SOMA: Subject-, object-, and modality-adapted precision atlas approach for automatic anatomy recognition and delineation in medical images

Abstract

Purpose: In the multi-atlas segmentation (MAS) method, a large enough atlas set, which can cover the complete spectrum of the whole population pattern of the target object will benefit the segmentation quality. However, the difficulty in obtaining and generating such a large set of atlases and the computational burden required in the segmentation procedure make this approach impractical. In this paper, we propose a method called SOMA to select subject-, object-, and modality-adapted precision atlases for automatic anatomy recognition in medical images with pathology, following the idea that different regions of the target object in a novel image can be recognized by different atlases with regionally best similarity, so that effective atlases have no need to be globally similar to the target subject and also have no need to be overall similar to the target object.

Methods: The SOMA method consists of three main components: atlas building, object recognition, and object delineation. Considering the computational complexity, we utilize an all-to-template strategy to align all images to the same image space belonging to the root image determined by the minimum spanning tree (MST) strategy among a subset of radiologically near-normal images. The object recognition process is composed of two stages: rough recognition and refined recognition. In rough recognition, subimage matching is conducted between the test image and each image of the whole atlas set, and only the atlas corresponding to the best-matched subimage contributes to the recognition map regionally. The frequency of best match for each atlas is recorded by a counter, and the atlases with the highest frequencies are selected as the precision atlases. In refined recognition, only the precision atlases are examined, and the subimage matching is conducted in a nonlocal manner of searching to further increase the accuracy of boundary matching. Delineation is based on a U-net-based deep learning network, where the original gray scale image together with the fuzzy map from refined recognition compose a two-channel input to the network, and the output is a segmentation map of the target object.

Results: Experiments are conducted on computed tomography (CT) images with different qualities in two body regions - head and neck (H&N) and thorax, from 298 subjects with nine objects and 241 subjects with six objects, respectively. Most objects achieve a localization error within two voxels after refined recognition, with marked improvement in localization accuracy from rough to refined recognition of 0.6-3 mm in H&N and 0.8-4.9 mm in thorax, and also in delineation accuracy (Dice coefficient) from refined recognition to delineation of 0.01-0.11 in H&N and 0.01-0.18 in thorax.

Conclusions: The SOMA method shows high accuracy and robustness in anatomy recognition and delineation. The improvements from rough to refined recognition and further to delineation, as well as immunity of recognition accuracy to varying image and object qualities, demonstrate the core principles of SOMA where segmentation accuracy increases with precision atlases and gradually refined object matching.

Keywords: anatomy recognition; multi-atlas segmentation; precision atlas selection.

FIGURE 1

Schematic representation of the SOMA approach

FIGURE 2

Two-dimensional example of a 5 × 5 subimage V_5,J*(v) (middle), where v is shown highlighted, the 5 × 5 region around v in FM_ℓ (left), and the same 5 × 5 region around v in the resulting FM_ℓ (right) after the update in Step R6

FIGURE 3

Deep learning network architecture for SOMA delineation procedure. A case of right parotid gland (RPG) is taken as an example

FIGURE 4

Image examples for objects in the head and neck (H&N) region. Two-dimensional images for reference masks (first column), fuzzy maps from RoR and ReR procedures (second and third columns), and delineation masks (fourth column) overlapped on gray scale images and overlapped by reference contours are shown in first two rows. The corresponding 3D surface or volume renditions are arranged at the bottom

FIGURE 5

Image examples for objects in the thorax region. Similar to Figure 4, reference masks, recognition maps, and delineation masks are shown from left to right

FIGURE 6

Illustration of why subject-level similarity should not be considered in the recognition process. (a–c) Three slices of the novel image are overlayed by the atlas with best subject-level similarity; this will not guarantee the regional similarity/match. (d) Initial region of interest (R_in). (e) Atlas map that indicates indexes of best-match atlases of each region. (f) Different regions of the novel image are matched by different atlases, and their binary masks are combined as in Figure 2 to generate a recognition map as in (g)