Subject-specific shoulder muscle attachment region prediction using statistical shape models: A validity study

Salhi, Asma; Burdin, Valérie; Mutsvangwa, Tinashe; Sivarasu, Sudesh; Brochard, Sylvain; Borotikar, Bhushan

doi:10.1109/embc.2017.8037154

Cited by 7 publications

(7 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because the scapular SSM and fitting method showed good accuracy for the prediction of landmarks and measurements in previous studies, 25,26 we did not re-evaluate this ability for the current muscle attachment points. Moreover, as demonstrated by Salhi et al, 29 the muscle attachment points on the scapula are expected to be indicated with higher accuracy than the muscle attachment points on the humerus because of the more distinct anatomic regions of the scapula. Third, we defined the muscle attachment points instead of using the complete muscle attachment regions.…”

Section: Discussionmentioning

confidence: 95%

“…17,24 Our results are comparable to those of other SSM-based landmark identification methods described in the literature. Salhi et al 29 investigated the prediction accuracy of subjectspecific muscle attachment regions using a healthy scapular and humeral SSM. For the humerus, they reported an average RMS error between 0.4 and 1.7 mm and a Hausdorff distance between 1.6 and 4.8 mm for all muscle attachment regions.…”

Section: Discussionmentioning

confidence: 99%

“…25,26,31,33 Statistical shape models (SSMs) are fitted to a target bone model, thereby better maintaining point correspondence and being less dependent on bone shape in comparison to morphing methods. Salhi et al 29 evaluated the prediction accuracy of muscle attachment regions on the scapula and humerus using an SSM approach. Although good accuracy was reported, the study used only a limited evaluation data set and did not report the errors that can be expected on muscle elongation measurements.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Automated muscle elongation measurement during reverse shoulder arthroplasty planning

Pitocchi

Plessers

Wirix-Speetjens

et al. 2021

Journal of Shoulder and Elbow Surgery

View full text Add to dashboard Cite

Background: Adequate deltoid and rotator cuff elongation in reverse shoulder arthroplasty is crucial to maximize postoperative functional outcomes and to avoid complications. Measurements of deltoid and rotator cuff elongation during preoperative planning can support surgeons in selecting a suitable implant design and position. Therefore, this study presented and evaluated a fully automated method for measuring deltoid and rotator cuff elongation. Methods: Complete scapular and humeral models were extracted from computed tomography scans of 40 subjects. First, a statistical shape model of the complete humerus was created and evaluated to identify the muscle attachment points. Next, a muscle wrapping algorithm was developed to identify the muscle paths and to compute muscle lengths and elongations after reverse shoulder arthroplasty implantation. The accuracy of the muscle attachment points and the muscle elongation measurements was evaluated for the 40 subjects by use of both complete and artificially created partial humeral models. Additionally, the muscle elongation measurements were evaluated for a set of 50 arthritic shoulder joints. Finally, a sensitivity analysis was performed to evaluate the impact of implant positioning on deltoid and rotator cuff elongation. Results: For the complete humeral models, all muscle attachment points were identified with a median error < 3.5 mm. For the partial humeral models, the errors on the deltoid attachment point largely increased. Furthermore, all muscle elongation measurements showed an error < 1 mm for 75% of the subjects for both the complete and partial humeral models. For the arthritic shoulder joints, the errors on the muscle elongation mea-surementswere<2mmfor75%ofthesubjects.Finally,thesensitivityanalysisshowedthatmuscleelongationswereaffectedbyimplantpositioning. Discussion: This study presents an automated method for accurately measuring muscle elongations during preoperative planning of shoulder arthroplasty. The results show that the accuracy in measuring muscle elongations is higher than the accuracy in indicating the muscle attachment points. Hence, muscle elongation measurements are insensitive to the observed errors on the muscle attachment points. Related to this finding, muscle elongations can be accurately measured for both a complete humeral model and a partial humeral model. Because the presented method also showed accurate results for arthritic shoulder joints, it can be used during preoperative shoulder arthroplasty planning, in which typically only the proximal humerus is present in the scan and in which bone arthropathy can be present. As the muscle elongations are sensitive to implant positioning, surgeons can use the muscle elongation measurements to refine their surgical plan.

show abstract

Section: Discussionmentioning

confidence: 95%

Section: Discussionmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Automated muscle elongation measurement during reverse shoulder arthroplasty planning

Pitocchi

Plessers

Wirix-Speetjens

et al. 2021

Journal of Shoulder and Elbow Surgery

View full text Add to dashboard Cite

show abstract

“…For clinical use, the key property of the SSM lies in the dense correspondence established during the registration process, which identifies the points sharing the same anatomical characteristics [2], [16], [17]. This feature has been effectively used to embed bony SSM with landmark-based anatomical information such as muscle insertions [18] or identify cortical bone thickness [13]. Moreover, the generative capabilities of these models enable the exploration of the shape coefficient representation within the valid anatomical shape variation [2], [5].…”

