The aim of this project is to develop a software tool that allows radiotherapy oncologists to annotate patient’s images and embeds the annotations directly into the DICOM standard images. DART tool can help oncologists to avoid medical errors that might be overlooked when the patient returns for treatment, by reconciling MDT’s text-based annotations to the exact anatomical changes observed in the patient’s imaging. The tool will be used to interactively create 2D and 3D markings with text-labels.
Modern cancer treatment is a multi-disciplinary process that typically occurs in the setting of multidisciplinary team (MDT) review. This process often centres on patient’s multimodality medical imaging. The demands of the modern NHS mean that only a few minutes may be available for the discussion of each patient’s individual case. At present, comments and annotations on the patients imaging findings are stored purely in text format in an electronic record of MDT meeting. This means that there is no easy way to reconcile the text-based annotations from the MDT to the exact anatomical changes observed in the patient’s imaging. Also, changes in workspace and tools (such as electronic medical records and radiotherapy planning systems) may introduce safety hazards. Communication mistakes are ubiquitous in radiotherapy, and some studies identify them as a source of serious errors. Errors in radiotherapy have been reported due to wrong contouring, mislabelling of left and right structures, inadequate treatment planning instructions, and incomplete annotations taken at the MDT. This represents a weak link in the chain of information flow within the integrated care framework, and triggers the demand for a reliable and universal approach that enhances safety and minimises risk factors.
What is novel about DART?
The novel aspect of our approach is that we encode these annotations in a fully DICOM compliant technique and independently from hospital systems. This allows for seamless integration of the markup tool and annotated data into existing hospital systems. We have already translated this approach into a prototype tool for rapid markup of annotations over DICOM image datasets.
For further information about IRIS, please contact Mark Hayes, Raj Jena or Mohammad Al Sa'd.
Project funded by