Machine learning hype is increasing among imaging professionals, report finds

According to a new market report from Reaction Data, 77 percent of medical imaging professionals believe that machine learning is “important," compared to 65 percent in 2017. Additionally, 59 percent of respondents reported that they “understand” machine learning in 2018 compared to 52 percent in 2017.  

“Machine Learning is one of the hottest topics in healthcare today, especially in medical imaging,” according to the report. “It promises to provide efficiencies in several areas with the expectation of improving quality while decreasing costs.”  

A total of 152 organizations are represented in the report, of which 36 percent were academic medical centers. Community hospitals (22 percent), integrated delivery networks (15 percent) and imaging centers (12 percent) were also included. Other facilities which represented five percent or less of total input were access hospitals, specialty clinics, cancer hospitals, children's hospitals and VA hospitals.  

Radiologists made up most of the respondents (32 percent), followed by directors of radiology (19 percent), imaging technologists (18 percent), PACS administrators (15 percent), managers of radiology (7 percent), chiefs of radiology (7 percent) and clinical informatics managers, CIOs and CFOs (less than one percent).  

However, 16 percent said they don’t plan on utilizing machine learning in the workplace.

Additionally, machine learning utilization in breast imaging decreased from more than 35 percent in 2017 to roughly 22 percent in 2018 and increased in lung cancer imaging from about 11 percent in 2017 to more than 20 percent in 2018. 

Cardiovascular imaging, chest imaging, bone imaging and neural imaging also saw increases in the application of machine learning with liver and pulmonary imaging slightly decreasing, according to the report.