Industry Trends: AI Revolutionizes Imaging Analysis
The radiology industry is embracing artificial intelligence (AI) to enhance image analysis, and it’s having a profound impact on patient care.
Despite constant advances in the medical imaging space, almost one in four (23%) patients experience false positives on image readings[i]. This can lead to unnecessary invasive procedures and follow-up scans that add cost and stress for patients. And while false negatives happen less often, the impact can be catastrophic. The surprisingly high rate of false positives is due in part to concerns among radiologists about missing a diagnosis. Late detection of disease significantly drives up treatment costs and reduces survival rates. One study showed that 70% of lung cancer patients survive for at least a year if diagnosed at the earliest stage, compared to just 14% for people diagnosed with the most advanced stage of disease[ii].
While radiologists have extensive experience and training in how to read these scans, a visual interpretation of an isolated image can be time-consuming, difficult, and result in a high rate of failure. Fortunately, advances in artificial intelligence (AI) and analytics is addressing this problem.
AI applications for radiology use deep learning algorithms and analytics to systematically assess images for tumors or suspicious lesions, and to instantly provide detailed reports on their findings. These systems are trained on labelled data to identify anomalies by reviewing and processing thousands of related images. When a new image is submitted, the algorithm applies its training to differentiate normal vs. abnormal structures (e.g., benign/malignant).
This deep learning capability needs AI tools to deliver more consistent analysis than visual assessments alone, which has an immediate impact on accuracy and health outcomes. Studies have shown that computer aided screening can decrease false negatives by ~45%,[iii] which allows patients to get diagnosed earlier reducing mortality, and decreasing treatment costs. Robust reporting with lesion tracking and snapshots over time also promotes greater accuracy, transparency and collaboration across care team through the patient’s care.
As these tools become more sensitive, they will also potentially enable earlier diagnosis of disease because they will be able to identify small variances in an image that are not easily spotted by the human eye. They can also be used to track treatment progress, recording changes in the size and density of tumors over time, which can inform treatment, and verify progress in clinical studies.
All the Buzz
The application of AI to imaging analysis has become one the hot trends among radiology professionals. AI for imaging was one of the most popular topics at the 2017 European Congress of Radiology (ECR), and the 2018 conference lived up to its promise to provide even more insights into this trend. The 2018 Congress featured an entire track on artificial intelligence and machine learning, including a challenge session on ‘Artificial intelligence and big data in medical imaging,” and sessions on the principles of deep learning, and how radiologists can integrate AI into their practice.
AI and deep learning where also leading trends at the Radiological Society of North America’s (RSNA) 2017 Congress, which offered sessions on how machine learning will change the radiology practice, how to leverage the value of AI in medical imaging, and how to overcome barriers to integrating AI tools into clinical practice.
Good News for Radiologists
While some industry experts fear that AI will replace radiologists, I see these advances as providing an opportunity rather than a risk. Radiologists can spend ten or more minutes reading each scan, and as the number of scans produced increases, they face increasing pressure to speed the process. AI tools can offer them a safety net, providing them a fast and effective way to speed the process without compromising quality.
These tools can rapidly review a number of images to extract the most meaningful insights, which can then be verified by the radiologist. This provides radiologists with more data to support their findings and frees them to spend more time consulting with physicians and to take a more active role in the patient journey.
As cloud-based storage solutions for medical images become cheaper and new biomarkers are identified, physicians will only increase their use of medical images, putting pressure on radiologists to read more scans in less time.
To adapt, radiologists will need to embrace AI and other automated tools to help them keep up with demands. AI is going to disrupt the radiology industry, the questions now are who will be the leaders, and how will progress be achieved. To make the most of this innovation, the industry must work collaboratively to identify new applications and to demonstrate to regulators, providers, and patients that AI can be a powerful tool in the diagnosis and treatment of cancer patients while delivering economic and healthcare benefits for all industry stakeholders.
[i] Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening N Engl J Med. 2011 August 4; 365(5): 395–409. http://www.nejm.org/doi/full/10.1056/NEJMoa1102873
[ii] Why is early diagnosis important? Cancer Research UK http://www.cancerresearchuk.org/about-cancer/cancer-symptoms/why-is-early-diagnosis-important
[iii] Evaluation of a Real-time Interactive Pulmonary Nodule Analysis System on Chest Digital Radiographic Images van Beek, Edwin J.R. et al. Academic Radiology , Volume 15 , Issue 5 , 571 – 575 https://www.ncbi.nlm.nih.gov/pubmed/18423313
Post Authored By:
Michael Auffret, Senior Director, iBiopsy® Program Management, Median Technologies