Improved Workflows. Improved Patient Care.

ACR DSI Releases TOUCH-AI Use Cases

The American College of Radiology Data Science Institute™ (DSI) is working with radiology professionals, industry leaders, government agencies and patients to foster the adoption of AI in health care.

The starting point for an effective AI ecosystem in radiology is establishing clinical context and technical requirements by providing the framework, strategy and focus for moving AI from concept to radiological practice. We’re laying the foundation for successful AI, so it can begin to impact patient care.

Visit the TOUCH-AI Directory »

Here’s How It Works

Use Cases

Use cases are scenarios where the use of artificial intelligence (AI) may help improve medical imaging care. ACR DSI Use Cases are available in the TOUCH-AI Directory.


Each use case provides narrative descriptions and standards that specify the health care goal of AI and how it will interface with users and tools.            


Standardized terminology ensures interoperability and will result in algorithms compatible with other radiology tools.                        

Visit the TOUCH-AI Directory »

Example Use Cases

LV Functional Assessment (for conditions like heart disease)

Heart disease is the leading cause of death in the United States — claiming more than 600,000 lives each year. This use case will help developers produce algorithms that provide fast, accurate measurements of the most important factors in heart healthiness. This gives clinicians a reliable and quick indicator to a patient’s heart health which holds promise for better patient care over time.

Suspicious Breast Microcalcification Analysis (for breast cancer)

Although breast calcifications are usually noncancerous, certain patterns of calcifications — such as tight clusters with irregular shapes — may indicate breast cancer or precancerous changes to breast tissue. Algorithms resulting from this use case would automatically classify microcalcifications to help physicians better determine which patients may need to return for further testing. This can help prevent “false positive” exams and help patients avoid unwarranted procedures (including biopsies), with less risk of underdiagnoses.

Scoliosis Detection

More than 4 million Americans are stricken with scoliosis. It remains the most common spinal deformity in school-age children. This use case will aid developers in creating algorithms that speed capture and analysis of common relevant metrics that are typically time-intensive to do manually. Though metrics like the Cobb angle are commonly reported today, they have their limitations in accounting for vertebrae rotation. So the use case also defines a series of secondary metrics that would best help radiologists call and treat scoliosis. This can help avoid variance in reporting and managing scoliosis patients.

Pneumothorax Detection (collapsed lung)

Trauma, cigarette smoking, drug abuse and certain lung diseases can lead to a “tension” (fully) or “simple” (partially) collapsed lung. This condition — which strikes nearly 74,000 Americans each year — can be deadly if not diagnosed quickly and accurately. Algorithms resulting from this use case can be used to speed diagnosis and intervention from a variety of end points — including prioritizing worklists, queuing up diagnostic tools for radiologists and alerting referring providers of the condition.