Surgical planning system with automated defect sizing

Abstract

Certain aspects of the present disclosure provide techniques for preparing a medical treatment plan, including: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; determining one or more quantitative measurements based on the fitted statistical shape model; and classifying a defect associated with the anatomical structure of the patient based on the one or more quantitative measurements.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of and priority to U.S. Provisional Patent Application No. 62 / 845,676, filed May 9, 2019, the entire contents of which are incorporated herein by reference.

[0003] introduce

[0004] Various aspects of the present disclosure relate to surgical planning systems, including surgical planning systems having automated defect sizing and population-based decision support capabilities.

[0005] Conventional surgical planning tools deal with the preoperative planning process. They solve routine problems associated with a specific surgery, such as the size and design of various components, including surgical instruments and implants, the location and orientation of implants and fixation devices. They usually take medical images of the patient as input, thus allowing the user (medical professional or non-medical professional, such as technicians or engineers) to make decisions based solely on the information available in these images. Summary of the invention

[0006] Certain aspects provide a method for preparing a medical treatment plan, comprising: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; determining one or more quantitative measurements based on the fitted statistical shape model; and classifying a defect associated with the anatomical structure of the patient based on the one or more quantitative measurements.

[0007] Some other aspects provide a method for determining a treatment for an anatomical defect, including: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; identifying a defect based on the three-dimensional anatomical model and the statistical shape model; determining a default treatment based on the identified defect; receiving patient population data associated with multiple other patients having the identified defect, wherein the patient population data includes multiple subsets of patient population data associated with different treatments for the identified defect; generating a visualization including: a representation of each subset of the patient population data based on at least one patient characteristic; and a representation of the patient based on at least one patient characteristic; and selecting a final treatment for the patient.

[0008] Some other aspects provide: a processing system configured to perform the aforementioned methods and those described herein; a non-transitory computer-readable medium comprising instructions that, when executed by one or more processors of the processing system, cause the processing system to perform the aforementioned methods and those described herein; a computer program product implemented on a computer-readable storage medium comprising code for performing the aforementioned methods and those further described herein; and a processing system comprising components for performing the aforementioned methods and those further described herein.

[0009] The following description and the associated annexed drawings set forth in detail certain illustrative features of the one or more embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings depict certain aspects of one or more embodiments and therefore are not to be considered limiting of the scope of the disclosure.

[0011] Figure 1 An example of a statistical shape model fitted to a 3D image of a patient's anatomy is depicted.

[0012] Figure 2 Depicted is an example of a statistical shape model divided into six regions.

[0013] Figure 3 An example of an anatomical measurement technique is depicted.

[0014] Figure 4 An example of an anatomical measurement technique is depicted.

[0015] Figure 5 An example of a method for measuring parameters associated with bone loss is depicted.

[0016] Figure 6 An example of a surgical planning workflow is depicted.

[0017] Figure 7 Another example of a surgical planning workflow is depicted.

[0018] Figure 8 Another example of a surgical planning workflow is depicted.

[0019] Fig. 9 An example of historical data-based analysis of a patient population being used to evaluate treatment options is depicted.

[0020] Fig.10 An example of historical data-based analysis of a patient population being used to evaluate treatment options is depicted.

[0021] Fig.11An example of historical data-based analysis of a patient population for evaluating parameter values ​​is depicted.

[0022] Fig.12 An example of using historical data-based analysis of a patient population to assess device size is depicted.

[0023] Fig.13 An example of historical data-based analysis of a patient population for evaluating parameter values ​​is depicted.

[0024] Fig.14 An example of defect quantification is depicted.

[0025] Fig.15 An example of historical data-based analysis of a patient population being used to evaluate treatment options is depicted.

[0026] Fig.16 An example of the use of historical data-based analysis of a patient population to evaluate implant options is depicted.

[0027] Fig.17 An example of the use of historical data-based analysis of a patient population to evaluate implant options is depicted.

[0028] Fig.18 An example of an interactive surgical planning system is depicted.

[0029] Fig.19 Depicted is an example of a representation of defect quantification using patient imaging data and SSM.

[0030] Fig. 20 An example representation of treatment options in a three-dimensional patient anatomy model is depicted.

[0031] FIG. 21A to FIG. 21D An example representation of treatment options in a three-dimensional patient anatomy model is depicted.

[0032] Fig. 22 Another example representation of treatment options in a three-dimensional patient anatomy model is depicted.

[0033] Fig.23 Another example representation of treatment options in a three-dimensional patient anatomy model is depicted.

[0034] Fig.24 Another example representation of treatment options in a three-dimensional patient anatomy model is depicted.

[0035] Fig.25 Another example representation of treatment options in a three-dimensional patient anatomy model is depicted.

[0036] Fig.26Another example representation of treatment options in a three-dimensional patient anatomy model is depicted.

[0037] Fig. 27 An example method for classifying defects using a statistical shape model is depicted.

[0038] Fig.28 An example decision support method is depicted.

[0039] Fig.29 An example method for determining a treatment for an anatomical defect is depicted.

[0040] Fig.30 An example processing system is depicted that may be configured to perform the various methods described herein.

[0041] To facilitate understanding, identical reference numerals have been used, where possible, to designate identical elements that are common to the figures. It is contemplated that elements and features of one embodiment may be beneficially incorporated in other embodiments without further recitation. DETAILED DESCRIPTION

[0042] Various aspects of the present disclosure provide apparatus, methods, processing systems, and computer-readable media for surgical planning systems, including surgical planning systems with automated defect sizing and population-based decision support capabilities.

[0043] The surgical planning system described herein solves several problems with conventional surgical planning tools.

[0044] For example, conventional planning tools do not provide information about healthy anatomical structures and therefore do not allow the user to correctly assess the size and location of the injury. In contrast, the surgical planning tool described herein provides an automated defect classification system that characterizes healthy anatomical structures as well as damaged anatomical structures. Thus, the surgical planning system described herein overcomes the problem of designing a preoperative plan based solely on damaged anatomical structures (such as bone and cartilage, etc.). Relatedly, the surgical planning system described herein provides a better, more detailed, and automated visual representation of damaged bone anatomical structures based on defect classification.

[0045] As another example, while providing planning support for specific surgeries, conventional planning tools provide little support for picking between such specific surgeries. The surgical planning system described herein has different starting points, allowing users to make more important high-level surgical decisions as well. Therefore, the surgical planning system described herein is more transparent to users such as surgeons. Specifically, the surgical planning system described herein provides statistics that allow surgeons to assess the patient's position within the patient population, enabling surgeons to make informed decisions while creating preoperative plans. The transparency of the system allows the user to trace each decision by providing the user with a complete patient profile. The surgical planning system is also intended to reduce the number of manual interactions required to create a preoperative surgical plan.

[0046] The systems and methods disclosed in the present invention are composed of interconnected parts.

[0047] Defect quantification and classification

[0048] Embodiments of the defect sizing system may use medical images as a starting point to implement methods for calculating characteristics of a defect or deformity in a patient's body, such as a bone, organ, musculoskeletal region, or any other anatomical site. In some embodiments, the defect sizing system and methods described herein may be a subsystem, module, or an integral part of a surgical planning system.

[0049] For example, if necessary, the shape and size of the bone defect remains useful information for surgeons, implant or surgical instrument manufacturers, implant positioning software providers, educational institutions and patients. Many classification systems are used to describe the shape and size of bone defects, such as the Paprosky hip classification system, the Dorr, Insall and Rand knee classification system, the Wallace, Walsch and Antuna shoulder classification system, etc.

[0050] Conventional methods use qualitative measurements of standard radiography or two-dimensional (2D) computed tomography (CT) scans. They rely on users to visually identify anatomical landmarks and guess the starting position of the defect and the appearance of conventional (i.e., healthy) anatomical structures. For example, in the case of bone defects or cartilage defects, such as erosion of the glenoid, acetabulum, tibial plateau, vertebrae, craniomaxillofacial region or any other bone anatomical structure or cartilage surface, the prior art will make the user rely on anatomical landmarks or observe unusual bone geometry to assess which parts of the anatomical structure have been eroded. However, in the absence of the shape of the undamaged anatomical structure as a reference, this is usually unable to go beyond a simple assessment. Similarly, when estimating soft tissues or organs (such as the heart, lungs, kidneys, brain, etc.), underdeveloped or overdeveloped parts, leaves, regions, chambers, and blood vessels can be identified by visual assessment or empirical rules, but without the shape of normal or healthy anatomical structures as a reference, it is impossible to carry out truly meaningful quantification of such underdevelopment or overdevelopment.

[0051] In addition, conventional methods use qualitative measurements based on 2D images. These measurements are not accurate because some information is lost when converting 3D objects to their 2D representations. That is, the actual patient anatomy exists in 3D, but the images used to plan surgery are captured in 2D. Due to their qualitative nature and variations in imaging protocols and environments, these 2D techniques have poor reliability. For example, the scale of objects in a 2D X-ray depends on the distance between the source and the acquisition plane and between the object and the acquisition plane. Similarly, parallax effects also depend on these distances and whether the light source is static or moving. Further, the orientation of the patient relative to the source and acquisition plane affects the projection of the anatomical structure.

[0052] In the system described herein, the defect or deformity is measured from the medical image of the patient using the model of the healthy body part as a reference (or as a template). By measuring the distance between the topological counterparts of the points or surfaces of the actual, damaged or deformed patient anatomical structure and these points or surfaces on the reference model, the size of the defect can be calculated in a variety of ways. For example, the distance can be measured by projecting rays from the virtual model of the healthy anatomical structure and calculating the distance from the healthy body part to the damaged body part along these rays. The virtual model of the patient's anatomical structure can be obtained by segmenting the medical image of the actual patient's anatomical structure. The virtual model of the corresponding healthy anatomical structure can be obtained in different ways, as explained below. In order to allow the user to visually assess the damage or deformity, the 2D or 3D virtual models of the damaged or deformed body parts and the healthy body parts can be superimposed and shown to the user. One or both of these models can be shown in a semi-transparent manner.

[0053] The reference model of the normal or healthy anatomical structure can come from different sources. For example, a mirror image of a healthy contralateral anatomical part can be used. To this end, a medical image of the contralateral anatomical part can be segmented and the resulting virtual model is mirrored.

[0054] In some embodiments, the methods disclosed herein use a 3D statistical shape model (SSM) to make quantitative measurements and predict the nature of a defect, defect or deformity by reconstructing a healthy body part. Statistical shape modeling can be used to predict native (i.e., healthy) anatomical shapes without the need for (an image of) an actual healthy skeleton. In such embodiments, a virtual model of a healthy anatomical structure can be obtained by fitting an SSM of a healthy anatomical structure to a portion of a (medical image or virtual model) of a patient's anatomical structure.

[0055] In general, an SSM is a mathematical model that represents the average shape and shape variation within a population. Each shape generated by an SSM can be represented by a number of shape coefficients, which may be referred to as SSM parameters.

[0056] In some embodiments, the method is performed on, for example, a 3D virtual model, a 3D biomechanical model (musculoskeletal model), an SSM and / or an SSM instance, such that there is no approximation or conversion of measurements between the 2D representation and the 3D world.

[0057] As an example, a fully automated defect classification system can be used to describe glenoid bone loss using three-dimensional measurements of a scapula and / or humerus model without requiring a healthy contralateral reference scapula. In other embodiments, similarly, the automated defect classification system can be used to measure defects or deformities in other body parts, such as the heart, knee, hip, spine, foot, lungs, other joints, etc.

[0058] An example method may include: (1) acquiring medical images of a patient with a glenoid bone defect or joint replacement; (2) segmenting the scapula to obtain a virtual three-dimensional surface model, such as using Mimics; and (3) fitting a statistical shape model (SSM) of a healthy scapula to a healthy surface area of ​​the patient's scapula, such as Figure 1 Depicted.

[0059] Consistent with this example, the SSM should describe the healthy scapula shape within the population to which the patient belongs. By fitting the SSM to the healthy portion of the patient's anatomy, the unhealthy surface of the scapula (such as the glenoid in this example) will also be reconstructed. The shape correlations embedded in the SSM will produce a reconstructed glenoid that is statistically most likely to resemble the original, healthy, or native shape of the now unhealthy region.

[0060] Example SSM fitted to a healthy bone region

[0061] Figure 1 An example of an SSM 102 fitted to a healthy region of a scapula (eg, 104 ) to reconstruct its native glenoid surface 106 is depicted.

[0062] Different techniques can be used to fit the SSM to partial data, such as healthy anatomy, so that missing data (e.g. bone loss due to bone erosion) can be predicted, such as a posteriori shape modeling. However, such techniques require a priori identification of healthy and damaged or deformed regions. This step is known to exhibit high inter- and intra-user variability. Therefore, it is beneficial to automate this step.

[0063] Divide the SSM into regions to improve the fitting error

[0064] In one embodiment of the automated method, the SSM is subdivided into topological regions such as Figure 2 202 to 212 in the example of FIG. 20A. For each of these regions, it is tested whether including the region in the regions used to fit the SSM would result in a decrease or increase in the fit error. When including a region results in an unacceptable or increased fit error, the region is assumed to be damaged or deformed and is excluded from the fit. The SSM is then fit to a subset of the remaining regions to obtain an SSM instance representing the patient's anatomy in a healthy or non-deformed condition.

[0065] exist Figure 2 In the example of , the surface of the SSM representing the scapula is divided into six regions: base region 202, acromion region 204, coracoid region 206, neck region 208, acromion tip region 210, and glenoid region 212. This is just one example, and other subdivisions are possible, such as subdivision into different regions, or into more or fewer regions.

[0066] Thus, an example method may proceed as follows: First, the SSM shape is fit to the target shape based only on the points in the base region 202. After the shape coefficients converge, the fit error is calculated as the root mean square error (RMSE) between the points on the SSM shape used for the fit and the corresponding points identified on the target shape.

[0067] If the fit error remains below the selected threshold, a second fit is performed using points in the acromion region 204. If the fit error then exceeds the threshold, the acromion region 204 of the target shape is deemed unhealthy and the acromion region 204 is excluded from the subset of topological regions. In this example, the same selection procedure is then repeated for points in the coracoid region 206, acromion tip 210, and cervical region 208. In some cases, the glenoid region 212 may be expected to be eroded and therefore not used for fitting.

[0068] If both the acromion region 204 and the coracoid region 206 are excluded for fitting, for example based on a fitting error exceeding a threshold, then in some cases the acromion tip region 210 and the neck region 208 are not further tested.

