Patient-Specific Planning Literature Roundup

Posted on 15 March 2023 by Kerim Genc

 

Following our recent post about point of care (POC) 3D printing work conducted by users of Simpleware software, we are pleased to provide another digest, this time focusing on research into patient-specific planning. 3D models derived from patient data (such as MRI and CT) provide valuable insights into pathologies and implant designs. As we move forward, the use of automated segmentation in our Simpleware AS Ortho/CMF and AS Cardio modules is showing great potential in reducing traditional bottlenecks associated with image data, leading to greater innovation. Below are some highlights of recent papers featuring Simpleware software.

Patient-Specific Brain Arteries Modeled as a Flexible Phantom Model Using 3D Printed Water-Soluble Resin

Nilsson, D.P.G., Holmgren, M., Holmlund, P., Wåhlin, A., Eklund, A., Dahlberg, T., Wiklund, K., Andersson, M., 2022. Patient-specific brain arteries molded as a flexible phantom model using 3D printed water-soluble resin. Nature: Scientific Reports, 12.

PDMS phantom model of blood vessels (CC BY 4.0)

PDMS has a flexible characteristic that can mimic real blood vessels and is therefore used in the phantom model. In panel (a) and (b), we show the effect on the model before and during the surface is depressed with a thumb, respectively. The channels are filled with a blood like substance (water and food dye) and an external force reduces the size of the channel significantly (as indicated by arrows). Showing possible applications of these compliant phantom models, for example, as to simulate flow restrictions in blood vessels (Image by Nilsson et al. / CC BY 4.0).

Context

"Visualizing medical images from patients as physical 3D models (phantom models) have many roles in the medical field, from education to preclinical preparation and clinical research. However, current phantom models are generally generic, expensive, and time-consuming to fabricate. Thus, there is a need for a cost- and time-efficient pipeline from medical imaging to patient-specific phantom models. In this work, we present a method for creating complex 3D sacrificial molds using an off-the-shelf water-soluble resin and a low-cost desktop 3D printer. This enables us to recreate parts of the cerebral arterial tree as a full-scale phantom model ([Formula: see text] cm) in transparent silicone rubber (polydimethylsiloxane, PDMS) from computed tomography angiography images (CTA). We analyzed the model with magnetic resonance imaging (MRI) and compared it with the patient data. The results show good agreement and smooth surfaces for the arteries. We also evaluate our method by looking at its capability to reproduce 1 mm channels and sharp corners. We found that round shapes are well reproduced, whereas sharp features show some divergence. Our method can fabricate a patient-specific phantom model with less than 2 h of total labor time and at a low fabrication cost."

Use of Simpleware Software

"The segmentation was performed with Synopsys’ Simpleware™ software (ScanIP P-2019.09, Synopsys, Inc., Mountain View, USA) and the exported CAD generated with the Simpleware FE module. The original image data had a resolution of 510×500μm in the trans-axial plane, and a slice thickness of 400μm. The image data originally covered the entire cranium but was before the segmentation cropped to only include the cerebral arteries of interest. The image volume was thereafter resampled with linear interpolation to obtain an isotropic resolution of 300μm. Before segmentation, we used an edge-preserving bilateral filter to reduce background noise from surrounding tissue. We only analysed the anterior part of the circle of Willis, as motivated in the previous study28. A coarse segmentation of the cerebral arterial tree was extracted by a threshold filter, from which the software specific gradient-based filter ’Local surface correction’ was used for wall detection. This filter used the image background intensity to adjust the segmented surface. Since we were only interested in the major arteries, smaller branches were manually removed, in addition to remaining parts of the scull-bone. The segmentation was finalized by applying a volume and topology smoothing filter on the mask. The CAD-file was exported with a target maximum and minimum element size of 600μm and 300μm, respectively, aiming for the interpolated image resolution."

Outcomes and Impact

"In conclusion, the newly developed water-soluble resin for 3D printers has provided a quick and easy way to fabricate patient-specific phantom models in flexible and transparent silicone rubber. These have suitable properties for a range of applications. For example, for medical education, the ability to rapidly produce cheap phantom models is an important step forward both in theoretical teaching (visualization) and for practical (rehearsal) training. Also, these phantom models can be used for training in MRI/CT imaging or Doppler ultrasound measurements. Finally, phantom models can help with dynamic flow simulations in clinical neuroscience as well as in personalized medicine."

Development of Bony Range of Motion (B-Rom) Boundary for Total Hip Replacement Planning

Palit, A., King, R., Pierrepont, J., Williams, M.A., 2022. Development of bony range of motion (B-ROM) boundary for total hip replacement planning. Computer Methods and Programs in Medicine, 222.

