Comparison between endoscopic ultrasound and magnetic resonance imaging for the staging of pancreatic cancer

Pancreas. 2011 May;40(4):567-70. doi: 10.1097/MPA.0b013e3182153b8c.

Abstract

Objective: Preoperative staging of pancreatic cancer is crucial for proper therapy. Through this study, we aimed to compare the ability of endoscopic ultrasound (EUS) and magnetic resonance imaging (MRI) to effectively detect and stage pancreatic cancer.

Methods: One hundred twenty-seven patients undergoing EUS-fine-needle aspiration and MRI for the workup of pancreatic cancer were captured in a prospective database for comparison. The final surgical stage was recorded in patients who went to surgery.

Results: Of 127 patients, 48 were surgically explored, and of these, 22 (46%) underwent pancreaticoduodenectomy. Agreement in the patients' staging between EUS and MRI was 94 (74%) of 127. Magnetic resonance imaging was more likely to report metastatic disease or arterial involvement. The overall correlation between EUS and MRI was marginal (κ = 0.42; 95% CI, 0.26-0.58). Of the 48 surgically explored patients, 12 (25%) were understaged by MRI, 13 (27%) were understaged by EUS, and 1 (2%) were overstaged. Endoscopic US and MRI had a sensitivity of 34 (97.2%) of 35 for stage II tumors and 35 (100%) of 35 for lower-stage tumors, respectively.

Conclusion: Endoscopic US and MRI had marginal correlation for staging, especially the more advanced tumors. Although EUS has the added advantage of tissue acquisition for confirmation, the tumors understaged by both the modalities were different. Therefore, both tests should be performed for accurate staging.

Publication types

  • Comparative Study

MeSH terms

  • Aged
  • Biopsy, Needle
  • Endosonography / methods*
  • Female
  • Humans
  • Magnetic Resonance Imaging / methods*
  • Male
  • Middle Aged
  • Neoplasm Staging / methods*
  • Pancreas / diagnostic imaging
  • Pancreas / pathology
  • Pancreatic Neoplasms / diagnosis*
  • Pancreatic Neoplasms / pathology
  • Radiography
  • Reproducibility of Results
  • Sensitivity and Specificity