Automatic localization of the anterior commissure, posterior commissure, and midsagittal plane in MRI scans using regression forests

Triangular fossa of the third cerebral ventricle - an original 3D model and morphometric study

Introduction

The triangular recess (TR), also called triangular fossa or vulva cerebri, represents the anterior extension of the diencephalic ventricle, located between the anterior columns of the fornix and the anterior white commissure. Over time, this structure of the third cerebral ventricle generated many disputes. While some anatomists support its presence, others have opposite opinions, considering that it only becomes visible under certain conditions. The aim of the study is to demonstrate the tangible structure of the triangular recess. Secondly, the quantitative analysis allowed us to establish an anatomical morphometric standard, as well as the deviations from the standard.

Materials and methods

Our study is both a quantitative and a qualitative evaluation of the triangular fossa. We dissected 100 non-neurological adult brains, which were fixed in 10% formaldehyde solution for 10 weeks. The samples are part of the collection of the Institute of Anatomy, "Grigore T. Popa" University of Medicine and Pharmacy, Iasi. We highlighted the triangular fossa by performing dissections in two stages at the level of the roof of the III ventricle.

Results

The qualitative analysis is a re-evaluation of the classical data concerning the anatomy of the fossa triangularis. We proposed an original 3D model of the triangular recess in which we described a superficial part called vestibule and a deep part called pars profunda. We measured the sides of the communication between the two proposed segments, as well as the communication with the III ventricle. By applying the Heron's formula, we calculated the area of the two communications. Statistical evaluations have shown that these communications are higher than they are wide. In addition, there is a statistical difference between the surfaces of the two communications: 34.07 mm2 ± 7.01 vs. 271.43 mm2 ± 46.36 (p = 0.001).

Conclusion

The outcome of our study is both qualitative and quantitative. Firstly, we demonstrated the existence of the triangular fossa and we conceived a spatial division of this structure. Secondly, the measurements carried out establish an anatomo-morphometric norm of the triangular recess, which is useful in assessing the degree of hydrocephalus during the third endoscopic ventriculoscopy.

A new approach to symmetric registration of longitudinal structural MRI of the human brain

BACKGROUND

This paper presents the Automatic Temporal Registration Algorithm (ATRA) for symmetric rigid-body registration of longitudinal T1-weighted three-dimensional MRI scans of the human brain. This is a fundamental processing step in computational neuroimaging.

NEW METHOD

The notion of leave-one-out consistent (LOOC) landmarks with respect to a supervised landmark detection algorithm is introduced. An automatic algorithm is presented for identification of LOOC landmarks on MRI scans. Multiple sets of LOOC landmarks are identified on each volume and a Generalized Orthogonal Procrustes Analysis of the landmarks is used to find a rigid-body transformation of each volume into a common space where the volumes are aligned precisely.

RESULTS

Qualitative and quantitative evaluations of ATRA registration accuracy were performed using 2012 volumes from 503 subjects (4 longitudinal volumes/subject), and on a further 120 volumes acquired from 3 normal subjects (40 longitudinal volumes/subject). Since the ground truth registrations are unknown, we devised a novel method for showing that ATRA's registration accuracy is at least better than 0.5 mm translation or 0.5° rotation.

COMPARISON WITH EXISTING METHOD(S)

In comparison with existing methods, ATRA does not require any image preprocessing (e.g., skull-stripping or intensity normalization) and can handle conditions where rigid-body motion assumptions are not true (e.g., movement in eyes, jaw, neck) and brain tissue loss over time in neurodegenerative diseases. In a systematic comparison with the FSL FLIRT algorithm, ATRA provided faster and more accurate registrations.

CONCLUSIONS

The algorithm is symmetric, in the sense that any permutation of the input volumes does not change the transformation matrices, and unbiased, in that all volumes undergo exactly one interpolation operation, which precisely aligns them in a common space. There is no interpolation bias and no reference volume. All volumes are treated exactly the same. The algorithm is fast and highly accurate.

