# US 11,308,612 B2 · Systems and methods for detection of infectious respiratory diseases

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 11,308,612 B2
- Document ID: US11308612B2
- Status: Granted
- Kind: B2
- Application number: 16/889,412
- Filed: 2020-06-01
- Published: 2021-10-21
- Granted: 2022-04-19
- Earliest claimed: 2020-06-01
- Assignees: Qure.ai Technologies Private Limited
- Applicants: Qure.ai Technologies Private Limited
- Inventors: Preetham Putha; Manoj Tadepalli; Bhargava Reddy; Tarun Raj; Ammar Jagirdar; Pooja Rao; Prashant Warier
- Web page: https://patentgrep.com/patent/US11308612B2

## CPC

- G06F 40/30: Semantic analysis
- A61B 6/503: {for diagnosis of the heart}
- A61B 6/505: {for diagnosis of bone}
- A61B 6/5217: {extracting a diagnostic or physiological parameter from medical diagnostic data}
- C12Q 1/689: for bacteria
- G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F 18/2431: Multiple classes
- G06F 40/20: Natural language analysis (semantic analysis of natural language G06F40/30)
- G06F 40/216: using statistical methods
- G06F 40/247: Thesauruses; Synonyms
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06F 40/295: Named entity recognition
- G06T 7/0012: {Biomedical image inspection}
- G06T 7/11: Region-based segmentation
- G06T 7/70: Determining position or orientation of objects or cameras (camera calibration G06T7/80)
- G06V 10/255: {Detecting or recognising potential candidate objects based on visual cues, e.g. shapes}
- G06V 10/774: Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
- G16H 10/40: for data related to laboratory analysis, e.g. patient specimen analysis
- G16H 30/40: for processing medical images, e.g. editing
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 50/50: for simulation or modelling of medical disorders
- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- G16H 50/80: for detecting, monitoring or modelling epidemics or pandemics, e.g. flu
- C12Q 1/701: {Specific hybridization probes}
- G06T 2207/10116: X-ray image
- G06T 2207/20076: Probabilistic image processing
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30061: Lung
- G06V 2201/03: Recognition of patterns in medical or anatomical images
- Y02A 90/10: Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- C12Q: MEASURING OR TESTING PROCESSES INVOLVING ENZYMES, NUCLEIC ACIDS OR MICROORGANISMS (immunoassay G01N33/53); COMPOSITIONS OR TEST PAPERS THEREFOR; PROCESSES OF PREPARING SUCH COMPOSITIONS; CONDITION-RESPONSIVE CONTROL IN MICROBIOLOGICAL OR ENZYMOLOGICAL PROCESSES
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- Y02A: TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE

## Abstract

This disclosure generally pertains to systems and methods for detection of infectious respiratory diseases by implementation of an automated X-rays-based triage approach alongside algorithmic clinical sample pooling for molecular diagnosis. Certain embodiments relate to methods for the development of deep learning algorithms that perform machine recognition of specific features and conditions in chest X-ray imaging data. The chest X-ray imaging data is used to guide the pooling strategy of clinical samples for a molecular test.

## Claims

### Claim 1

A method for detection of an infectious respiratory disease by implementation of an automated X-rays-based triage approach alongside algorithmic clinical sample pooling for molecular diagnosis, comprising:
receiving a clinical sample and chest X-ray scan imaging procedure data from a subject;
processing images from the chest X-ray scan imaging procedure data;
detecting and localizing medical abnormalities of the images using a deep learning system,
generating a first score that corresponds to a level of recognition of each of the medical abnormalities and outputting a bounding box representing the precise location and extent of the medical abnormalities;
generating a second score for each subject that corresponds to a level of recognition of an infectious respiratory disease based on the first scores of the medical abnormalities that suggest a typical or atypical symptom of the infectious respiratory disease;
mapping the second score to a probability of a subject having the infectious respiratory disease;
ranking the clinical sample from lowest to highest based on the second score; and
selecting a pooling size of clinical samples for a molecular test based on a collective probability of all the subjects having the infectious respiratory disease in a pool, wherein the collective probability is a sum of the probability of each subject having the infectious respiratory disease in the pool and wherein the probability is lesser than or equal to 1.