[0069] In various embodiments, different fitting error thresholds may be used. For example, sensitivity studies show that a fitting error threshold of 1.7 mm produces good results. Other values ​​of fitting error thresholds can also be selected, such as 0.5 mm, 0.6 mm, 0.7 mm, 0.8 mm, 0.9 mm, 1.0 mm, 1.1 mm, 1.2 mm, 1.3 mm, 1.4 mm, 1.5 mm, 1.6 mm, 1.8 mm, 1.9 mm, 2.0 mm, 2.5 mm, 3.0 mm, to name a few.

[0070] Similar methods can be applied to other anatomical structures. Therefore, to generalize the process, the anatomical structure can be subdivided into multiple topological regions (e.g. Figure 2 202 to 212 in FIG. 1 ). The first region (e.g. Figure 2 The base region 202 in the figure can be selected to start a topological region subset, which in some cases can be a region away from a defect or deformity. The first region can then be fitted, and the fitting error can be calculated and compared with a threshold, such as described above. Subsequently, additional topological regions can be added to the subset, which can then be fitted to the target model.

[0071] After each topological region is added to the subset, the fitting error can be recalculated, and depending on whether the fitting error exceeds a set threshold (such as the threshold mentioned above), the additional topological region can be removed from the subset or retained in the subset. To speed up the process, topological regions that are not directly connected to the base region can be ignored if one or more regions in between are classified as damaged or deformed. To further speed up the process and improve the results, topological regions that are known to be damaged or deformed can also be ignored.

[0072] Analyzing a bone defect (e.g., a glenoid bone defect) by comparing its shape to the predicted native shape (e.g., of undamaged glenoid bone) results in quantitative measurements, such as in the case of glenoid bone, glenoid vault loss, percentage of glenoid vault loss, glenoid erosion area, percentage of glenoid erosion area, maximum erosion depth, etc.

[0073] Distance measurement techniques for comparing anatomical shapes

[0074] In one embodiment, to compare anatomical shapes (e.g., between predicted and actual shapes), distances between topologically equivalent points on each shape model can be measured, such as between the closest points or between points along a ray emanating from one model to the other, to name just a few options.

[0075] For example, for anatomical parts that are substantially spherical or hemispherical, such as Figure 3 302 in the acetabulum, rays 304 may be emitted outward from the center 306 of the sphere in a concentric manner, such as Figure 3 The example depicted.

[0076] As another example, for a substantially flat or planar anatomical region, rays 404 may be emitted perpendicular to best fit plane 402 in a parallel manner, such as Figure 4 Depicted.

[0077] As yet another example, for elongated anatomical parts, rays may be shot outward and perpendicular to the central axis of the anatomical part. For other anatomical parts, rays may be shot perpendicular to the surface of the SSM instance. It is worth noting that these are just a few options, and other ray casting strategies or combinations of strategies are possible.

[0078] Thus, the methods described herein can automatically calculate metrics based on the SSM, such as: glenoid vault loss (total volume of the glenoid vault lost due to bone erosion), percentage of glenoid vault loss (percentage of the volume of the glenoid vault lost due to bone erosion), percentage of local vault loss (in the superior, inferior, anterior, and posterior zones), erosion area (surface area of ​​the glenoid cavity affected by bone erosion), maximum erosion depth (maximum distance measured between the actual anatomical surface and a healthy reference model), percentage of erosion area (percentage of the surface area of ​​the glenoid cavity affected by bone erosion), subluxation distance, etc. It is worth noting that while the glenoid is used as an example herein, similar metrics can be calculated for other anatomical sites, such as other bones, joints, etc. Based on this calculation, the system described herein can automatically classify the defect.

[0079] Examples of Measuring Metrics Associated with Bone Loss

[0080] Using the glenoid bone as an example, the glenoid vault loss percentage metric indicates how much of the glenoid vault volume has been eroded and represents the severity of the glenoid bone defect. The upper, anterior, lower, and posterior vault loss percentages represent how much of the vault has been eroded in each anatomical region or quadrant of the glenoid, thereby better understanding the shape of the defect. The maximum erosion depth describes the amount of bone erosion at the deepest point of erosion. This metric can help surgeons decide whether they should ream or use a bone graft during surgery. The erosion area percentage shows how much of the native glenoid surface is no longer intact, thereby indicating the possible amount of implant bone support. Finally, the subluxation distance and area describe the amount and direction of humeral subluxation, thereby better understanding the cause of the glenoid bone defect.

[0081] Figure 5 Depicted are examples of metrics used to measure metrics associated with bone loss in the glenoid bone.

[0082] To measure these metrics, a ray casting algorithm (as described above) can be used. For example, first, a plane (e.g., 506) is fitted through the glenoid surface of the fitted SSM, and parallel rays are cast from the glenoid points of the fitted SSM shape in the opposite direction of the plane normal. The distance that ray i intersects the fitted SSM shape is called the vault depth (d i vault )(e.g. 502), where d max (e.g. 504) as the selected maximum value.

[0083] The amount of bone erosion is then assessed by shooting rays (e.g., 502 and 506) from the glenoid point of the fitted SSM shape toward the bone defect and parallel to the glenoid plane normal. The measured distance that the ray intersects the bone defect is defined as the erosion depth (d i ero ) (e.g. 506), is limited to d max (e.g. 504). If the erosion depth is infinite, then there is no bone at that location. Next, the loss depth (d i loss ) is defined as the lost vault depth. Loss depth is similar to erosion depth, but cannot exceed vault depth.

[0084] So, in one example, for each ray i:

[0085] If d i vault >d max :Then i vault =d max

[0086] If d iero >d max And d i ero ≠inf: then d i ero =d max

[0087] If d i ero ≤d i vault :Then i loss =d i ero

[0088] If d i ero >d i vault :Then i loss =d i vault

[0089] Based on the depth measurement, nine parameters describing the glenoid bone defect can be calculated.

[0090] For example, the volume of the fornix is ​​calculated as the sum of all fornix depths multiplied by the corresponding surface element (A i ) size. Similarly, the amount of vault loss is calculated as the sum of the loss depths multiplied by the corresponding surface area. The percentage of vault loss is then calculated as the percentage of the amount of vault loss compared to the vault volume.

[0091] The glenoid surface was divided into four quadrants using the glenoid center point for the superior (sup), anterior (ant), inferior (inf), and posterior (post) vault loss percentages. The vault loss percentages in these regions were equal to the amount of local vault loss divided by the local vault volume.

[0092] Next, in one example, the maximum erosion depth is calculated as the 95th percentile of all erosion depth values. The erosion area is calculated as the value of all surface elements A i The area where the erosion depth exceeds one-third of the maximum erosion depth. To obtain the percentage erosion area, in one example, the erosion area is divided by the total area of ​​the glenoid. The subluxation distance is calculated as the in-plane distance from the center point of the humeral head to the center point of the glenoid after projecting the center point of the humeral head onto the glenoid plane. The subluxation area is defined as the area where the center point of the humeral head is projected onto the glenoid (sup, ant, inf, post).

[0093] So, in an example:

[0094]

[0095]

[0096] Vault loss percentage = (vault loss volume) / (vault volume)

[0097] For all i in the region,

[0098] For all i in the region,

[0099] Local vault loss percentage = (local vault loss volume) / (local vault volume) Maximum erosion depth = p95 (d i ero )

[0100] Targeting d i ero For all i > 1 / 3 of the maximum erosion depth, erosion area = ∑ i A i

[0101] Erosion area percentage = (erosion area) / (∑ i A i )

[0102] In some examples, multiple classification systems may be combined, such as the Wallace classification in axial view and the Antuna classification in frontal view (as above), which beneficially provides a user (eg, a surgeon) with a three-dimensional classification of the defect compared to conventional two-dimensional classifications.

[0103] It is noteworthy that similar quantification can be performed on other anatomical parts, such as other joints, other bones, organs (heart, lungs, kidneys, brain, etc.) to assess injury, deformity or disease. Based on this quantification, similar classification systems can be defined. The systems and methods use appropriate and / or known classification systems or combinations thereof based on the body part in need of treatment.

[0104] Preoperative surgical planning tool

[0105] Existing preoperative planning tools (such as The SurgiCase Knee Planner from , offers the possibility to generate a preoperative surgical plan for a specific type of surgery (usually involving a specific type, brand or product line of implant). Preoperative planning usually starts after the surgeon has made important surgical decisions, such as: type of surgical treatment, type of implant and type of surgical instrument to be used, matching of standard implants to the patient, etc.

[0106] Further, these decisions are based on medical images taken from the patient. For example, for orthopedic treatments, those medical images may depict the damaged bone / cartilage anatomy. Existing planners generate initial or default plans based on the damaged anatomy (e.g., bone and / or cartilage), which are then reviewed by the surgeon. Upon review, the surgeon may propose certain changes, such as: the location or size of the implant, which are then merged by the planner, and a new preoperative plan is generated for use during the actual surgical procedure.

[0107] Unfortunately, because existing planners only take medical images as input, preoperative planning only considers information visible in these medical images. Preoperative planning does not address any aspects that cannot be easily derived from medical images or all the complexities associated with surgery that surgeons encounter in the operating room, which may affect surgical outcomes, the risk of intraoperative or postoperative complications, or patient satisfaction.

[0108] A surgical planning system can use multiple patient-specific medical images by using aggregated prediction techniques based on one or more known preoperative plan sets. For example, such a planning system can obtain historical data from preoperative plans, data collected during surgery, and data collected after surgery. Further, the planning system can select preoperative plans from the preoperative plan set and then apply prediction techniques, such as machine learning, deep learning, neural networks, or other artificial intelligence (AI)-based techniques to create an aggregated preoperative plan and recommend changes to the user. However, this preoperative plan generation method is generally not transparent to the surgeon, that is, the surgeon cannot know in advance how or why the planner incorporates the proposed changes, what characteristics of a particular patient lead to the proposed changes, how sensitive the system is to those characteristics, or the impact of these changes on the patient. Therefore, although the system itself can be self-learning, it does not allow the surgeon to make informed decisions.

[0109] By providing more information to the surgeon and acting as a guide for the surgeon, the system disclosed herein overcomes the shortcomings of existing surgical planning tools. As a guide, the embodiments of the system described herein provide timely suggestions, opinions and warnings and detailed information to confirm these suggestions, opinions and warnings, allowing the surgeon to make wise decisions. The control of the system lies with the surgeon, enabling the surgeon to make each decision consciously, making it a transparent and user-friendly system. The system disclosed herein beneficially reduces the time spent in the operating room and the changes that the surgeon must deal with in the operating room, and increases the possibility of positive surgical results, thereby generally reducing the number of revision surgeries that the patient may need.

[0110] The systems described herein may use multiple feedback loops to provide information to the surgeon in the form of suggestions, warnings, advice, and / or default preoperative plans that also involve the establishment of one or more interconnected databases.

[0111] Surgical planning workflow

[0112] A surgical planning method (e.g., performed by a surgical planning system described herein) may include multiple steps, including: (1) loading a medical image; (2) processing the medical image, e.g., to identify anatomical landmarks and / or create one or more virtual 3D models of anatomical structures; (3) automatically creating a default surgical plan, which is typically based on a number of geometric calculations based on the identified landmarks, and which typically includes selecting one or more implants, implant sizes, locations, and orientations of all implants, corresponding resection or reaming steps, etc.; (4) allowing a clinician to change the default plan to obtain an approved preoperative plan; and (5) making the preoperative plan available for execution during surgery. In some embodiments, the preoperative plan can be used, for example, in a navigation system, in a robotic system (to design patient-specific guidance), in an augmented and / or virtual reality system, and for other purposes.

[0113] For example, Figure 6 A workflow of a conventional surgical planning method and tool including steps 602 to 618 is depicted.

[0114] A database or other data repository may be used to store the approved plan in the database along with relevant patient data, such as medical images and any virtual 3D models and landmark information. Additionally, the system described herein provides Figure 6 Add one or more feedback loops to the workflow depicted in .

[0115] For example, Figure 7 The depicted first feedback loop 702 may mine information from an approved preoperative surgical plan for use before or during the planning step, and store it in a database 708. A second feedback loop 704 may collect information intraoperatively, store the data in a database 708, and mine the information for use before or during the planning step. A third feedback loop 706 may collect information postoperatively, store the data in a database 708, and mine the information for use before or during the planning step.

[0116] right Figure 6 and Figure 7 Further improvements to the described data flow are in Figure 8, where historical data available in a database is used to perform historical data analysis 802, correlating patient characteristics with planning decisions, or correlating one or more planning parameters with surgical outcomes. Further, the results of the historical data analysis can be presented to a user (e.g., a surgeon) in such a way that the location of the patient or planning parameters within a population is shown together with the distribution of planning decisions or surgical outcomes, respectively, over the population (e.g., in 804). Beneficially, presenting this information does not force the user to blindly choose between accepting and rejecting the proposed plan changes. Instead, it shows the user which planning decision options or parameter values ​​are appropriate and to what extent they are more suitable for possible postoperative scenarios than other options or values.

[0117] For example, when considering options A, B, and C, instead of simply suggesting: "Take option A," the system might show how the patient population is distributed across options A, B, and C and where the patients are located within the population. Through the presentation of the results of the historical data analysis, the user can see not only whether the patient is sitting in option A, or more precisely on the border between options A and B, but also whether that border is a clearly defined boundary or a wide range with a smooth transition.

[0118] Surgical planning data and database

[0119] The systems described herein may utilize one or more databases that are connected to different parts of the surgical planning system via one or more feedback loops. For example, data may be collected at one or more stages of the workflow, as described above with respect to Figure 7 described and stored in a database.

[0120] In some embodiments, the collected data can be divided (logically or physically) into subsets, such as patient data, preoperative data (including collecting pre-existing plans (i.e., preoperative plans that have been used for optimization of future preoperative plans)), retrospective data, intraoperative data, and postoperative data. Links between data in different subsets but related to individual patients are maintained; in other words, the database tracks which patient data, preoperative plans, intraoperative data, and postoperative data belong to the same patient. As above, the data can be stored in a single database, or in different databases.

[0121] Which of these types of data is stored in the database(s) depends on which feedback loops are implemented in the system. Some subsets of patient data are always stored. However, a base system may, for example, implement only a feedback loop for an approved preoperative plan. Other systems may also implement feedback loops for intraoperative data and / or postoperative data. Other combinations are also possible. One or more feedback loops may be called at a time. In some embodiments, in the case of revision surgery, all feedback loops may be called to obtain the entire patient profile from the previous surgery.

[0122] In some embodiments, the systems described herein may be run locally or "on-site", in which case the database(s) may contain only data relevant to one or more local users (such as surgeons, physicians, or clinicians or their teams). In other embodiments, the system may be a network-based system, such as a web-based system or a cloud-based system, in which case the database(s) may contain data relevant to a larger user base.

