Median Technologies, The Imaging Phenomics Company®, announces today the results of a preliminary retrospective study on the evaluation of the risk of recurrence for patients with primary liver cancer (HCC – hepatocellular carcinoma) based on a non-invasive biomarker. The iBiopsy® imaging biomarker discovery platform, which integrates advanced technologies in artificial intelligence has produced these results on one of the three indications on which it is positioned. Bearing in mind, the various clinical development plans for iBiopsy® were communicated on April 20, and relate to (1) the evaluation of severity of hepatic fibrosis in non-alcoholic fatty liver disease (NASH – Non Alcoholic Steatohepatitis), (2) the identification of responders/non-responders to certain immunotherapies in oncology, and finally (3) the detection, characterization and prognosis of primary liver cancer (HCC).
The level of hepatic fibrosis  is of great prognostic value and helps inform therapeutic intervention with regard to liver disease. The diagnosis of fibrosis is classically based on the anatomo-pathological examination of a fragment of liver collected by a hepatic puncture, a painful invasive act, presenting risks for the patient and which turns out to be expensive.
There was a two-fold objective for the preliminary retrospective study: first, to accurately quantify liver fibrosis on CT images, and second, to correlate the score of fibrosis with the risk of recurrence in post-operative patients with HCC.
Therefore, a learning algorithm was used on the pre-operative CT scans of 94 patients that were separated in two sets (training and validation) in order to model the relationship between image features of liver, spleen and hepatic fibrosis characterized by histological METAVIR scoring system (F0-F4). The performance of iBiopsy® testing characterized by the area under the curve (AUROC) is 0.91 with a specificity of 1 for the diagnosis of severe fibrosis (F3-F4). Using this quantification methodology, iBiopsy® non-invasive fibrosis biomarker permits the stratification of pre-operative patients at high-risk of tumor recurrence (HR = 4.1 (CI: [1.2,13.9], p-value<0.01). Such knowledge can positively impact the therapeutic approach in patients to undergo hepatic resection.
Based solely on the threshold values of advanced fibrosis, the iBiopsy® fibrosis test would avoid biopsy in 100% of cases, unlike other imaging techniques such as ultrasonic elastography and elasto-MR which have lower specificity and sensitivity values. These very encouraging first results need to be confirmed in larger independent patient cohorts. This will be one of the objectives of the Liver iBiopsy study, conducted in collaboration with the AP-HP, as part of the AP-HP Median partnership announced on March 2.
“These initial results are exciting and show the relevance of our approach. With the integration of the AI technologies, cloud computing and our data science expertise, we are positioning iBiopsy® as a platform for the discovery of imaging biomarkers, on which we will deploy our entire non-invasive biomarkers portfolio, ” said Fredrik Bragg, CEO and co-founder of Median.
“In this first study, the success of the quantification of HCC-related fibrosis via routine scanner imaging opens the prospective to the evaluation of the severity of fibrosis on NASH patients using the same modality. Our objective is to exploit the information content of the images widely used in clinical routine and clinical trials. iBiopsy® allows for the comprehensive analysis of features of images on whole organ, which permits a true evaluation of the entire tumoral landscape and increases the amount of information that we can extract from images, and thus the performance of the biomarkers we calculate,” he added.