Examining the accuracy and clinical usefulness of intraoperative frozen section analysis in the management of pancreatic lesions

Abstract

Background: Intraoperative frozen section analysis is often performed in the surgical management of pancreatic lesions. This test is used to obtain histologic diagnosis, to assess resectability because of unanticipated locoregional spread, and to ensure negative margins after resection. We sought to define the accuracy and clinical usefulness of intraoperative frozen section analysis in patients with pancreatic lesions and to determine the impact on long-term outcomes.

Methods: A retrospective database review was performed for all patients who underwent pancreatic resection at our institution from 2002 to 2011. Patient demographics, indications for frozen section analysis, final pathology, and long-term outcomes were analyzed. Five-year survival was compared using the Kaplan-Meier method.

Results: Sixty-eight patients were identified (mean age 65 ± 14 years, 52% female). Malignancy was identified on final pathology in 38 (56%) patients. Intraoperative frozen section analysis was performed in 59 (87%) patients. Frozen section analysis was performed for histologic diagnosis in 6 (10%) cases, to determine resectability in 15 (25%) cases, and to evaluate margin status in 58 (98%) cases. Frozen section analysis for histologic diagnosis was associated with a sensitivity of 80%, a specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 50% with an overall accuracy of 83%. Frozen section analysis for the determination of resectability was associated with a sensitivity of 38%, specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 58% with an overall accuracy of 66%. Intraoperative frozen section analysis for the determination of the final margin status was associated with a sensitivity of 33%, specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 97% with an overall accuracy of 97%. There were no false-positive results on frozen section analysis. Errors on frozen section analysis interpretation did not negatively impact survival outcomes (mean survival = 2.2 years in those with concordant frozen section analysis vs 1.7 years in those with discordant frozen section analysis, P = .69).

Conclusions: Although intraoperative frozen section analysis is helpful for ensuring negative final margins, its usefulness for obtaining histologic diagnosis and determining resectability is limited by low negative predictive potential. These results highlight the importance of preoperative staging and intraoperative surgical judgment for questionable resectable disease.