a crimped state with a restraining device (such as a sheath covering the valve). When the valve is positioned at or near the target site, the restraining device can be removed or retracted to allow the valve to self-expand to its expanded functional size, or to self-expand to a deployment configuration.

[0031] The sealing portion or skirt 116 may include a single piece or multiple pieces or materials (e.g., cloth, polymer, etc.) having opposing ends that are secured to one another to form a Figure 1In some embodiments, the upper edge of the sealing portion or skirt 116 may have an undulating shape, which generally follows the shape of the pillars of the frame 112. In this way, the upper edge portion of the sealing portion or skirt 116 can be tightly fixed to the corresponding pillar with sutures 156. The sealing portion or skirt 116 can be placed on the outside of the frame 112, or placed on the inside of the frame 112 (as shown), and the upper edge portion of the sealing portion or skirt 116 can be surrounded around the upper surface of the frame pillar and fixed in place with sutures. Sutures 156 can be used to provide a durable attachment of the sealing portion or skirt 116 to the frame 112.

[0032] In some embodiments, the leaflet structure 193 may include three leaflets (eg, Figure 1 ), which can be arranged to fold in a tricuspid arrangement. Although a three-leaflet embodiment is shown, it should be understood that the valve implant inspected and / or evaluated according to the embodiments disclosed herein can have any number of leaflets, such as, for example, two or four. The leaflets 193 can be formed from separate flaps of material or tissue such as xenograft tissue (e.g., bovine pericardium), or all three leaflets can be derived from a single xenograft valve (e.g., a porcine valve). The lower edge of the leaflet structure 193 can have a variety of shapes. In certain embodiments, the lower edge of the leaflet structure 193 can have a wavy, curved, and / or fan-shaped shape that can be sutured to the frame 112. The leaflets 193 can be secured to each other at their adjacent sides to form a commissure 184 of the leaflet structure, where the edges of the leaflets meet together. The leaflet structure 193 can be secured to the frame 112 using any suitable technique and / or mechanism. For example, the commissures 184 of the leaflet structure can be aligned with and secured to the support posts 118, such as by using sutures, adhesives, clamps, crimps, and / or other attachment means. In one embodiment, the point of attachment of the leaflet 193 to the post 118 can be reinforced, for example, with a bar comprising a relatively rigid material, such as stainless steel. According to aspects of the present disclosure, any of the various attachment points of the valve 110 can be inspected for defects in a quality inspection.

[0033] In some embodiments, the lower edge portion (not shown) of the skirt 116 is configured to be folded upward on the frame 112 from below the bottom of the frame and sutured thereto. The folded portion of the skirt can be sutured to the corresponding struts of the frame 116. Folding and securing the skirt 116 to the frame from the outside can allow excess material to fill the gap between the frame 116 and the surrounding native annulus to assist in forming a fluid-tight seal between the valve and the native annulus. This configuration can at least partially prevent paravalvular leakage after implantation of the valve.

[0034] The quality inspection of the valve device 110 may involve the assessment or evaluation of the various attachment points between the leaflets and the posts 118 and / or the stent / frame 112. In addition, the inspection may involve the evaluation of the various sutures performed to attach the leaflets 193, the frame 112, and / or other components of the valve 110 to each other. In addition, the inspection of the valve 110 may involve identifying the presence of foreign matter or other contaminants in various areas of the valve 110, such as in the folds or wedges thereof. The quality of the leaflet tissue or material 193 may also be evaluated as part of the quality inspection, such as to identify areas of delamination in the leaflets.

[0035] Figure 2 is a perspective view of a prosthetic human heart valve 210 according to one or more embodiments. The heart valve 210 can be considered an embodiment of a surgical prosthetic heart valve. The heart valve 210 can include a peripheral sealing ring structure 291 that is configured to provide support for nesting the heart valve 210 in a heart valve cavity and / or resting it on or attaching it to a native valve annulus or other structure of the heart. The valve 210 can further include a frame member 292, such as a metal and / or plastic frame, which can provide support for a plurality of flexible leaflets 293 and can define three (or two) upstanding commissure posts 294, wherein the leaflets 293 can be supported between the commissure posts 294. In one embodiment, as Figure 2 As shown in , a sealing ring 291 can be attached around the periphery of a frame member 292 at the inflow end 280 of the valve 210 , with commissure posts 294 protruding in the outflow direction.

[0036] The leaflets 293 can be formed from separate flaps of material or tissue, such as, for example, xenograft tissue (e.g., bovine pericardium), or all three leaflets can be derived from a single xenograft valve (e.g., a porcine valve). The leaflets 293 can be secured and supported by commissure posts 294 and arcuate cusps 295 along the frame members between the commissure posts. Figure 1 Similar to the prosthetic heart valve 110 of , the prosthetic heart valve 210 has a valve orifice (not shown) at the inflow end 280 of the valve, wherein blood can flow through the valve orifice in one direction, but the valve leaflets 293 substantially prevent blood flow in the opposite direction.

[0037] As shown, one or more parts of the heart valve 210 can be covered with cloth. Covering valve 210 with cloth can involve bringing together cloth parts to form joints or seams, which can be completed by suturing or other known cloth edge joining techniques. As shown, cloth can be wrapped around frame 292. The cloth cover can be formed by any biocompatible fabric (such as, for example, polyethylene terephthalate). Cloth to frame and / or to other parts when sutured improperly can be used as the basis for scrapping heart valves during quality inspection. In some embodiments or situations, after cloth has been sutured to frame and / or other parts, it is impossible or impractical to manually check some defects and / or areas or parts of the device.

[0038] Figure 3 An embodiment of an annuloplasty ring 310 is shown according to one or more embodiments. The annuloplasty ring 310 can be used for repair of a native heart valve. For example, the annuloplasty ring 310 can provide a surgical device that can be used for repair of a leaking valve, such as, for example, repair of a mitral valve or a tricuspid valve. As described above, due to various factors, the leaflets that normally seal the native valve and allow blood flow to reverse may not coapt properly in certain circumstances. Surgical repair of such a valve can involve implantation of an annuloplasty ring to reshape the native valve annulus, wherein the annuloplasty ring when implanted can be configured to draw the leaflets together to promote coaptation and help reestablish proper native valve function.

