Method, system, software and medium for advanced intelligent image analysis and display of medical images and information
Computerized interpretation of medical images for quantitative analysis of multi-modality breast images including analysis of FFDM, 2D/3D ultrasound, MRI, or other breast imaging methods. Real-time characterization of tumors and background tissue, and calculation of image-based biomarkers is provided for breast cancer detection, diagnosis, prognosis, risk assessment, and therapy response. Analysis includes lesion segmentation, and extraction of relevant characteristics (textural/morphological/kinetic features) from lesion-based or voxel-based analyzes. Combinations of characteristics in several classification tasks using artificial intelligence is provided. Output in terms of 1D, 2D or 3D distributions in which an unknown case is identified relative to calculations on known or unlabeled cases, which can go through a dimension-reduction technique. Output to 3D shows relationships of the unknown case to a cloud of known or unlabeled cases, in which the cloud demonstrates the structure of the population of patients with and without the disease.
1. A non-transitory computer-readable medium storing thereon executable instructions, that when executed by a computer, cause the computer to execute a process for determining a probability of a disease state of a patient, the process comprising:
obtaining medical data including at least one of a medical image, medical image data, and data representative of a clinical examination of the patient;
segmenting a lesion indicated in the medical data;
extracting features of the lesion based on the segmented lesion;
calculating features based on lesion analysis or voxel analyses, the calculated features establishing a high-dimensional feature space for the lesion which is beyond three-dimensions;
reducing the high-dimensional feature space to a reduced dimensional space of two or three dimensions; and
displaying a data point on the reduced dimensional space which represents the lesion for the patient, together with data points which represent lesions for other patients according to the reduced dimensional space.
2. The medium according to claim 1 , the process further comprising:
determining the probability of the disease state of the patient by comparing the location of the data point for the patient with the data points for the other patients, the data points for the other patients identifying a combination disease and non-disease state lesions shown differently in the reduced dimensional space.
3. The medium according to claim 2 , wherein the disease and non-disease state lesions are shown different in the reduced dimensional space by utilizing different colors.
4. The medium according to claim 1 , the process further comprising:
defining a region around the data point for the patient about which defines those of the other patients having similar cases to that of the patient, wherein the region is displayed on the reduced dimensional space as a two-dimensional or three-dimensional region depending on whether the reduced dimensional space is two-dimensional or three-dimensional.
5. The medium according to claim 1 , the process further comprising:
rotating, according to a user input, the reduced dimensional space about an axis to visually identify relationships and similarities between the patient and the other patients.
6. The medium according to claim 1 , further comprising:
obtaining labeled and unlabeled medical data of the other patients;
segmenting disease and non-disease state lesions in the labeled medical image data and lesions in the unlabeled medical image data;
extracting features of the segmented labeled and unlabeled medical image data;
performing dimension reduction on the extracted features from the high-dimensional feature space to the reduced dimensional space; and
training a classifier, based on only the labeled medical image data, which is utilized in determining the probability of the disease state of the patient.