Atypia of Undetermined Significance in Thyroid Fine Needle Aspiration Biopsy Evaluated by Whole Slide Image Analysis with Risk Assessment of Malignancy
Brian T Collins, Laura E Collins. Washington University in St Louis, St. Louis, MO
Background: Atypia of undetermined significance (AUS) is a specific category of thyroid FNA biopsy that presents significant interpretive and patient management challenges. The cytomorphologic features are subjective and difficult to apply consistently. Individual and institutional rates of AUS can vary widely from low single digits up to 20%. There is an increased intermediate incidence of malignancy in the AUS category. Defining objective criteria could allow for more uniform and consistent categorization and risk assessment of malignancy of thyroid AUS.
Design: In order to investigate objective morphologic criteria for the classification of thyroid AUS and asses for malignancy risk, whole slide images (WSI) (Aperio) were prepared and analyzed by image analysis for a set of morphologic criteria. For one calendar year, all thyroid FNA AUS cases were identified and those with subsequent thyroid excision selected. Each FNA case was reviewed and a single representative Diff-Quick stained smear selected and subjected to WSI. The WSI were evaluated in a blinded fashion. All individual follicular groups on the WSI were manually delineated, and measured for total group number, group area (um), nuclear area (um), and nuclear to cytoplasmic ratio (N:C).
Results: There were 44 cases of thyroid AUS with surgical excision and defined outcome. Surgical resection categorized 15 benign non-neoplastic, 8 follicular adenomas, and 21 malignant entities. For malignant cases, 17/21 (81%) had an N:C ratio of less than 0.50 and 4/21 (19%) had an N:C ratio of greater than 0.50. For benign non-neoplastic cases, 6/15 (40%) had an N:C ratio of less than 0.50 and 9/15 (60%) had an N:C ratio of greater than 0.50. In thyroid AUS patients, a N:C ratio of less than 0.50 had a risk ratio for malignancy of 2.4. For malignancies, 17/21 (81%) had over 20 cell groups while the benign non-neoplastic category had 8/15 (53%) with greater than 20 cell groups. The benign non-neoplastic category had an average cell group count of 46 compared to 67 for the malignant category.
Conclusions: WSI image analysis can provide objective morphologic measurements. Utilizing WSI with image analysis on thyroid AUS cases with known outcomes demonstrated that an N:C ratio of less than 0.50 had a 2.4 risk ratio for malignancy. Determining subsequent clinical follow up on patients with AUS is clinically evolving. Establishing defined objective WSI criteria has the potential to provide a risk assessment of malignancy for AUS thyroid FNA cases.
Tuesday, March 5, 2013 1:00 PM
Poster Session IV # 92, Tuesday Afternoon