Preoperative multidetector row computed tomography for evaluation and assessment of resection criteria in patients with pancreatic masses

Acta Radiol. 2010 Dec;51(10):1067-77. doi: 10.3109/02841851.2010.520023. Epub 2010 Oct 7.

Abstract

Background: preoperative assessment of pancreatic masses is still challenging as regards the characterization and assessment of irresectability. The opportunities of modern multidetector computed tomography (MDCT) with image postprocessing can be expected to enhance the diagnostic performance if accurate criteria are elaborated.

Purpose: to estimate the accuracy of MDCT and multiplanar image reconstructions with the use of standardized imaging criteria for preoperative evaluation of pancreatic masses with respect to irresectability.

Material and methods: a total of 105 consecutive patients who underwent exploratory laparoscopy or pancreatic resection and had preoperative 3-phase MDCT (4-64 rows) were enrolled retrospectively. First, transverse sections and secondly additional 3Ds were reviewed by two independent blinded observers (O1/O2). Preoperative imaging findings were correlated with intraoperative and histopathologic results.

Results: among all 105 patients, 70 malignant pancreatic tumors and 35 benign pancreatic diseases were found (accuracy of 93% for O1 and 91% for O2). For arterial tumor invasion, receiver operator characteristic (ROC) analysis (values averaged from the results of O1 and O2) revealed an area under the curve (AUC) of 0.931 for transverse sections and 0.986 for 3Ds. Regarding irresectability, positive predictive values were 97% (with 3Ds, 97%) for O1/O2; negative predictive values were 84% (with 3Ds, 89%) for O1 and 86% (with 3Ds, 91%) for O2.

Conclusion: MDCT with 3Ds was highly accurate for evaluation and assessment of irresectability criteria in patients with pancreatic masses. However, due to the limited specificity regarding arterial tumor infiltration, the indication for surgical exploration should be made generously in case of inconclusive findings.

Publication types

  • Evaluation Study

MeSH terms

  • Area Under Curve
  • Contrast Media
  • Female
  • Humans
  • Image Processing, Computer-Assisted / methods
  • Imaging, Three-Dimensional / methods
  • Iohexol / analogs & derivatives
  • Male
  • Middle Aged
  • Observer Variation
  • Pancreas / diagnostic imaging
  • Pancreas / surgery
  • Pancreatic Diseases / diagnostic imaging
  • Pancreatic Diseases / surgery
  • Pancreatic Neoplasms / diagnostic imaging*
  • Pancreatic Neoplasms / surgery*
  • Predictive Value of Tests
  • Preoperative Care / methods*
  • ROC Curve
  • Radiographic Image Enhancement / methods
  • Reproducibility of Results
  • Retrospective Studies
  • Sensitivity and Specificity
  • Tomography, X-Ray Computed / methods*

Substances

  • Contrast Media
  • Iohexol
  • iopromide