A novel approach for semi-quantitative assessment of reliability of blood flow values in DCE-CT perfusion

In the last few years, cancer treatments have improved significantly with the introduction of new therapies aiming at reducing tumour angiogenesis, a process leading to disease progression and metastasis formation. Computed tomography perfusion (CTp) is being emerged as a promising functional technique for assessing tumour response to these new treatments, which yield a reduction of perfusion heterogeneity, occurring long before morphological reduction. However, several factors, such as noise induced by respiratory and physiological involuntary motion, prevent a reliable quantitative assessment, hence the clinical use of CTp. Currently, the assessment strategies rely on global measurements that fail in discriminating between noise and heterogeneity of tumour perfusion, both characterized by a wide value dispersion. This paper presents a new approach for reliability estimation by introducing a novel local-based index, which is able to discriminate between tumour heterogeneity, featured by locally structured patterns, and noise, characterised by sparse and unstructured values. This index enables a proper comparison between perfusion maps and can replace the parameters based on the global mean, thus improving the overall reliability of CTp studies and favouring the translation into clinical routine.