Author + information
- Daniel Nguyen,
- Aditi Nayak,
- Christopher Pray,
- Christopher Link,
- Andrew Althouse and
- Prem Soman
Background: Prior studies have established that the selection of patients for stress myocardial perfusion imaging based on appropriate use criteria results in superior test performance. However, the repeatability of AUC category allocation has not been tested.
Methods: Two reviewer pairs, blinded to each other, each comprising of a cardiology fellow and an internal medicine resident, analyzed the electronic medical records of 300 consecutive patients undergoing stress SPECT imaging, after initial participation in a joint training session on 100 cases with an attending physician for standardization. The 2013 Astellas multimodality AUC mobile phone application was used to allocate standard appropriateness categories, and identify uncategorized indications.
Results: Reviewer pairs were concordant on 82.5% of studies (kappa statistic = 0.59). Both pairs of reviewers scored 67.4%, 7.0% and 6.0% of studies as appropriate, maybe appropriate and rarely appropriate, respectively. However, the AUC score was identical only 57% of cases. Failure modes were identified. 2% of studies could not be categorized.
Discussion: The allocation of AUC category by trained operators is highly repeatable. However, failure modes and uncategorized indications are encountered relatively frequently. Even when the allocated AUC category was the same, there was often a difference in the actual AUC score indicating heterogeneity of the algorithm chosen for score derivation. These results may help improve subsequent iterations of AUC.
Poster Hall, Hall C
Saturday, March 18, 2017, 3:45 p.m.-4:30 p.m.
Session Title: Nuclear Cardiology: Quality
Abstract Category: 30. Non Invasive Imaging: Nuclear
Presentation Number: 1246-224
- 2017 American College of Cardiology Foundation