[0039] The annuloplasty ring 310 can have an elliptical or somewhat D-shaped configuration with a relatively straight anterior segment 334 opposed to a relatively curved posterior segment 336. The annuloplasty ring 310 further includes a cloth covering 316 that can be sewn onto or otherwise disposed on an inner frame (not shown) of the annuloplasty ring. The annuloplasty ring 310 can further include one or more seams, sutures, or other features 338 associated with the cloth covering 316, which in some embodiments can function as commissures and / or triangular markers. The shaped posterior segment 336 can be shaped to generally follow the shape of the mitral valve annulus. For example, the posterior segment 336 can be at least partially curved downward relative to the anterior segment 334 (e.g., into the page relative to the orientation shown). Alternatively, the annuloplasty ring 310 can have a substantially planar configuration. Some tricuspid valve annuli include multiple tortuosity. Generally, inspection of three-dimensionally shaped elements (such as components of prosthetic valves and annuloplasty rings) can be relatively complex. For example, in some cases, two-dimensional imaging techniques may not provide sufficient information for inspecting three-dimensional shaped components.

[0040] like Figure 3The annuloplasty ring shown in may be subject to quality inspection at various stages of manufacture. Quality inspection may be implemented to identify defects in shape or incorrect or unsatisfactory stitch work, among other aspects. In addition to the prosthetic heart valves and annuloplasty rings described above, quality inspection may also be performed on the following components: delivery catheters and / or systems for delivering medical implant devices such as heart valve devices to a target site, and subassemblies and / or accessories for such devices (including sizers, retainers, crimpers, etc.).

[0041] With further reference to prosthetic heart valve devices, such devices may include various components, the assembly of which may result in certain defects that justify an inspection failure determination. For example, Figure 4 An exploded perspective view of a prosthetic heart valve 410 is shown, which is similar in some respects to Figure 2 of surgical heart valves 210. Despite Figure 4 A surgical valve is shown in FIG. 1 , but it should be understood that transcatheter heart valves, annuloplasty rings, and other medical implant devices may similarly include the assembly of multiple separate components, the attachment / assembly of which may lead to the introduction of various defects, and thus may be difficult to obtain. Figure 4 The relevant following description is applicable to the assembly of other types of prosthetic implant devices.

[0042] The valve assembly 410 includes a frame member 489, which can include a metal or other at least partially rigid frame structure, such as a wire structure 489. In some embodiments, the frame member 489 can be at least partially covered with a fabric 492 or other material. The frame 489 and fabric 492 assembly can define a relatively narrow, upwardly protruding arcuate commissure region 494 between downwardly protruding arcuate cusps 495. The valve assembly 410 can further include a support member 497 designed to fit over the sealing ring 491. In some embodiments, the support member 497 includes a plastic strip 496 (e.g., polyester, polyethylene terephthalate (PET), or biaxially oriented PET, such as Mylar BoPET, DuPont Teijin Films, Wilmington, Delaware, USA), which can also be covered with cloth.

[0043] The frame 489 and the support member 497 can be at least partially fixed or attached to the leaflet assembly 493, for example, by sandwiching the leaflet assembly between the frame and the support member and sewing through the fabric covering of the frame, the leaflet assembly and the fabric covering of the support member. Attachment to the frame 489 can damage the leaflet 493, such as tearing it, or improper attachment of the leaflet 493 can lead to a failed quality inspection. The leaflet 493 can be made of any suitable material that provides flexibility and structure for blocking fluid flow through the valve 410, which, as described above, is, for example, at least partially a tissue of biological origin. When assembled, the leaflet 493 extends inwardly from the surrounding frame 492 into the flow orifice defined thereby. In some embodiments, there are three bioprosthetic leaflets that are bent toward the outflow direction and joined in the middle of the valve orifice to promote unidirectional flow through the valve. Improper engagement of the leaflet 493 can be used as a basis for failure of the quality inspection.

[0044] The valve assembly 410 further includes a support component 497 designed to fit over the sealing ring 491. In some embodiments, the support component 497 includes a plastic strip 496 (e.g., polyester, polyethylene terephthalate (PET)) (e.g., PET, Invista, Wilmington, Del.) or biaxially oriented PET (e.g. The support member 497 may be provided with a plastic strap 496 (BoPET) wherein the leaflet 493 may be sewn to or otherwise attached to the plastic strap. Improper suturing of the leaflet 493 to the support member 497 may be identified during a quality inspection and may be used as a basis for a failure determination. Thus, visibility and identification of stitches coupled to the support member 497 and / or the leaflet 493 may aid in quality assessment of the device assembly 410. The support member 497 may further include a rigid reinforcement strap 499, which may include, for example, metal or other rigid material. The plastic strap 496 may include a commissure support portion 498, which may be at least partially assembled within an upwardly protruding commissure region 494 of the frame member 492. Proper coupling of the support member 497 to the frame structure 492 may be confirmed during a quality inspection, and improper coupling or damage associated therewith may be used as a basis for a failure determination.

[0045] The sealing ring 491 of the heart valve assembly 410 can be configured to at least partially stabilize the valve annulus and / or support functional changes that occur during the cardiac cycle, such as by maintaining coaptation and valve integrity to prevent reverse flow while allowing good hemodynamics during forward flow. The sealing ring 491 can include an inner at least partially rigid substrate (e.g., a metal such as stainless steel or titanium, or a flexible material such as silicone rubber, fabric (e.g., Fiberweb, Wilmington, Delaware, Nonwoven polyester fabric)) or rope (e.g. PET cord) and may be at least partially covered with a biocompatible fabric or cloth (not shown) to allow the ring to be sutured to the heart tissue. The sealing ring 491 can be rigid / hard or flexible and can have a variety of shapes, including circular, D-shaped, kidney-shaped, or C-shaped. In some embodiments, the heart valve 410 can include various visual markers (not shown; for example, radiopaque markers and / or visual markers) to aid in quality inspection.

[0046] When coupled or stitched Figure 4 When the various components of the valve assembly 410 shown in the figure are inspected, certain particles, debris or other contaminants (e.g., oil) may adhere to one or more components, or may be trapped between coupled components or regions. The presence of such particles or other contaminants may negatively affect the performance of the heart valve and / or may cause harm to the patient recipient of the valve implant. Therefore, the quality inspection process and mechanism can advantageously provide visibility or indication for identifying such particles, debris or other contaminants, and the presence of such contaminants can be used as a basis for fault determination in quality inspection. In some cases, the visibility of particles, debris or (one or more) other contaminants may require the use of one or more image enhancement tools, such as microscopes, cameras or the like. In addition, as described above, manual inspection techniques may cause damage to (one or more) components of the device being inspected. For example, manual manipulation of the heart valve leaflets can cause damage to the device during inspection and / or increase the risk of failure / scrap. Although certain embodiments are described herein in the context of a heart valve device, it should be understood that such descriptions may be applied to other types of prosthetic implant devices and certain non-biological applications.

