Developing & Validating a Clinical Decision Support Tool for ER-Targeted PET Imaging With 16α-18F-Fluoro-17β-Fluoroestradiol

Abstract

Background

Estrogen receptor (ER) status in breast cancer (BC) is routinely determined by immunohistochemistry (IHC); however, this technique is not without limitations, including false results. Imaging of FES-PET (fluoroestradiol F18) injection provides high diagnostic accuracy of ER expression, supplementing information from biopsy. A Clinical Decision Support (CDS) tool was developed to better assess its clinical usefulness in metastatic and recurrent breast cancer management. This study evaluated a conceptual tool that reflects clinical practice variables.

Methods

Individual patient characteristics - candidacy for therapeutic treatment and rate of recurrence - determined initial eligibility. The CDS tool uses rules (IF-THEN statements) to produce an output on the diagnostic accuracy of ER status based on tumor burden, anatomical location(s) of metastasis, heterogeneity, and confidence in sample collection & pathology accuracy (CSC & PA). An Excel-based probability decision tree calculates the accuracy of ER expression.

Results

360 oncologists in the United States participated in the survey study. 223 respondents identified as medical oncologists (62%), 77 as clinical oncologists (21%), and 60 as hematologic oncologists (17%). 93% of respondents found the CDS tool intuitive and easy to follow with medical and clinical oncologists favoring the tool more than hematologic oncologists. Individual CDS attributes - clinical criteria, diagnostic comparator, true positive and true negative, patient inclusion and exclusion, and clinical patient level inputs - were tested with overall positive feedback.

Conclusions

Based on respondent feedback, further development of CDS tools are warranted for potential use in patients’ diagnostic workup.