### Claim 2 (depends on 1)

The method of claim 1, wherein the deep learning system carried out by a computer for analyzing chest X-ray scans is developed by the steps of:
selecting medical imaging scans and extracting the medical abnormalities using natural language processing (NLP) algorithms to generate extracted findings, wherein the extracted findings are used as labels for training a deep learning algorithm, wherein the medical abnormalities comprise blunted costophrenic angle, calcification, cardiomegaly, cavity, cervical rib, consolidation, hyper inflation, fibrosis, prominence in hilar region, opacity, pleural effusion, and scoliosis;
segmenting, via an anatomy segmenter, the selected chest X-ray imaging scans to generate segmentation masks corresponding to chest cavity, lungs, diaphragm, mediastinum and ribs;
outputting, via a region of interest (ROI) generator, a plurality of ROIs that are relevant for detecting a particular abnormality, wherein the ROI generator uses the chest X-ray imaging scan at full resolution and the corresponding anatomy segmentation masks; and
detecting the abnormalities, via abnormality detector, and outputting a low-resolution probability map per each ROI, a confidence score per each ROI, and a confidence score for an entire chest X-ray scan by combining the confidence scores per each ROI, wherein the abnormality detector is a hybrid classification plus segmentation network.

### Claim 3 (depends on 1)

The method of claim 1, wherein the infectious respiratory disease comprise tuberculosis, pertussis, Swine influenza (H1N1), Avian influenza (H5N1), Enterovirus infectious diseases, Influenza A, Influenza B, Bronchitis, Coronavius COVID-19 (SARS-CoV-2), Severe Acute Respiratory Syndrome (SARS), and Middle East Respiratory Syndrome (MERS).

### Claim 4 (depends on 2)

The method of claim 2, wherein the extracted findings comprise locations, severity, size, shape and texture.

### Claim 5 (depends on 2)

The method of claim 2, wherein the labels comprises scan level labels, ROI level labels, and pixel level labels.

### Claim 6 (depends on 2)

The method of claim 2, wherein the NLP algorithms are rule-based.

### Claim 7 (depends on 1)

The method of claim 1, wherein the deep learning algorithm comprises convolutional neural networks (CNNs).

### Claim 8 (depends on 2)

The method of claim 2, wherein the anatomy segmenter uses a U-Net based neural network.

### Claim 9 (depends on 2)

The method of claim 2, wherein the ROI generator is rule-based.

### Claim 10 (depends on 2)

The method of claim 2, wherein the confidence scores per each ROI are combined using a pooling operator that is a convex approximation of LogSumExp (LSE) function.

### Claim 11 (depends on 1)

The method of claim 1, wherein the system uses a weighted sum of cross entropy loss at three levels comprising ROI level predictions, pixel level probability maps, and final pooled chest X-ray scan level prediction.

### Claim 12 (depends on 2)

The method of claim 2, wherein the abnormality detector is based on ResNeXT-50 with Squeeze-excitation modules (SE-Res NeXT-50).

### Claim 13 (depends on 1)

The method of claim 1, wherein the medical abnormalities comprise blunted CP angle, calcification, cardiomegaly, cavity, consolidation, fibrosis, hilar enlargement, opacity and pleural effusion.

### Claim 14 (depends on 1)

The method of claim 1, wherein the molecular diagnosis comprises detecting the presence of a nucleic acid disease marker and a protein disease marker.

### Claim 15 (depends on 1)

The method of claim 1, wherein the clinical sample is a sputum sample.

### Claim 16 (depends on 1)

The method of claim 1, wherein the molecular diagnosis is a real-time PCR assay that facilitates semiquantitative detection of infectious agents comprising Mycobacterium tuberculosis (MTB) and rifampin (RIF) resistance.

### Claim 17

A system for detection of an infectious respiratory disease by implementation of an automated X-rays-based triage approach alongside algorithmic clinical sample pooling for molecular diagnosis, comprises
a first subsystem for automating detection and localization of medical abnormalities on chest X-ray imaging scans using a deep learning algorithm carried out by a computer, wherein the deep learning algorithm is developed by the steps of:
receiving and processing images from the chest X-ray scan imaging procedure data;
detecting and localizing medical abnormalities of the images using a deep learning system, wherein the deep learning system carried out by a computer is developed by the steps of:
generating a first score that corresponds to a level of recognition of each of the medical abnormalities and outputting a bounding box representing the precise location and extent of the medical abnormalities; and
generating a second score for each subject that corresponds to a level of recognition of an infectious respiratory disease based on the first scores of the medical abnormalities that suggest typical or atypical the infectious respiratory disease; and
a second subsystem for algorithmic pooling of clinical samples from each subject who also receives chest X-ray scans, wherein the algorithmic pooling is carried out by a computer following the steps of:
mapping the second score derived from the first subsystem to a probability of a subject having the infectious respiratory diseases;
ranking the clinical samples from lowest to highest based on the second score; and
selecting a pooling size of the clinical samples for a molecular test based on a collective probability of all the subjects having the infectious respiratory disease in a pool, wherein the collective probability is a sum of the probability of each subject having the infectious respiratory disease in the pool and wherein the probability is lesser than or equal to 1.