Determination of HJC using best fit sphere on the femoral head (CC BY 4.0)

Determination of HJC using best fit sphere on the femoral head. (a) Manual selection of points on the spherical articular surface of femoral head where the coordinates of each selected points are shown in square boxes with x, y and z values; (b) Construction of best fit sphere (blue) using the selected points. The centre of the sphere is defined as HJC (yellow) (Image by Palit et al. / CC BY 4.0).

Context

"Pre-operative surgical planning using computer simulation is increasingly standard practice before Total Hip Arthroplasty (THA), in order to determine the optimal implant positions, and thereby minimise post-operative complications such as dislocation, wear and leg length discrepancy. One of the limitations of current methods, however, is the lack of information on the subject-specific reference range of motion (ROM) that could be used as targets for surgical planning. Only a limited number of hip motions are considered, which are neither subject-specific, nor representative of all the hip motions associated with all the activities of daily livings (ADLs). In this paper, therefore, a method was developed to calculate subject-specific representative bony range of motion (B-ROM) that would cover all the possible joint motions and presented in terms of pure joint motions."

Use of Simpleware Software

"Construction of bony geometries from CT images and identification of landmarks was performed using Simpleware™ ScanIP software (Synopsys, Inc., Mountain View, USA). The DICOM CT images were imported in ScanIP followed by cubic/isotropic resampling and cropping the region of interest. The pelvis and femur were then semi automatically segmented using lower and upper greyscale threshold values, and subsequently, morphological close operation was performed to close the holes within the segmented masks. Finally, the 3D geometry was constructed from the mask, and subsequently, exported as STL files for the simulation. Four landmarks were identified for pelvis as follows: (a) right anterior superior iliac spines (ASISRight), (b) left anterior superior iliac spines (ASISLeft), (c) right pubic tubercles (PTUBRight), (d) left pubic tubercles (PTUBLeft). In case of femur geometry, two landmarks were identified: (a) lateral Epicondyle. (EPILateral), (b) medial Epicondyle. (EPIMadial)."

Outcomes and Impact

"The method encompasses every potential ADL, and as a result, more comprehensive surgical planning is possible, as the implant positions can be optimised in order to maximise impingement-free ROM, and consequently minimise clinical complications."

Intramuscular Fat in Gluteus Maximus for Different Levels of Physical Activity

Belzunce, M.A., Henckel. J., Di Laura, A., Hart, A., 2022. Intramuscular fat in gluteus maximus for different levels of physical activity. Nature: Scientific Reports, 11.

Intramuscular fat in gluteus maximus (CC BY 4.0)

FF measurement of GMAX. (A) An automated tool is used to label the in-phase Dixon image. (B) A FF image is created from fat and water images where the labels are overlaid and landmarks for the bulk muscle are identified. (C) FF of right and left GMAX are estimated by masking the FF image with the labels. A coronal view of the masked FF image is shown, where the central region between dashed lines is used to estimate the FF. On the left an axial slice of the masked image (highlighted in the coronal view) is found. On the right, a 3D view of the FF in both gluteus maximus is shown where a dark signal indicates fat content (Image by Belzunce et al. / CC BY 4.0).

Context

"We aimed to determine if gluteus maximus (GMAX) fat infiltration is associated with different levels of physical activity. Identifying and quantifying differences in the intramuscular fat content of GMAX in subjects with different levels of physical activity can provide a new tool to evaluate hip muscles health. This was a cross-sectional study involving seventy subjects that underwent Dixon MRI of the pelvis. The individuals were divided into four groups by levels of physical activity, from low to high: inactive patients due to hip pain; and low, medium and high physical activity groups of healthy subjects (HS) based on hours of exercise per week. We estimated the GMAX intramuscular fat content for each subject using automated measurements of fat fraction (FF) from Dixon images. The GMAX volume and lean volume were also measured and normalized by lean body mass. The effects of body mass index (BMI) and age were included in the statistical analysis. The patient group had a significantly higher FF than the three groups of HS (median values of 26.2%, 17.8%, 16.7% and 13.7% respectively, p < 0.001). The normalized lean volume was significantly larger in the high activity group compared to all the other groups (p < 0.001, p = 0.002 and p = 0.02). Employing a hierarchical linear regression analysis, we found that hip pain, low physical activity, female gender and high BMI were statistically significant predictors of increased GMAX fat infiltration."

