Change Assessment Method
AssigneeKitware, Inc.
InventorsRicardo S. Avila, Luis Ibanez
Life of this patent
Estimating the term…
Abstract
This invention relates to a method of assessing the change in size of cancerous lesions in humans utilizing CT scans through the application of a boundary detection algorithm capable of detecting subtle changes in CT images. The change assessment method may include the steps for defining boundary regions on a CT scan image, analyzing said boundaries utilizing a computational program containing statistical analysis techniques, and producing a medically useful output showing any change in lesion size. The change assessment method may also be interfaced with other medical diagnostic techniques or devices that are known in the art.
Claims12
A method for assessing the amount of change of an object in a longitudinal series of medical images consisting of the following steps performed at each imaging time point:
- a. a human operator using a tool for defining a boundary region completely inside the object being assessed;
- b. a human operator using a tool for defining a boundary region completely outside the object being assessed;
- c. a human operator using a tool for identifying corresponding landmarks surrounding the object being assessed;
- d. a computational analysis producing signal intensity distribution descriptive statistics for inside and outside the object being assessed;
- e. obtaining the point spread function, signal to noise ratio, and other characteristics of the acquisition device;
- f. the computation of object boundary regions using an edge detector utilizing knowledge of inside/outside object statistics and acquisition device characteristics;
- g. the calculation of a set of surfaces that exist between the inside and outside boundaries representing only those boundaries which can be measured with high confidence;
- h. the calculation of a correspondence between the set of high confidence surfaces and the set of high confidence surfaces at the first imaging time point using a deformable registration method;
- i. the use of a rigid registration technique based on the landmarks obtained in step (c);
- and
- j. the calculation of local and global change of the object being assessed along with a calculated precision and bias.
A method for estimating the Point Spread Function (PSF) and noise characteristics of a CT scan based on an analysis of patient anatomy present within the images consisting of the following steps:
- a. detecting large homogeneous regions consisting of air;
- b. identifying the trachea and the main bronchial airways;
- c. estimating the CT density of the region just outside the airway way;
- d. sampling the perpendicular change in CT density along the entire main airway, and;
- e. estimating the 3D PSF needed to produce the partial volume intensities sampled across the airway.
A method for identifying the progression of a lesion in a patient in a computer implemented system, comprising the following steps:
- a. identifying on a CT scan the anatomical position at time t=0 of one or more high confidence boundary patches on the lesion and collecting a longitudinal dataset relative to the position at t=0;
- b. identifying on a CT scan at a later time, t=1, changes in the anatomical position of the one or more high confidence boundary patches and collecting a longitudinal dataset relative to the position at t=1;
- and
- c. assessing lesion progression by comparing the longitudinal datasets from t=0 and t=1.
Dependencies are inferred from the claim text.