[0123] Patient data may be stored in the form of one or more of medical images, personal information such as age, gender, weight, height, ethnicity, lifestyle, activity level, medical history, and any data collected during preoperative examinations, such as complaints, pain scores, gait measurements, range of motion measurements, degenerative or congenital defects, sports or age-related injuries, genetic information, dental casts, etc. In some embodiments, patient data may be anonymized to protect patient privacy or to comply with various patient privacy regulations, such as the Health Insurance Portability and Accountability Act (HIPAA) or the General Data Protection Regulation (GDPR).

[0124] Preoperative data may be used, for example, in the form of preoperative treatment planning (e.g. Figures 6 to 8 614), which may alternatively be referred to as a preoperative plan or preoperative surgical plan. The preoperative data may capture some or all medical decisions related to the treatment of the patient's medical condition, such as one or more of the following: type of treatment (interventional and non-invasive treatment); type, brand, product line, size, implant site, and orientation (if any) of the planned implant; delivery system and method of any implant; design of patient-specific instrumentation (if any); details of any reaming steps; type or design of any defect filling component, such as autograft, allograft, porous structure, and other aspects.

[0125] Intraoperative data (e.g. Figures 7 and 8710) may be stored in the form of any data captured during surgery, such as measurements, locations of anatomical landmarks identified intraoperatively, observations, or occurrences of intraoperative complications. Intraoperative data may involve information that cannot be easily derived from medical images or preoperative examinations, such as information related to soft tissue, muscles, muscle attachments, muscle ruptures, tendons, ligaments, ligament tension, etc. Intraoperative data may also include any changes made during surgery relative to the preoperative plan. Intraoperative data may also include synthetic data, which in one example may be data that cannot be quantified but can be recorded due to its effect on the outcome of the surgery (such as ligament forces in the case of the knee joint). This may be stored in the form of a biomechanical model.

[0126] Postoperative data (e.g. Figures 6 to 8 618) may be stored in the form of any data captured after the surgery, such as the occurrence of any complications, any data captured during the post-operative examination, pain scores, patient satisfaction, functional scores, revision surgeries, post-operative imaging, recovery time, rehabilitation time, rehabilitation treatment methods, details and observations of the physical therapist; range of motion measurements (if any), etc.

[0127] The data may be entered into the system manually or automatically through a surgical planning system, through any device used during surgery (such as a navigation system, a robotic system, or an augmented reality (AR) or virtual reality (VR) system), through an electronic access device, through a wearable device, or through a sensor embedded in an implant or a chip embedded in the patient's body. It is worth noting that these are just a few examples.

[0128] The example surgical planning system described herein may implement the automated defect sizing system discussed above.Based on the defect classification and description, the surgical planner provides the surgeon with additional valuable information to help plan and perform the surgery.

[0129] Acquisition of patient data

[0130] Patient data can be loaded from a file, storage medium or database, or manually entered into the system. If the patient has undergone surgery previously, his old file can be called up from the database. If not, a new case file or record is generated.

[0131] For many applications, medical images will be a valuable part of the patient data.

[0132] Data processing: Patient data can be processed. For example, a medical image can be converted into one or more virtual 3D models of anatomical parts, such as bone anatomy, cartilage, organs, organ walls, blood pool volume, etc. Anatomical landmarks can be determined or indicated in the medical image or virtual 3D model. This can be done manually or automatically, for example through feature recognition technology. Other information can be derived from the medical image, such as bone density information, bone loss, impact of bone-to-bone contact, spread / extent of defects on surrounding anatomical structures, adjacent and attached soft tissue properties, such as muscles, ligaments, cartilage, tendons, menisci, thickness of soft tissue, etc. Additionally, a biomechanical model can also be generated to display musculoskeletal data, such as bone anatomy and soft tissue data that can be further simulated.

[0133] In some embodiments, the defect or deformity is quantified and / or classified as described above.

[0134] Default treatment plan creation

[0135] In some embodiments, the surgical planning system described herein can be associated with a specific surgery and / or a specific type, brand, or product line of implants. Additionally, unlike conventional systems, the system described herein can support more important advanced treatment decisions, such as: treatment type, including interventional treatment, non-invasive treatment, or referral. Further treatment decisions can include the type of implant, as many pathologies can be treated with different types of implants, such as off-the-shelf, custom-made, or customized implants, or a combination thereof. For example, for joints: cartilage repair, remodeling, or replacement; partial or full (e.g., unicompartmental / total distal femoral implants, unicompartmental / total proximal tibial implants); fixation strategies (cemented / non-cemented, stemmed / stemless, press fit, screws); functional strategies (e.g., posterior stabilized / cross-retained femoral implants, anatomical / reverse shoulder implants); acceptable range of motion, etc. can be considered. For cardiac applications: valve repair, suturing, replacement, annuloplasty, stent type, and other aspects can be considered. For craniomaxillofacial applications: orthognathic, reconstruction, trauma, TMJ, surgical procedures on alveolar bone types, treatment of the maxilla or mandible or both, orbital floor or parts of the skull and other related aspects can all be considered.

[0136] For pulmonary applications: intraluminal and extraluminal stents, valve types, and other aspects may be considered.

[0137] Type of instrument or guidance: Conventional instruments, patient-specific guidance, navigation systems, AR systems, robotic systems, etc. can be considered.

[0138] To support these decisions, the surgeon may be presented with additional relevant information to understand the defect in more detail, such as information or models derived from medical images and / or the results of defect or deformity quantification and classification, as described above. For example, the surgeon may be presented with the results of quantification and classification and / or a visual representation of the defect or deformity by superimposing a virtual 3D model of the actual patient anatomy and a model representing the healthy anatomy (such as a portion of the SSM fitted to the patient's anatomy). One or more models may be shown in a semi-transparent manner, such as described above. A biomechanical model simulation may also be shown along with the virtual 3D SSM model.

[0139] As further support for these decisions, the system can run one or more population analyses based on historical data collected in a database through one or more feedback loops. This analysis can make one or more patient characteristics related to one or more treatment decisions. Therefore, the system can use 1) the selection of the population, 2) the selection of the treatment decision to be supported, and 3) the selection of one or more patient characteristics to characterize the members of the population and the patients to be treated. These choices can be left to the user, such as through a drop-down box or a check box in the user interface. Alternatively, the system can present one or more pre-programmed selection combinations to the user, such as in a wizard-like process. Correlation analysis can reveal which patient characteristics may be related to which treatment decisions. Alternatively, the system can first track user behavior and then present the most common combinations by default. For example, an AI-based system can understand the decision-making factors that are frequently selected and display them to the surgeon at the appropriate time during the future preoperative planning stage. Alternatively, an AI-based system can learn the correlation between certain characteristics (especially the 'best characteristics') and their impact on treatment decisions, and use them for optimization, thereby providing treatment options based on the 'best' characteristics or based on the surgeon's preference for the 'best characteristics'.

[0140] With respect to the selection of a population, the historical data analysis may be based on all or a subset of records in the database. For example, a population may be limited to only those records that are sufficiently complete, i.e., records that contain the appropriate data required for the analysis. A population may also be limited to patients who have one or more common characteristics (e.g., gender, age, race, etc.) with the patient to be treated. A population may also be limited to only those patients who have been treated in the same country, the same hospital, or by the same clinician, physician, surgeon, school of thought, etc.

[0141] Analysis of historical data may reveal how the selected population is distributed across the different options for the selected treatment decision. Population members are characterized by selected patient characteristics (multiple). The patient to be treated can be located by his / her specific patient characteristics (multiple) within the analysis population, thus revealing which decision option seems to be most suitable for this particular patient based on the historical data in the database. Alternatively, in the same instance, if the (multiple) "best" characteristics selected by the system are different from the selected (multiple) patient characteristics, the system can show a comparative analysis based on the (multiple) "best" characteristics selected by the system, allowing the user to re-evaluate his decision.

[0142] In some embodiments, historical data analysis can correlate one or more treatment decisions with the expected occurrence of an intraoperative or postoperative event, observation, or outcome. For example, historical data analysis can reveal how the chance or risk of a certain event, observation, or outcome occurring increases or decreases with a certain preoperative planning parameter.

[0143] For example, analysis of historical data may correlate a selected heart valve size with risk of leakage, or may correlate a selected amount of lateralization of a shoulder implant with risk of acromion fracture.

[0144] In some embodiments, historical data analysis can utilize retrospective data including data obtained from senior surgeons or key opinion leaders (KOLs) and provide it to new or junior surgeons to guide their decision making, such as bone defect data, to model treatment options or provide junior surgeons with treatment plans that they use or prefer. In some embodiments, retrospective data can include information provided and used by schools of thought (e.g., surgeons using the same plan or treatment options or otherwise).

[0145] By limiting the group to those patients that show similarities to the patient to be treated, for example, with respect to one or more patient characteristics, this type of analysis can be made more accurate or relevant to the patient to be treated. This type of historical data analysis may require: 1) selecting zero or more patient characteristics to limit the group; 2) selecting one or more types of events, observations, or results; and 3) the selection of one or more treatment decisions. As mentioned above, these choices can be left to the user, such as through a drop-down box or check box in a user interface. Alternatively, the system can present one or more preprogrammed selection combinations to the user, such as in a wizard-like process. Correlation analysis can reveal which events, observations, or results may be related to which treatment decisions. Alternatively, the system can first track user behavior and then present the most common combinations by default. For example, an AI-based system can understand frequently selected decision-making factors and display them to the surgeon at the appropriate time during the future preoperative planning stage.

[0146] Regarding the selection criteria for the group, the group should preferably be limited to members that show similarity to the patient to be treated. This similarity may be related to one or more patient characteristics.

[0147] For example, in the case of a leaky heart valve, those patient characteristics may be a set of measurements describing the shape of the anatomy surrounding the valve, such as the minimum and maximum diameters of the annulus.

[0148] As another example, in the case of an acromion fracture, patient characteristics may include information about bone density, as derived from a CT scan or from the results of the defect quantification and classification described above.

[0149] For those analyses where the outcome is known or suspected to depend on the shape of the patient's anatomy, the patient characteristics can include parameters or a subset of parameters of an SSM fitted to a portion of the patient's anatomy. These parameters or such a subset form an n-dimensional vector that describes the shape of the patient in n-dimensional space, covering all possible shape variations. Thus, the population used for historical data analysis may be limited to all members whose corresponding n-dimensional vectors fall within a certain preset distance from the patient to be treated.

[0150] The results of the historical data analysis can be presented in different ways, some examples of which are described below. Example embodiments of population analysis are also described below.

[0151] As an alternative to historical data analysis, the system may also locate the member within the selected population that best matches the patient characteristics of the patient to be treated and display the decision options selected for that member.

[0152] Once a high-level treatment decision is made, whether or not the above-mentioned decision support process is used, the system described herein can create a default preoperative plan for the patient to be treated. The plan will typically rely on one or more algorithms or heuristics to calculate treatment parameters, such as: implant position and orientation based on patient data and processed patient data.

[0153] For example, SurgiCase Knee Planner uses a geometric algorithm based on the anatomical landmarks identified on the virtual 3D model of patient's femur and tibia, to calculate the local anatomy coordinate system and with respect to the default size, place and orientation of the patient's anatomical structure of femur implant and tibia implant.For some input parameters of this algorithm, the value of general colony scope can be used.Alternatively, value can be manually or automatically selected based on the support from above-mentioned decision support process.

[0154] For example, for total knee replacement, a default value of bow leg correction to 3° bow leg may be used for all patients, historical data analysis may suggest a certain value for bow leg correction, or the value for bow leg correction of the most matching member of the population may be used. Thus, the decision support program of the present invention may be used for both high-level treatment decisions and low-level treatment specific decisions.

[0155] Historical data collected through one or more feedback loops can also be used to improve automatically created default plans or create new default plans. For example, AI-based techniques (such as machine learning, deep learning, neural networks, etc.) can be used to incorporate changes that are frequently or consistently made during planned steps or treatments into the default plan. In addition, information about intraoperative or postoperative complications can be used to include some changes and ignore other changes.

[0156] Modifying treatment plans

[0157] Once a default plan is made, it is presented to the user for further fine-tuning. The user may be presented with the possibility of changing one or more treatment plan parameters. For example, the user may have the possibility to change the implant size, implant location, or implant orientation.

[0158] In the planning step, the system can support the user's decision making through the above-mentioned decision support process.

[0159] The outcome of the planning step is an approved preoperative plan, i.e., the treatment plan that the clinician has decided to perform.

[0160] In some embodiments, the system includes a feedback loop that stores all approved preoperative plans in a database. The information collected in this way can be used as historical data to feed the decision support process. For example, running a population analysis on the preoperative plans approved by a user will tell the user which changes or parameter values ​​exist within his past practice or experience. In contrast, running a population analysis on the preoperative plans approved by all users will allow the user to learn from the cumulative experience of a larger population, or to compare his personal practice with the average practice of all users. Other options are also possible, such as limiting the historical data to preoperative plans approved by all users in the same hospital or all users in the same country.

[0161] Patient treatment according to treatment plan

[0162] Once an approved preoperative plan has been made, the clinician can proceed with its execution, i.e., treat the patient. In some (primarily non-invasive) treatments, the pre-approved treatment plan may take the form of a prescription, such as for medication or exercise. In other (primarily interventional) treatments, the pre-approved treatment plan may take the form of a data file that can be used in a surgical guidance system. For example, the plan can be used to design and manufacture patient-specific instruments to help the surgeon achieve the planned surgical outcome during surgery. Alternatively, the plan can be loaded into a surgical navigation system or AR system to display guidance information to the surgeon during surgery. Alternatively, the plan can be loaded into a robotic system to automatically or semi-automatically perform part of the surgery.

[0163] The system may include a feedback loop to store intraoperative data in a database. This may include any of the intraoperative data described above. The data can be automatically collected by sensors in the operating room, specialized surgical equipment, surgical guidance systems (such as navigation systems, AR systems, or robotic systems), or can be manually entered through an electronic access device.

[0164] For example, the system can prompt the surgeon to store any intraoperative changes or complications encountered during surgery. This information can be about the implant, the surrounding patient anatomy, the actual implant and surgical instruments used, synthetic data that cannot be measured but is critical, etc. The system can also act as a notebook for the surgeon to write down any relevant information about the patient's anatomy, which may be useful at a later stage. This data is stored in the database for two purposes: 1) to complete the patient case file; and 2) to optimize future preoperative plans.

