The radiology community is abuzz with talk of artificial intelligence (AI) systems that can assist physicians with image interpretation and perform other tasks. Like any new technology, it will take time before AI gains widespread acceptance due to the cost of implementation. This is analogous to the early-day PACS, where the benefits of efficiency had to be proven in order to justify the expenditure for such a costly system. Today almost all imaging is interpreted on computer systems … when was the last time anyone looked at a piece of film?
Artificial Intelligence appears in many different systems that are designed to do many different things. The imaging news source AuntMinnie.com has a section dedicated to AI, and recently listed articles describing systems that
- Detect physical child abuse
- Identify orthopedic implants make and model on radiographs
- Speed up quantitative volumetric MRI
- Predict complications from CT-guided biopsy
- Perform radiation therapy contouring, and
- Analyze chest CT to shorten turnaround time.
Our own Deep Dive Analytics solution has been used to identify patients in need of follow-up to remind them to return for prescribed imaging and implanted device removal, among other uses. Our blog contains a number of articles describing the use of the system.
Medicare will now pay for the use of artificial intelligence in only two limited cases – diagnosis of diabetic retinopathy under the Medicare Physician Fee Schedule (MPFS) and diagnosis and treatment of large-vessel occlusion strokes under the Inpatient Prospective Payment System (IPPS). The latter is a payment to hospitals that does not provide any professional component reimbursement to radiologists. The Imaging Wire pointed us to the article Who Will Pay for AI? by Drs. Melissa M. Chen, Lauren Parks Golding, and Gregory N. Nicola, that was published March 3, 2021 by the Radiology Society of North America (RSNA). The authors conclude that “Payment for AI in the current fee-for-service environment may be challenging, and sustained adoption of AI may not occur within the framework of the IPPS and Physician Fee Schedule.” They suggest that AI might have a bigger role in value-based systems such as the Alternative Payment Models (APM) that are part of Medicare’s Quality Payment Program.
Direct reimbursement for the use of AI is only one piece of the puzzle. In order to gain widespread usage, AI systems will have to provide a Return On Investment (ROI) in terms of cost savings or increased volume rather than direct reimbursement. For example, the use of Deep Dive Analytics not only provides improved patient care, but when patients return for regular follow-up imaging the practice benefits from the scan volume. When an implanted device such as an IVC filter is removed in a timely manner, it provides appropriate procedural revenue and also helps to establish the radiologist’s value in the medical community.
One upcoming use of AI is in clinical decision support (CDS). The use of CDS has been mandated by Medicare for certain imaging procedures beginning in 2022, with financial penalties applied to radiologists who perform exams without the ordering physician having consulted a CDS. The use of AI in a medical community with an integrated information system can make this required process easier for ordering physicians, thereby reducing the risk of denied payment to radiologists.
The American College of Radiology, along with several other radiology interest groups, have proposed new radiology-specific CPT® codes that include quantitative ultrasound tissue characterization (non-elastographic) and artificial intelligence analysis for detection of vertebral fractures. If they are approved, they would be temporary Category III codes that are usually assigned to emerging or experimental technologies and not generally reimbursed by most payers until permanent codes and RVU values are assigned. The proposed codes could be released in July for potential use in 2022, although final adoption of permanent codes could take several years.
We no doubt will hear much more about the integration of AI into private practice. The Journal of the American College of Radiology has called for case studies on the topic, and we look forward to learning about the ways practices have implemented cost-effective systems. Subscribe to this blog to get the latest on the state-of-the-art in radiology AI.