Cardiac MR analysis packages prone to intersoftware variability

Semiquantitative cardiac MR perfusion analysis software provides intra- and inter-observer repeatability and reproducibility, but variability among software programs was significant, according to a study published online Dec. 13, 2012, in Radiology. The findings led the authors to caution that data produced with different software packages are not interchangeable.

Although multiple vendors offer software packages for myocardial perfusion analysis, few studies have examined the reproducibility of semiquantitative cardiac MR perfusion analysis software.

Astri Handayani, MSc, from the department of radiology at University Medical Center in Groningen, the Netherlands, and colleagues sought to characterize the repeatability and reproducibility of semiquantitative MR perfusion analysis by comparing different software packages.

The study focused on four dedicated software packages. Two independent reviewers analyzed two cardiac MR exams conducted one day prior to and one day after percutaneous coronary intervention for eight patients with single-vessel disease. Reproducibility and repeatability were evaluated for perfusion upslope and myocardial perfusion reserve index.

The researchers observed “good and comparable” agreement for repeat measurements within each software platform, with no significant variability by observer after poor-quality data were discarded.

However, they found significant differences in most values produced by the different software packages. “Our study shows that intersoftware variability of perfusion parameters are significantly higher than intra- and inter-observer variability, the last two being comparable in magnitude,” wrote Handayani et al.

The findings suggest relatively consistent variability among packages across users but not across packages. After calling for intersoftware calibration and noting that it remains unavailable, the researchers offered several suggestions to providers:  

• Perform within group and/or sequential observations of quantitative perfusion parameters on one platform;
• Specify and hold constant measurement settings;
• Acknowledge that semiquantitative measurements require interpretation of values as estimates; and
• Take into account the spectrum of the patient’s condition in clinical decision making.

Finally, the low concordance among software packages highlights the importance of resolving such issues prior to incorporation of semiquantitative perfusion analysis in clinical evaluation, concluded Handayani and colleagues.