Real-life Performance Evaluation of the LiFlow X-ray Platform

The thoraco-abdomino-pelvic (TAP) scanner is crucial for assessing and monitoring solid cancers. However, advancements in scanner technology have led to a significant increase in data volume, from 100 images per exam 20 years ago to 2,000 today. The rising number of cancer cases and treatments requiring closer monitoring further strain the workload, prolonging interpretation time and causing delays in therapeutic management and adjustments. The limited number of radiologists contributes to this saturation, increasing the risk of missing metastatic lesions, especially in the lungs, liver, bones, peritoneum, and lymph nodes. The RECIST 1.1 criteria, introduced 15 years ago for standardized follow-up, are useful but time-consuming to implement, resulting in only a small fraction of oncology CT reports using them.

The LiFE study consists in retrospectively interpreting the TAP scans of patients who have agreed to take part in the study at two separate times, once without the LiFlow tool and once with it, at least 8 weeks after the scan date. To avoid any memory effect, a washout period of 8 weeks will be observed between these two interpretations. The study will involve 10 radiologists and 400 patients. Each scan will be interpreted twice by the same doctor (with and without the LF tool). During the second retrospective interpretation, the radiologist will assess the impact of of the additional lesions discovered by LiFlow on the patient's therapeutic management. The notion of change in management is defined by a change in treatment, such as a change in surgical technique, a change in chemotherapy or an indication for a treatment not initially planned. In addition, for these lesions, a gold standard will be obtained (an opinion from one or two other radiologists and/or other characterization imaging and/or biopsy/histology and/or follow-up of the lesion). The aim of the study is to validate the time saved on interpretations and reports, and to improve the accuracy of TAP scan interpretation.