Advanced structured reporting templates could improve diagnostic performance

Implementing a structured template with dropdown menus can improve a radiologist’s adherence to the Prostate Imaging Reporting and Data System (PI-RADS) version 2 lexicon, according to a new study published in the Journal of the American College of Radiology. The authors added that improving adherence to practice guidelines such as PI-RADS may have a positive influence on diagnostic performance.

The authors retrospectively studied data from patients who underwent prostate MRI followed by MRI-ultrasound fusion biopsy at a single facility from October 2015 to October 2017. The facility had started using PI-RADS version 2 in June 2015 through “a basic structured reporting template with subheadings.” In June 2016, a much more advanced reporting template with “dropdown menus for describing and scoring lesions using the PI-RADS version 2 rules and lexicon” was introduced.

Overall, adherence to the PI-RADS lexicon jumped from more than 32 percent before implementation of the more advanced reporting template to more than 88 percent after implementation. Sensitivity of prostate MRI for clinically significant prostate cancer (CS-PCa) in peripheral zone (PZ) increased from 53 percent to 70 percent. The specificity of prostate MRI for CS-PCa in the PZ changed from 60 percent to 55 percent, viewed as an insignificant change.

“The results of this study suggest that using a structured template with dropdown menus that incorporate the PI-RADS version 2 lexicon and classification rules improves radiologist adherence to PI-RADS version 2 and may help increase the sensitivity of MRI to detect CS-PCa in the PZ without significantly compromising specificity,” wrote lead author Hiram Shaish, MD, department of radiology at Columbia University Medical Center in New York City, and colleagues.

Shaish et al. noted the significance of the group’s improvement in sensitivity after implementation. What does this say about the potential impact advanced structured reporting templates can have on patient care?

“The improvement in sensitivity was not due to varying frequencies of MRI interpretation by the six radiologists, years of radiologist experience reading prostate MRI, different MRI scanners, different US-MRI fusion biopsy operator, or any additional PI-RADS version 2 training,” the authors wrote. “In fact, the only significant difference we detected between the two cohorts was the adherence to PI-RADS version 2 rules and lexicon as determined by retrospective review of the original radiology reports. This suggests that when radiologists are forced to adhere to PI-RADS version 2, an example of expert-formed consensus guidelines, diagnostic performance may improve.”

Shaish and colleagues noted that there are three tiers of structured reporting—the first contains simple headings, the second contains subheadings and then the third makes use of a specific set of rules or guidelines. In their study, radiologists began with a template that fell under the second tier before moving on to a template in the third tier.

“Structured reporting is encouraged by the American College of Radiology as a way to promote standardization and communication of results and to improve adherence to expert-formed consensus guidelines,” the authors wrote. “More recently an increase in reimbursement has been tied to complete reporting, which is facilitated through structured reporting. Data mining and research are also facilitated through the use of structured reporting.”