Quantified ADC histogram analysis: a new method for differentiating mass-forming focal pancreatitis from pancreatic cancer

Acta Radiol. 2014 Sep;55(7):785-92. doi: 10.1177/0284185113509264. Epub 2013 Oct 28.

Abstract

Background: As their prognosis and management are different, differentiation of mass-forming focal pancreatitis (FP) from pancreatic adenocarcinoma (PC) is important. However, the similar clinical presentations and imaging features of these conditions, along with inconclusive biopsy results can make such differentiation difficult.

Purpose: To determine whether apparent diffusion coefficient (ADC) histogram analysis can discriminate between a normal pancreas, FP, and PC.

Material and methods: In a retrospective study, 25 PC patients, 14 FP patients, and 25 subjects with a normal pancreas underwent breath-hold diffusion-weighted imaging (DWI) on a 3.0 T magnetic resonance (MR) scanner. Regions of interest (ROIs) were drawn on the normal pancreases and on the entire focal lesions of both PC and FP. The ADC value was averaged from the lowest to 10th, 30th, 50th, and 100th percentile of the histogram (i.e. ADC10, ADC30, ADC50, and ADC100, respectively), and the results were analyzed statistically.

Results: There were no significant differences among the head, body, and tail of normal pancreases for any of the mean ADC values (P > 0.05). ADC10, ADC30, and ADC50 values demonstrated significant differences between lesion and non-lesion areas of both PC (P < 0.05) and FP (P < 0.05). Differences in lesion areas between PC and FP were found with ADC50 and ADC100 values (P < 0.05), and helped differentiate a normal pancreas from FP and PC, and FP from PC.

Conclusion: Quantified ADC histogram can specifically reflect tissue heterogeneity and help differentiate a normal pancreas from FP and PC.

Keywords: Diffusion-weighted imaging; apparent diffusion coefficient; focal pancreatitis; histogram analysis; pancreatic adenocarcinoma.

MeSH terms

  • Adult
  • Aged
  • Algorithms*
  • Diagnosis, Differential
  • Diffusion Magnetic Resonance Imaging / methods*
  • Female
  • Humans
  • Image Enhancement / methods*
  • Image Interpretation, Computer-Assisted / methods*
  • Male
  • Middle Aged
  • Numerical Analysis, Computer-Assisted
  • Observer Variation
  • Pancreatic Neoplasms / pathology*
  • Pancreatitis / pathology*
  • Reproducibility of Results
  • Sensitivity and Specificity