Computer Vision News - April 2019

The algorithm operates following several steps: The first step is the automatic detection and segmentation of the tumors in the scan . Then, according to the standard, the algorithm chooses the two largest tumors as targets and it measures the RECIST score for them. In the second step, a sophisticated registration process locates the same tumors in the later CTs that are scanned along the treatment period. It makes it fast and easy to compare different time points and state with certainty whether the tumor has grown or recessed. Tracking the RECIST scores of those tumors along time gives the researcher the necessary data for a very accurate clinical evaluation of the medicine and its effect on the sampled patients. In addition, the software can measure the volume of the tumor (not only the longest diameter) and its progression over time, providing more meaningful insights regarding the severity of the pathology and its response to different drugs and therapies. Once the Automated RECIST algorithm is properly trained by RSIP Vision engineers, it can make the radiologist’s work easier and less error prone. The results are presented to the radiologists in a sequential mode, in order to enable them to give a final approval, after viewing the tumors progression or regression along the treatment. This fully transparent and fully logged procedure gives the Pharma company whole traceable data along with the final results - all of which is performed in a much more efficient way. RECIST algorithms allow to review very high volume of scans, enabling to make the most of large databases. Read about more projects for the Pharma industry Project 17 Computer Vision News A project by RSIP Vision Take us along for your next Deep Learning project! Request a call here