[0165] The information collected in this manner can be used as historical data to feed the decision support process described above. For example, capturing intraoperative measurements and observations allows the user to be presented with statistical information about patient characteristics that cannot be inferred from available medical images or can only be measured interventionally, such as ligament tension, the occurrence of infection, or soft tissue damage, etc., in a step prior to approving the preoperative plan. As another example, capturing information about intraoperative complications allows the user to be presented with statistical information about the likelihood of such complications in a step prior to approving the preoperative plan. Finally, capturing any changes made to the operating plan or any deviation from the approved preoperative plan allows the decision support process described in the "Planning Step" to be replaced or expanded from presenting information about choices made during the planning step to choices made during the operation.

[0166] Post-treatment data collection

[0167] After treatment, more information can be collected and captured through feedback loops, such as post-operative medical images, virtual 3D models based on such images, post-operative measurements, functional measurements, pain scores, functional scores, patient satisfaction information, information about post-operative complications, activity data, information about revision surgery… The data can be automatically collected, for example, through sensors embedded in one or more implants or wearable devices, or manually entered into an electronic access device.

[0168] The information collected in this way can be used as historical data to feed the above-mentioned decision support process. For example, it allows to present statistical information about actual surgical results, potential complication risks, implant life expectancy or patient satisfaction to the user in the steps before approving the preoperative plan.

[0169] Elimination of ineffective treatment plans

[0170] Special forms of intraoperative or postoperative feedback loops capture intraoperative and postoperative information about complications and use it to classify, label or flag preoperative plans that are less effective, for example based on how much the performance of the surgery deviated from the preoperative plan based on a certain threshold (which may be user-defined), the severity of the complication, or the longevity of the implant. This feedback loop allows for further optimization of the automatically created default plans by eliminating the least effective treatment plans from the training data that generated such default plans using AI-based techniques. This feedback loop also allows for improvement of the decision support system by eliminating the least effective treatment plans from data used in historical data analysis.

[0171] A very basic form of eliminating feedback loops allows the user to manually mark preoperative plans or treatment plans that should not be included in any training data or historical data analysis.

[0172] Presenting the results of historical data analysis

[0173] The information generated as a part of the decision support process can be presented to the user in any practical manner. For example, when supporting a decision (such as a selection between a plurality of treatment options or available implant sizes) involving a limited number of discrete options or discrete parameter values, a distribution curve graph or histogram can be shown for each of these options with a patient characteristic as an independent variable. The patient characteristic values ​​of the specific patient to be treated can be indicated on the graph by the marks on the independent axis, thereby showing to the user which decision-making option based on historical data seems to be most suitable for the patient.

[0174] For example, Fig. 9 An example of the results of an analysis of historical data is depicted in the form of distribution curves 902 to 906. The location of the patient to be treated within the patient population is indicated by vertical line 908. From this the user can infer that treatment B appears to be most appropriate.

[0175] This represents a significant improvement over conventional systems that only present suggestions of discrete treatment options or discrete parameter values ​​to the user. Fig. 9 In the example, the results of the historical data analysis are preferably presented to the user in an intuitive manner. Fig. 9 In the example above, the user does not simply get the suggestion "Treatment B". The user also sees where the patient is within the patient population and whether there are sharp or smooth transitions between the different options. For example, the user can deduce from the graph that Treatment B seems most appropriate, but Treatment A may be a possible competitor, while Treatment C is not. If the surgeon has other medical or non-medical reasons for preferring Treatment A over Treatment B, such as the cost of the treatment or his own lack of experience with Treatment B, the system of the present invention does not simply suggest Treatment B, but also teaches the user to consider Treatment A as a viable option and then provides the user with all relevant information about Treatment A.

[0176] Alternative representations are possible. For example, the data of the above graph can also be shown as an area graph or a bar graph. Alternatively, it can be shown in a gradient (e.g., color gradient) graph, where each of the decision options is represented by a specific color, pattern, or intensity (e.g., grayscale), and the distribution of the population across the decision options is represented by mixing proportional amounts of the corresponding colors, patterns, or intensities.

[0177] For example, Fig.10 Depicted is an example of the results of a historical data analysis represented in the form of a color map. The location of the patient to be treated within the patient population is indicated by a white dot. From this, the user can infer that treatment B seems most appropriate.

[0178] pass Fig.10 , the user can derive similar information from the distribution plots described above. Specifically, the user can infer where the patients are within the population, how the population is distributed across different treatment options or parameter values, and whether there are smooth or sharp transitions between these options and values ​​by looking at the color gradient. It may be more difficult to derive numerical values ​​from a colormap, but a colormap may be more intuitive to interpret.

[0179] In other embodiments, the analysis correlating the discrete options to the two patient characteristics may be presented in other visual ways, such as a 3D bar graph or other 2D graph (eg, using color, pattern, intensity, or other visual reference).

[0180] As another example, the results of an analysis of historical data that supports selection of a continuous valued parameter, such as bow leg correction for a knee implant, lateralization for a shoulder implant, implant depth for a heart valve, or patient satisfaction scores, can be presented via a line graph.

[0181] For example, Fig.11Depicted is an example of the results of a historical data analysis in the form of a line graph 1002. The location of the patient to be treated within the patient population is indicated by a vertical line 1004. From this the user can deduce that a value for parameter A between 0.1 and 0.2 appears to be most appropriate.

[0182] Examples (such as Fig.11 ) represents an improvement over conventional systems that only present suggestions for continuous parameter values ​​to the user. Fig.11 , the user does not just get a suggestion of "0.15". Instead, the user sees where the patient is located within the patient population and whether the parameter is very sensitive to patient characteristics within the general location of the patient. For example, the user can deduce from the line graph that a value of 0.15 for parameter A seems most appropriate, but that the value of parameter A does not vary greatly among patients similar to the patient to be treated. To give more information, the line graph can also show confidence intervals, for example by vertical bars (so-called "whiskers") or shaded areas around the value curve.

[0183] As another example, the result of historical data analysis linking intraoperative or postoperative events, observations or results, chances or risks with treatment decisions or parameters can also be shown with curve graphs, area graphs or bar graphs (optionally with confidence intervals) or color maps or pattern diagrams. The same way as the patient's location in other places is shown in the above examples, the current selection of decision options or parameter values ​​can be displayed. In certain embodiments, curve graphs, charts or color maps can be displayed together with the depiction of any equipment, instrument or implant that forms a part for planned treatment of the patient's anatomical structure and / or plan, such as 2D or 3D images, line graphs, medical images or virtual models. Curve graphs, charts, color maps and depictions can all be interactive, and changes in one of them may be automatically reflected in another.

[0184] For example, Fig.12 An example of the results of a historical data analysis represented in the form of a bar graph is depicted. Here, the risk of two complications is correlated with the selected device size. The currently chosen device size is indicated by a circle 1202, but other ways are possible, such as by the opacity, saturation, color, pattern, etc. of the bars in the graph. From this, the user can infer that device sizes 3 and 4 appear to be most appropriate.

[0185] Fig.13 A further example of the results of a historical data analysis in the form of a color map is depicted. Here, the risk of two complications is associated with selected parameter values. The currently selected parameter value is indicated by circle 1302. From this the user can deduce that the selected value is within the safety zone.

[0186] The various methods of displaying decision support data in the example figures described herein represent important improvements over conventional systems that merely present suggestions of decision options or parameter values ​​to a user. Fig.12 and Fig.13 In the representation shown, the user does not simply get a recommendation "Device size 3" or "Parameter X = x." Instead, the user sees what it means to deviate from the recommendation, how likely the outcome or risk of complications is, how much that chance or risk increases or decreases when changing the decision option or parameter value, and therefore how much wiggle room the user has with different decision options or parameter values. For example, the user can deduce from the plot that the current value of parameter X is within the safety zone, and that while it may be safe to increase the value slightly, decreasing it seems undesirable. One or more such historical data analysis representations can be displayed to the user at any given moment.

[0187] Sample Application: Shoulder Treatment Decision Support

[0188] Different treatments are available for shoulder-related conditions, depending on the pathology. For example, shoulder arthritis can be treated with rest, medications, corticosteroid injections, arthroscopic debridement, hemiarthroplasty, resection joint replacement, total (anatomical) shoulder replacement (TSA), reverse shoulder replacement (RSA), etc. Depending on the complexity of the pathology or treatment, some physicians may also choose to refer the patient to a colleague or another hospital, or follow the treatment options of one of their known peers.

[0189] The systems and methods of the present invention can assist physicians in deciding treatment based on patient characteristics and historical data.

[0190] For example, based on medical images of the patient's bone and / or cartilage anatomy, such as CT or MRI images, a virtual 3D model of the patient's shoulder anatomy can be created. The defect can be quantified in the above manner. The quantitative results can be presented to the user, for example, using Fig.14 The description in .

[0191] Specifically, Fig.14 An example of a representation of the results of defect quantification of the glenoid is depicted. On the left, a virtual 3D model 1402 of the bony anatomy of a patient's scapula, with the glenoid in the center. On the right, the anatomical structure 1404 is shown superimposed on an SSM 1406 representing the healthy anatomy, fitted to a portion of the patient's scapula. Different results of defect quantification are shown. The erosion depth (the distance from the actual bone surface to what that surface would have been in a healthy state, represented here by the surface of the SSM instance) is shown in the form of a gradient map 1408.

[0192] exist Fig.14In the example of , the erosion depth is calculated perpendicular to the best fit plane through the glenoid cavity surface of the SSM instance. Other measurement directions are also possible, such as locally perpendicular to the glenoid cavity surface of the SSM instance.

[0193] Additional measurements are calculated and shown, such as the percentage of vault loss (the percentage of the volume of the glenoid vault lost due to bone erosion), the percentage of eroded area (the percentage of the surface area of ​​the glenoid cavity affected by bone erosion), and the maximum erosion depth. In this example, the glenoid is also subdivided into four quadrants, and quantitative measurements such as the anterior, posterior, upper or lower vault loss percentages are shown in each quadrant. In addition, the subluxation distance is calculated. To this end, the rotation center of the humeral head is calculated by best fitting a sphere to the articular surface of the humeral head; the center point of the sphere is vertically projected onto the best fit plane through the glenoid cavity surface of the SSM instance; the distance between the projection point and the geometric center of the glenoid cavity is measured and displayed. Moreover, the subluxation area is shown, that is, the quadrant in which the rotation center of the humeral head is projected.

[0194] Fig.14 This represents a significant improvement over conventional systems because with this information and depiction, the user now has reproducible and objective information to assess the extent and location of the bone defect. This information is important in deciding the most appropriate treatment.

[0195] As described above, the system can also support decision making by presenting statistical information based on historical data. For example, a system including a feedback loop for an approved preoperative plan can run an analysis to correlate any of the metrics described above with treatments selected in previous cases. The results of this historical data analysis can be presented to the user in any of the ways described above. For example, the results can be presented in a graph, such as in Fig.15 middle.

[0196] Specifically, Fig.15 An example of a representation of the results of the historical data analysis is depicted and indicates the percentages of patients treated in different ways, sorted by percentage of vault loss. Patients to be treated are indicated by vertical lines 1502.

[0197] All records in the database can be used as the basis for historical data analysis. Alternatively, the group selected as the basis for historical data analysis may be limited in a variety of ways. For example, by limiting the group to only cases treated by the user, the user will gain insight into how the patient to be treated relates to his or her past experience. Including more or all of the user's cases will provide insight into the practice of a larger community of surgeons, such as all surgeons in a particular hospital, country, or world.

[0198] The group may also be limited to patients who show some similarity to the patient to be treated. This similarity may be based on one or more patient characteristics, such as gender, age, race, activity level, etc.

[0199] Referral to a colleague or another hospital may be one of the options. Based on the information stored in the database, the system may have the functionality to suggest clinicians who are open to referral. Based on historical data in the database, the system may even suggest clinicians who have more experience with similar patients, i.e., patients presenting with similar pathology and / or other patient characteristics, or suggest following the referring surgeon's treatment plan.

[0200] The historical data analysis has now been described based on approved preoperative plans collected and stored via feedback loops. However, similar and potentially more relevant analysis can be performed on intraoperative or postoperative data collected via other feedback loops. Such data may not represent the treatment that the surgeon intended to give, but rather the treatment that was actually administered.

[0201] Example application: Shoulder surgery implant type decision support

[0202] Similar to the previous example, the system described herein can provide support for decision making regarding which type of implant to use in a shoulder replacement, including, for example, an off-the-shelf implant versus a custom implant.

[0203] For example, the system can provide decision support in the form of historical data analysis, making the selection between standard or off-the-shelf implants and custom implants quantitatively related to the above-mentioned bone defects. The results of the analysis can be presented to the user in the form of a graph, chart, color map or pattern map or other examples such as described herein.

[0204] For example, Fig.16 Depicted is an example of a presentation of the results of an analysis of historical data correlating the selection between a standard implant and a custom implant to the percentage of vault loss of a patient's glenoid. The patient to be treated is indicated by a vertical line 1602.

[0205] Fig.17 An example of a presentation of the results of an analysis of historical data correlating the selection between a standard implant and a custom implant to the percentage of vault loss of a patient's glenoid is depicted. The patient to be treated is indicated by a circle 1702.

[0206] As another example, the system may be provided with a library of implants, and the analysis may correlate the selection of an implant to one or more defect characteristics calculated by defect quantification.

[0207] Example applications: Shoulder surgery, RSA lateralization

[0208] In reverse shoulder arthroplasty, lateralization of the center of rotation is often used as a way to improve the torque generated by the rotator cuff and increase internal and external rotation. However, excessive lateralization may lead to excessive muscle lengthening and even acromion fracture due to increased loading. Inadequate lateralization may lead to joint instability due to decreased muscle loading.

[0209] Therefore, the system described in this article can provide decision support by simulating muscle lengthening due to lateralization.

[0210] For example, the system can provide a 2D or 3D depiction of the patient's anatomy and implant. The depiction can include a virtual model of the bony anatomy of the scapula and humerus, the implant, and one or more shoulder muscles. The shoulder muscles can be shown in their actual shape, or more precisely, schematically, such as by a straight line, curve, broken line, or cylindrical shape. The depiction can simulate how the muscle trajectory changes with the lateralization of the implant and is displayed as a biomechanical model. For reference, the depiction may display the muscle trajectory and bone model in a native (i.e., preoperative or healthy) situation in an overlay manner. The preoperative situation can be derived from a medical image. The health status can be approximated by fitting the SSM representing the healthy shoulder anatomy to a portion of the patient's anatomy.

[0211] The system can be interactive.For example, as shown in 18 figures, the system can allow the user to manually shift the center of rotation from first position 1802 to second position 1804 by for example clicking and dragging the model of the implant manipulated by an input device (such as a computer mouse).Alternatively, the system can provide user interface controls, such as buttons or slide blocks, to adjust lateralization.Describing can be automatically updated to reflect the adjustment done.For example, the relative position of scapula, humerus and implant assembly and corresponding muscle track are updated.

