Application of deep learning for medical imaging evaluation
Inventors
Chilamkurhy, Sasank • Ghosh, Rohit • Tanamala, Swetha • Rao, Pooja • Warier, Prashant
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-10504227-B1
Publication Date
2019-12-10
Expiration Date
Abstract
This disclosure generally pertains to methods and systems for processing electronic data obtained from imaging or other diagnostic and evaluative medical procedures. Certain embodiments relate to methods for the development of deep learning algorithms that perform machine recognition of specific features and conditions in imaging and other medical data. Another embodiment provides systems configured to detect and localize medical abnormalities on medical imaging scans by a deep learning algorithm.
Core Innovation
The invention streamlines a head scan interpretation workflow by automating an initial screening and triage process for CT head scan imaging procedure data. It receives and processes images from CT head scan imaging procedure data and uses a deep learning system carried out by a computer to detect and localize medical abnormalities of the images. The system is configured to output a mask representing the precise location and extent of the medical abnormalities, and to generate a score that corresponds to a level of recognition of the medical abnormalities.
The deep learning system is developed by selecting medical imaging scans and extracting the medical abnormalities using natural language processing (NLP) algorithms, where each type of the medical abnormalities is annotated at scan, slice and pixel levels. A slice-wise deep learning algorithm is trained with the selected medical imaging scans to segment pixel-level annotated scans and to generate slice-level confidences. The system then predicts confidences for a presence of each type of medical abnormalities.
The workflow prioritizes an assignment of a medical evaluation to an evaluator based on the score generated for the images. In the described approach, slice-level confidences are combined into scan-level confidence, and localization is produced as a mask output. The described implementation focuses on non-contrast head CT and targets medical abnormalities including intracranial hemorrhage and additional findings such as midline shift, mass effect, and skull/calvarial fractures.
Claims Coverage
The document includes two independent claims. Each independent claim centers on an automated head CT workflow with NLP-based abnormality annotation, slice-wise deep learning segmentation and slice-level confidences, prediction of confidences for presence of each abnormality type, generation of a recognition score and an output mask, and triage prioritization of evaluator assignment based on the score.
Automated initial screening and triage prioritization for head scan interpretation
receiving and processing images from CT head scan imaging procedure data; detecting and localizing medical abnormalities of the images using a deep learning system; outputting a mask representing the precise location and extent of the medical abnormalities and generating a score that corresponds to a level of recognition of the medical abnormalities; prioritizing an assignment of a medical evaluation to an evaluator based on the score generated for the images.
NLP-based scan, slice and pixel level abnormality annotation development for deep learning
selecting medical imaging scans and extracting the medical abnormalities using natural language processing (NLP) algorithms, wherein each type of the medical abnormalities is annotated at scan, slice and pixel levels.
Slice-wise segmentation and slice-level confidence generation
training a slice-wise deep learning algorithm with the selected medical imaging scans to segment pixel-level annotated scans; training the deep learning algorithm with the selected medical imaging scans to generate slice-level confidences.
Abnormality presence confidence prediction and recognition score masking output
predicting confidences for a presence of each type of medical abnormalities; and generating a score that corresponds to a level of recognition of the medical abnormalities and outputting a mask representing the precise location and extent of the medical abnormalities.
Automated head CT screening and triage system configuration for receiving, detecting/localizing, and prioritizing
a deep learning system carried out by a computer to detect and localize medical abnormalities on non-contrast head CT scans, wherein the automated head CT scan screening and triage system is configured to receive and process images from CT head scan imaging procedure data; configured to detect and localize the medical abnormalities of the images using the deep learning system; and configured to prioritize an assignment of a medical evaluation to an evaluator based on the score generated for the images.
Across the two independent claims, the inventive coverage is directed to an automated head CT workflow that uses an NLP-developed deep learning system for scan, slice and pixel abnormality annotation, slice-wise segmentation and slice-level confidences, prediction of confidences for presence of each abnormality type, and generation of a recognition score together with a precise localization mask. The score is then used to prioritize assignment of a medical evaluation to an evaluator.
Stated Advantages
Automates an initial screening and triage process to streamline a head scan interpretation workflow.
Enables prioritizing assignment of a medical evaluation to an evaluator based on the generated score.
Documented Applications
Automated head CT scan screening and triage workflow to streamline a head scan interpretation workflow for CT head scan imaging procedure data.
Use of deep learning for non-contrast head CT detection and localization of medical abnormalities to support evaluator assignment prioritization.
Interested in licensing this patent?