Section: Introductionmentioning

confidence: 99%

Anatomically Parameterized Statistical Shape Model: Explaining Morphometry through Statistical Learning

Boutillon,

Salhi,

Burdin

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Statistical shape models (SSMs) are a popular tool to conduct morphological analysis of anatomical structures which is a crucial step in clinical practices. However, shape representations through SSMs are based on shape coefficients and lack an explicit oneto-one relationship with anatomical measures of clinical relevance. While a shape coefficient embeds a combination of anatomical measures, a formalized approach to find the relationship between them remains elusive in the literature. This limits the use of SSMs to subjective evaluations in clinical practices. We propose a novel SSM controlled by anatomical parameters derived from morphometric analysis. Methods: The proposed anatomically parameterized SSM (ANAT SSM ) is based on learning a linear mapping between shape coefficients (latent space) and selected anatomical parameters (anatomical space). This mapping is learned from a synthetic population generated by the standard SSM. Determining the pseudo-inverse of the mapping allows us to build the ANAT SSM . We further impose orthogonality constraints to the anatomical parameterization (OC-ANAT SSM ) to obtain independent shape variation patterns. The proposed contribution was evaluated on two skeletal databases of femoral and scapular bone shapes using clinically relevant anatomical parameters within each (five for femoral and six for scapular bone). Results: Anatomical measures of the synthetically generated shapes exhibited realistic statistics. The learned matrices corroborated well with the obtained statistical relationship, while the two SSMs achieved moderate to excellent performance in predicting anatomical parameters on unseen shapes. Conclusion: This study demonstrates the use of anatomical representation for creating anatomically parameterized SSMs and as a result, removes the limited clinical interpretability of standard SSMs. Significance: The proposed models could help analyze differences in relevant bone morphometry between populations, and be integrated in patientspecific pre-surgery planning or in-surgery assessment.

show abstract

“…Statistical shape models (SSMs) provide a valuable way to describe shape variability within a training dataset. Since their introduction (Cootes and Taylor 2001), these models have been used for multiple applications: to automatically segment bone structures (Lamecker et al 2004;Ma et al 2017), to study the shapes of human anatomy (Sarkalkan et al 2014;Salhi et al 2017;Sintini et al 2018;Casier et al 2018), to virtually reconstruct large bone defects (Vanden Berghe et al 2017;Poltaretskyi et al 2017;Plessers et al 2018;Abler et al 2018), to build 3D models starting from 2D information (Grassi et al 2017;Mutsvangwa et al 2017). The main concept behind SSM techniques is to perform principal component analysis (PCA) on corresponding landmarks derived from the dataset objects and to extract the main modes of variation.…”

Section: Introductionmentioning

confidence: 99%

Integration of cortical thickness data in a statistical shape model of the scapula

Pitocchi

Wirix-Speetjens

Lenthe

et al. 2020

Computer Methods in Biomechanics and Biomedical Engineering

View full text Add to dashboard Cite

Knowledge about bone morphology and bone quality of the scapula throughout the population is fundamental in the design of shoulder implants. In particular, regions with the best bone stock (cortical bone) are taken into account when planning the supporting screws, aiming for an optimal fixation. As an alternative to manual measurements, statistical shape models (SSMs) have been commonly used to describe shape variability within a population. However, explicitly including cortical thickness information in an SSM of the scapula still remains a challenge. Therefore, the goal of this study is to combine scapular bone shape and cortex morphology in an SSM. First, a method to estimate cortical thickness, based on HU (Hounsfield Unit) profile analysis, was developed and validated. Then, based on the manual segmentations of 32 healthy scapulae, a statistical shape model including cortical information was created and evaluated. Generalization, specificity and compactness were calculated in order to assess the quality of the SSM. The average cortical thickness of the SSM was 2.0 ± 0.63 mm. Generalization, specificity and compactness performances confirmed that the combined SSM was able to capture the bone quality changes in the population. In this work we integrated information on the cortical thickness in an SSM for the scapula. From the results we conclude that this methodology is a valuable tool for automatically generating a large population of scapulae and deducing statistics on the cortex. Hence, this SSM can be useful to automate implant design and screw placement in shoulder arthroplasty.

show abstract

Subject-specific shoulder muscle attachment region prediction using statistical shape models: A validity study

Cited by 7 publications

References 12 publications

Automated muscle elongation measurement during reverse shoulder arthroplasty planning

Automated muscle elongation measurement during reverse shoulder arthroplasty planning

Anatomically Parameterized Statistical Shape Model: Explaining Morphometry through Statistical Learning

Integration of cortical thickness data in a statistical shape model of the scapula

Contact Info

Product

Resources

About