Subject-specific models for image-guided cardiac surgery

A Persistent Homology-Based Topological Loss for CNN-Based Multiclass Segmentation of CMR

Multi-class segmentation of cardiac magnetic resonance (CMR) images seeks a separation of data into anatomical components with known structure and configuration. The most popular CNN-based methods are optimised using pixel wise loss functions, ignorant of the spatially extended features that characterise anatomy. Therefore, whilst sharing a high spatial overlap with the ground truth, inferred CNN-based segmentations can lack coherence, including spurious connected components, holes and voids. Such results are implausible, violating anticipated anatomical topology. In response, (single-class) persistent homology-based loss functions have been proposed to capture global anatomical features. Our work extends these approaches to the task of multi-class segmentation. Building an enriched topological description of all class labels and class label pairs, our loss functions make predictable and statistically significant improvements in segmentation topology using a CNN-based post-processing framework. We also present (and make available) a highly efficient implementation based on cubical complexes and parallel execution, enabling practical application within high resolution 3D data for the first time. We demonstrate our approach on 2D short axis and 3D whole heart CMR segmentation, advancing a detailed and faithful analysis of performance on two publicly available datasets.

A method for optimizing planning target volume margins for patients receiving lung stereotactic body radiotherapy

Lung stereotactic-body radiotherapy (SBRT) places additional requirements on targeting accuracy over standard approaches. In treatment planning, a tumour volume is geometrically expanded and the resulting planning target volume (PTV) is covered with the prescribed dose. This ensures full dose delivery despite various uncertainties encountered during treatment. We developed a retrospective technique for optimizing the PTV expansion for a patient population. The method relies on deformable image registration (DIR) of the planning CT to a treatment cone-beam CT (CBCT). The resulting transformation is used to map the planned target onto the treatment geometry, allowing the computation of the achieved target/PTV overlap. Basic validation of the method was performed using an anthropomorphic respiratory motion phantom. A self-validation technique was also implemented to allow estimation of the DIR error for the data being analyzed. Our workflow was used to retrospectively optimize PTV margin for 25 patients treated over 93 fractions. Targets for these patients were contoured on 4D CT images. SBRT delivery followed CBCT acquisition and a couch correction. A post-treatment CBCT was also acquired in some cases. Our basic validation demonstrated that the DIR-based technique is capable of transforming target volumes from planning CTs to treatment CBCTs with sub-mm accuracy. Our clinical analysis showed that the minimum percentages of target volumes covered for 3, 4, and 5 mm PTV margins were 92.1, 97.6, and 99.2, respectively. Analyzing data acquired before and just after treatment demonstrated that margins exceeding 5 mm did not significantly improve coverage. Finally, a 5 mm PTV margin achieved  ⩾95% target volume coverage with  ⩾95% probability. Our technique is accurate, automated, self-validating, and incorporates complex ITV shapes/deformations to allow PTV margin optimization. The analysis of clinical data indicates a 5 mm PTV margin is optimal for our process. This approach is generalizable to other disease sites and treatment strategies.

Spatiotemporal Strategies for Joint Segmentation and Motion Tracking From Cardiac Image Sequences

Although accurate and robust estimations of the deforming cardiac geometry and kinematics from cine tomographic medical image sequences remain a technical challenge, they have significant clinical value. Traditionally, boundary or volumetric segmentation and motion estimation problems are considered as two sequential steps, even though the order of these processes can be different. In this paper, we present an integrated, spatiotemporal strategy for the simultaneous joint recovery of these two ill-posed problems. We use a mesh-free Galerkin formulation as the representation and computation platform, and adopt iterative procedures to solve the governing equations. Specifically, for each nodal point, the external driving forces are individually constructed through the integration of data-driven edginess measures, prior spatial distributions of myocardial tissues, temporal coherence of image-derived salient features, imaging/image-derived Eulerian velocity information, and cyclic motion model of myocardial behavior. The proposed strategy is accurate and very promising application results are shown from synthetic data, magnetic resonance (MR) phase contrast, tagging image sequences, and gradient echo cine MR image sequences.

A review of heart chamber segmentation for structural and functional analysis using cardiac magnetic resonance imaging

Cardiovascular magnetic resonance (CMR) has become a key imaging modality in clinical cardiology practice due to its unique capabilities for non-invasive imaging of the cardiac chambers and great vessels. A wide range of CMR sequences have been developed to assess various aspects of cardiac structure and function, and significant advances have also been made in terms of imaging quality and acquisition times. A lot of research has been dedicated to the development of global and regional quantitative CMR indices that help the distinction between health and pathology. The goal of this review paper is to discuss the structural and functional CMR indices that have been proposed thus far for clinical assessment of the cardiac chambers. We include indices definitions, the requirements for the calculations, exemplar applications in cardiovascular diseases, and the corresponding normal ranges. Furthermore, we review the most recent state-of-the art techniques for the automatic segmentation of the cardiac boundaries, which are necessary for the calculation of the CMR indices. Finally, we provide a detailed discussion of the existing literature and of the future challenges that need to be addressed to enable a more robust and comprehensive assessment of the cardiac chambers in clinical practice.

