Medical College of Wisconsin
CTSIResearch InformaticsREDCap

Correlation of SMAD4 Expression in Pretreatment Cytologic Specimens and Postneoadjuvant Treatment Surgical Specimens in Patients With Pancreatic Cancer International Journal of Radiation Oncology, Biology, Physics ORAL SCIENTIFIC SESSION| VOLUME 96, ISSUE 2, SUPPLEMENT , S141, OCTOBER 01, 2016

Date

10/01/2016

Abstract

Purpose/Objective(s)

Pancreatic cancers which have loss of SMAD4 protein have been correlated with the metastatic phenotype. However, accuracy of SMAD4 immunostaining based on pretreatment needle biopsy, and the effect of neoadjuvant therapy on the prognostic value of such immunostaining has not been described in patients with operable pancreatic cancer. We hypothesized that neoadjuvant therapy may result in the loss of SMAD4 expression (treatment effect), but that gain of SMD4 expression found in the resected tumor after neoadjuvant therapy was likely due to the failure to detect expression in the FNA sample.

Materials/Methods

In an ongoing prospective clinical trial for patients with localized pancreatic cancer, we have used molecular profiling of endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) biopsy specimens (prior to neoadjuvant therapy) and resected surgical specimens (after neoadjuvant therapy) to direct chemotherapy selection. We evaluated a subset of patients with paired cytologic and surgical specimens for SMAD4 expression using immunohistochemical staining. SMAD4 expression was scored by a single pathologist using both staining intensity and the percent of cells with immunochemical reactivity. Lack of nuclear SMAD4 staining was considered loss of SMAD4 expression.

Results

SMAD4 expression was evaluable in 46 patients from EUS-FNA specimens; expression was lost in 33 (72%) and present in 13 (28%). Neoadjuvant therapy and surgery was completed in 41 (89%) of 46 patients; 5 patients did not undergo surgery due to the development of metastatic disease. SMAD4 expression was evaluable in 38 (93%) of the 41 surgical specimens; expression was lost in 22 (58%) and present in 16 (42%). Of the 38 patients with both FNA biopsy specimens and surgical specimens, SMAD4 expression was concordant in 25 (66%); 19 (50%) were negative and 6 (16%) were positive. Discordant expression was observed in 13 (34%) patients; expression changed from negative to positive in 10 (29%) patients and from positive to negative in 3 (8%) patients (spearman r = 0.15, P = 0.32).

Conclusion

SMAD4 analysis of EUS-FNA cytology specimens for molecular diagnostics is feasible. However, cytology assessment of SMAD4 expression will incorrectly classify up to 29% of tumors as having lost expression. Therefore, the absence of SMAD4 expression on cytology specimens should be interpreted with caution. Disconcordance may reflect tumor heterogeneity and EUS-FNA sampling limitations.

Author List

J.R. Kharofa C. Barnes A. Mackinnon B. George P. Ritch K. Christians D. Evans B.A. Erickson

Author

Beth A. Erickson MD Professor in the Radiation Oncology department at Medical College of Wisconsin


View Online