Use of Simpleware Software

"We estimated the GMAX IMF content using an automated tool that computes the FF of GMAX from Dixon images32,36. The tool is an in-house plugin for Simpleware ScanIP (Version 2020.6; Synopsys, Inc., Mountain View, USA) based on a multi-atlas segmentation method that employs a library with 15 manually segmented Dixon scans. Different to most of the studies looking at fat-content, where cross-sectional areas (CSA)37,38,39 of the muscles were used, our tool obtains the mean FF of the full 3D muscle volume. The method automatically generates 3D labels for left and right GMAX in the in-phase Dixon image, being a label the sets of voxels that represent each muscle in the image. Next, the mean GMAX FF value is computed by averaging the label voxel values of the FF image, being the latter the ratio image between the fat image and the sum of the water and fat images. In average, 1.2 × 106 voxels were averaged per GMAX label. In32, we showed that the FF error for the comparison of two small groups of subjects was lower than 0.6%. The resulting mean FF value of each label quantifies GMAX IMF. The latter, which is the adipose tissue depot between and among skeletal muscle fibres in the skeletal muscle bed18,40, needs to be distinguished from IMAT, which is fat underneath the deep fascia and between adjacent muscle groups."

Outcomes and Impact

"In this work, we used an automated tool to compute FF in GMAX muscle as a measure of GMAX IMF content in four groups of subjects with different levels of physical activity. We report novel quantitative data of GMAX IMF content for these groups that ranged from patients with reduced physical activity due to hip pain to highly active subjects reporting more than 8 h of physical activity per week, included experienced marathon runners. The group of patients with hip pain had a significantly higher IMF content than the three groups of HS and the group of high levels of activity had a significantly lower IMF content than the low activity group. We found that hip pain, low physical activity, female gender, BMI and left side are predictors of GMAX fat infiltration, even within healthy subjects. These results show that automated measurements of GMAX FF could be used to quantitatively assess muscle health, mobility and sarcopenia in clinical trials and clinical practice."

The In Vivo Location of Edge-Wear in Hip Arthroplasties: Combining Pre-Revision 3D CT Imaging with Retrieval Analysis

Bergiers, S., Hothi, H., Henckel, J., Di Laura, A., Belzune, M., Skinner, J., Hart, A., 2021. The in vivo location of edge-wear in hip arthroplasties: combining pre-revision 3D CT imaging with retrieval analysis. Bone Joint Research, 10, pp. 639-649.

Models of a pelvis and Birmingham hip replacement (BHR) implant in Simpleware ScanIP (CC BY 4.0)

Computational models of a pelvis and Birmingham hip replacement (BHR) implant, generated within Synopsys Simpleware software. The "For Review Only" standardized reference system defined to measure the in vivo location of edge-wear is illustrated (Cup-APP (CAPP)), which is parallel to the anterior pelvic plane (APP) and intersects the centre of the cup opening (Image by Bergiers et al. / CC BY 4.0).

Context

"Acetabular edge-loading was a cause of increased wear rates in metal-on-metal hip arthroplasties, ultimately contributing to their failure. Although such wear patterns have been regularly reported in retrieval analyses, this study aimed to determine their in vivo location and investigate their relationship with acetabular component positioning."

Use of Simpleware Software

"First, a pre-revision, full pelvis CT scan of each patient was imported into Synopsys Simpleware ScanIP (Synopsys, USA) as an anonymized stack of Digital Imaging and Communications in Medicine (DICOM) images. These were compiled to form a 3D image composed of voxels, each with their own grayscale intensity. Material density dictated X-ray attenuation during CT imaging and consequently the grayscale values of these voxels, allowing computational models of the patient’s pelvis (bone) and BHR hip implant (CoCrMo) to be segmented from other materials."

"This was achieved using a semi-automated process called ‘thresholding’, which involved defining a range of grayscale values representative of each material. A grayscale range of approximately 130 to 1,200 was used to isolate bone from each image stack, while the BHR implants were segmented using a range of grayscale values often above 1,800. The authors’ discretion was required as the grayscale varied in magnitude, based on scanning parameters. Computational post-processing tools were used sparingly to remove metal artifacts from the models, without impacting anatomical dimensions. An automated registration function was used to co-register the BHR CAD model to the actual acetabular component segmented from the CT scan. Manual input was required to refine the resulting fit, which involved selecting landmarks such as the rim and stabilizing fins to inform the alignment."

"The in vivo location of acetabular edge-wear was determined within the Simpleware ScanIP software. As with the measurements of component positioning, the APP was used as a standardized reference between patients. A plane termed the Cup-APP (CAPP) was defined, which was parallel to the APP and intersected the centre of the BHR cup opening. The two points at which the CAPP intersected the acetabular component rim were used to define the 0° and 180° limits of the measurement system. Vertical and horizontal axes were formed from these points, within the acetabular rim plane, dividing the articulating surface into four quadrants. As a result of its relationship to the APP, 0° was considered representative of the vertical standing pelvic position.14,15 Component position was then neutralized, achieving a perpendicular view of the acetabular component rim, with the vertical axis positioned appropriately. The angle between both limits of the edge-wear scar were measured, with respect to the vertical axis. As both right- and left-sided implants were included in this study, the anterior and posterior halves of the acetabular rim were represented by positive and negative angles, respectively."