[0212] The system can display a numerical value (such as a percentage) quantifying the amount of elongation of a single muscle, or a decrease in thickness according to a straight line, curve, polyline, or cylindrical shape, or an average for some or all muscles. These values ​​can be superimposed on a depiction of the anatomical structure, or listed elsewhere in the user interface.

[0213] The system according to the present invention can provide additional decision support by analyzing the historical data of past cases. As mentioned above, the data of this analysis can be collected in the form of an approved preoperative plan or an actual surgical plan collected during or after surgery through one or more feedback loops. The colony can be based on all available records, or can be limited in different ways as described above. In a preferred embodiment, the colony is limited to patients who show a certain similarity with the patient to be treated on one or more patient characteristics. For example, bone density can be derived from a CT scan and may play an important role in assessing the risk of acromion fracture. Alternatively or additionally, shape characteristics (such as the thickness of the acromion) may play an important role. As mentioned above, these shape characteristics can be quantified by some measured values ​​or by the parameter values ​​of the SSM fitted to the patient's anatomical structure. The results of the analysis can be displayed in the form of a curve graph, a chart or a color map, showing the frequency of these quantities that have been planned or implemented before for different lateralization amounts. The current lateralization can be indicated on a curve graph, a chart or a color map by a marker (such as a line, a point, a diamond, etc.).

[0214] Alternatively, the analysis may investigate how often the amount of muscle lengthening has been planned or performed before. This can be the muscle lengthening of a single muscle and an average of a selection of all shoulder muscles or a weighted average of a selection of all shoulder muscles.

[0215] Finally, in embodiments where the system collects and stores information about intraoperative or postoperative complications, the analysis can additionally include the risk of such complications, such as acromion fracture or instability. The user can then see from a graph, chart, or color map whether the selected lateralization is not only within customary practice, but also within a safe zone.

[0216] In addition to or in place of the above-mentioned interactive features, a graph, chart or color map can be interactive. For example, a user can select the lateralization amount by clicking on a graph, chart or color map or by sliding a marker representing the current lateralization amount. Any depiction of the anatomical structure and (a plurality of) planned implants can be automatically updated to reflect the change of lateralization.

[0217] The systems and methods described herein can be operated and executed by, for example, computing devices, such as desktop computers, portable computers, portable electronic devices, tablet computers, smart phones, and other computerized devices. In some embodiments, the methods described herein can be executed by native software applications, while in other embodiments, they can be executed in server-client embodiments. For example, in some embodiments, software configured to perform the methods described herein can be hosted by a remote server or a cloud-based system. In some cases, various aspects of the systems and methods described herein can be distributed on different computing devices.

[0218] Further, the systems and methods described herein are capable of being operated and performed by, for example, a medical professional (such as a surgeon, physician, or nurse) or a non-medical professional (such as a clinical technician, a design engineer, an implant manufacturer (e.g., to whom is outlined the types of implants used by a particular surgeon and a drawing is generated depicting the same to him), a resident student, or a patient (e.g., someone scheduling a surgery prior to the actual surgery)).

[0219] Example application: CMF treatment decision support

[0220] The defect sizing system described herein can also be used to detect and further classify one or more defects in the craniomaxillofacial (CMF) region, such as trauma; orbital reconstruction; distraction osteogenesis; temporomandibular joint; skull vault reconstruction; congenital craniofacial malformations, such as craniosynostosis; alveolar bone surgery; or any other cosmetic or reconstructive surgery involving one or more sites in the craniomaxillofacial region.

[0221] As an example embodiment, the defect sizing system described herein may use patient data (eg, imaging data) and one or more feedback loops to detect the type of defect to be sized and then classify the defect, such as an orthognathic defect.

[0222] In one example of the method, one or more medical images or scans (typically imaging data) of a patient's anatomical structure that requires correction may be acquired. For example, the imaging data may be associated with a jaw deformity of the patient. In this example, the imaging data may include, for example, image data of one or more of the patient's mandible, maxilla, or chin. As described above, the anatomical structure in the imaging data may be segmented (e.g., between the mandible, maxilla, and / or chin) to obtain a virtual three-dimensional surface model. A statistical shape model of a healthy anatomical structure (e.g., a healthy jaw) may then be fit to the three-dimensional surface model to identify healthy and damaged portions of the patient's anatomical structure (e.g., a lost portion of the patient's jaw).

[0223] Fig.19 An example of a representation of defect quantification using patient imaging data and SSM is depicted. In this example, the imaging data includes a three-dimensional model of the patient's skeletal mandibular anatomy 1902 superimposed on an SSM 1904 of an original healthy mandible.

[0224] It is clear that in this example, this patient only requires treatment of the mandible and not the maxilla.

[0225] Comparing the patient's actual anatomy (e.g., via a 3D model created from medical imaging data) to a healthy anatomy model (e.g., an SSM model) allows the surgeon to visualize possible surgical approaches. For example, in this case, the surgeon is able to manipulate the positioning of the mandible while providing the healthy anatomy as a reference. In this example, Fig.19 The defect shown and proposed surgical treatment may be identified, such as mandibular reconstruction.

[0226] During the planning phase, the system guides the surgeon by showing the parts of the anatomy that can be resected 2002, such as Fig. 20 Specifically, the system shows clear resection margins and can alert the surgeon if the surgeon decides to resect more or less bone than is necessary based on the quantified defect.

[0227] Further, as described above, the three-dimensional patient anatomical model can be accompanied by historical data associated with the patient, and patient-specific implants for planned treatment can be suggested. For example, the treatment plan can include the use of a bone graft, and based on the patient's medical history, the patient's left fibula can be selected for the graft. The system can also indicate healthy sites on the fibula and show postoperative results. It is worth noting that these are just a few examples.

[0228] In another example, a proposed treatment plan may involve treatment of additional CMF areas including maxillary, mandibular, and genioplasty.

[0229] Fig.21A An example is depicted where the defect is classified as LeFort I, a type of skull fracture involving the maxilla and surrounding structures in a horizontal, pyramidal, or transverse direction. For this classification, the treatment plan may involve bilateral sagittal split osteotomy (BSSO) and genioplasty osteotomy. As above, the model of the patient's anatomy is segmented into various regions 2102 to 2108, which can be used for defect sizing and considered during preoperative planning.

[0230] Fig. 21B Depicted in Fig.21A In particular, Fig. 21B The recommended maxillary movement distance to treat the defect is depicted. In some cases, the recommended distance may be based on historical data, and the system may also show the range 2110A and 2110B of possible maxillary movement (eg, in mm).

[0231] Fig. 21C An example of a proximal overlap and resection margin 2112 is depicted. Specifically, Fig. 21C Reduction of the identified bone anatomy in 2114 is necessary.

[0232] Fig.21D Another example of the proposed treatment of a defect is depicted. In this example, the system displays a warning 2116 that the gap needs to be filled.

[0233] In some embodiments, based on the quantified defect and the initial treatment plan, the system can also suggest relevant implant types and sizes to connect different bone parts, such as to fill Fig.21D For example, the system can suggest guidelines for placing a mandibular implant. The system can also allow the surgeon to visualize different implant options before making a selection and update the treatment plan accordingly.

[0234] Sample application: Jaw surgery decision support

[0235] Another example application of the surgical planning system described herein is orthognathic surgery decision support. In one example, a preoperative planning tool (e.g. Proplan CMF, etc.) can be used to generate a preoperative surgical plan for a specific craniomaxillofacial surgery. The imaging data from the preoperative planning tool can then be used by a defect sizing system, such as described herein.

[0236] In one example, the defect may be classified as a jaw deformity requiring orthognathic surgery to remedy the defect. In this example, the defect sizing system may quantify the defect based on various existing osteotomy classifications that surgeons are familiar with, such as Limberg's oblique subcondylar osteotomy, Moose's mandibular reduction procedure, Caldwell and Letterman's vertical ramus osteotomy, Trauner and Obwesefer's sagittal split osteotomy (SSO), bilateral sagittal split osteotomy (BSSO), Winstanley's intraoral vertical ramus osteotomy (IVRO), etc. If the defect is in the maxilla, the defect sizing system may also allow the user to visualize different fractures of the skull, such as Lefort I, Lefort II, modified LeFort I, etc.

[0237] Alternatively or additionally, the defect sizing system can also classify the defect based on the type of incision or surgery. For example, based on the defect and the generated SSM model, the user can choose to perform a bimaxillary (maxilla + mandible), multi-section maxilla, maxilla only, mandible only, or genioplasty surgery.

[0238] Once the defect is quantified, a default treatment plan can be created as described above. In some cases, three-dimensional cephalometric data (measuring deviations from norms), asymmetry assessments, and records of previous surgeries and other types of patient data stored with the patient file can be taken into account.

[0239] In one example, if the defect is in the mandible, a bilateral sagittal split osteotomy (BSSO) may be proposed by the surgical planning system. In some cases, this treatment may be performed without any treatment of the maxilla (upper jaw). The surgical planning system may allow a user (e.g., a surgeon or other practitioner) to visualize a mandibular surgical approach with appropriate changes in the maxilla, and allow the user to decide on the best approach.

[0240] In some embodiments, the system may also assist the surgeon in selecting the exact type of BSSO to be selected, such as Dalpont, Obwegeser, short ramus osteotomy, inverted L, and vertical ramus. Depending on the type of defect and osteotomy, the system may provide warnings, such as proximity to or damage to peripheral nerves, and suggest an appropriate osteotomy. The surgical planning system may also warn the user when too much or too little bone is being removed in the planned treatment. The surgical planning system may also prompt the user for appropriate resection margins and issue a warning when the margin is exceeded compared to historical data for a selected patient population (e.g., a population of which the patient being planned is a part).

[0241] Based on the type of osteotomy, the surgical planning system may also help the user decide on an appropriate fixation method (such as patient-specific or standard) and the area in which the fixation method will be placed. Some of the options available to the user may include selecting one or more plates, the type of plate (patient-specific or standard), the use of guidance, and / or the use of lag screws, etc.

[0242] In some embodiments, if the treatment plan involves treatment of the maxilla, the user may choose between two or more bone plates based on patient history, and may be able to compare the types and selections of multiple bone plates selected for similar patients using historical data analysis and / or patient population graphs.

[0243] In some embodiments, if the treatment plan involves treatment of the mandible, the surgical planning system may allow the user to visualize the positioning and orientation of the bone plates or lag screws, upper or lower fixation areas, etc. The surgical planning system may also allow for selection of thickness and width of the bone plates, fixation materials (e.g., CPTi, TAIV, bioresorbable) based on available bone volume, number and location of fixation screws on each side of the osteotomy, guidance for use with patient-specific bone plates or standard bone plates, etc. All of the above selections and configurations may become part of the treatment plan generated by the surgical planning system.

[0244] In an example embodiment, the defect sizing system may classify a patient as having a Class 2 narrow maxillary defect requiring treatment of mandibular advancement and maxillary impaction. A default treatment plan may include treatment of the maxilla, such as a multi-stage Lefort I osteotomy and a BSSO of the mandible. The default treatment plan may also recommend the use of a patient-specific bone plate for the maxilla and three lag screws on each side for the mandible. A user of the surgical planning system (e.g., a surgeon) may approve the default treatment plan, or may explore modifications to the plan through the surgical planning system's ability to visualize the treatment plan.

[0245] The user can then approve the treatment plan and use it during surgery (e.g., in the operating room). While in the operating room, changes or deviations from the treatment plan may be entered into the surgical planning system, such as the time required to perform surgical steps, anastomosis, ischemic time for bone graft harvest, required surgical equipment inspection before the start of surgery, blood loss, timed inspection of pathological tissue to determine accurate resection margins, etc.

[0246] After the surgery is completed, the patient file can be updated and certain data about the treatment can be generated for future preoperative surgical planning as well as historical data analysis, which can be used as described above. Likewise, other postoperative data may also be included in the patient file, such as infection rates, stability and recurrence rates, pain scores, discharge and related notes, orifice scans and notes, recurrence and recurrence rates in oncological cases, flap survival rates in reconstructive surgery, functional results and aesthetic results, etc.

[0247] Sample Application: Reconstructive Surgery Decision Support

[0248] Another example application of the surgical planning system described herein is in reconstructive surgery decision support. In one example, a preoperative planning tool (e.g. Proplan CMF, etc.) can be used to generate a preoperative surgical plan for a specific craniomaxillofacial surgery. The imaging data from the preoperative planning tool can then be used by a defect sizing system, such as described herein.

[0249] In one example, a defect may be classified as a deformity involving the mandible or midface, including reconstructive surgery. Based on patient archival data, such as patient history and patient imaging data, a three-dimensional SSM model of the patient may be generated. The imaging data (showing the defect) and the SSM may then be compared to generate a defect classification. Based on the defect classification, a default treatment plan may be generated by a surgical planning system, such as described herein.

[0250] In the case of cancer patients, the defect sizing system can quantify the defect based on the type of cancer and / or lesion (benign or malignant), the area of ​​lesion to be removed and treated during surgery, the number of surgeries required, and other factors. Any other patient information (such as other treatments, such as chemotherapy, radiation therapy, etc.) is also included in the patient file.

[0251] In the case of corrective surgery, the patient history can be taken into account during treatment planning. For example, based on the patient imaging data, the user (e.g., surgeon) is able to assess asymmetry and its deviation from the normal native anatomy. Using a three-dimensional model based on the patient data, the defect is simulated in comparison with the healthy anatomy.

[0252] In the case of trauma, the visualization capabilities of the surgical planning system can be used together with patient population and historical data analysis to effectively create an appropriate treatment plan. In some embodiments, the surgical planning system can recommend a default plan based on characteristics identified in trauma patients.

[0253] Further, historical data analysis performed by the surgical planning system may allow a user to compare the success rates of various surgical approaches for a specific indication, such as vascularized grafts versus bone non-vascularized grafts, autologous versus bone substitutes, and the like.

[0254] In some embodiments, the system may also store relevant information needed to match a donor with a recipient, and the surgical planning system may also provide information about other users to contact (e.g., other surgeons) or other facilities to contact with potential donors (e.g., other hospitals). In cases where tissue has already been harvested, the surgical planning system may display information about donor site morbidity, such as in the case of a harvested bone graft. In cases of trauma involving larger bone defects, the surgical planning system may prompt the user to use larger, stronger bone plates, and in some cases even patient-specific bone plates. It is worth noting that these are just some examples, and other examples are possible.

