Scientific Session 17 — Breast -MRI/New TechnologiesWednesday, May 3, 2017
2487. Optoacoustic Breast Imaging: Imaging-Pathology Correlation in Benign and Malignant Breast Masses
Butler R1*, Tucker F2, Lavin P3, Neuschler E4 1. Yale University, New Haven, CT; 2. Virginia Biomedical Laboratories, LLC, Wirtz, VA; 3. Boston Biostatistics Research Foundation Inc, Framingham, MA; 4. Northwestern University Feinberg School of Medicine, Chicago, IL
Address correspondence to R. Butler (email@example.com)
Objective: The purpose of this study was to elucidate the histopathologic basis of optoacoustic breast imaging and investigate its ability to differentiate benign and malignant tumors.
Materials and Methods: A HIPAA-compliant, institutional review board–approved clinical phase III pilot study of optoacoustic imaging was undertaken as a prologue for a larger, 2000+ subject U.S. Food and Drug Administration pre–market approval multisite, multireader pivotal trial for the Imagio device (Seno Medical Instruments). One hundred patients with 103 solid or complex cystic and solid breast masses assessed as BI-RADS category 3, 4, or 5 on conventional diagnostic ultrasound were enrolled with informed consent. Of these, 83 masses were biopsied, yielding 39 malignant and 44 benign results. Thirteen masses were followed as BI-RADS category 3 lesions and seven were lost to follow-up or excluded. Optoacoustic imaging features of all masses confirmed as malignant or benign by biopsy at 12-month follow-up were scored on five criteria by seven trained independent readers blinded to the study. Criteria were defined as three internal features (tumor vessels, blush, and hemoglobin) and two external features (boundary zone and peripheral zone vessels) on the basis of patterns of tumor vascularity and deoxygenation seen on optoacoustic images. Optoacoustic scores were correlated with histopathology and differences between benign and malignant lesions analyzed for statistical significance using an unpaired two-sided t test.
Results: Mean optoacoustic scores were higher for malignant masses compared with benign masses for all individual internal and external optoacoustic features, as well as for combined internal and external optoacoustic features (p < 0.0001). External optoacoustic features were more predictive of malignancy than internal features, with a greater difference in their means and 99% CIs between malignant and benign masses. Internal mean scores (and 99% CIs) for malignant and benign masses were 3.6 (3.4–3.8) and 2.6 (2.4–2.8), respectively, for internal vessels; 3.5 (3.3–3.7) and 2.7 (2.3–2.9) for internal blush; and 3.2 (3.0–3.4) and 2.2 (2.0–2.4) for internal hemoglobin (p < 0.0001). External mean scores (and 99% CI) were 4.4 (4.2–4.6) and 2.5 (2.2–2.7) for boundary zone vessels and 3.4 (3.2–3.7) and 1.5 (1.3–1.7) for peripheral zone vessels (p < 0.0001). The total internal score means were 10.3 (9.8–10.8) and 7.3 (6.8–7.8) and the total external score means were 7.9 (7.5–8.2) and 3.8 (3.5–4.2), respectively (p < 0.0001). The overall combined total internal and external optoacoustic score was a strong predictor of malignancy, with mean scores (99% CIs) of 18.1 (17.5–18.8) for malignant and 11.1 (10.3–11.9) for benign masses (p < 0.0001).
Conclusion: Correlation of optoacoustic features of vessel patterns and deoxygenation in breast tumors with histopathology shows a significant difference in scores between benign and malignant lesions. External vessel features within the tumor boundary zone and periphery may be more predictive of malignancy than those in the tumor interior. Further data may support a role for optoacoustic imaging in improving the distinction between benign and malignant masses.