"To visualize the in vivo location of wear with respect to the pelvic anatomy, the implant and its hemipelvis were exported from Simpleware ScanIP. This was re-imported into SolidWorks, where the original CAD model and co-registered wear map could be re-aligned to the stabilizing fins."

Outcomes and Impact

"In conclusion, acetabular edge-wear was found to be predominantly centred about an anterosuperior location in vivo, in agreement with previous reports of hip joint forces being directed anteriorly during a greater portion of walking gait. As edge-wear was consistently located at the superior acetabular edge, it also supports the contribution of clearance, arc of coverage angle, and inclination to instances of edge-loading in MOM hip arthroplasties as previously theorized. For the first time, retrieval evidence was found to suggest the influence of acetabular anteversion on the AP location of the bearing contact patch and on edge-wear. Further adoption of this novel method could provide an insight into the distribution of load through hip arthroplasties and improve the current definition of the optimal and safe zones for acetabular component positioning."

Predicting Risk of Sudden Cardiac Death in Patients with Cardiac Sarcoidosis Using Multimodality Imaging and Personalized Heart Modeling in a Multivariable Classifier

Shade, J.K., Prakosa, A., Popescu, D.M., Yu, R., Okada, D.R., Chrispin, J., Trayanova, N.A., 2021. Predicting risk of sudden cardiac death in patients with cardiac sarcoidosis using multimodality imaging and personalized heart modeling in a multivariable classifier. Science Advances, 7, 31.

CHAI Risk Predictor method for predicting risk of sudden cardiac death (CC BY 4.0)

The multivariable CHAI Risk Predictor synergistically combines mechanistic modeling and ML. In the first step (left), LGE-MRI and FDG-PET images are used, together with electrophysiological data, to create personalized MRI-PET fusion models. Mechanistic computational modeling of cardiac electrical function is performed to investigate the arrhythmia propensity of the CS patient’s heart. In the second step (right), a supervised ML algorithm is trained and optimized to predict the risk of SCD due to ventricular arrhythmia using features selected by a random forest algorithm from three types of inputs: (i) features characterizing the patient’s arrhythmogenic propensity extracted from the results of mechanistic simulations (yellow arrow), (ii) features extracted from clinical images characterizing heterogeneity in disease remodeling in the patient’s ventricles (orange arrow), and (iii) baseline patient data (red arrow). (Image by Shade et al. / CC BY 4.0).

Context

"Cardiac sarcoidosis (CS), an inflammatory disease characterized by formation of granulomas in the heart, is associated with high risk of sudden cardiac death (SCD) from ventricular arrhythmias. Current “one-size-fits-all” guidelines for SCD risk assessment in CS result in insufficient appropriate primary prevention. Here, we present a two-step precision risk prediction technology for patients with CS. First, a patient’s arrhythmogenic propensity arising from heterogeneous CS-induced ventricular remodeling is assessed using a novel personalized magnetic-resonance imaging and positron-emission tomography fusion mechanistic model. The resulting simulations of arrhythmogenesis are fed, together with a set of imaging and clinical biomarkers, into a supervised classifier. In a retrospective study of 45 patients, the technology achieved testing results of 60% sensitivity [95% confidence interval (CI): 57-63%], 72% specificity [95% CI: 70-74%], and 0.754 area under the receiver operating characteristic curve [95% CI: 0.710-0.797]. It outperformed clinical metrics, highlighting its potential to transform CS risk stratification."

Use of Simpleware Software

"A 3D mesh with a mean edge length of 350 μm was created from the segmented images using Synopsys’ Simpleware ScanIP software. Segmented regions of fibrosis and inflammation were interpolated to 350-μm3 voxel resolution and mapped onto the mesh using a previously described method (9, 26). Fiber orientations in the mesh were assigned on a per-element basis using a validated approach (30) used in a number of previous publications (9, 10, 20, 26). Briefly, the Laplace-Dirichlet method was used to define transmural and apicobasal directions at every point in the ventricles, and then bidirectional spherical linear interpolation was applied to assign fiber orientations based on a set of rules."

Outcomes and Impact

"ML and mechanistic modeling have historically been distinct approaches. The CHAI Risk Predictor illustrates how they can be used synergistically and suggests how to overcome concerns over clinical decisions being informed by “black-box” algorithms that lack explainability. The technology developed here paves the way for the use of integrative approaches in precision medicine that predict adverse events in complex diseases with a high degree of accuracy and mechanistic underpinning. The merging of computational modeling and data science with medicine, as exemplified by the CHAI technology, is poised to lead to major advances in the improvement of patient care."

Learn More

Any Questions?

If you would like to know more, please contact us. Our technical specialists will be happy to show you Simpleware software solutions for any anatomy.