[0047] Equipment inspection / assessment

[0048] As described above, the quality inspection of medical implant devices and other devices and equipment can be performed in various ways. For example, the inspection of the device can be implemented by manual inspection performed by one or more human inspectors of the device at one or more stages of manufacturing or assembly. Regarding the manual inspection process, due to the relatively high dependence on the inspector's judgment, such a process may involve the risk of inconsistent results of different inspectors. That is, although different inspectors can inspect the device according to certain parameters or standards, in execution, the subjective judgment of such inspectors in applying such parameters or standards can lead to inconsistent results determined by such operators. In addition, manual visual inspections performed by human inspectors may be relatively time-consuming and / or expensive, and may increase the risk of damaging the device under test, thereby increasing the risk of inspection failures. In addition, in the case where a fault determination is performed by an engineer, trainer and / or supervisor's review committee to determine whether the identified product is in a fault state and should be disposed of or subjected to (one or more) corrective processes, such additional reviews may also cause undesirable or excessive costs in some cases. In addition, additional reviews of the device may increase the risk of damage and / or failure of the device.

[0049] Manual human inspection of a medical implant device may involve the use of one or more microscopes or other visual aids or enhancements to identify relatively minor defects or conditions that may be used as a basis for product rejection. For example, for a prosthetic heart valve, a visual aid or enhancement may be used to help identify: tissue damage to the valve leaflets or other components; particles that have become stuck, deposited, or otherwise undesirably disposed on or elsewhere on the heart valve; or sutures or threads that have been improperly sutured or omitted. Additionally, for a heart valve device, its frame components may not be properly welded or closed, which may similarly represent a condition that may be used as a basis for rejecting the heart valve at one or more stages of the manufacturing process. As described herein, any of a variety of defect conditions may result in the product under test being determined to be rejected.

[0050] For a heart valve device that incorporates biological tissue (such as for valve leaflets or the like), it may be necessary or desirable to maintain such a component of the heart valve device in a relatively moist state to prevent its tissue from becoming undesirably dehydrated or drying out, which may result in rupture, breakage or other damage to the biological tissue. Therefore, in some embodiments, the heart valve being inspected is maintained in a wide-mouthed bottle or other container, which has a fluid or other moisturizing element contained therein, wherein a human inspector may be required to physically remove the valve device from the container for inspection. The inspector can then use a microscope or other visual enhancement tool to inspect the withdrawn device. However, during the time period in which the valve is removed from the moisturizing container, the inspector may have the need or desire to soak or spray one or more components of the valve in a moisturizing fluid or gas in order to maintain the desired or necessary moisture in the biological tissue of the heart valve device. In some embodiments, the re-soaking of the valve must be performed every minute or during other time periods in which the device is exposed to the outside of the container. Once the inspection is complete, the prosthetic valve device can be returned to the container, which can be sealed to retain moisture therein and / or otherwise protect the device from contamination or damage, and the inspector can then fill out certain paperwork or provide other input indicating that the inspection of the device has been completed. If a defect or error in the inspected device is observed, the inspector can identify such defect / error in some manner, at which point the inspected product can be ultimately disposed of.

[0051] Inspection of medical implant devices (such as prosthetic heart valve devices) can be performed at various stages in the manufacturing / assembly process. For example, the device can be inspected when a biological tissue component (e.g., a leaflet) is mounted to a frame or other component of the implant device. An inspection of tissue installation can be performed to confirm that there are no defects relative to tissue installation. After tissue installation, a commissure cover and / or one or more other components may be further connected to or coupled to the device, which typically obscures the visibility of certain features or aspects of the device. Therefore, inspection of such obscured features may typically need to be performed before the attachment of the cover. Inspection of the device can be further performed after final assembly, after which a retainer device or the like can be connected to the device, wherein certain aspects of the device may no longer be visible or available for inspection after the retainer device is connected to the device. Additional inspections can be performed after the connection of the retainer device. In addition to the aforementioned inspections, certain other intermediate inspections can be performed in some embodiments.

[0052] Manual visual inspection of the device by a human inspector may further incorporate the use of certain probing or manipulating instruments or tools that may be used to manipulate tissue or other components (e.g., leaflets of a prosthetic heart valve) for inspection thereof. For example, for a heart valve device, visual inspection of an edge or interface between biological leaflet tissue and a sewing ring may require the use of a tool to separate the tissue from the ring for inspection of its coupling and / or associated components. Such tools may typically be rigid and have certain features for prying or prying between device components, which may be used to open up spacing between tissue and / or other components to some extent to provide visibility of certain aspects and features of the device. Additionally, caliper tools or other measuring instrument tools may be used to assist in determining certain parameter measurements associated with the device during inspection.

[0053] A heart valve device can be inspected to determine if delamination is present relative to the biological leaflet tissue of the device. One or more inspection tools can be used to confirm or demonstrate the presence of delamination in the tissue. Generally, delamination can represent a condition in which one or more layers of biological tissue making up the valve leaflets or other components can become at least partially loose and separate from one or more other layers of biological tissue. The occurrence of delamination can result in tissue weakening or damage, and thus the presence of delamination can represent a parameter of inspection that can be used as a basis for fault determination. As described above, the use of rigid tools in inspecting implanted devices can result in damage to the tissue or other components of the device in certain circumstances. For example, due to operator error and / or sensitivity of device components, sharp or rigid features of (one or more) tools may tear or press into one or more portions or components of the device.

[0054] In some embodiments, manual inspection of medical implant devices can be assisted by the use of one or more robotic or automated devices and / or cameras or imaging devices. Such devices can at least partially relieve the burden on human inspectors when inspecting devices. However, manipulation of devices by robots or other automated tools may pose a risk of damage to sensitive components of the device being inspected by such devices.

[0055] As described above, in some cases, disadvantages associated with time, cost, and / or damage to inspection equipment associated with manual human inspection and / or robotic / automated device assisted inspection may make such inspection processes and systems undesirable. For example, in addition to the relatively excessive time that may be associated with certain manual human inspections, manipulation of biological tissue of an implanted device to inspect a wedge or features thereof for delamination or other conditions may result in an increased risk of damage to the valve device through manipulation of tissue or other components.