Application of the anatomical fiducials framework to a clinical dataset of patients with Parkinson's disease

Establishing spatial correspondence between subject and template images is necessary in neuroimaging research and clinical applications such as brain mapping and stereotactic neurosurgery. Our anatomical fiducial (AFID) framework has recently been validated to serve as a quantitative measure of image registration based on salient anatomical features. In this study, we sought to apply the AFIDs protocol to the clinic, focusing on structural magnetic resonance images obtained from patients with Parkinson's disease (PD). We confirmed AFIDs could be placed to millimetric accuracy in the PD dataset with results comparable to those in normal control subjects. We evaluated subject-to-template registration using this framework by aligning the clinical scans to standard template space using a robust open preprocessing workflow. We found that registration errors measured using AFIDs were higher than previously reported, suggesting the need for optimization of image processing pipelines for clinical grade datasets. Finally, we examined the utility of using point-to-point distances between AFIDs as a morphometric biomarker of PD, finding evidence of reduced distances between AFIDs that circumscribe regions known to be affected in PD including the substantia nigra. Overall, we provide evidence that AFIDs can be successfully applied in a clinical setting and utilized to provide localized and quantitative measures of registration error. AFIDs provide clinicians and researchers with a common, open framework for quality control and validation of spatial correspondence and the location of anatomical structures, facilitating aggregation of imaging datasets and comparisons between various neurological conditions.

Automatic Symmetry Detection From Brain MRI Based on a 2-Channel Convolutional Neural Network

Symmetry detection is a method to extract the ideal mid-sagittal plane (MSP) from brain magnetic resonance (MR) images, which can significantly improve the diagnostic accuracy of brain diseases. In this article, we propose an automatic symmetry detection method for brain MR images in 2-D slices based on a 2-channel convolutional neural network (CNN). Different from the existing detection methods that mainly rely on the local image features (gradient, edge, etc.) to determine the MSP, we use a CNN-based model to implement the brain symmetry detection, which does not require any local feature detections and feature matchings. By training to learn a wide variety of benchmarks in the brain images, we can further use a 2-channel CNN to evaluate the similarity between the pairs of brain patches, which are randomly extracted from the whole brain slice based on a Poisson sampling. Finally, a scoring and ranking scheme is used to identify the optimal symmetry axis for each input brain MR slice. Our method was evaluated in 2166 artificial synthesized brain images and 3064 collected in vivo MR images, which included both healthy and pathological cases. The experimental results display that our method achieves excellent performance for symmetry detection. Comparisons with the state-of-the-art methods also demonstrate the effectiveness and advantages for our approach in achieving higher accuracy than the previous competitors.

DeepNavNet: Automated Landmark Localization for Neuronavigation

Functional neurosurgery requires neuroimaging technologies that enable precise navigation to targeted structures. Insufficient image resolution of deep brain structures necessitates alignment to a brain atlas to indirectly locate targets within preoperative magnetic resonance imaging (MRI) scans. Indirect targeting through atlas-image registration is innately imprecise, increases preoperative planning time, and requires manual identification of anterior and posterior commissure (AC and PC) reference landmarks which is subject to human error. As such, we created a deep learning-based pipeline that consistently and automatically locates, with submillimeter accuracy, the AC and PC anatomical landmarks within MRI volumes without the need for an atlas. Our novel deep learning pipeline (DeepNavNet) regresses from MRI scans to heatmap volumes centered on AC and PC anatomical landmarks to extract their three-dimensional coordinates with submillimeter accuracy. We collated and manually labeled the location of AC and PC points in 1128 publicly available MRI volumes used for training, validation, and inference experiments. Instantiations of our DeepNavNet architecture, as well as a baseline model for reference, were evaluated based on the average 3D localization errors for the AC and PC points across 311 MRI volumes. Our DeepNavNet model significantly outperformed a baseline and achieved a mean 3D localization error of 0.79 ± 0.33 mm and 0.78 ± 0.33 mm between the ground truth and the detected AC and PC points, respectively. In conclusion, the DeepNavNet model pipeline provides submillimeter accuracy for localizing AC and PC anatomical landmarks in MRI volumes, enabling improved surgical efficiency and accuracy.

