Earlier – and Better – Cancer Detection

July 31, 2020

The earlier cancer is detected, the more effectively it can be treated. However, cancer can be difficult to detect in the early stages. New artificial intelligence (AI) and IoT tools are in development to give doctors the edge in detection.

IoT Devices to Detect Breast Cancer

iTBra, Eva Bra and Smart Bra are in development to aid breast cancer detection. These IoT devices are more comfortable than a mammogram and more accurate, with the ability to detect cancer in earlier stages and with fewer false positives and negatives.

Cancer cells generate more heat than normal cells. Sensors in the iTBra and Eva Bra detect cell temperatures and produce thermal images within two hours of wear. An app alerts users if an abnormality is found. Images can also be sent to doctors for further examination.

The Smart Bra uses ultrasound and echography scans to detect cancer cells. Researchers also hope the technology can be adapted to emit low-dose ultrasound waves to destroy cancer cells while patients wear the device.

AI for Skin Cancer Detection

AI is proving better than dermatologists at detecting skin cancer, outperforming in a study with 58 dermatologists. Since skin cancer is usually discovered through sight, doctors must be highly adept at recognizing unusual moles. However, through machine learning, AI accurately detected skin cancer at a 6% higher rate than experts. Dermatologists could use the technology to aid their diagnoses. General practitioners and other non-experts could use the AI via smart phone app to aid in detecting cancers and avoid unnecessary referrals for benign cases.

Lung Cancer Detection with AI

AI is able to detect lung cancer with the same accuracy as pathologists. Treatment methods for lung cancer depend on the grade of tumor, but pathologists can disagree on classification based on the appearance of tissue. AI can highlight features of a tumor for pre-diagnosis or diagnostic aid. It can also suggest genetic testing for patients with detected histological patterns.

Another AI program has been trained to detect lung cancer in slides of tissue samples. Researchers believe the technology could aid surgeons in assessing the quality of biopsy samples. It can also detect gene mutations that cause cancer from 64% to 86% accuracy.

Classifying Brain Tumor Tissue with AI

AI can distinguish healthy brain tissue from tumor tissue in biopsies in less than three minutes. This allows tissue to be analyzed within the operating room, saving time for both surgeon and patient. In a clinical trial it was found both pathologists and the AI classified tumor tissue correctly 95% of the time. In the instances where one party was wrong, the other was correct, making the AI an excellent crosschecking tool. The AI can also visualize tumor cells during surgery, helping surgeons avoid cutting healthy tissue.

Using AI to Test DNA Samples

Researchers have developed AI that can identify cancer in DNA samples. The AI was able to detect later cancers at a far more successful rate than early stage cancers, but further development of the technology may improve those numbers. When the AI did detect cancer, it was able to identify the location of the cancer within the body with 93% accuracy. The AI has a low false positive rate and could be a boon for health professionals in quickly identifying cancers and developing treatment plans.