[0056] Inspection / evaluation using imaging devices

[0057] Certain embodiments disclosed herein provide improved inspection of prosthetic heart valves and other medical devices or equipment, wherein implementation of such embodiments may involve a reduction or elimination of the need for manual inspection by a human inspector. In some implementations, embodiments of the present disclosure utilize imaging devices and / or data, such as three-dimensional (3D) computed tomography (CT), magnetic resonance imaging (MRI), molecular imaging, and / or other imaging devices and / or data, to inspect medical implant devices, such as prosthetic heart valves and / or annuloplasty rings, wherein such imaging data may be processed to generate a pass / fail determination, or some other determination or decision. In some embodiments, the pass / fail determination may be based at least in part on an analysis of data associated with a large number of previously inspected reference devices or products and associated imaging data, such reference devices being similar in one or more aspects to the target device under inspection, wherein correlations between features or characteristics of such imaging data of a plurality of previously inspected reference devices and failure determinations are determined based on the results of the inspection of such reference devices. In some embodiments, such previous inspections may have been performed at least in part by one or more currently trained inspectors. The determination of the correlation between imaging data features and inspection results may be part of a machine learning process or mechanism, as described in detail below. For the purpose of performing imaging-based inspection of a device according to aspects of the present disclosure, a machine learning process can be implemented or utilized by an inspection decision engine. The machine learning process can be enhanced by the input and / or use of inspection results generated by sampling and / or manual inspection. Such sampling and / or manual inspection input can provide a learning element for the machine learning process.

[0058] Figure 5 An imaging-based inspection system 500 is shown in accordance with one or more embodiments. The system includes an imaging device 520 for generating imaging data associated with a target device 530, such as a prosthetic heart valve device, as described in detail above. Although a single imaging device 520 is shown, it should be understood that any number of imaging devices and / or any combination of imaging device technologies may be implemented in the system 500. Thus, the imaging device 520 should be understood to represent any number of imaging devices and / or types of imaging device technologies. The target device 530 (e.g., a prosthetic heart valve) may be manually placed in the imaging device 520 for imaging thereof, or may be fed into the imaging device 520 by a conveyor or mechanism 521. Imaging data 522 generated by the imaging device 520 and associated with the target device 530 is provided to certain control circuits 510 of the system 500, which may be configured to perform certain inspection-related functions as described herein.

[0059] The control circuit 510 may include any number of hardware and / or software or firmware components. For example, the control circuit 510 may include one or more application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), circuits, digital logic circuits, analog circuits, combinations of discrete circuits, gates, or any other type of hardware or combination thereof. Additionally or alternatively, the control circuit 510 may include one or more processors or processor circuits, and / or instructions executable by one or more processors or processor circuits to implement one or more of the features of the control circuit 510. The control circuit 510 may take the form of a processing circuit, a microprocessor or processor and / or a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by a processing circuit (e.g., a microprocessor), and / or one or more logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller, or an embedded microcontroller.

[0060] The control circuit 510 is a component or module configured to generate an inspection decision 527 for an imaged target device (e.g., a prosthetic implant device). The control circuit 510 includes a machine learning circuit 517 configured to generate an inspection prediction model 519 for evaluating / inspecting imaging data (e.g., imaging data 522). The control circuit 510 further includes an imaging data evaluation circuit 515 configured to evaluate and / or parse the imaging data to identify certain features and / or characteristics suitable for fault determination and can be reflected in the inspection prediction model 519.

[0061] The control circuit 510 further includes an inspection decision engine circuit 512 configured to make an inspection / failure determination based on imaging data 522 and / or image features 523, which may be provided as output from the imaging data evaluation circuit 515. Figure 5 The diagram of FIG. 5 shows the decision engine circuit 512, the machine learning circuit 517, and the imaging data evaluation circuit 515 as separate functional components, but it should be understood that such components may be part of a single computing device, or may be combined in some manner in some embodiments, or their functions may be divided into Figure 5510. In addition, although the decision engine circuit 512, the machine learning circuit 517, and the imaging data evaluation circuit 515 are shown as part of the control circuit 510, it should be understood that such functional units / circuits may be distributed in any manner, and one or more components of the control circuit 510 may be part of a separate and / or remote system relative to one or more other components of the control circuit 510. As used herein, one or any combination of the control circuit 510, the decision engine circuit 512, the machine learning circuit 517, the imaging data evaluation circuit 515, the data storage device 502, and / or the imaging device 520 may be referred to as a "control circuit."

[0062] As described above, in some embodiments, the present disclosure provides for the use of scanning / imaging instrumentation to generate imaging data for a target implant device at one or more stages of the manufacturing process. Figure 5 As shown in , the imaging device 520 can be configured to implement electromagnetic or other imaging techniques to generate imaging data 522 for a target device 530 (e.g., a medical implant device). In some embodiments, the imaging device 520 works with a delivery system 521 (such as a transport or the like) that can be configured to feed the medical implant device 530 through a scanning area or channel of the imaging device 520. Alternatively, an operator can manually place the target device 530 in the scanning area of ​​the imaging device 520.

[0063] The imaging device 520 can incorporate any suitable or desired imaging technology. In addition, the imaging device 520 can advantageously provide sufficient resolution and / or detail to identify particles / contaminants between certain components of the target device 530. For example, with respect to embodiments related to inspection of a prosthetic heart valve device, the imaging device 520 can advantageously provide sufficient resolution to identify particles or other contaminants disposed between components of the heart valve device (such as between tissue portions (e.g., valve leaflets) and suture / sealing ring structures or other components). In addition, the imaging device 520 can be further configured to provide imaging data that can indicate the positioning of sutures performed in the target device, which imaging data can be used to determine whether such sutures are in the correct position or whether such sutures are incorrectly sutured.

[0064] The imaging device 520 may be configured to implement one or more of a variety of imaging techniques known to those skilled in the art that provide direct or indirect imaging of a structure. For example, in some embodiments, the imaging device 520 is configured to implement a two-dimensional or three-dimensional computed tomography (CT) or computed axial tomography (CAT) scan, which typically involves the use of x-rays from multiple different directions. The CT scan may involve the use of certain control circuits to perform numerical integration calculations on the measured x-ray sequence to estimate how much of the x-ray beam is absorbed in certain volumes of the target device being scanned. The imaging data generated in a CT-type embodiment of the imaging device 520 may include one or more cross-sectional images of the target device.

[0065] In some embodiments, imaging device 520 is configured to implement diffuse optical imaging (DOI) or diffuse optical tomography (DOT), which may generally involve generating an image of a target device using near infrared light. In such embodiments, imaging device 520 may be configured to measure optical absorption in an area of ​​the target device. In some embodiments, imaging device 520 is configured to implement magnetic resonance imaging (MRI), which may generally involve generating a high quality two-dimensional or three-dimensional image of a target device using magnetic fields and radio waves.

