CHLOE BONARELLI, Presenter: Nothing to Disclose

PEDRO TEIXEIRA, Abstract Co-Author: Nothing to Disclose

GABRIELA HOSSU, Abstract Co-Author: Nothing to Disclose

CHEN BAILIANG, Abstract Co-Author: Nothing to Disclose

Alain Gilbert Blum MD, Abstract Co-Author: Research Consultant, Toshiba Corporation Research Consultant, General Electric Company

Highlight the impact of the ADC calculation method on the performance of diffusion weighted imaging in soft tissue tumors' characterisation. Assess two practical methods of ADC calculation in the tumour characterisation of soft tissue lesions.

69 consenting patients displaying a soft tissue tumour with histological proof have been included prospectively between November 2009 and October 2012. The study had been approved by the local ethical committee. Two radiologists calculated several ADCs (minimal: ADCmin and average: ADCavg) for each lesion according to two methods (manual and semi-automated).

We demonstrated a significant relationship between the ADC value and the lesion’s nature (malignant or benign), regardless of the ADC calculation method chosen (p = 0.02). A significant difference emerged between ADC values of benign tumours and malignant ones (p < 0.01), and between the ADCmin values obtained manually versus the ones obtained through the semi-automated method (p < 0.0001). The inter-observer reproducibility was excellent for ADCmin (ICC = 0.82) and good for ADCavg (ICC = 0.77). The manually obtained ADC yielded the best results for tumour characterisation (Se = 83%, NPV = 88%). The manually obtained ADCavg yielded the best results in the solid-lesion subgroup (Se = 80%, NPV = 92%), while the ADCmin obtained though semi-automated method yielded the best results for mixt lesions (Se = 80%, NPV = 93%).

The choice of ADC has a significant impact on the method’s performance and is influenced by the tumour’s own morphology. The manually-obtained ADCmin has yielded the best results overall.

Correlate the ADC calculation method with the tumor morphology to improve the performance of diffusion weighted imaging and increase diagnostic confidence.

BONARELLI, C, TEIXEIRA, P, HOSSU, G, BAILIANG, C, Blum, A, Diffusion of Non-fatty Soft Tissues Tumors though Magnetic Resonance Imaging: Impact of ROI Positioning and Lesion Morphology in the Performance of ADC Value Analysis.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14006208.html