[0255] Example Application: Cardiac Therapy

[0256] Another example application of the surgical planning system described herein is cardiac therapy. In one example, a preoperative planning tool (e.g. MIMICS and MIMICS Enlight, etc.) can be used to generate preoperative surgical plans for structured cardiac and other vascular interventions. The imaging data from the preoperative planning tool can then be used by a defect quantification system, such as described herein.

[0257] For example, patient data including images, scans, patient history, etc. is stored by the surgical planning system. As described above, the imaging data can be converted into a three-dimensional model of the patient's anatomy. The SSM model can then be used by the defect quantification system to classify cardiac defects based on congenital or acquired diseases. In some examples, the defects can be classified as septal defects, valvular heart disease (such as heart disease of the aortic or mitral valve), vascular occlusions, fistulas, and other conditions. Each category can be further divided into several categories based on severity. Once the defect is quantified, a default treatment plan can be generated such as described above.

[0258] In one example, a patient may be identified as having a defective aortic valve, indicating a transcatheter aortic valve replacement (TAVR) procedure. Several factors can be determined from the 3D anatomical model and the SSM model as part of the defect classification system, which assists the user (e.g., surgeon) in generating a treatment plan, such as aortic valve morphology, assessment of the aortic root, annular assessment (size and height), LVOT calcification, sinotubular junction height, coronary ostium assessment (height), vulvar sinus assessment (diameter and height), coronary artery occlusion risk assessment, prediction of optimal fluoroscopic projection angles for device deployment, transfemoral access route assessment for TAVR devices, assessment of alternative routes when transfemoral is not feasible, assessment of feasibility of carotid artery protection devices, etc. These factors may influence treatment planning decisions, such as catheter planning, device selection, route planning in cases where a traditional transfemoral route is not accessible, incision size, device type, etc.

[0259] In one example, a patient may be identified as having a defect in the mitral valve, indicating a transcatheter mitral valve replacement (TMVR) procedure. As part of the defect classification system, multiple factors can be determined from the three-dimensional anatomical model and the SSM model, which assist the surgeon in generating a treatment plan, such as assessing the landing zone, including assessing the size of the mitral annulus (diameter, height, APML, leaflets), calcification, estimating the risk of left ventricular outflow tract (LVOT) obstruction, assessing the risk of interaction with other intercardiac devices (new or recently implanted or to be implanted), distance from such devices, determining the optimal transseptal puncture site or transapical route, assessing the optimal perspective angle, height of the papillary muscles, volume and size of the left ventricle, assessing the delivery device and route, angulation of the mitral valve, entry site, extent of the transseptal crossing (e.g., fossa ovalis), etc. These factors may affect, for example, the entry point, incision size, type and size of surgical devices, etc. For example, a user (e.g., a surgeon) can determine the entry point so that the apex / apical puncture is perpendicular to the mitral annulus to place the device. Using the defect sizing system along with historical data, surgeons may be able to predict the outcome of the neoLVOT procedure by placing the patient within a selected patient population using one or more visualization methods.

[0260] In one example, a patient may be identified as having a defect in the left atrial appendage (LAA), indicating that the LAA is closed. Multiple factors can be determined from the 3D anatomical model and the SSM model as part of the defect classification system, which assists the surgeon in generating a treatment plan, such as assessing the landing zone for device placement, determining the optimal transseptal puncture site, determining and assessing the optimal perspective projection angle for device delivery, selecting and planning the delivery device, catheter selection, and angulation to the LAA. Based on the diameter, height, depth, and shape of the LAA, the appropriate device and its size can be selected for the treatment plan.

[0261] Using historical data and patient populations, the surgical planning system can prompt the user for the type and size of equipment, catheter selection, and delivery routes as a few examples. Based on the severity of the disease, the age of the patient, the health risks involved, and the availability and feasibility of the equipment, the surgical planning system can prompt the user for alternative treatments. For example, open-heart surgery may be considered last. Based on historical data, the system can also store relevant information about catheter delivery and the path used, such as catheter deformation percentage, and alert the user to consider a more suitable catheter (if available).

[0262] Other structural heart interventions such as paravalvular leaks, atrial septal defects (ASD), patent foramen ovale (PFO) can also be planned using the surgical planning system described herein.

[0263] Further, intraoperative measurements may also be suggested, such as optimal viewing angles for fluoroscopy or C-arm angles to properly position the patient during surgery, and appropriate warnings may be provided during preoperative planning and via one or more navigation systems during the planned treatment (e.g., surgery).

[0264] For example, based on imaging data of the patient's anatomy (eg, CT or MRI images), a virtual three-dimensional model of the patient's heart anatomy can be made.The defect can be quantified in the manner described above.

[0265] exist Fig. 22 In the depicted example, the patient is identified as having a defective mitral valve.

[0266] Structural heart interventions (such as TMVR) involve placement of a mitral valve device 2202, such as Fig. 22 Based on a 3D model of the patient's heart, Fig. 22As depicted, a user (e.g., a surgeon) can determine the size, type, location, and place of an implant. Patient metrics such as angulation, the available cross-sectional area corresponding to the fluid passageway can be taken into account. Further, using one or more visualization tools of the surgical planning system, the risk of leakage can be determined while considering the implant type. Further, a delivery method and entry point may also affect the selection of an implant.

[0267] In another example, the current delivery route of the implant to be delivered may need to be determined for a patient who needs a LAA procedure. In this case, the selection of the catheter based on the patient's anatomy and its entry point and delivery trajectory needs to be planned so that during the procedure, the implant is safely delivered to the patient.

[0268] Fig.23 A target trajectory 2302 for delivering the implant is depicted. A user of the surgical planning system can experiment with different catheters before making a final treatment plan. Further, if the delivery path selected for the patient will lead to further complications, the surgical planning system can warn the user to reconsider the delivery path.

[0269] Sample application: Knee treatment decision support

[0270] Another example application of the surgical planning system described herein is for joint defects (e.g., ankle, hip), such as knee treatment decision support. In one example, a preoperative planning tool (e.g., SurgiCase Knee Planner can be used to generate preoperative surgical plans for joint replacements, such as knees. Patient data can be used by the surgical planning system, including medical imaging data (e.g., MRI and CT scans), patient history, preoperative PROM scores, new or revised surgical information, patellar height, axis, deformity type, etc., to generate a treatment plan.

[0271] For example, the defect quantification system can be used to classify the severity of the defect as requiring a total knee replacement or a partial knee replacement. As above, the defect quantification system can compare the three-dimensional model of the patient's anatomy with the SSM model to help quantify the defect. Before or during planning, information such as the type (standard or patient-specific) and size of the implant and information about bow legs / valgus angles, cartilage wear and other soft tissue data can be presented to the user so that the preoperative plan can be determined.

[0272] In some cases, the user can compare the generated default preoperative plan with a selected patient population and use historical data analysis, as described above. Specifically, the surgical planning system can present the user with information about why a certain type of implant is recommended, the location and position of the implant, the bow leg / valgus angle to be considered, and the system can enable the user to visualize how changing the implant characteristics will affect the patient's expected postoperative outcome.

[0273] For example, if the patient is young and active, the surgical planning system may extract data about treatment options for young patients and suggest that the user consider a partial knee replacement (PKA) instead of a total knee replacement (TKA). The surgical planning system may also suggest to the user to use guidance as well as patient-specific implants while showing the most suitable treatment option with minimal cartilage wear and tear.

[0274] The surgical planning system may also enable a user to view the treatment plan on a biomechanical model that includes bone and cartilage information and soft tissue data, such as ligament and muscle attachments. Further, the surgical planning system may also be configured to simulate the biomechanical model by rotation and translation and present data such as ligament elongation and knee joint loading, so that minimal damage is caused to the soft tissue surrounding the knee joint as a result of the treatment.

[0275] In some cases, the biomechanical model may be stored using one of the feedback loops described above and may be used as a reference (along with the navigation system) during surgery (in real time) so it can prompt the user (e.g., surgeon) and warn if the actual treatment deviates from the treatment plan or encounters other complications.

[0276] In some embodiments, intraoperative measurements may be stored, such as deviations from preoperative planning, soft tissue information, etc., to complete the patient profile and also create future preoperative planning and historical treatment data.

[0277] In some embodiments, intraoperative measurements including deviations from plan can be recorded by the surgical planning system, such as: need for cementation (tibia / femur), patella, approach, alignment technique, femoral rotation, femoral valgus, patellar release, medial and lateral release, level of satisfaction with balance achieved after surgery (e.g., unhappy, happy, very happy), blood loss, surgical time, range of motion during closure, use of robotic or other navigation systems, bone quality, diagnosis, PCL incision and size, limb alignment (bowed legs / neutral / valgus), joint spacing before incision opening (medial / lateral), joint compartment opening after implant placement (medial / lateral), laxity score (e.g., high / good / low), flexion contracture, ligament release, patellar remodeling, use of tibial and / or femoral guidance and guide fit, tibial slope, proximal tibial cut, tibial implant, confirmation that the planned implant is used or another size and type, tibial insert type and thickness, distal femoral cut, posterior femoral cut, AP femoral shift, anterior femoral cut, femoral implant rotation, ROM: maximum flexion, flexion balance, extension balance, etc.

[0278] Further, postoperative data (such as PROM scores currently used by the surgeon, such as KSS, KOOS, OKS, EQ5D, FJS), etc., as well as other input provided by the patient or his or her therapist during follow-up may also be recorded by the surgical planning system.

[0279] In one example, medical images of the patient's bone and / or cartilage anatomy (such as CT or MRI images) can be used to generate a three-dimensional model of the patient's knee joint 2402. The defect can be quantified in the manner described above.

[0280] For example, Fig.24 A representation of cartilage thickness on the bony anatomy of the knee joint (tibia and femur) is depicted. Certain identified areas (e.g., 2404) are considered healthy, such as where adequate amounts of cartilage are found. Other areas (e.g., 2406) indicate deficiencies, such as areas of weaker cartilage. This information can be used by a user (e.g., a surgeon) in deciding which treatment option to select for a treatment plan.

[0281] For example, a user may decide to treat a patient with a partial knee replacement rather than Fig.24 Based on this decision, the surgical planning system may suggest an implant, size, brand, and type from a variety of implants for this patient.

[0282] Further, the surgical planning system can be configured to allow the user to visualize the type and size of the implant for cartilage wear before making a final decision on the treatment plan. In some embodiments, the user can also use historical data and patient population analysis to compare the type of implant, such as described above.

[0283] Further, the surgical planning system may also be configured to display the bow leg / valgus angle 2502 for limb alignment, such as Fig.25 Depicted.

[0284] Similarly, the surgical planning system can be configured to display other patient metrics such as tibial slope, implant location and position, resection values, etc. via the three-dimensional model.

[0285] Historically, implants have been selected based on the patient's anatomy, such as Fig.26 The implant shown, for the patient's tibia, the user (e.g., surgeon) can also refine the position of the implant within the three-dimensional model. For example, if the implant is overhanging (as depicted in 2602), the surgical planning system can warn the user and can suggest that the user re-estimate the position of the implant. In some cases, if the appropriate position is not established, the surgical planning system may suggest a different implant.

[0286] Example Method

[0287] Fig. 27 An example method 2700 for classifying defects using a statistical shape model is depicted.

[0288] Method 2700 begins at step 2702 by acquiring medical image data associated with an anatomical structure of a patient.

[0289] Method 2700 then proceeds to step 2704 to create a three-dimensional anatomical model based on the medical image data.

[0290] The method 2700 then proceeds to step 2706 to fit the statistical shape model to the three-dimensional anatomical model.

[0291] The method 2700 then proceeds to step 2708 to determine one or more quantitative measurements based on the fitted statistical shape model.

[0292] The method 2700 then proceeds to step 2710 to classify a defect associated with the patient's anatomical structure based on the one or more quantitative measurements.

[0293] In some embodiments of method 2700, fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model.

[0294] In some embodiments of method 2700, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: when a corresponding topological region among the multiple topological regions is included in the subset of topological regions, if the fitting error exceeds a threshold, excluding the corresponding topological region.

[0295] In some embodiments of method 2700, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: selecting a first topological region from the multiple topological regions; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; and calculating a first fitting error based on the first fit of the statistical shape model, the first fit being based on the first topological region.

[0296] In some embodiments of method 2700, the first fitting error is calculated as a root mean square error (RMSE) between a plurality of points on the statistical shape model and a plurality of corresponding points on the three-dimensional anatomical model.

[0297] In some embodiments of method 2700, determining a subset of topological regions from a plurality of topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: determining that a first fitting error is below a threshold; selecting a second topological region from the plurality of topological regions; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; and calculating a second fitting error based on the second fit of the statistical shape model, the second fit being based on the second topological region.

[0298] In some embodiments of method 2700, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: determining that a first fitting error is above a threshold; and excluding a second topological region from the multiple topological regions from the subset of topological regions based on the first fitting error being above the threshold.

[0299] In some embodiments, the method 2700 further includes: excluding a third topological area from the plurality of topological areas from the subset of topological areas based on excluding the second topological area.

[0300] In some embodiments of method 2700, the threshold is about 1.7 mm. In some embodiments of method 2700, the threshold is in the range of 0.5 mm to 3 mm.

[0301] In some embodiments of method 2700, determining a subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model further comprises: excluding known damaged or deformed topological regions from the plurality of topological regions from the subset of topological regions.

[0302] In some embodiments of method 2700, classifying the defect based on the one or more quantitative measurements further comprises combining two or more classification systems to generate a three-dimensional classification, wherein each of the two or more classification systems is based on a different perspective of the patient's anatomy.

[0303] In some embodiments, method 2700 further includes creating a default treatment plan based on the classified defects associated with the patient's anatomy.

[0304] In some embodiments, method 2700 further includes acquiring patient data associated with a plurality of patients having the classified defects; selecting patient population data based on characteristics associated with the patients; and displaying a treatment options analysis comparing a plurality of treatment options based on the patient population data.

[0305] In some embodiments, method 2700 further includes displaying a patient reference on the treatment options analysis based on a characteristic associated with the patient.

[0306] In some embodiments, method 2700 further includes modifying a default treatment plan based on the treatment options analysis.

[0307] In some embodiments of method 2700, multiple treatment options relate to treatment of shoulder defects.

[0308] In some embodiments of method 2700, multiple treatment options relate to the treatment of a joint defect.

[0309] In some embodiments of method 2700, the plurality of treatment options relate to treatment of a diseased site of an anatomical structure.

[0310] In some embodiments of method 2700, the plurality of treatment options relate to treatment of a defect site in an anatomical structure.

[0311] Fig.28 An example method 2800 for determining a treatment for an anatomical defect is depicted.

[0312] Method 2800 begins at step 2802 by acquiring medical image data associated with an anatomical structure of a patient.