[0066] In some embodiments, the imaging device 520 is configured to implement millimeter wave imaging, which may generally involve the use of millimeter waves transmitted simultaneously from multiple antennas as the multiple antennas rotate around a target device. Wave energy reflected back from components of the target device on the body can be used to construct a three-dimensional image. Alternatively or in addition, the imaging device 520 may be configured to implement backscattered X-ray technology, which may involve transmitting X-ray radiation through the target device and examining the radiation reflected back from the target device. Other types of active and / or passive imaging techniques may also be implemented by the imaging device 520. In some embodiments, the imaging device 520 utilizes ultrasound technology, which involves the use of sound waves generated at a relatively high frequency (e.g., between approximately 1 MHz and 10 MHz) and measuring the reflected sound waves to produce imaging data.

[0067] As described above, the imaging device 520 can be implemented or operated in conjunction with a conveyor device 521 (e.g., a tunnel) or the like to feed the implant device through the conveyor device. The imaging data 522 generated by the medical implant device 530 can be interpreted by the control circuit 510 substantially automatically and / or immediately, or can be evaluated by a trained human inspector. The imaging device 520 can be advantageously configured to provide (one or more) cross-sectional images or other data images of the target device 530 without rotating or otherwise repositioning the device 530 during scanning. For example, in some embodiments, the imaging device 520 includes an imaging component that rotates or otherwise moves around the target device to rotate or otherwise move. Using the imaging device 520 as an alternative to manual inspection by an operator can allow the implant device 530 to be inspected without using a device manipulation tool or instrument as described above, which may cause damage to the implant device in some cases. In some examples, the target device can be rotated and / or translated within the imaging field (e.g., under X-ray irradiation or within a millimeter antenna array) (e.g., on a conveyor belt or by a robotic arm).

[0068] Although generally described herein in the context of scanning of a single medical implant device, it should be understood that in some embodiments, the imaging device 520 can be used to simultaneously scan multiple implant devices in batches, wherein the control circuit 510 can be configured to parse or decipher imaging data from a batch of scanned devices to generate imaging information and / or features associated with each of the corresponding scanned devices. In some embodiments, the system 500 includes a device remover mechanism 560 that can be designed to remove scrapped products or devices from a process flow that has been determined by the inspection decision engine 512 of the control circuit 510 to meet certain failure criteria based on the inspection prediction model 519. For example, the device remover component 560 can include one or more robotic arms or components configured to pull or push the scanned medical device 530 from a conveyor or other component of the system 500.

[0069] In some embodiments, when the control circuit 510 makes a failure determination with respect to a scanned target device, processes according to the present disclosure may involve a follow-up inspection by a trained human inspector to confirm the scrapping of the scanned device prior to disposal.

[0070] As described above, the control circuit 510 may include an imaging data evaluation circuit 515 that may be configured to analyze the imaging data 522 to identify certain features or characteristics thereof. For example, the imaging data evaluation circuit 515 may be configured to generate image feature data 523 that indicates the presence of particles or other contaminants, color or shading / shading in one or more components or regions, shape information, and / or size information relative to one or more components of the target device 530. In some embodiments, weights may be derived from the imaging data characteristics, or alternatively, measurements may be made using a scale or other weight measurement mechanism of the imaging device 520. In some embodiments, size information may be obtained from the imaging data by locating or identifying one or more markers or other identifiers of the target device.

[0071] As described above, the control circuit 510 may include a machine learning circuit 517, which may be configured to generate an inspection prediction model 519 based on an analysis of certain reference device data 529. For example, the reference device data 529 may include imaging data 503, image feature data 509, and / or decision data 504 associated with a set of previously inspected and imaged reference devices (e.g., heart valve implant devices). The imaging data 503, image feature data 509, and decision data 504 may be maintained in one or more data storage devices referred to herein as data storage devices 502. The imaging data 503 may include image data of multiple previously scanned devices (not shown). The imaging data 503 may include actual images generated by an imaging device (e.g., imaging device 520), and the image feature data 509 may include certain identified image parameters or characteristics associated with the actual imaging data 503 (e.g., data indicating the presence and / or location, color, weight, size, etc. of contaminants). Although Figure 5 Both imaging data 503 and image feature data 509 are shown, but system 500 may include only one of imaging data 503 and image feature data 509, or any combination of portions thereof. Decision data 504 may include data indicating actual inspection results resulting from human inspection of a reference device as associated with a corresponding reference device. Data storage device 502 may further store product data 505, which may include identifiers or other data associated with the type(s) of reference devices.

[0072] The initial or early stages of the machine learning process for generating the inspection prediction model 519 may involve analyzing the imaging data 503, image feature data 509, and / or decision data 504 associated with the reference devices. The machine learning circuit 517 (and / or one or more other components of the control circuit 510) may determine a correlation with certain characteristics or features of the imaging data associated therewith (such as device model, size, dimensional information, and / or other parameters) for devices that ultimately failed to pass the manual / trained reference batch. Such features may be provided in the image feature data 509. The machine learning circuit 517 may generate a prediction model 519 for predicting a pass / fail decision or determination based on the evaluated imaging data 503, image feature data 509, and / or decision data 504 of the reference device data 529. In some embodiments, the generated inspection prediction model 519 may be validated against an additional batch of reference devices, which are then manually inspected to confirm the accuracy of the prediction model.

[0073] Although the machine learning circuit 517 is shown near the decision engine 512 and the imaging data evaluation circuit 515, it should be understood that such components may be distributed among any number of devices at any number of local or remote physical locations. In addition, the functions of each component of the control circuit 510 may be performed at different times by their corresponding functional components or by any other component of the system 500. For example, the machine learning circuit 517 may temporarily generate the inspection prediction model 519 before the decision engine circuit generates a fault determination regarding the target device 530, or even significantly before the decision engine circuit generates a fault determination.

[0074] Inspection decision generation based on imaging data according to embodiments of the present disclosure can provide accelerated and / or less labor-intensive inspections of medical implant devices. Thus, image scanning devices and / or systems can be used in conjunction with other information from relatively large data sets (e.g., reference inspection devices / data) that have been computationally analyzed to reveal correlations and / or associations between device failures at one or more manufacturing stages and various parameters or characteristics of the device.

[0075] In some embodiments, the imaging data generated by the imaging device 520 can be used to assist a human inspector in visually inspecting the target device 530. That is, in addition to the decision engine 512 of the control circuit 510 providing a final inspection decision regarding the target device 530, in some embodiments, a human inspector can inspect the imaging data 522 and / or the image feature data 523 to make a pass / fail determination for the target device 530. By providing the imaging data 522 for inspection by the inspector, such a process can advantageously allow the human inspector to make a pass / fail determination for the target device 530 without requiring manual manipulation of the target device 530 by the inspector. That is, in addition to performing a tedious manual inspection process as described above, the inspector can interpret the imaging data alone or in addition to the visual inspection of the target device itself. The imaging data 522 can be further used to assist a manual assembler or inspector, thereby potentially reducing the need for inspectors.

