1959. Can Ultrasound Molecular Imaging Characterize Ovarian Masses into Benign or Malignant?
Authors * Denotes Presenting Author
  1. Neha Antil *; Stanford University
  2. Ahmed El Kaffas; Stanford University
  3. Ann Folkins; Stanford University
  4. Huaijun Wang; Stanford University
  5. Teri A. Longacre; Stanford University
  6. Jonathan S. Berek ; Stanford University
  7. Amelie Lutz; Stanford University
To study if ultrasound molecular imaging (USMI) can characterize ovarian masses (OMs) into benign or malignant when combined to a clinical-grade contrast agent (Kinase insert domain receptor [KDR]-targeted contrast microbubble [MBKDR] and correlate USMI results with gold standard surgical histopathology (HPE) findings.

Materials and Methods:
USMI was performed in 24 women (age 48 - 79 years) with complex OMs after meeting the eligibility criteria and obtaining informed consent. While imaging in dual mode with B-mode and contrast mode side-by-side, MBKDR (BR55, Bracco, [0.03, 0.05 and 0.08 mL/kg of body weight]) were manually injected over 10 seconds followed by 10 mL saline flush. Imaging was obtained starting with initial 45 seconds acquisition to capture the wash-in phase of MBKDR, followed by 10 second acquisition every 2 minutes until 30 minutes. Surgery for these OMs was performed within a 2-week interval following imaging. Qualitative and quantitative USMI analysis was performed by two radiologists in consensus who were blinded to final HPE diagnosis. Both radiologists visually assessed the presence of focal enhancement and graded using a three-grade visual scale: strong (well-defined and strong visual stationary targeted imaging signal), weak (enhancement is weak but considered stationary), or no enhancement (no focal stationary targeted imaging signal). For quantitative USMI analysis, the ROI was drawn over the target lesion and mean contrast signal intensity was calculated and recorded using the in-house built software (ITK-SNAP). USMI analysis was then correlated with HPE results.

Out of 24 lesions, 22 (92%) were confirmed in the ovaries on surgical HPE; 2/24 (8%) lesions were intraoperatively found outside the ovaries (leiomyoma and hematosalpinx) and, were excluded from the analysis. A total of 13/22 (59%) lesions were malignant, with a mean size of 5.38 cm (range 1.2 - 10.0 cm) ; and 1 of the malignant lesions was a metastasis from a neuroendocrine gastrointestinal tumor. Nine out of 22 ( 41%) lesions were benign, with a mean size of 1.86 cm (range, 0.7 - 3.7 cm). Overall, qualitative USMI analysis showed a sensitivity of 76.92% (10/13), specificity of 77.78% (7/9), PPV of 83.33% (10/12) and NPV of 70% (7/10), while quantitative USMI analysis showed a sensitivity of 100% (13/13), specificity of 88.89% (8/9), PPV of 92.85% (13/14) and NPV of 100% (8/8), when compared to gold standard HPE. None of the malignant lesions was miscategorized as a benign lesion. In conclusion, KDR-targeted USMI showed promising results in noninvasively characterizing OMs without any malignant lesions miscategorized as benign. However, this requires further validation.

A targeted USMI strategy provides scope for further development of US targeted microbubbles, to address imaging biomarkers such as KDR that can help detect and characterize other cancers as well.