Armelle Merle, director of clinical trial business at Intrasense, looks at how dedicated software can improve radiological assessment in multicentre clinical trials.
There has been a dramatic increase in cancer cases and the World Health Organization estimates that the current number will have doubled by 2030. Consequently, the effort to develop new oncology drugs is intensifying. In clinical practice and clinical trials, imaging is one of the key elements for the diagnosis and evaluation of response to treatment.
During trials, image acquisition is performed at investigator sites. The investigators record the results of their assessments in case report forms, possibly using electronic data capture (EDC) tools. Images are also sent to the imaging contract research organisation (CRO) when centralised imaging is required. This workflow tends to evolve with the increasing importance of information systems in hospital radiology workstations, picture archiving and communication systems (PACS), radiology information systems (RISs) and hospital information systems (HISs). Another trend is the requirement for quantitative assessments according to guidelines such as Response Evaluation Criteria in Solid Tumors (RECIST) and Cheson criteria for malignant lymphoma, as part of the data collected by investigators.
Risks management
There are a number of risks associated with the imaging workflow, which should be carefully considered. These include missing data, heterogeneity in image acquisition and interpretation of images; for example, reader variability is over 30% for tumour burden evaluation.
Oncology and radiology workloads, the increasing number of clinical trials managed in hospitals, potentially different requirements of radiology routine and clinical trials, as well as the variety in the type of assessments, make errors more likely to happen. Long-duration clinical trials and multicentre studies worsen the situation.
These risks are well known and they have led to the publication of guidelines by health authorities such as the Food and Drug Administration or European Medicines Agency, which include recommendations for the handling of imaging data in clinical trials. Dedicated medical imaging software is useful to implement these recommendations.
The software should be usable in clinical routines and hence fully integrated within the hospital information system. It should help radiologists save time, produce standard measurements and facilitate communication with colleagues such as oncologists.
For the evaluation of standard response criteria such as RECIST and Cheson, the software should provide tools for lesion quantification and labelling. The tracking of target and non-target lesions over time can be greatly simplified with the automatic retrieval of prior patient images and assessments, and the registration of images across time-points. The software should check compliance with guidelines and automatically calculate the response. It should also produce standardised reports for oncologists.
In order to streamline clinical trial logistics, the software should be easy to install, easy to learn and simple to use. It must be compatible with all imaging modalities, vendor-neutral, seamlessly integrated to the PACS and RIS infrastructures, and able to export results directly to the EDC. It needs to be approved by health authorities, and compliant with 21CFR part 11 and HIPAA. In the context of multicentre clinical trials, the software should be available in local languages, remotely maintainable and compliant with local regulation.
Such software benefits investigators and imaging CROs alike. It facilitates compliance with guidelines, reinforces traceability, and improves the standardisation and quality of imaging assessments in clinical trials. Tight integration of the software into the hospital workflow and with the EDC systems is also key for a successful deployment in multicentre trials.