Semiautomatic Detection of Scoliotic Rib Borders From Posteroanterior Chest Radiographs

Abstract

Abstract-3D assessment of scoliotic deformities relies on an accurate 3D reconstruction of bone structures from biplanar Xrays, which requires a precise detection and matching of anatomical structures in both views. In this paper, we propose a novel semi-automated technique for detecting complete scoliotic rib borders from PA-0º and PA-20º chest radiographs, by using an edge-following approach with multiple-path branching and oriented filtering. Edge-following processes are initiated from user starting points along upper and lower rib edges and the final rib border is obtained by finding the most parallel pair among detected edges. The method is based on a perceptual analysis leading to the assumption that no matter how bent a scoliotic rib is, it will always present relatively parallel upper and lower edges. The proposed method was tested on 44 chest radiographs of scoliotic patients and was validated by comparing pixels from all detected rib borders against their reference locations taken from the associated manually delineated rib borders. The overall 2D detection accuracy was 2.64  1.21 pixels. Comparing this accuracy level to reported results in the literature shows that the proposed method is very well suited for precisely detecting borders of scoliotic ribs from PA-0º and PA-20º chest radiographs.

Personalized 3D reconstruction of the rib cage for clinical assessment of trunk deformities

“…Our method makes use of a semi-automatic detection of the ribs [16] in the PA-0 view and an interactive selection of a sufficient number of points in the LAT, which is about three time faster. This is an important factor favoring the new method over the current one for the clinical setting.…”

“…To do so, we use the algorithm developed by Plourde et al [16]. This allows us to extract the ribs' edges in the PA-0, not just their midlines, by manually inputting only four points per rib.…”

“…These models are deformed so that their projections match the silhouettes extracted in each radiograph. More recently, Koehler and Wischgoll [15] proposed a new technique that uses the semi-automatic rib detection algorithm proposed by Plourde et al [16]. First, the ribs are detected in the frontal radiograph, and the rib cage boundary is segmented interactively in the lateral view.…”

A Novel Method for the 3-D Reconstruction of Scoliotic Ribs From Frontal and Lateral Radiographs

“…Thus, the total number of primitives per rib in our method is 14. Furthermore, a semi-automatic detection of rib contours [51] in the PA (which requires the identification of 4 points per rib), combined with an automatic extraction of the rib midlines, could reduce the manual intervention to only 7 points per rib (3 in the LAT and 4 in the PA). This would, first, increase the repeatability of ribs identification step and, second, it would reduce the processing time by about two thirds (40 minutes approximately), making our method more suitable in a clinical setting.…”

Multimodal image fusion of anatomical structures for diagnosis, therapy planning and assistance

“…To reconstruct the bone structures of the trunk, a calibrated system based on three digital multiplanar radiographs (postero-anterior view (PA-0), postero-anterior view with the x-ray tube lifted up and angled down 20 (PA-20) and a lateral view (LAT) (see Figure 3) is employed [23]. We have also developed and validated novel techniques using digital multiplanar radiographs for 3D reconstruction and analysis of the bone structures of the human trunk, including the spine, pelvis and ribcage [24][25][26][27][28][29].…”