This study aims to determine the accuracy of a novel ultrasonography (US) scanning and reporting protocol to detect recurrences. The secondary aim is to compare US and MRI accuracy and agreement.
In this IRB-approved prospective study, consecutive patients presenting for MRI surveillance after resection were enrolled and underwent same-day US. Blinded to clinical information and the MRI, the US scanner characterized lesions using a proposed novel lexicon. Outcome was defined either by histology or a subsequent MRI scan confirming the presence or absence of recurrence. Fisher's exact test and Kappa test were performed to assess of the significance and agreement between US, MRI, and outcome.
A total of 68 US scans were performed on 55 patients. The overall accuracy to diagnose recurrence was the same for US and MRI (92.6%) while US was less sensitive (75.0% vs. 91.7%) but more specific (97.6% vs. 92.9%) than MRI. The two lesions missed by US but not MRI were an entirely intraosseous metastasis and a subcentimeter skin nodule. There was strong agreement between US and MRI with outcome (k = 0.787 and 0.801, respectively).
These pilot data suggest the accuracy of this novel US local recurrence surveillance method is comparable to MRI. A multi-institutional prospective trial would increase power and determine reproducibility.