[0313] Method 2800 then proceeds to step 2804 to create a three-dimensional anatomical model based on the medical image data.

[0314] Method 2800 then proceeds to step 2806 to fit the statistical shape model to the three-dimensional anatomical model.

[0315] Method 2800 then proceeds to step 2808 to identify defects based on the three-dimensional anatomical model and the statistical shape model.

[0316] Method 2800 then proceeds to step 2810 to determine a default treatment based on the identified defect.

[0317] The method 2800 then proceeds to step 2812 to receive patient population data associated with a plurality of other patients having the identified impairment, wherein the patient population data includes a plurality of subsets of the patient population data associated with different treatments for the identified impairment.

[0318] The method 2800 then proceeds to step 2814 to generate a visualization including: a representation of each patient population data subset based on at least one patient characteristic; and a representation of the patients based on at least one patient characteristic.

[0319] Method 2800 then proceeds to step 2816 to select a final treatment for the patient.

[0320] In some embodiments of method 2800, the final treatment comprises a modified default treatment.

[0321] In some embodiments of method 2800, the final treatment comprises a default treatment.

[0322] In some embodiments, method 2800 further includes generating a new patient population data entry based on the treatment outcome and the selected treatment associated with the patient.

[0323] In some embodiments of method 2800, fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model.

[0324] In some embodiments of method 2800, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: when a corresponding topological region among the multiple topological regions is included in the subset of topological regions, if the fitting error exceeds a threshold, excluding the corresponding topological region.

[0325] In some embodiments of method 2800, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: selecting a first topological region from the multiple topological regions; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; and calculating a first fitting error based on the first fit of the statistical shape model, the first fit being based on the first topological region.

[0326] In some embodiments of method 2800, the first fitting error is calculated as a root mean square error (RMSE) between a plurality of points on the statistical shape model and a plurality of corresponding points on the three-dimensional anatomical model.

[0327] In some embodiments of method 2800, determining a subset of topological regions from a plurality of topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: determining that a first fitting error is below a threshold; selecting a second topological region from the plurality of topological regions; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; and calculating a second fitting error based on the second fit of the statistical shape model, the second fit being based on the second topological region.

[0328] In some embodiments of method 2800, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: determining that a first fitting error is above a threshold; and excluding a second topological region from the multiple topological regions from the subset of topological regions based on the first fitting error being above the threshold.

[0329] In some embodiments, the method 2800 further includes: excluding a third topological area from the plurality of topological areas from the subset of topological areas based on excluding the second topological area.

[0330] In some embodiments of method 2800, the threshold value is approximately 1.7 mm.

[0331] In some embodiments of method 2800, the threshold is in the range of 0.5 mm to 3 mm.

[0332] In some embodiments of method 2800, determining a subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model further comprises: excluding known damaged or deformed topological regions from the plurality of topological regions from the subset of topological regions.

[0333] In some embodiments of method 2800, the final treatment involves treatment of the shoulder defect.

[0334] In some embodiments of method 2800, the final treatment involves treatment of the joint defect.

[0335] In some embodiments of method 2800, the final treatment involves treatment of a diseased site of the anatomical structure.

[0336] In some embodiments of method 2800, the final treatment involves treatment of the defect site of the anatomical structure.

[0337] Fig.29 Example methods for determining treatment of an anatomical defect are depicted.

[0338] Method 2900 begins at step 2902 by acquiring medical image data associated with an anatomical structure of a patient.

[0339] Method 2900 then proceeds to step 2904 to create a three-dimensional anatomical model based on the medical image data.

[0340] Method 2900 then proceeds to step 2906 to fit the statistical shape model to the three-dimensional anatomical model.

[0341] The method 2900 then proceeds to step 2908 to identify defects based on the three-dimensional anatomical model and the statistical shape model.

[0342] The method 2900 then proceeds to step 2910 to receive a default treatment plan using analysis of historical data, wherein the historical data includes previously used pre-operative treatment plans for the identified defect.

[0343] The method 2900 then proceeds to step 2912 to optionally generate a visualization comprising: a representation of the treatment plan based on at least one patient characteristic; and a representation of the patient based on at least one patient characteristic.

[0344] Method 2900 then proceeds to step 2914 to approve final treatment for the patient.

[0345] In some embodiments of method 2900, the final treatment comprises a modified default treatment.

[0346] In some embodiments of method 2900, the final treatment comprises a default treatment.

[0347] In some embodiments, method 2900 further includes generating a new patient population data entry based on the treatment outcome and the selected treatment associated with the patient.

[0348] In some embodiments of method 2900, fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the statistical anatomical model.

[0349] In some embodiments of method 2900, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: when a corresponding topological region among the multiple topological regions is included in the subset of topological regions, if the fitting error exceeds a threshold, excluding the corresponding topological region.

[0350] In some embodiments of method 2900, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: selecting a first topological region from the multiple topological regions; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; and calculating a first fitting error based on the first fit of the statistical shape model, the first fit being based on the first topological region.

[0351] In some embodiments of method 2900, the first fitting error is calculated as a root mean square error (RMSE) between a plurality of points on the statistical shape model and a plurality of corresponding points on the three-dimensional anatomical model.

[0352] In some embodiments of method 2900, determining a subset of topological regions from a plurality of topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: determining that a first fitting error is below a threshold; selecting a second topological region from the plurality of topological regions; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; and calculating a second fitting error based on the second fit of the statistical shape model, the second fit being based on the second topological region.

[0353] In some embodiments of method 2900, determining a subset of topological regions from multiple topological regions for fitting a statistical shape model to a three-dimensional anatomical model also includes: determining that a first fitting error is above a threshold; and excluding a second topological region from the multiple topological regions from the subset of topological regions based on the first fitting error being above the threshold.

[0354] In some embodiments, the method 2900 further includes: excluding a third topological area from the plurality of topological areas from the subset of topological areas based on excluding the second topological area.

[0355] In some embodiments of method 2900, the threshold value is approximately 1.7 mm.

[0356] In some embodiments of method 2900, the threshold is in the range of 0.5 mm to 3 mm.

[0357] In some embodiments of method 2900, determining a subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model further comprises: excluding known damaged or deformed topological regions from the plurality of topological regions from the subset of topological regions.

[0358] In some embodiments of method 2900, the final treatment involves treatment of the shoulder defect.

[0359] In some embodiments of method 2900, the final treatment involves treatment of the joint defect.

[0360] In some embodiments of method 2900, the final treatment involves treatment of a diseased site of the anatomical structure.

[0361] In some embodiments of method 2900, the final treatment involves treatment of the defect site of the anatomical structure.

[0362] Example Processing System

[0363] Fig.30 An exemplary processing system 3000 configured to perform a method for detecting and removing personally identifiable information is depicted.

[0364] The processing system 3000 includes a CPU 3002 connected to a data bus 3008. The CPU 3002 is configured to process computer executable instructions stored, for example, in the memory 3010 or the storage device 3030, and causes the processing system 3000 to perform the functions described herein, for example, with respect to Figure 27 to Figure 29 CPU 3002 is included to represent a single CPU, multiple CPUs, a single CPU with multiple processing cores, and other forms of processing architectures capable of executing computer-executable instructions.

[0365] The processing system 3000 also includes input / output devices and interfaces 3004 that allow the processing system 3000 to interface with input / output devices, such as, for example, a keyboard, display, mouse device, pen input, touch-sensitive input device, camera, microphone, medical imaging device, and other devices that allow interaction with the processing system 3000. It is noted that although not depicted with separate external I / O devices, the processing system 3000 can interface with external I / O devices (e.g., external display devices) through physical and wireless connections.

[0366] The processing system 3000 also includes a network interface 3006 that provides the processing system 3000 with access to external computing devices, such as via a network 3009 .

[0367] The processing system 3000 also includes a memory 3010, which in this example includes various components configured to perform the functions described herein. In this embodiment, the memory 3010 includes an imaging component 3012, a modeling component 3014, a fitting component 3016, a quantification component 3018, a classification component 3020, a determination component 3022, a selection component 3024, a display 3026, and an identification component 3028. These various components may, for example, include computer executable instructions configured to perform the various functions described herein.

[0368] Note that although for simplicity in Fig.30 3010, various aspects stored in memory 3010 may be stored in different physical memories, but all accessible by CPU 3002 via internal data connections, such as bus 3012. For example, some components of memory 3010 may reside locally on processing system 3000, while in other embodiments, other components may be executed on a remote processing system or in a cloud-based processing system. This is just one example.

[0369] Processing system 3000 also includes storage device 3030, which in this example includes patient data 3032, medical imaging data 3034, patient population data 3036, treatment data 3038, surgical equipment data 3040, default planning data 3042, preoperative planning data 3044, intraoperative planning data 3046, postoperative planning data 3048, historical data and drawings 3050, and SSM model data 3052. Although not shown in FIG. Fig.30 , but other aspects may be included in storage device 3030.

[0370] As with memory 3010, single storage device 3030 is shown in FIG. Fig.30 3008 , but the various aspects stored in storage device 3030 may be stored in different physical storage devices, but all accessible by CPU 3002 via internal data connections (such as bus 3008 ) or external connections (such as network interface 3006 ).

[0371] Additional considerations

[0372] The previous description is provided so that any person skilled in the art can practice the various embodiments described herein. The examples discussed herein do not limit the scope, applicability or embodiments stated in the claims. Various modifications of these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments. For example, without departing from the scope of the present disclosure, changes can be made to the functions and arrangements of the elements discussed. Various examples can appropriately omit, replace or add various programs or components. For example, the described method can be performed in a different order than described, and various steps can be added, omitted or combined. Moreover, the features described with respect to some examples can be combined in some other examples. For example, a device can be implemented, or a method can be practiced using any number of aspects stated herein. In addition, the scope of the present disclosure is intended to cover such a device or method practiced using other structures, functionality, or structures and functionality other than or different from the various aspects of the present disclosure stated herein. It should be understood that any aspect of the present disclosure disclosed herein can be implemented by one or more elements of the claims.

[0373] As used herein, the word “exemplary” means “serving as an example, instance, or illustration.” Any aspect described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other aspects.

[0374] As used herein, a phrase referring to "at least one of" a list of items refers to any combination of those items, including single members. As an example, "at least one of the following: a, b, or c" is intended to cover a, b, c, ab, ac, bc, and abc, as well as any combination with multiples of the same item (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc, and ccc, or any other ordering of a, b, and c).

[0375] As used herein, the term "determining" encompasses a variety of actions. For example, "determining" may include calculating, computing, processing, deriving, investigating, searching (e.g., searching in a table, a database, or another data structure), ascertaining, etc. Also, "determining" may include receiving (e.g., receiving information), accessing (e.g., accessing data in a memory), etc. Also, "determining" may include parsing, selecting, choosing, establishing, etc.

[0376] The methods disclosed herein include one or more steps or actions for implementing these methods. Without departing from the scope of the claims, method steps and / or actions can be interchanged with each other. In other words, unless the specific order of steps or actions is specified, the order and / or use of specific steps and / or actions can be modified without departing from the scope of the claims. Further, the various operations of the above-mentioned methods can be performed by any suitable components that can perform corresponding functions. The components may include (multiple) various hardware and / or software components and / or (multiple) modules, including but not limited to circuits, application specific integrated circuits (ASICs) or processors. Typically, in the case of operations illustrated in the accompanying drawings, these operations may have corresponding corresponding components plus functional components, which have similar numbers.

[0377] The various illustrative logical blocks, modules, and circuits described in conjunction with the present disclosure may be implemented or performed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device (PLD), discrete gate or transistor logic, discrete hardware components, or any combination designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in an alternative, the processor may be any commercially available processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices (e.g., a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration).

[0378] The processing system can be implemented with a bus architecture. Depending on the specific application and overall design constraints of the processing system, the bus may include any number of interconnecting buses and bridges. The bus may link various circuits including a processor, a machine-readable medium, and an input / output device, etc. together. A user interface (e.g., a keypad, a display, a mouse, a joystick, etc.) may also be connected to the bus. The bus may also link various other circuits, such as timing sources, peripherals, voltage regulators, power management circuits, and other circuit elements well known in the art, and therefore will not be described further. The processor may be implemented with one or more general and / or special purpose processors. Examples include microprocessors, microcontrollers, DSP processors, and other circuit systems capable of executing software. Those skilled in the art will recognize how to best implement the described functionality of the processing system depending on the specific application and the overall design constraints imposed on the entire system.

[0379] If implemented in software, the function may be stored as one or more instructions or codes on a computer-readable medium or sent via a computer-readable medium. Software should be broadly interpreted as representing instructions, data, or any combination thereof, whether referring to software, firmware, middleware, microcode, hardware description language, or other. Computer-readable media include computer storage media and communication media, such as any medium that facilitates the transfer of computer programs from one place to another. The processor may be responsible for managing the bus and general processing, including the execution of software modules stored on a computer-readable storage medium. The computer-readable medium may be coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. In an alternative, the storage medium may be essential to the processor. By way of example, the computer-readable medium may include a transmission line, a carrier modulated by data, and / or a computer-readable storage medium separated from a wireless node on which instructions are stored, all of which may be accessed by the processor via a bus interface. Alternatively or in addition, the computer-readable medium or any part thereof may be integrated into the processor, such as the case where there may be a cache and / or a general register file. By way of example, examples of machine-readable storage media may include RAM (random access memory), flash memory, ROM (read only memory), PROM (programmable read only memory), EPROM (erasable programmable read only memory), EEPROM (electrically erasable programmable read only memory), registers, magnetic disks, optical disks, hard drives, or any other suitable storage media or any combination thereof. The machine-readable medium may be implemented in a computer program product.

[0380] A software module may include a single instruction or many instructions, and may be distributed over multiple different code segments, between different programs, and across multiple storage media. A computer-readable medium may include multiple software modules. A software module includes instructions that cause a processing system to perform various functions when executed by a device such as a processor. A software module may include a sending module and a receiving module. Each software module may reside in a single storage device or be distributed over multiple storage devices. By way of example, when a triggering event occurs, a software module may be loaded from a hard drive into a RAM. During the execution of a software module, the processor may load some instructions into a cache to increase access speed. One or more cache lines may then be loaded into a general register file for execution by the processor. When referencing the functionality of a software module, it is to be understood that this functionality is implemented by the processor when executing instructions from the software module.

