International Congress Series 1281 (2005) 11 – 14
www.ics-elsevier.com
Is PACS research and development still necessary? Davide Caramella Diagnostic and Interventional Radiology, University of Pisa, Italy
Abstract. Picture Archiving and Communication Systems (PACS) has been introduced in clinical practice too soon and with too much hype. More research and development would have been needed before starting actual deployment of these complex systems and R&D is still very much in need now, although many hospitals are running successful PACS installations for several years now. In fact, new areas of PACS development include improved reporting strategies, optimized distribution of radiological information to referring doctors, and knowledge management applications ranging from e-learning to computer aided diagnosis. D 2005 Published by Elsevier B.V. Keywords: PACS; Radiological reporting; e-Learning; Computer aided diagnosis
1. Introduction Many authoritative radiologists and computer scientists have reportedly started to wonder if PACS is still a hot topic for research, or if it is just an issue for vendors that are interested to refine and update the technology of their commercial products. Why is there such a pessimistic opinion? There are several answers, but probably the main reason is that when it was born, PACS was already announced as the ultimate tool who would have revolutionized radiology by turning hospitals into filmless operations. This initial bold definition granted to PACS immediate interest and it soon became a highly investigated research topic. In the mid1980s, the problem that everybody was aiming to solve was: bhow is PACS going to work?Q. However, after the first pioneering installations, in the 1990s the problem became: bwhy is PACS not working as expected?Q. In more recent times, as technology made it
E-mail address:
[email protected]. 0531-5131/ D 2005 Published by Elsevier B.V. doi:10.1016/j.ics.2005.03.173
12
D. Caramella / International Congress Series 1281 (2005) 11–14
finally possible to have viable PACS in clinical use, many are questioning if any additional research is useful at all. I believe that PACS research is suffering from the initial emotional approach that I described and that prevented our community to see the PACS as it was just another radiological modality. By bmodalityQ I mean a technological tool that enables radiologists to acquire and manipulate images in order to perform diagnostic interpretations or image-assisted interventions. According to this definition, there is no doubt that PACS should have been considered as a bmodalityQ since the beginning, and that its technical development should have been patiently waited for, as we did for the most innovative modalities. Let us go back, say to 1985, when the PACS acronym and concept were created. In the same period, MR was starting to become a clinically viable radiological modality. Try to imagine what would have happened to MR research if at that time there was a widespread expectation that this new modality already had the means to acquire real-time volumetric datasets of the human body from head to toes with interventional capabilities. With these expectations, the last 20 years of MR research would have been totally different, and now that the prospect of real-time whole-body MR imaging is finally a reality, people would start to ask whether MR research is still useful. 2. PACS: a radiological modality From the point of view of research, PACS should be seen as a modality. Researchers, on both the technological and clinical sides, are expected to envision possible improvements and innovative applications, that have to be implemented and tested and subsequently evaluated in clinical settings. Why should it be more exciting to compare a 64-slice MDCT to a 4-slice MDCT and not to compare a fault-tolerant storage using grid technology to a traditional backup on offline media? If you think that the latter does not have clinical implications, this means you never had the experience of retrieving a corrupted file from your backup archive, or no file at all, when there was the need to visualize a former exam. It is therefore obvious that PACS research, as any modality research, will never be over. And it is also obvious that such research cannot be let to vendors alone, who will always need the cooperation of academia to challenge their projects, by suggesting less evident technological approaches or envisioning new clinical breakthroughs. 3. Radiological reporting The production of the report is one of the main goals of the entire radiological process: PACS has a huge (and rapidly evolving) role in this area. The introduction of soft copy reporting and voice recognition has changed the way radiology is practiced. Still many questions are waiting to be fully addressed in terms of diagnostic performance of the radiologist (i.e. reduction of mistakes), ergonomy (i.e. workstation fatigue), etc. Without soft copy reporting, it would not be possible to effectively read the large volumes of data produced by up-to-date acquisition modalities, nor we would have
D. Caramella / International Congress Series 1281 (2005) 11–14
13
experienced a seamless integration in the radiological workflow of advanced image processing, 3-4-5D image analysis, multimodality image fusion, computer aided diagnosis. Finally, reporting at the PACS workstation has made it possible to foresee systematic use of the structured report, that holds the promise to reduce the indeterminate nature of many radiological reports, making data included in our reports bdatabasableQ items, that must be unambiguously entered for being later available for all kind of data mining. The impact of these structural changes on the outcome of the radiological process has not been fully understood, yet, and the risk exists that we passively accept the changes that are being imposed to us by technology, rather than plan them in advance and steer their implementation. 4. Communication of radiological information The structured report is going to change the way we interact with our referring colleagues, streamlining the information flow between us and them. PACS has already changed this complex relationship. Let us take as an example the electronic image distribution within the hospital: it is clearly perceived that the pattern of utilization of the images on the part of the clinicians has changed, but the details are still to be fully elucidated. Radiologists are often satisfied to say that they are sure that by expediting the distribution of images and reports throughout the hospital, they provide added value to the clinicians. But how is this quantified? Are there outcome studies that justify this assumption? Was it proved the ability of a more efficient hospital-wide PACS communication to reduce medical errors and to achieve better patient care? It is surprising the superficiality that many PACS users show when they declare that PACS increases the added clinical value of their reporting without submitting any hard evidence. Would it be accepted to make similar statements in other radiological domains? For instance imagine some author claim that a 3-T MR system is better than a 0.5-T system, without conducting a rigorous study on the diagnostic performance of the two systems in controlled clinical settings. The distribution of radiological images is no longer confined within the hospital, since in many instances, regional PACS are emerging as the best solution for a rapidly consolidating healthcare sector. This trend will make the term bteleradiologyQ obsolete, since teleradiology is progressively becoming just another PACS function. From a technical point of view, image distribution has been implemented using the web architectural model or the integration of off-line devices, such as CD that are often involved in the communication with general practitioners. Moreover, many innovative products coming out from the consumers market (i.e. last-generation mobile phones, tablet PC’s, hard-drive based devices) can be potentially tested and adopted for improving the distribution of radiological data. 5. PACS beyond radiology and as an e-learning tool Radiologists are medical professionals that work with images. However, many other medical disciplines are based on imaging or require doctors to conduct image-intensive
14
D. Caramella / International Congress Series 1281 (2005) 11–14
tasks (namely the surgeons). Most PACS include Nuclear Medicine and Radiotherapy, but a great deal of investigational effort will be necessary for defining and then for meeting the specific requirements of integrating into the PACS environment pathological images (interactive display of the microscopic fields), endoscopic images (video-sequences), dermatological images. And still more research will be needed to specify and implement the so-called bsurgical PACSQ. All hospitals that have a PACS have experienced its ability to improve the quality of teaching, both in terms of continuing medical education (i.e. clinico-radiological conferences) and in the specific area of radiological training. The availability of huge amount of image and clinical data allows ready access to pathological examples, facilitates the construction of multimedia teaching files, and prepares the physicians to use the powerful resources of e-learning. The radiological archive is no longer the deposit of our past exams that we keep for legal reasons, but the active repository of our professional knowledge that is updated at every encounter with the pathologies that we correctly diagnose. Therefore CAD systems are now to be fully integrated into the PACS, and the knowledge base that they use for decision support has to be fed by the results of our daily clinical work, following the paradigm of the back-propagation algorithms used in neural networks. This is only a brief summary of the many considerations that can be made on the topic of PACS research and development. Many aspects would need a more detailed discussion, but I feel confident in saying to those that might have expressed doubts about the future of PACS research: . . .bwe are just at the beginning!Q.