A framework for evaluating correspondence between brain images using anatomical fiducials

Accurate spatial correspondence between template and subject images is a crucial step in neuroimaging studies and clinical applications like stereotactic neurosurgery. In the absence of a robust quantitative approach, we sought to propose and validate a set of point landmarks, anatomical fiducials (AFIDs), that could be quickly, accurately, and reliably placed on magnetic resonance images of the human brain. Using several publicly available brain templates and individual participant datasets, novice users could be trained to place a set of 32 AFIDs with millimetric accuracy. Furthermore, the utility of the AFIDs protocol is demonstrated for evaluating subject-to-template and template-to-template registration. Specifically, we found that commonly used voxel overlap metrics were relatively insensitive to focal misregistrations compared to AFID point-based measures. Our entire protocol and study framework leverages open resources and tools, and has been developed with full transparency in mind so that others may freely use, adopt, and modify. This protocol holds value for a broad number of applications including alignment of brain images and teaching neuroanatomy.

[Design and implementation of postoperative evaluation pipeline of deep brain stimulation by multimodality imaging]

Deep brain stimulation (DBS) surgery is an important treatment for patients with Parkinson's disease in the middle and late stages. The accuracy of the implantation of electrode at the location of the nuclei directly determines the therapeutic effect of the operation. At present, there is no single imaging method that can obtain images with electrodes, nuclei and their positional relationship. In addition, the subthalamic nucleus is small in size and the boundary is not obvious, so it cannot be directly segmented. In this paper, a complete end-to-end DBS effect evaluation pipeline was constructed using magnetic resonance (MR) data of T1, T2 and SWI weighted by DBS surgery. Firstly, the images of preoperative and postoperative patients are registered and normalized to the same coordinate space. Secondly, the patient map is obtained by non-rigid registration of brain map and preoperative data, as well as the preoperative nuclear cluster prediction position. Then, a three-dimensional (3D) image of the positional relationship between the electrode and the nucleus is obtained by using the electrode path in the postoperative image and the result of the nuclear segmentation. The 3D image is helpful for the evaluation of the postoperative effect of DBS and provides effective information for postoperative program control. After analysis, the algorithm can achieve a good registration between the patient's DBS surgical image and the brain map. The error between the algorithm and the expert evaluation of the physical coordinates of the center of the thalamus is (1.590 ± 1.063) mm. The problem of postoperative evaluation of the placement of DBS surgical electrodes is solved.

Evaluating reinforcement learning agents for anatomical landmark detection

Automatic detection of anatomical landmarks is an important step for a wide range of applications in medical image analysis. Manual annotation of landmarks is a tedious task and prone to observer errors. In this paper, we evaluate novel deep reinforcement learning (RL) strategies to train agents that can precisely and robustly localize target landmarks in medical scans. An artificial RL agent learns to identify the optimal path to the landmark by interacting with an environment, in our case 3D images. Furthermore, we investigate the use of fixed- and multi-scale search strategies with novel hierarchical action steps in a coarse-to-fine manner. Several deep Q-network (DQN) architectures are evaluated for detecting multiple landmarks using three different medical imaging datasets: fetal head ultrasound (US), adult brain and cardiac magnetic resonance imaging (MRI). The performance of our agents surpasses state-of-the-art supervised and RL methods. Our experiments also show that multi-scale search strategies perform significantly better than fixed-scale agents in images with large field of view and noisy background such as in cardiac MRI. Moreover, the novel hierarchical steps can significantly speed up the searching process by a factor of 4-5 times.

Localizing landmark sets in head CTs using random forests and a heuristic search algorithm for registration initialization

Cochlear implants (CIs) use electrode arrays that are surgically inserted into the cochlea to stimulate frequency-mapped nerve endings to treat patients with hearing loss. CIs are programmed postoperatively by audiologists using behavioral tests without information on electrode-cochlea spatial relationship. We have recently developed techniques to segment the intracochlear anatomy and to localize individual contacts in clinically acquired computed tomography (CT) images. Using this information, we have proposed a programming strategy that we call image-guided CI programming (IGCIP), and we have shown that it significantly improves outcomes for both adult and pediatric recipients. One obstacle to large-scale deployment of this technique is the need for manual intervention in some processing steps. One of these is the rough registration of images prior to the use of automated intensity-based algorithms. Although seemingly simple, the heterogeneity of our image set makes this task challenging. We propose a solution that relies on the automated random forest-based localization of multiple landmarks used to estimate an initial transformation with a point-based registration method. Results show that it produces results that are equivalent to a manual initialization. This work is an important step toward the full automation of IGCIP.