[0381] The following claims are not intended to be limited to the embodiments shown herein, but to conform to the full scope consistent with the language of the claims. Within the claims, unless so expressly provided, reference to an element in the singular is not intended to mean "one and only one", but "one or more". Unless specifically provided, the term "some" refers to one or more. According to the provisions of 35 U.S.C. § 112 (f), no claim element is interpreted unless the element is explicitly stated using the phrase "a component for...", or in the case of a method claim, the element is stated using the phrase "a step for...". All structures and functions equivalent to elements throughout the various aspects of the present disclosure that are known or later known to those of ordinary skill in the art are expressly incorporated herein by reference and are intended to be covered by the claims. Moreover, whether or not such disclosure is explicitly stated in the claims, the content disclosed herein is not intended to be dedicated to the public.

Claims

1. A method for classifying an anatomical defect, comprising: acquiring medical image data associated with an anatomical structure of a patient from an electronic device; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; determining one or more quantitative measurements based on the fitted statistical shape model; as well as classifying a defect associated with the anatomical structure of the patient based on the one or more quantitative measurements, wherein the statistical shape model and the three-dimensional anatomical model are superimposed, and at least one of the statistical shape model and the three-dimensional anatomical model is shown to a user in a semi-transparent manner, Wherein fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model, Wherein determining the subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model further comprises: selecting a first topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; calculating a first fitting error based on a first fit of the statistical shape model, the first fit being based on the first topological region, determining that the first fitting error is below a threshold; selecting a second topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; and A second fit error is calculated based on a second fit of the statistical shape model, the second fit being based on the second topological region.

2. The method of claim 1 , wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: When a corresponding topological region among the plurality of topological regions is included in the topological region subset, if a fitting error exceeds a threshold, the corresponding topological region is excluded.

3. The method of claim 1, wherein the first fitting error is calculated as a root mean square error (RMSE) between a plurality of points on the statistical shape model and a plurality of corresponding points on the three-dimensional anatomical model. The method according to claim 1 , wherein the threshold value is 1.7 mm. The method according to claim 1 , wherein the threshold value is in the range of 0.5 mm to 3 mm.

6. The method of claim 1 , wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: Known damaged or deformed topological regions among the plurality of topological regions are excluded from the subset of topological regions.

7. The method of claim 1 , wherein classifying the defect based on the one or more quantitative measurements further comprises: Combining two or more classification systems in order to generate a three-dimensional classification, Each of the two or more classification systems is based on a different perspective of the anatomy of the patient.

8. The method according to claim 1, further comprising: A default treatment plan is created based on the classified defects associated with the anatomical structure of the patient.

9. The method according to claim 8, further comprising: acquiring patient data associated with a plurality of patients having the classified defects; selecting patient population data based on characteristics associated with the patient; as well as A treatment options analysis is displayed comparing a plurality of treatment options based on the patient population data.

10. The method according to claim 9, further comprising: A patient reference is displayed on the treatment options analysis based on the characteristics associated with the patient.

11. The method according to claim 10, further comprising: The default treatment plan is modified based on the treatment options analysis.

12. The method of claim 9, wherein the plurality of treatment options relate to treatment of a shoulder defect.

13. The method of claim 9, wherein the plurality of treatment options relate to treatment of a joint defect.

14. The method of claim 9, wherein the plurality of treatment options relate to treatment of a diseased portion of the anatomical structure.

15. The method of claim 9, wherein the plurality of treatment options relate to treatment of a defect site of the anatomical structure.

16. A method for determining treatment of an anatomical defect, comprising: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; identifying a defect based on the three-dimensional anatomical model and the statistical shape model; determining a default treatment based on the identified defect; receiving patient population data associated with a plurality of other patients having the identified impairment, wherein the patient population data includes a plurality of subsets of the patient population data associated with different treatments for the identified impairment; Generate visualizations, including: a representation of each patient population data subset based on at least one patient characteristic; and a representation of the patient based on the at least one patient characteristic; and Selecting the definitive treatment for the patient, wherein the statistical shape model and the three-dimensional anatomical model are superimposed, and at least one of the statistical shape model and the three-dimensional anatomical model is shown to a user in a semi-transparent manner, Wherein fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model, Wherein determining the subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model further comprises: selecting a first topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; calculating a first fitting error based on a first fit of the statistical shape model, the first fit being based on the first topological region, determining that the first fitting error is below a threshold; selecting a second topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; and A second fit error is calculated based on a second fit of the statistical shape model, the second fit being based on the second topological region.

17. The method of claim 16, wherein the final treatment comprises a modified default treatment. The method of claim 16 , wherein the final treatment comprises the default treatment.

19. The method according to claim 16, further comprising: A new patient population data entry is generated based on the treatment outcome associated with the patient and the treatment selected.

20. The method of claim 16, wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: When a corresponding topological region among the plurality of topological regions is included in the topological region subset, if a fitting error exceeds a threshold, the corresponding topological region is excluded.

21. The method of claim 16, wherein the first fitting error is calculated as a root mean square error (RMSE) between a plurality of points on the statistical shape model and a plurality of corresponding points on the three-dimensional anatomical model.

22. The method of claim 16, wherein the threshold is 1.7 mm.

23. The method of claim 16, wherein the threshold is in the range of 0.5 mm to 3 mm.

24. The method of claim 16, wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: Known damaged or deformed topological regions among the plurality of topological regions are excluded from the subset of topological regions.

25. The method of claim 16, wherein the final treatment involves treatment of a shoulder defect.

26. The method of claim 16, wherein the final treatment involves treatment of a joint defect.

27. The method of claim 16, wherein the final treatment involves treatment of a diseased portion of the anatomical structure.

28. The method of claim 16, wherein the final treatment involves treatment of a defect site of the anatomical structure.

29. An electronic system comprising: at least one memory; as well as at least one processor, the at least one processor being configured to: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; determining one or more quantitative measurements based on the fitted statistical shape model; as well as classifying a defect associated with the anatomical structure of the patient based on the one or more quantitative measurements, wherein the statistical shape model and the three-dimensional anatomical model are superimposed, and at least one of the statistical shape model and the three-dimensional anatomical model is shown to a user in a semi-transparent manner, Wherein fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model, Wherein determining the subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model further comprises: selecting a first topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; calculating a first fitting error based on a first fit of the statistical shape model, the first fit being based on the first topological region, determining that the first fitting error is below a threshold; selecting a second topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; and A second fit error is calculated based on a second fit of the statistical shape model, the second fit being based on the second topological region.

30. A non-transitory computer readable medium comprising instructions that, when executed by a system, cause the system to: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; determining one or more quantitative measurements based on the fitted statistical shape model; as well as classifying a defect associated with the anatomical structure of the patient based on the one or more quantitative measurements, wherein the statistical shape model and the three-dimensional anatomical model are superimposed, and at least one of the statistical shape model and the three-dimensional anatomical model is shown to a user in a semi-transparent manner, Wherein fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model, Wherein determining the subset of topological regions from the plurality of topological regions for fitting the statistical shape model to the three-dimensional anatomical model further comprises: selecting a first topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; calculating a first fitting error based on a first fit of the statistical shape model, the first fit being based on the first topological region, determining that the first fitting error is below a threshold; selecting a second topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; and A second fit error is calculated based on a second fit of the statistical shape model, the second fit being based on the second topological region.

31. A method for classifying an anatomical defect, comprising: acquiring medical image data associated with an anatomical structure of a patient from an electronic device; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; determining one or more quantitative measurements based on the fitted statistical shape model; as well as classifying a defect associated with the anatomical structure of the patient based on the one or more quantitative measurements, wherein the statistical shape model and the three-dimensional anatomical model are superimposed, and at least one of the statistical shape model and the three-dimensional anatomical model is shown to a user in a semi-transparent manner, Wherein fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model, Wherein determining the subset of topological regions from the multiple topological regions for fitting the statistical shape model to the three-dimensional anatomical model further includes: when a corresponding topological region among the multiple topological regions is included in the subset of topological regions, if the fitting error exceeds a threshold, excluding the corresponding topological region.

32. The method of claim 31 , wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: selecting a first topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; as well as A first fitting error is calculated based on a first fit of the statistical shape model, the first fit being based on the first topological region.

33. The method of claim 32, wherein the first fitting error is calculated as a root mean square error (RMSE) between a plurality of points on the statistical shape model and a plurality of corresponding points on the three-dimensional anatomical model.

34. The method of claim 32, wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: determining that the first fitting error is below a threshold; selecting a second topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; as well as A second fit error is calculated based on a second fit of the statistical shape model, the second fit being based on the second topological region.

35. The method of claim 32, wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: determining that the first fitting error is above a threshold; as well as Based on the first fitting error being higher than the threshold, a second topological region among the plurality of topological regions is excluded from the subset of topological regions.

36. The method of claim 35, further comprising: Based on excluding the second topological area, a third topological area among the plurality of topological areas is excluded from the subset of topological areas.

37. The method of claim 34, wherein the threshold is 1.7 mm.

38. The method of claim 34, wherein the threshold is in the range of 0.5 mm to 3 mm.

39. The method of claim 31 , wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: Known damaged or deformed topological regions among the plurality of topological regions are excluded from the subset of topological regions.

40. The method of claim 31 , wherein classifying the defect based on the one or more quantitative measurements further comprises: Combining two or more classification systems in order to generate a three-dimensional classification, Each of the two or more classification systems is based on a different perspective of the anatomy of the patient.

41. The method of claim 31 , further comprising: A default treatment plan is created based on the classified defects associated with the anatomical structure of the patient.

42. The method of claim 41 further comprising: acquiring patient data associated with a plurality of patients having the classified defects; selecting patient population data based on characteristics associated with the patient; as well as A treatment options analysis is displayed comparing a plurality of treatment options based on the patient population data.

43. The method of claim 42, further comprising: A patient reference is displayed on the treatment options analysis based on the characteristics associated with the patient.

44. The method of claim 43, further comprising: The default treatment plan is modified based on the treatment options analysis.

45. The method of claim 42, wherein the plurality of treatment options relate to treatment of a shoulder defect.

46. ​​The method of claim 42, wherein the plurality of treatment options relate to treatment of a joint defect.

47. The method of claim 42, wherein the plurality of treatment options relate to treatment of a diseased portion of the anatomical structure.

48. The method of claim 42, wherein the plurality of treatment options relate to treatment of a defect site of the anatomical structure.

49. A method for determining treatment for an anatomical defect, comprising: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; identifying a defect based on the three-dimensional anatomical model and the statistical shape model; determining a default treatment based on the identified defect; receiving patient population data associated with a plurality of other patients having the identified impairment, wherein the patient population data includes a plurality of subsets of the patient population data associated with different treatments for the identified impairment; Generate visualizations, including: a representation of each patient population data subset based on at least one patient characteristic; and a representation of the patient based on the at least one patient characteristic; and Selecting the definitive treatment for the patient, wherein the statistical shape model and the three-dimensional anatomical model are superimposed, and at least one of the statistical shape model and the three-dimensional anatomical model is shown to a user in a semi-transparent manner, Wherein fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model, Wherein determining the subset of topological regions from the multiple topological regions for fitting the statistical shape model to the three-dimensional anatomical model further includes: when a corresponding topological region among the multiple topological regions is included in the subset of topological regions, if the fitting error exceeds a threshold, excluding the corresponding topological region.

50. The method of claim 49, wherein the final treatment comprises a modified default treatment.

51. The method of claim 49, wherein the final treatment comprises the default treatment.

52. The method of claim 49, further comprising: A new patient population data entry is generated based on the treatment outcome associated with the patient and the treatment selected.

53. The method of claim 49, wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: selecting a first topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based only on the first topological region; as well as A first fitting error is calculated based on a first fit of the statistical shape model, the first fit being based on the first topological region.

54. The method of claim 53, wherein the first fitting error is calculated as a root mean square error (RMSE) between a plurality of points on the statistical shape model and a plurality of corresponding points on the three-dimensional anatomical model.

55. The method of claim 53, wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: determining that the first fitting error is below a threshold; selecting a second topological area from the plurality of topological areas; fitting the statistical shape model to the three-dimensional anatomical model based on the second topological region; as well as A second fit error is calculated based on a second fit of the statistical shape model, the second fit being based on the second topological region.

56. The method of claim 53, wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: determining that the first fitting error is above a threshold; as well as Based on the first fitting error being higher than the threshold, a second topological region among the plurality of topological regions is excluded from the subset of topological regions.

57. The method of claim 56, further comprising: Based on excluding the second topological area, a third topological area among the plurality of topological areas is excluded from the subset of topological areas.

58. The method of claim 55, wherein the threshold is 1.7 mm.

59. The method of claim 55, wherein the threshold is in the range of 0.5 mm to 3 mm.

60. The method of claim 49, wherein determining the subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model further comprises: Known damaged or deformed topological regions among the plurality of topological regions are excluded from the subset of topological regions.

61. The method of claim 49, wherein the final treatment involves treatment of a shoulder defect.

62. The method of claim 49, wherein the final treatment involves treatment of a joint defect.

63. The method of claim 49, wherein the final treatment involves treatment of a diseased portion of the anatomical structure.

64. The method of claim 49, wherein the final treatment involves treatment of a defect site of the anatomical structure.

65. An electronic system comprising: at least one memory; as well as at least one processor, the at least one processor being configured to: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; determining one or more quantitative measurements based on the fitted statistical shape model; as well as classifying a defect associated with the anatomical structure of the patient based on the one or more quantitative measurements, wherein the statistical shape model and the three-dimensional anatomical model are superimposed, and at least one of the statistical shape model and the three-dimensional anatomical model is shown to a user in a semi-transparent manner, Wherein fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model, Wherein determining the subset of topological regions from the multiple topological regions for fitting the statistical shape model to the three-dimensional anatomical model further includes: when a corresponding topological region among the multiple topological regions is included in the subset of topological regions, if the fitting error exceeds a threshold, excluding the corresponding topological region.

66. A non-transitory computer readable medium comprising instructions that, when executed by a system, cause the system to: acquiring medical image data associated with an anatomical structure of a patient; creating a three-dimensional anatomical model based on the medical image data; fitting a statistical shape model to the three-dimensional anatomical model; determining one or more quantitative measurements based on the fitted statistical shape model; as well as classifying a defect associated with the anatomical structure of the patient based on the one or more quantitative measurements, wherein the statistical shape model and the three-dimensional anatomical model are superimposed, and at least one of the statistical shape model and the three-dimensional anatomical model is shown to a user in a semi-transparent manner, Wherein fitting the statistical shape model to the three-dimensional anatomical model further comprises: subdividing the statistical shape model into a plurality of topological regions; and determining a subset of topological regions from the plurality of topological regions for use in fitting the statistical shape model to the three-dimensional anatomical model, Wherein determining the subset of topological regions from the multiple topological regions for fitting the statistical shape model to the three-dimensional anatomical model further includes: when a corresponding topological region among the multiple topological regions is included in the subset of topological regions, if the fitting error exceeds a threshold, excluding the corresponding topological region.

Images