### Claim 18 (depends on 17)

The system of claim 17, wherein the deep learning system carried out by a computer for analyzing chest X-ray scans is developed by the steps of:
selecting medical imaging scans and extracting the medical abnormalities using natural language processing (NLP) algorithms to generate extracted findings, wherein the extracted findings are used as labels for training a deep learning algorithm, wherein the medical abnormalities comprise blunted costophrenic angle, calcification, cardiomegaly, cavity, cervical rib, consolidation, hyper inflation, fibrosis, prominence in hilar region, opacity, pleural effusion, and scoliosis;
segmenting, via an anatomy segmenter, the selected chest X-ray imaging scans to generate segmentation masks corresponding to chest cavity, lungs, diaphragm, mediastinum and ribs;
outputting, via a region of interest (ROI) generator, a plurality of ROIs that are relevant for detecting a particular abnormality, wherein the ROI generator uses the chest X-ray imaging scan at full resolution and the corresponding anatomy segmentation masks; and
detecting the abnormalities, via abnormality detector, and outputting a low-resolution probability map per each ROI, a confidence score per each ROI, and a confidence score for an entire chest X-ray scan by combining the confidence scores per each ROI, wherein the abnormality detector is a hybrid classification plus segmentation network.

### Claim 19 (depends on 17)

The system of claim 17, wherein the infectious respiratory diseases comprise tuberculosis, pertussis, Swine influenza (H1N1), Avian influenza (H5N1), Enterovirus infectious diseases, Influenza A, Influenza B, Bronchitis, Coronavius COVID-19 (SARS-CoV-2), Severe Acute Respiratory Syndrome (SARS), and Middle East Respiratory Syndrome (MERS).

### Claim 20 (depends on 18)

The system of claim 18, wherein the extracted findings comprise locations, severity, size, shape and texture.

### Claim 21 (depends on 18)

The system of claim 18, wherein the labels comprises scan level labels, ROI level labels, and pixel level labels.

### Claim 22 (depends on 18)

The system of claim 18, wherein the NLP algorithms are rule-based.

### Claim 23 (depends on 17)

The system of claim 17, wherein the deep learning algorithm comprises convolutional neural networks (CNNs).

### Claim 24 (depends on 18)

The system of claim 18, wherein the anatomy segmenter uses a U-Net based neural network.

### Claim 25 (depends on 18)

The system of claim 18, wherein the ROI generator is rule-based.

### Claim 26 (depends on 18)

The system of claim 18, wherein the confidence scores per each ROI are combined using a pooling operator that is a convex approximation of LogSumExp (LSE) function.

### Claim 27 (depends on 17)

The system of claim 17, wherein the system uses a weighted sum of cross entropy loss at three levels comprising ROI level predictions, pixel level probability maps, and final pooled chest X-ray scan level prediction.

### Claim 28 (depends on 18)

The system of claim 18, wherein the abnormality detector is based on ResNeXT-50 with Squeeze-excitation modules (SE-ResNeXT-50).

### Claim 29 (depends on 17)

The system of claim 17, wherein the medical abnormalities comprise blunted CP angle, calcification, cardiomegaly, cavity, consolidation, fibrosis, hilar enlargement, opacity and pleural effusion.

### Claim 30 (depends on 17)

The system of claim 17, wherein the clinical sample is a sputum sample.

### Claim 31 (depends on 17)

The system of claim 17, wherein the molecular diagnosis comprises detecting the presence of a nucleic acid disease marker and a protein disease marker.

### Claim 32 (depends on 17)

The system of claim 17, wherein the molecular diagnosis is a real-time PCR assay that facilitates semiquantitative detection of infectious agents comprising Mycobacterium tuberculosis (MTB) and rifampin (RIF) resistance.