Diffusion kurtosis imaging in the characterisation of rectal cancer: utilizing the most repeatable region-of-interest strategy for diffusion parameters on a 3T scanner

Eur Radiol. 2018 Dec;28(12):5211-5220. doi: 10.1007/s00330-018-5495-y. Epub 2018 May 24.

Abstract

Objectives: Our goal was to investigate the correlation between histopathology and diffusion parameters by utilising the most repeatable region-of-interest (ROI) strategy for diffusion parameters in rectal cancer on a 3T scanner.

Methods: 113 patients underwent DKI-MR and 66 of these patients received surgery without neoadjuvant chemoradiotherapy. Two readers independently measured the parameters using three slice protocols including single slice, three slices and whole-tumour slice (WTS), combined with one of two ROIs, including outline and round ROI. ANOVA, Kruskal-Wallis, a paired sample t-test, interclass correlation coefficient (ICC), Bland-Altman, Student's t-tests, receiver operating characteristic curves and z statistic were used for statistical analysis.

Results: There were no significant differences among the three slice protocols in ADC values (p = 0.822, 0.987), K values (p = 0.842, 0.859) and D values (p = 0.917, 0.988) using round and outline ROI, respectively. The ADC and D values derived from outline ROIs were higher than those from round ROIs (all p < 0.001 for ADC, all p < 0.001 for D), while K values derived from outline ROIs were lower than those from round ROIs (p < 0.001, p = 0.001, p < 0.001) using three slice protocols, respectively. The WTS-outline ROI resulted in the best intra- and inter-observer ICC. Utilising the WTS-outline ROI method, the AUC for assessment of well-differentiated tumours was 0.871 by K and 0.809 by ADC; and the AUC for T2 was 0.768 by K.

Conclusions: The most repeatable strategy was the WTS-outline ROI method. In addition to DWI, DKI also have diagnostic value for rectal cancer histopathological characteristics utilising the WTS-outline ROI on a 3T scanner.

Key points: • DKI using a 3T scanner is feasible for assessing rectal cancer. • ROI and slice protocol show considerable influence on DKI parameters. • DKI parameters exhibit excellent repeatability using whole-tumour slice-outline ROI on 3T scanner. • DKI has considerable diagnostic value for the estimation of rectal cancer characteristics.

Keywords: Biomarkers; Diffusion magnetic resonance imaging; Feasibility studies; Rectal neoplasms; Reproducibility of results.

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Diffusion Tensor Imaging / instrumentation*
  • Equipment Design
  • Female
  • Humans
  • Male
  • Middle Aged
  • ROC Curve
  • Rectal Neoplasms / diagnosis*
  • Rectal Neoplasms / therapy
  • Reproducibility of Results