[0076] The imaging data 521 may provide sufficient detail for identifying particles that may be undesirably disposed or captured in one or more components of the target device 530. For example, such particles may produce certain shadows or other visually identifiable features that may be identifiable by the imaging data evaluation circuit 515. Using certain imaging techniques, the imaging data 522 may further allow for visual identification of defects and various metal components (such as metal frames, bands, or the like) of the medical implant device 530, as described in detail above. For example, certain metal components may be welded together or otherwise joined or coupled according to certain implant device manufacturing processes. Thus, the imaging data 522 may advantageously indicate where metal components are not properly closed or coupled, or where there are undesirable openings / gaps, such as gaps, cracks, or defects.

[0077] In addition, the imaging data 522 can advantageously allow for the identification of the positioning and / or other features of the sutures performed or set in the target device 530. For example, the imaging data 522 can indicate whether the sutures are in the proper orientation or the desired orientation, whether some sutures are omitted or there are redundant sutures, whether the correct type of sutures are used, and / or whether the correct suture method is used. In some embodiments, the target device 530 is sutured using radiopaque sutures, which can be visible under fluoroscopy or by using various types of electromagnetic radiation imaging. Alternatively, for other types of imaging techniques, the sutures and / or other components of the implant device may include materials designed to be visible using such imaging techniques. For example, such materials may be configured to reflect sound waves or the like. In embodiments related to heart valves or other medical implant devices as (one or more) target devices, the characteristics of the biological tissues of the medical implant device can be distinguished from plastics and / or other components based on their material characteristics, so that the imaging data 522 relatively clearly presents different elements composed of different materials.

[0078] In some embodiments, the pass / fail determination relative to the target device 530 can be based at least in part on the overlap of multiple different images. Such images can be generated using the same imaging technology, or can include images from multiple different imaging technologies. In some embodiments, the target device 530 is configured with one or more markers or other identifiers that are at least partially visible under imaging according to the technology of the imaging device 520. Therefore, such markers or identifiers can be captured or indicated in the imaging data 521. In the case where the target device 530 is not completely symmetrical relative to one or more cross-sections or planes, radiopaque or other markers visible under imaging can be particularly useful. In some embodiments, the imaging data 522 and / or the imaging device 520 are configured to provide imaging details for measuring the size of the target device 530 to a granularity of tens of millimeters or millimeters. Generally, the lower the resolution of the imaging device 520 and / or the imaging data 522, the less accurate the prediction decision of the decision engine 512.

[0079] With respect to a heart valve device, the imaging data 503 and / or the imaging data 522 may include information identifying or relating to the presence of particles and / or a suture / sewing pattern, which may be identified based on its different material properties relative to surrounding tissue and / or material. For example, the image feature 523 and / or the image feature 509 may identify the number of threads and / or whether enough associated grids have been used. That is, with respect to certain heart valve sewing processes, the open-pore fabric may be woven into a grid-like finish or arrangement to produce a grid-type braid (e.g., a "grid") in which the suture needle is threaded once every x grids. In some heart valve inspection embodiments, the image feature 523 and / or the image feature 509 may be associated with a manufacturing stage after the commissure cover described in detail above has been attached. The reference device data 520 maintained in the data storage device 502 may further include product data 505, which may include other data associated with the reference device, such as a product ID or the like.

[0080] In some embodiments, the reference device data 529 includes imaging data, decision data, and / or product data for tens of thousands or more reference devices that have been inspected, at least in part, by trained human inspectors. The machine learning circuit 517 may be configured to associate the identified image features 509 of the imaging data 503 and / or the product data 505 with the results embodied in the decision data 504 for the corresponding inspection reference device. Once the machine learning circuit 517 has generated an inspection (e.g., pass / fail) prediction model 519, the decision engine circuit 512 may be able to use the model 519 to determine whether a given imaging target device (e.g., a medical implant device) fits certain criteria of the model associated with failure prediction. Therefore, as long as the imaging data 522 captured by the imaging device 520 is sufficient to identify certain features / characteristics, artifacts, particles, or the like used as the basis for the prediction model 519, the decision engine 512 may be able to make a product scrap determination.

[0081] Imaging features 503 and / or image features 523 may include any type of image data that may be derived from imaging data of a target device. For example, as shown, image features 509 (and / or image features 523) may include dimensional data that may provide measurements of various dimensions for a relevant target device. Such dimensions may be associated with a pass / fail determination as captured in decision data 504. The dimensional data may be associated with any suitable size or (one or more) desired size of the target device, such as with respect to a heart valve device, a size associated with an opening or gap between valve leaflets. The dimensional data may further indicate the height or other dimensions of the leaflets of a heart valve device and / or the relative vertical and / or horizontal offset between the leaflets. Additional dimensional data may include frame size and / or diameter measurements, commissure height and / or height, or the like.

[0082] Image feature 509 (and / or image feature 523) may further include shape data, which may include an indication or value indicating the shape of one or more parts or components of a heart valve or other target device. For example, shape data may indicate whether the valve has a desired circular shape or is undesirably deformed. Image feature 509 (and / or image feature 523) may further include color data, which may indicate the color of one or more components or regions of the associated target device. For example, a color mark associated with a heart valve leaflet may indicate that the tissue of the valve leaflet is undesirably dehydrated or brittle. Such color marks or differences captured in the color data may be associated with decision data 504 to indicate the importance of certain color components or regions thereof relative to a fault decision. With respect to the valve leaflet, discoloration may be at least partially due to delamination, where air may be undesirably trapped or set between tissue layers, which is visually apparent in the imaging data. Product data 505 may include model ID / number data or other product identifier information, as well as product size data or other specifications.

[0083] The prediction model 519 generated by the machine learning circuit 517 can combine multiple image features when making a failure decision. For example, a failure decision can be based on a combination of features, such as the presence of one or more colors in one or more parts or regions combined with the presence of certain spots or particles / contaminants and discontinuities or defects in the parts (such as frame welds or form defects). For example, in the case where only one of the multiple combined parameters exists for a particular target device, the device can still be scrapped according to the model 519 when other required parameters are not present. The machine learning circuit 517 can be configured to determine the dimensional value and / or other parameters used as thresholds for failure determination. In some embodiments, the model 519 includes or represents a three-dimensional complex curve. The model 519 advantageously describes the correlation between various parameters / features that are identified as statistically significant with respect to the results. The machine learning circuit 517 can be configured to ignore parameters or features that are determined to have no substantial or significant impact on the device scrapping / failure results. In some embodiments, the model 519 provides a polynomial equation suitable for reference device data maintained in the reference data data storage device 502. For example, the polynomial model or equation may include one or more variables corresponding to or representing image features or other criteria based on imaging data.

