The OXMI 2020 Conference was held on-line in Goteborg on 22 -24th June. This conference was the 5th in a series dating back to 1999, where the first event was organised as part of a coordinated research project pursued under the EC Fourth Framework research programme involving nine European partners.
This original project was concerned with predictivity and optimisations in diagnostic radiology, whose aim was to help to develop the scientific framework for the production, quantification and optimisation of image quality in diagnostic radiology. The purpose of the first conference was to disseminate the up to date research outcomes from the EC funded project to the wider scientific community. Subsequent conferences in the series have similarly been based upon the further development of the scientific basis of optimisation in medical imaging, where image quality and clinical outcomes both play a crucial role in the optimisation process. The presentations at all previous conferences were published as proceedings in Radiation Protection Dosimetry.
Due to the covid -19 pandemic, the 5th OXMI Conference was successfully held on-line due to the excellent and dedicated actions of the local organizing committee. The Conference was attended by over 100 delegates from approximately 30 different countries and comprised 10 presentation sessions together with a sizeable group of poster presentations. In all, 47 papers were presented together with 34 posters. Some of the sessions contained invited lectures by keynote speakers and social and exhibitor interaction sessions were also included between the presentation sessions. Equally, both verbal and written questions could be addressed to each speaker. In short, the overall experience as a delegate was excellent.
The OXMI conferences strive to present up-to-date clinical and scientific research outcomes in the field of medical imaging and the 2020 meeting was no exception. The first session was concerned with machine-learning based segmentation and detection in medical imaging and interesting applications were presented. Sessions dealing with CT, mammography and tomosynthesis also presented interesting new findings. Of particular note were the findings of 14 year’s experience of the use of breast tomosynthesis in screening programmes. Studies of patient doses in both X-ray and molecular imaging techniques were also well covered.
One of the great strengths of Medical Physics research in Sweden concerns the evaluation of image quality and the development of software tools to support this process. Indeed, a very interesting presentation concerned the development of virtual clinical trials in medical imaging optimisation. This involved the production of a virtual image data set, which included artificially implanted, clinically realistic abnormalities. The virtual image data set was shown to provide an ROC when assessed by human observers that was identical to that of actual clinical images with similar characteristics. This could raise the possibility of using virtual images produced with defined and controllable image characteristics to train AI and deep learning algorithms.
Education and training initiatives to enhance clinical performance were covered as well as quality control, quality assurance of medical imaging systems. AI and machine-learning for the optimisation of medical imaging provided interesting considerations as to how these techniques may affect future practice. Also, strategies for developing and implementing optimisation in clinical practice were presented.
Posters are an important, relevant and worthwhile component of any successful meeting and the range and quality of work presented at OXMI 2020 was no exception. Posters have the advantage of providing a snapshot of work that is under development or can provide interesting outcomes from a more detailed study. They always provide food for thought.
The organisers must be commended on organising a high-quality scientific meeting in the field of medical imaging that demonstrated that European scientific research in this relevant area is alive and well. As previously, the Proceedings of this 5th meeting, which will hopefully be available in 2021, will provide a welcome and relevant contribution to the evolving field of medical imaging sciences.