Probabilistic liver atlas construction

BACKGROUND

Anatomical atlases are 3D volumes or shapes representing an organ or structure of the human body. They contain either the prototypical shape of the object of interest together with other shapes representing its statistical variations (statistical atlas) or a probability map of belonging to the object (probabilistic atlas). Probabilistic atlases are mostly built with simple estimations only involving the data at each spatial location.

RESULTS

A new method for probabilistic atlas construction that uses a generalized linear model is proposed. This method aims to improve the estimation of the probability to be covered by the liver. Furthermore, all methods to build an atlas involve previous coregistration of the sample of shapes available. The influence of the geometrical transformation adopted for registration in the quality of the final atlas has not been sufficiently investigated. The ability of an atlas to adapt to a new case is one of the most important quality criteria that should be taken into account. The presented experiments show that some methods for atlas construction are severely affected by the previous coregistration step.

CONCLUSION

We show the good performance of the new approach. Furthermore, results suggest that extremely flexible registration methods are not always beneficial, since they can reduce the variability of the atlas and hence its ability to give sensible values of probability when used as an aid in segmentation of new cases.

Automatic localization of landmark sets in head CT images with regression forests for image registration initialization

Cochlear Implants (CIs) are electrode arrays that are surgically inserted into the cochlea. Individual contacts stimulate frequency-mapped nerve endings thus replacing the natural electro-mechanical transduction mechanism. CIs are programmed post-operatively by audiologists but this is currently done using behavioral tests without imaging information that permits relating electrode position to inner ear anatomy. We have recently developed a series of image processing steps that permit the segmentation of the inner ear anatomy and the localization of individual contacts. We have proposed a new programming strategy that uses this information and we have shown in a study with 68 participants that 78% of long term recipients preferred the programming parameters determined with this new strategy. A limiting factor to the large scale evaluation and deployment of our technique is the amount of user interaction still required in some of the steps used in our sequence of image processing algorithms. One such step is the rough registration of an atlas to target volumes prior to the use of automated intensity-based algorithms when the target volumes have very different fields of view and orientations. In this paper we propose a solution to this problem. It relies on a random forest-based approach to automatically localize a series of landmarks. Our results obtained from 83 images with 132 registration tasks show that automatic initialization of an intensity-based algorithm proves to be a reliable technique to replace the manual step.

Multi-modal Learning-based Pre-operative Targeting in Deep Brain Stimulation Procedures

Deep brain stimulation, as a primary surgical treatment for various neurological disorders, involves implanting electrodes to stimulate target nuclei within millimeter accuracy. Accurate pre-operative target selection is challenging due to the poor contrast in its surrounding region in MR images. In this paper, we present a learning-based method to automatically and rapidly localize the target using multi-modal images. A learning-based technique is applied first to spatially normalize the images in a common coordinate space. Given a point in this space, we extract a heterogeneous set of features that capture spatial and intensity contextual patterns at different scales in each image modality. Regression forests are used to learn a displacement vector of this point to the target. The target is predicted as a weighted aggregation of votes from various test samples, leading to a robust and accurate solution. We conduct five-fold cross validation using 100 subjects and compare our method to three indirect targeting methods, a state-of-the-art statistical atlas-based approach, and two variations of our method that use only a single modality image. With an overall error of 2.63±1.37mm, our method improves upon the single modality-based variations and statistically significantly outperforms the indirect targeting ones. Our technique matches state-of-the-art registration methods but operates on completely different principles. Both techniques can be used in tandem in processing pipelines operating on large databases or in the clinical flow for automated error detection.