[0084] In some embodiments, the imaging data 522 projects the imaged target device on a visual measurement scale or coordinate grid that provides additional information relative to the size or other features of the imaged target device when the image is superimposed on the grid. The use of a grid or other visual measurement representation can advantageously help make the visual features of the imaging data 522 quantifiable to its readers.

[0085] Cleaning of target device using imaging device

[0086] With respect to heart valves and other medical implants, non-sterile conditions can pose serious health risks. Therefore, in some embodiments, the imaging device 520 used to generate imaging data of the target device 530 can also be configured to perform a sterilization function to at least partially sterilize the target device 530. Generally, sterilization according to embodiments of the present disclosure can involve eliminating, removing, killing or inactivating undesirable biological agents (e.g., fungi, bacteria, viruses, spore forms, prions, and the like) present in the target device or its surface, volume, fluid, compound or component.

[0087] Sterilization can be achieved in system 500 using electron beams, chemical agents (e.g., ethylene oxide or nitric oxide), electromagnetic radiation (e.g., X-rays, gamma rays, ultraviolet (UV) light), or other techniques. Imaging device 520 can advantageously use the same radiation technology used for device imaging to sterilize the target device. In some embodiments, the technology used for cleaning is different from the technology used for imaging. Cleaning can be performed by imaging device 520 simultaneously with imaging of the target device, or before or after imaging.

[0088] Imaging-based inspection process

[0089] Figure 6 600 for generating an inspection prediction model according to one or more embodiments. Process 600 may be performed at least in part by a device inspection system or control circuitry of a device according to an embodiment of the present disclosure. With respect to the various methods and processes disclosed herein, although certain orders of operations or steps are shown and / or described, it should be understood that the various steps and operations shown and described may be performed in any suitable or desired temporal order. In addition, any operations or steps shown and / or described may be omitted from any given method or process, and the methods and processes shown / described may include additional operations or steps that are not explicitly shown or described.

[0090] At block 602, process 600 involves obtaining image data for a set of reference devices. For example, the image data may include images for a large number of previously inspected devices, such as heart valve implant devices or the like.

[0091] At block 604, process 600 involves determining image features of the imaging data of the reference device. For example, various sizes, colors, shapes, contaminants, and / or other features of the imaging data of the reference device may be determined and / or identified. For example, process 600 may involve accessing imaging data and / or image feature data corresponding to a plurality of imaged reference devices, which may be maintained in a data storage device including one or more data storage devices. Such imaging data and / or image feature data have been generated in order to check the calibration of the prediction engine. In some embodiments, determining the image features involves accessing image feature data stored in the data storage device. In some embodiments, determining the image features involves generating image features based at least in part on the imaging data. In some embodiments, process 600 does not involve obtaining imaging data at block 602, but only involves obtaining image feature data. The image feature data may include any number of features, such as various sizes, colors, weights, sizes, shapes, or the like of a plurality of reference devices (e.g., medical implant devices).

[0092] At block 606, process 600 involves obtaining inspection decision data for a reference device. For example, the inspection decision data may include data indicating inspection results associated with the reference device, where such inspection results may be determined by a trained human inspector or by some other means.

[0093] At block 608, process 600 involves determining, from the inspection decision data, a cross-correlation between image features of the reference device imaging data and a failure decision for the corresponding device. As described above, in some embodiments, a process for generating a pass / fail decision for a device (e.g., a medical implant device) may involve machine learning to determine a correlation between an imaging feature or characteristic and a pass / fail determination.

[0094] At block 610, process 600 involves generating an inspection prediction model or criterion based on determined cross-correlations between image features and failure decisions of reference inspection decision data. The prediction model can be used to evaluate image data of a target device subsequently imaged. For example, process 600 can involve analyzing imaging data and / or image feature data relative to reference inspection decision data to determine cross-correlations between image features associated with the imaging data and pass and / or fail decision results of the reference inspection decision data. The inspection prediction model can be generated without knowing the correlation or importance of various image features, in addition to the cross-correlations between these features and decisions made relative to their associated reference devices.

[0095] Figure 7 7 is a flow chart illustrating a process 700 for inspecting a device according to one or more embodiments. The process 700 may be performed at least in part by a device inspection system or a control circuit of a device according to an embodiment of the present disclosure.

[0096] At block 702, process 700 involves obtaining and / or generating imaging data for a target device. For example, as described in detail herein, the imaging data may be generated using an imaging device, such as a radiographic imaging device.

[0097] At block 704, process 700 involves determining certain image features of the imaging data of the target device. For example, certain colors, shapes, weights, sizes, and / or other features of the imaging data may be identified. Additionally or alternatively, certain other features may be assessed or derived from the imaging data (such as the presence of particles or other contaminants or the like). The image features may be correlated with (e.g., determined to be correlated with) failure decisions of previously inspected reference devices and / or may be correlated with known inspection prediction models or criteria.

[0098] At block 706, process 700 involves evaluating image features of a target device using an inspection prediction model or criteria. For example, the inspection prediction model can be used to identify features of imaging data of the target device that correspond to failure criteria of the inspection prediction model. At block 708, process 700 involves determining an inspection result based on the inspection prediction model. For example, the inspection prediction model can be used to determine a failure decision, which can result in scrapping and / or disposal of the target device.

[0099] Additional Examples

[0100] According to the embodiment, some actions, events or functions of any process or algorithm described herein may be performed in a different order, may be added, merged or omitted completely. Therefore, in some embodiments, not all described actions or events are necessary for the practice of the process.

[0101] Unless otherwise specified, or understood in context otherwise, conditional language used herein, such as "may", "can", "might", "could", "for example", etc., is used in a common sense and is generally intended to convey that certain embodiments include and certain embodiments do not include certain features, elements and / or steps. Therefore, such conditional language is generally not intended to imply that features, elements and / or steps are required in any way in one or more embodiments, or that one or more embodiments necessarily include a method for determining whether these features, elements and / or steps are included or to be performed in any particular embodiment with or without author input or prompting. The terms "include", "comprise", "have", etc. are synonyms, used in their common sense, and are used inclusively in an open-ended manner without excluding additional elements, features, actions, operations, etc. Similarly, the term "or" is used in its inclusive sense (rather than in its exclusive sense), so that when used, for example, to connect elements in a list, the term "or" means one, some, or all of the elements in the list. Unless otherwise explicitly stated, connecting language such as the phrase "at least one of X, Y, and Z" should be understood in conjunction with the context of common use to convey that an item, term, element, etc. can be X, Y, or Z. Thus, this linking language is generally not intended to imply that certain embodiments require the presence of at least one of X, at least one of Y, and at least one of Z.

