Artificial Intelligence in Diagnosing Diseases

Reading Time: 7 minutes
5 (100%) 3 votes

It was only a fortnight ago when artificial intelligence (AI) proved to diagnose skin cancer with more accuracy than a human doctor. This is just the tip of a massive iceberg, where AI is gearing up to become a strong tool in diagnosis, aiding better healthcare. From breath… (Featured image is for representational purpose and has been sourced from


It was only a fortnight ago when artificial intelligence (AI) proved to diagnose skin cancer with more accuracy than a human doctor. This is just the tip of a massive iceberg, where AI is gearing up to become a strong tool in diagnosis, aiding better healthcare. From breath analyzers to retina tests, AI is making headway across a gamut of diagnostic methods.

Breath analysis

Researchers from several institutes such as Israel Institute of Technology, Bar Ilan University, Bnai Zion Medical Center, University of Latvia and others conducted a research recently, wherein an artificially intelligent nanoarray was trained on breath to collected samples were from 1404 subjects having one of 17 different disease conditions. Blind experiments showed that 86% accuracy could be achieved with the artificially intelligent nanoarray, allowing both detection and discrimination between the different disease conditions examined. Analysis of the artificially intelligent nanoarray also showed that each disease has its own unique breathprint, and that the presence of one disease would not screen out others. This technology may have a huge impact on the medical industry and will be useful to detect diseases while saving time and costs.

Information in the eye

Google has developed an algorithm to analyze retinal images of the persons to predict smoking habits, blood pressure, age, etc. The algorithm uses artificial intelligence and machine learning to predict a probability of having a heart attack or a stroke.

DeepMind (acquired by Google) has developed an algorithm in collaboration with Moorfields, a specialist NHS eye hospital in London. The algorithm has been trained using anonymized 3D retinal scans provided by Moorfields. The algorithm can be used to detect eye diseases such as glaucoma, diabetic retinopathy, and age-related macular degeneration.

U.S. based Eyenuk Inc. has filed for a patent US9008391B1 disclosing a method to detect diseases such as diabetic retinopathy, cytomegalovirus retinitis, retinopathy of prematurity, clinical myopia, hypertensive retinopathy, stroke, cardiovascular disease, glaucoma, macular degeneration, Alzheimer’s, or macular edema by analyzing retinal images of a patient.

The above image from US9008391B1 shows a system capable of performing retinal image analysis can be applied.

CN107680683A, assigned to Shanghai Muqing Visual Tech Co Ltd. – a Chinese technology company – designs and programs computer graphics films and images, discloses a method to acquire a plurality of fundus images, analyzing the acquired images using artificial intelligence to generate a health report and output a health index.

AI reads X-rays for power-packed diagnosis

Stanford researchers have developed an algorithm that offers diagnosis of diseases by analyzing chest X-ray images. The algorithm has been claimed to diagnose up to 14 types of medical conditions, with the specific success of diagnosing pneumonia better than a radiologist.

Lunit Inc., an AI-powered medical image analysis software company, presented a chest X-ray solution at 2017 Radiological Society of North America (RSNA) annual meeting. The solution can detect major chest abnormalities, lung nodule/mass consolidation and pneumothorax with 97% accuracy in nodule detection and 99% for consolidation and pneumothorax.

Lunit Inc. has 17 patents spread across nine INPADOC families, as per Thomson Innovation database. Four of these nine families seem to be related to using of machine learning/artificial intelligence in medical diagnosis. Prominent among these are US20170061608A1, US20170236271A1, WO2017069596A1, and WO2017010612A1. These patent applications relate to methods of using AI/Machine Learning in pathological image analysis, X-Ray image analysis to detect abnormalities/diseases.

The above figure from US20170061608A1 shows a cloud-based pathological analysis system.

The above figure from US20170236271A1 shows a classification apparatus for pathologic diagnosis of a medical image.


Mumbai-based (India) health-tech start-up,, has developed an artificial intelligence (AI)-powered technology that can examine X-rays, MRIs and CT scans to diagnose diseases such as tuberculosis (TB) or stroke used 1.5 million X-rays and fed them to an algorithm. The system was trained to understand what an abnormality would look like. Deep learning automatically identifies multiple disease states from a medical image. The technology can make TB screening in the country faster, effective and economical. It may also help prevent many non-TB patients from having to go through the confirmatory GeneXpert test.’s algorithm can rule out brain bleed in seconds making the treatment protocol for stroke much more effective. These algorithms are unique because they clearly highlight the problem area in the patient’s X-rays or CT scans through a heat map.

