Click here to


Are you sure ?

Yes, do it No, cancel

Lung Feature Tracking in 4D-MRI Using a Scale-Invariant Feature Transform Method

E ColvillE Colvill1*, A Lomax1, O Bieri2, Z Celicanin2, D Weber1 , M Peroni1 , (1)Paul Scherrer Institut, Villigen-PSI, Switzerland, (2)University of Basel, Basel, Switzerland


(Sunday, 7/29/2018) 4:30 PM - 5:00 PM

Room: Exhibit Hall | Forum 8

Purpose: The purpose of this study is to demonstrate the feasibility of a scale invariant feature transform (SIFT) method for automated lung motion characterization from 2D navigator slices acquired during a 4D-MRI scan.

Methods: 4D-MRI scans of 20 healthy volunteers and 3 patients were acquired. The associated 2D navigator slices had a slice thickness of 6 mm and a frame rate of 3.33 Hz. A SIFT-based method was employed to automatically extract features and corresponding trajectories tracked on the majority (>75%) of frames with fewer than 20 misses. Misses and outlier data points were replaced by cubic interpolation. Outliers were identified as data points with a distance from the trajectory centroid greater than a threshold (threshold=P₇₅+1.5×P₇₅₋₂₅ where P₇₅ is the 75th percentile of the points-to-centroid distance distribution and P₇₅₋₂₅ is the interquartile range (IQR)).The locations of the features within the lung and motion trajectory information about each feature were calculated. Motion information included the mean range of the trajectory, IQR, and 5th and 95th percentiles.

Results: Preliminary results from five volunteer 4DMRI sets show that on average 5 features were tracked for each set, with the majority of features in the lower-to-middle portion of the lung with no features tracked in the apex of the lung. The average mean range, IQR and 5th and 95th percentiles of the feature trajectories were 4.8 (±2.7)mm, 3.5 (±2.0)mm, 2.9 (±1.5)mm and 3.3 (±1.6)mm respectively, with inferior features displaying greater motion ranges than superior.

Conclusion: Automatic lung motion characterization based on 2D navigator slices was successfully demonstrated using a SIFT-based method. The features extracted from different portions of the lung displayed different motion characteristics. The method could be further improved by image pre-processing to allow for extraction of features in the apex.


MRI, Feature Selection, Lung


IM- MRI : General (Most aspects)

Contact Email