[0102] It should be understood that in the above description of the embodiments, in order to simplify the present disclosure and help understand one or more of the various inventive aspects, various features are sometimes combined in a single embodiment, figure or description thereof. However, the method of the present disclosure should not be interpreted as reflecting the following intention: any claim requires more features than the features explicitly stated in the claim. In addition, any component, feature or step shown and / or described in the specific embodiments of this document can be applied to any other embodiment or used with any other embodiment. In addition, for each embodiment, no component, feature, step or group of components, features or steps is necessary or indispensable. Therefore, it is intended that the scope of the present invention claimed herein and below should not be limited by the above-mentioned specific embodiments, but should only be determined by a reasonable understanding of the attached claims.

[0103] It should be understood that certain ordinal numbers (e.g., "first" or "second") may be provided for ease of reference and that these ordinal numbers do not necessarily imply physical characteristics or order. Therefore, as used herein, ordinal numbers (e.g., "first", "second", "third", etc.) used to modify elements (such as structures, components, operations, etc.) do not necessarily have to indicate the priority or order of an element relative to any other element, but can generally distinguish the element from another element with a similar or identical name (but using an ordinal number). In addition, as used herein, indefinite articles ("a" and "a kind of") can indicate "one or more" instead of "one". In addition, operations performed "based on a condition or event" can also be performed based on one or more other conditions or events that are not explicitly stated.

[0104] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art to which the exemplary embodiments belong. It will be further understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and will not be understood in an idealized or overly formal manner unless explicitly defined herein.

[0105] For ease of description, the spatially relative terms "external," "internal," "upper," "lower," "below," "above," "vertical," "horizontal," and similar terms may be used herein to describe the relationship between one element or component shown in the figures and another element or component. It will be understood that the spatially relative terms are intended to cover different orientations of the device in use or operation in addition to the orientation shown in the drawings. For example, where the device shown in the drawings is flipped, a device positioned "below" or "beneath" another device may be placed "above" the other device. Thus, the illustrative term "below..." may include a lower orientation and an upper orientation. The device may also be oriented in another direction, and thus the spatially relative terms may be interpreted differently depending on the orientation.

[0106] Unless expressly stated otherwise, comparative and / or quantitative terms such as "less", "more", "greater", etc. are intended to encompass the concept of equality. For example, "less" may not only mean "less" in the strictest mathematical sense, but may also mean "less than or equal to".

Claims

1. An apparatus for inspecting a target device, the apparatus comprising: Data communication interface; and A control circuit, the control circuit being configured to: obtaining image characteristic data associated with a plurality of reference prosthetic heart valve devices; obtaining inspection decision data corresponding to the plurality of reference prosthetic heart valve devices, the inspection decision data including failure decision data; determining a cross-correlation between the fault decision data and the image feature data; as well as determining a fault prediction model based on the cross-correlation; receiving, using the data communication interface, imaging data of a target prosthetic heart valve device, the target prosthetic heart valve device comprising a sealing skirt having an upper edge, the upper edge of the sealing skirt having an undulating shape that follows a shape of struts of a frame and tightly secured to the struts of the frame with sutures; identifying a suture feature associated with the suture based at least in part on the imaging data; determining whether the seam feature is correctly positioned based at least in part on the failure prediction model; and An inspection result is determined based at least in part on determining whether the suture feature is properly positioned.

2. The apparatus of claim 1, wherein the imaging data comprises three-dimensional computed tomography imaging data.

3. The apparatus of claim 1 wherein the upper edge of the sealing skirt wraps around one or more frame struts.

4. The apparatus of claim 1, wherein the image feature data further comprises one or more of contaminant presence data, size data, color data, and shape data.

5. The apparatus of claim 1, wherein: The inspection result is a fault determination; and The control circuitry is further configured to send a command to a device remover device to remove the target prosthetic heart valve device from a manufacturing process flow based on the fault determination.

6. A method of determining an inspection decision for a target device, the method comprising: obtaining image characteristic data associated with a plurality of reference prosthetic heart valve devices; obtaining inspection decision data corresponding to the plurality of reference prosthetic heart valve devices, the inspection decision data including failure decision data; and generating a fault prediction model based on a cross-correlation between the fault decision data and the image feature data; obtaining imaging data of a target prosthetic heart valve device, the target prosthetic heart valve device comprising a sealing skirt having an upper edge, the upper edge of the sealing skirt having an undulating shape that follows the shape of struts of a frame and tightly secured to the struts of the frame with sutures; determining a location of the suture based on the imaging data; using the failure prediction model to assess whether the suture is correctly positioned; as well as An inspection result associated with the target prosthetic heart valve device is determined based on the evaluation.

7. The method of claim 6, wherein the imaging data comprises computed tomography imaging data.

8. The method of claim 6, wherein the image characteristic data further comprises one or more of contaminant presence data, size data, color data, and shape data.

9. The method of claim 6, further comprising removing the target prosthetic heart valve device from a process flow based on the inspection results.

10. A device inspection system, the system comprising: Imaging device; One or more non-volatile data storage devices, the one or more non-volatile data storage devices storing: image characteristic data associated with a plurality of reference prosthetic heart valve devices; and examination decision data associated with the plurality of reference prosthetic heart valve devices; and A control circuit, the control circuit being configured to: accessing the image feature data and the inspection decision data, the inspection decision data including fault decision data; determining a cross-correlation between the fault decision data and the image feature data; generating a fault prediction model based on the cross-correlation between the fault decision data and the image feature data; receiving imaging data of a target prosthetic heart valve device from the imaging device, the target prosthetic heart valve device comprising a sealing skirt having an upper edge, the upper edge of the sealing skirt having an undulating shape that follows the shape of struts of a frame and tightly secured to the struts of the frame with sutures; identifying a suture feature associated with the suture based at least in part on the imaging data; as well as determining whether the seam feature is correctly positioned based at least in part on the failure prediction model; and An inspection result is determined based at least in part on determining whether the suture feature is properly positioned.

11. The system of claim 10, further comprising a conveyor device configured to advance the target prosthetic heart valve device through a scanning region of the imaging device.