Pharmaceutical firms face many challenges in bringing their products through clinical trials — from enrollment delays to data quality issues — and every advantage can make a difference.
In image analysis, sponsors can gain a critical advantage by leveraging artificial intelligence (AI) or machine learning (ML) — computer solutions that mimic human intelligence and are continually enhanced and updated to become more accurate as they take in more data. One subset of machine learning, known as deep learning or convoluted neural networks, offers high accuracy in image recognition and has applications for medical image analysis and interpretation.1,2 For example, in liver and spleen volume measurement, AI and ML allow rapid and efficient assessment of every slice of an image series.3 They also can aid in more accurate identification, classification and measurement of tumors.4
These benefits are particularly pronounced for medical imaging in clinical trials, where eligibility decisions and efficacy evaluations can be accelerated, and data accuracy and quality improved. AI can reduce variability in imaging endpoints, enhance the reliability of those endpoints for use in statistical models of efficacy and safety, and reduce the cost of providing quantitative endpoints for a trial.5