Purpose: To test the hypothesis that the observed incidence of important noncardiac findings (NCFs) on myocardial perfusion imaging increases significantly when readers systematically review all rotating projectional images.
Methods: We prospectively studied NCF reported from a population of 7658 stress cardiac perfusion studies using a single-agent same-day protocol by a review of rotating projectional images. The incidence of NCF was evaluated for both an ensemble of general clinical readers and for a subset of readers who systematically review all rotating projectional images. The clinical validation of NCF was determined first from an evaluation of the electronic medical record and then after direct communication with the primary physicians.
Results: NCF incidence for general clinical readers was 0.08%, increasing significantly to 0.82% for readers who systematically review all rotating projectional images (P=0.0013). The spectrum of pathology was diverse, particularly including significant findings such as previously unknown cancers. A total of 53 NCFs were detected, of which clinical follow-up found a corresponding diagnosis in 35 cases (66%). Of these, 43% were previously known, 20% new, and 37% false positives. Thirty-four percent of all NCF had no follow-up, mainly because of lack of additional diagnostic tests.
Conclusion: NCFs are infrequent but often clinically important findings, whose detection is significantly enhanced by systematic reviews of all rotating projectional images.