On mixed reality environments for minimally invasive therapy guidance: systems architecture, successes and challenges in their implementation from laboratory to clinic

Mixed reality environments for medical applications have been explored and developed over the past three decades in an effort to enhance the clinician's view of anatomy and facilitate the performance of minimally invasive procedures. These environments must faithfully represent the real surgical field and require seamless integration of pre- and intra-operative imaging, surgical instrument tracking, and display technology into a common framework centered around and registered to the patient. However, in spite of their reported benefits, few mixed reality environments have been successfully translated into clinical use. Several challenges that contribute to the difficulty in integrating such environments into clinical practice are presented here and discussed in terms of both technical and clinical limitations. This article should raise awareness among both developers and end-users toward facilitating a greater application of such environments in the surgical practice of the future.

Accuracy considerations in image-guided cardiac interventions: experience and lessons learned

MOTIVATION

Medical imaging and its application in interventional guidance has revolutionized the development of minimally invasive surgical procedures leading to reduced patient trauma, fewer risks, and shorter recovery times. However, a frequently posed question with regard to an image guidance system is "how accurate is it?" On one hand, the accuracy challenge can be posed in terms of the tolerable clinical error associated with the procedure; on the other hand, accuracy is bound by the limitations of the system's components, including modeling, patient registration, and surgical instrument tracking, all of which ultimately impact the overall targeting capabilities of the system.

METHODS

While these processes are not unique to any interventional specialty, this paper discusses them in the context of two different cardiac image guidance platforms: a model-enhanced ultrasound platform for intracardiac interventions and a prototype system for advanced visualization in image-guided cardiac ablation therapy.

RESULTS

Pre-operative modeling techniques involving manual, semi-automatic and registration-based segmentation are discussed. The performance and limitations of clinically feasible approaches for patient registration evaluated both in the laboratory and in the operating room are presented. Our experience with two different magnetic tracking systems for instrument and ultrasound transducer localization is reported. Ultimately, the overall accuracy of the systems is discussed based on both in vitro and preliminary in vivo experience.

CONCLUSION

While clinical accuracy is specific to a particular patient and procedure and vastly dependent on the surgeon's experience, the system's engineering limitations are critical to determine whether the clinical requirements can be met.

Automatic image guidance for prostate IMRT using low dose CBCT

PURPOSE

Varian's On-Board Imager is a linac-integrated cone-beam CT (CBCT) system used at the authors' institution to acquire images prior to delivering each fraction of prostate intensity modulated radiotherapy. The images are used to determine a couch shift that realigns the tumor with the position obtained in the planning CT. However, this manual image-guided radiotherapy (IGRT) technique is operator dependent, time consuming, offers limited degrees of freedom, and requires significant imaging dose over the course of treatment. To overcome these problems, the authors propose two fully automatic IGRT techniques that require significantly less imaging dose.

METHODS

Dose is reduced by lowering the x-ray tube mA s during CBCT acquisition at the cost of increasing image noise. In "forward" IGRT, the CBCT image is automatically registered to the planning CT to obtain the necessary couch shift. The "reverse" technique offers additional degrees of freedom as it involves nonrigid registration of the planning CT to the CBCT. Both techniques were evaluated using images of an anthropomorphic phantom with simulated motion and by retrospectively analyzing data from ten prostate cancer patients.

RESULTS

IGRT error for the phantom data at 100% relative imaging dose was 8.2 +/- 3.7, 3.5 +/- 1.2,, and 2.1 +/- 0.6 mm for setup only, forward, and reverse techniques, respectively. For patient images acquired at 100% relative imaging dose, the errors were 5.4 +/- 1.7, 5.0 +/- 1.6, 5.0 +/- 2.0, and 4.0 +/- 1.6 mm for setup only, manual forward (performed clinically), automatic forward, and reverse IGRT, respectively. Furthermore, imaging dose could be reduced to 20% without a significant loss in image guidance accuracy.

CONCLUSIONS

The presented image guidance methods are accurate while requiring only 20% of the standard imaging dose. The combination of low dose, automation, and accuracy enables frequent corrections during treatment, possibly leading to reduced margins and improved treatment outcomes.

How accurate is accurate enough? A brief overview on accuracy considerations in image-guided cardiac interventions

Image-guided interventions have revolutionized the development of minimally invasive surgical procedures, leading to reduced patient trauma, fewer risks and shorter recovery times. However, one of the most frequently posed question with regards to an image guidance system is how accurate it is. In this work we provide a brief overview on accuracy considerations from our perspective on cardiac image-guided procedures: what are the clinically-imposed accuracy constraints, how do these measure against the limitations of the image-guidance system, and how can surgeons directly benefit from real-time accuracy feedback to ensure optimal navigation at all times during the intervention?