Grading and Staging of Pancreatic Endocrine Tumor: EUS-FNA-Based Compared to Surgical Pathology
Faisal Mukhtar, Shantel Hebert Magee, Shaoying Li, Isam Eltoum. University of Alabama at Birmingham, Birmingham; University of Vanderbilt, Nashville
Background: A TNM system for staging and grading of Pancreatic endocrine neoplasms (PENs) has been proposed and shown to stratify the risk of PENs. The aim of this study is to assess if EUS-FNA-based TNM satging and staging correlates with that of surgical pathology.
Design: This a retrospective study in which we reviewed medical records, cytology, histology EUS reporst trying to identify cases with adaquate follow-up. Maximum and minimum tumor diameters were extracted from EUS and the surgical pathology reports. Grading was based on Ki67 immunohistochemistry on the cell blocks and surgical specimens. The most proliferative area was photographed and scored using image Immunoratio™ program. Paired t-test was used to compare means of tumor diameters and Ki67 index. Bland-Altman (B-A) plot and kappa statistics were used to assess the differences and agreements between EUS-FNA-based vs. surgical pathology-based staging and grading of PENs.
Results: Of 138 patients with PENs, 61(44%) had surgical resection of whom 48 (79%) had their tumor measurement recorded and 30 (49%) had adequate material for Ki67 staining. The mean (SE) tumor diameter based on surgical pathology, 2.8 (0.26) cm, was not significantly different than that based on EUS, 2.7 (0.18) cm, Figure 1a. The B-A difference between the two methods was 0.1 (95% C.I. -2.4, 2.2), Figure 1b and there was a moderate agreement between the methods regarding staging, kappa statisitics 0.44. The mean (SE) Ki67 index based on histology was 9.6% (1.5%) significantly different than that based on cytology 3.0% (0.7%), p value .01, Figure 2a. The B–A difference between the two methods was 7.0 (95% C.I., -9.4, 23) and it varies with an increasing Ki67 index. This trend was attenuated when the difference was represented as a percentage of the means of the methods, Figure 2b. The two methods have poor agreement on PENs grading, kappa statistic 0.05.
Conclusions: There is moderate agreement between EUS-FNA-based and surgical pathology staging of PENs but poor agreement between the two methods on grading. There is a need for standardization of Ki67 staining and counting for both histology and cytology.
Tuesday, March 20, 2012 9:30 AM
Poster Session III # 54, Tuesday Morning