Journal of Clinical and Diagnostic Research (Apr 2019)

Does Normalised Apparent Diffusion Coefficient Further Improve the Diagnostic Accuracy of Diffusion-Weighted MRI at 3T in Characterising Breast Masses?

  • Varsha S Hardas,
  • Aparna H Chandorkar,
  • Aparna A Atre

DOI
https://doi.org/10.7860/JCDR/2019/40523.12800
Journal volume & issue
Vol. 13, no. 4
pp. TC01 – TC06

Abstract

Read online

Introduction: In recent times, a significant rise has been observed in the incidence of breast cancer among Indian women. Breast Magnetic Resonance Imaging (MRI) plays an adjuvant role in the high risk screening, diagnosis, staging and follow-up of breast cancer. It is used as a complementary tool to mammography and ultrasound for the diagnostic work-up of breast masses. An advanced MRI technique such as DiffusionWeighted Imaging (DWI) with quantitative absolute and normalised Apparent Diffusion Coefficient (ADC) measurements is employed to improve the diagnostic performance of ContrastEnhanced Breast MRI (CE-MRI). Aim: To assess whether glandular tissue-normalised Apparent Diffusion Coefficient (nADC) could further improve the diagnostic accuracy of MRI, in characterising benign versus malignant breast masses. Materials and Methods: This cross-sectional study included 39 patients with 51 focal breast masses. These patients underwent CE-MRI and DWI, on a 3T MR system. Absolute ADC values and glandular tissue-normalised ADC values were measured in the breast masses satisfying the inclusion criteria. The diagnostic accuracy and kappa measure of agreement between the diagnoses obtained from various imaging techniques (independently and in combination) and the final histopathology/ follow-up results were calculated. Results: Twenty six (51%) of the 51 breast masses were benign and 25 (49%) were malignant. The mean nADC value (0.619+0.21×10−3 mm2 /sec) obtained from malignant breast masses was significantly lower than the mean nADC value (0.98+0.26×10−3 mm2 /sec) measured from benign breast masses (p<0.05). Adding, normalised ADC to CE-MRI, increased the specificity of breast MRI in differentiating benign from malignant breast masses, from 88.5% to 92.3% and improved its kappa score of agreement with histopathology or follow-up results, from 0.883 to 0.960. Receiver Operating Characteristic (ROC) curve analysis indicated that the Area Under Curve (AUC) for nADC (0.870) was higher than the AUC for absolute ADC (0.809). Conclusion: Quantitative DWI with glandular tissue-normalised ADC mapping at 3T, improves the diagnostic performance of breast MRI in characterising breast masses; especially in a subset of masses with borderline CE-MRI findings and absolute ADC values.

Keywords