Jia Li, who leads research and development at Google Cloud, revealed new research on applying AI to radiology imaging. Li and his colleagues used deep learning to detect abnormalities in chest x-rays. CNN (Convolutional Neural Network) is applied to the input image so that the model learns the information of the entire image and implicitly encodes both the class and location information for the disease. The image is sliced into a patch grid to capture the local information of the disease.

IBM Watson also seems to provide solutions to analyze x-ray images. The project is called “Eyes of Watson”. The solution can detect breast cancer from x-ray images with great accuracy.

US7738683B2, assigned to Carestream Health Inc., discloses a method to detect abnormalities in a medical image of a patient. The method includes receiving multiple medical images from different modalities. Each image is analyzed to detect any abnormalities and detected abnormalities from each image are combined. Finally, it can be detected whether the patient is suffering from any disease.

The above images from US7738683B2 show exemplary abnormality detection methods.

Tissue image analysis

Researchers at Stanford University presented use deep Convolutional Neural Networks (CNNs) algorithms for dermatologist-level classification of skin cancer. The research was carried out using dataset, comprising 129,450 clinical images and consisting of 2,032 different diseases. The dataset was used to train a CNN algorithm. The research was successful at detecting the deadliest skin cancer with great accuracy.

US20160253466A1 filed by Stanford University, describes a method for performing multi-stage detection and classification of cancer regions from digitized images of biopsy slides

Google developed a Deep Neural Network, which was trained to detect cancer cells by analyzing images of human tissues. The algorithm is applied to a microscope to detect this in real time. The model was trained on images of human tissue and the testing results have been impressive, with the AUC as high as 0.98.

The above image shows Google’s Machine Learning Model

Infervision, a Beijing-based company, uses machine learning algorithms and computer vision methods to support lung cancer diagnosis. The product was promoted as a “second pair of eyes” for radiologists as it can identify more than 20 different cardiothoracic lesions and can be used in physical examinations to screen for lung cancer characteristics.

The Niramai, a start-up company, proposes a multi-patented solution, SMILE,. Big data analytics, artificial intelligence and machine learning are combined for reliable, early and accurate breast cancer screening. Early results, from data of 300 patients collected in two hospitals and one diagnostic center, demonstrate high accuracy. The company seems to have filed 7 unique families (total 16 patent publications). All families seem to be related to the technology described above.

Researchers from Showa University in Yokohama, Japan developed a computer-aided diagnostic system that uses an endocytoscopic image – a 500-fold magnified view of a colorectal polyp – to analyze about 300 features of the polyp after applying Narrow-Band Imaging (NBI) mode or staining with methylene blue to detect colorectal cancer in less than a second. Overall, 306 polyps were assessed real-time by using the AI-assisted system, providing a sensitivity of 94%, specificity of 79%, an accuracy of 86%, and positive and negative predictive values of 79% and 93% respectively, in identifying neoplastic changes.


A patent publication US20160235372A1 by Telebyte Inc. discloses a method of detecting cancer using artificial intelligence. A target area is illuminated using a plurality of light signals of the different wavelength. A processing device processes reflected signals from the target area to compute estimates of absorption and scattering and applies an artificial intelligence system to the absorption and scattering estimates to output a decision as to whether cancer is detected in the target area.

The above image from US20160235372A1 shows a block diagram of a system for cancer diagnosis.

The above image from US20160235372A1 shows a flowchart of a method for cancer diagnosis.

Artificial intelligence is transforming every industry, but the role it will play in healthcare is overwhelming. With more research and development being done by medical and healthcare companies on the improving computational power of AI and as medical practitioners are being trained on using the AI-powered tool; Artificial Intelligence has the potential to revolutionize medical diagnosis in the future, leading to better medical care and diagnosis for everyone.

(Featured image is for representational purpose and has been sourced from


Neha Garg
Neha Garg

From delving into the pages of history of enchanting places to mining infringement evidence from large patent landscapes, exploring the lesser known keeps Neha on her toes at all times.

Post a Comment

Your email address will not be published. Required fields are marked *