CONFERENCE & SEMINAR
March 14-16, 2013
Bangalore, India
ABSTRACT
Title of presentation: / Prefetch: Quick Access to Prior studies using DICOM ServicesAuthor(s) name(s): / Kishan Harwalkar
Author(s) title(s) / Sr. Domain Specialist – Interoperability
Affiliation (Company or organization) / Philips Electronics India Ltd.
Postal address / Philips Innovation Campus, Manyata Tech Park, Nagavara, Bangalore – 560045
E-mail address /
Telephone: / +91 80 4189 1026
Abstract: (200-750 words, in English) /
Introduction
Quick and easy access to prior DICOM Studies is very important during the Diagnostic scenarios requiring comparison of studies acquired over time. This is very useful in Oncology, Cardiology and in clinical applications such as Tumor Analysis, Colonoscopy and Lung Nodules to name a few. Manual retrieval of prior studies from PACS is time consuming and it is essential automate this workflow. A well known solution to automatically fetch prior studies, already exists in the PACS and VNA domain, using HL7 order messages. This paper describes the implementation of Prefetch feature in a DICOM based Advanced Visualization system. It covers the experiences and challenges faced in deploying such solution with various RIS and PACS Systems in the healthcare enterprise.Prefetch Functionality
The foremost aspect for a successful implementation of DICOM Prefetch is to decide on using a combination of DICOM services, some of which are not commonly implemented for Workstation. The ideal workflow will be to periodically query the RIS for scheduled examinations, using the DICOM MWL, and fetch the corresponding priors for matching patients from PACS using DICOM Q/R. Such a solution may not work in all situations, for example in Emerging Markets, where scheduling via MWL is not common. In such scenarios, an alternative approach will be to trigger the retrieval of priors from PACS using DICOM Q/R, when a newly scanned study is pushed to the system.The next important aspect to consider is that, not all new examinations require access to priors. So, we need a mechanism to identify the subset of studies for which priors are relevant. Arriving at this subset will depend on certain DICOM Attributes. Several discussions with radiologists and clinicians, lead to the finalization of the following attributes which can be used in arriving at a subset of studies needing Prefetch:
-Modality
-StudyDescription
-BodyPartExamined
-SeriesProtocol
Next aspect is to identify the right set of priors on the PACS to Prefetch; which is essential to avoid fetching too many unwanted studies. Identification of right priors again depends on the DICOM attributes mentioned above.
Challenges
During the validation and deployment of Prefetch, many challenges were encountered mainly due to varied implementation of the DICOM standard. Few of these challenges are identified below:-There is no consistent mechanism of mapping RIS Codes into DICOM attributes such as StudyDescription, Protocol and BodyPart. Prefetch using MWL query can only be based on Modality.
-Some RIS systems do not support matching on Modality.
-For identifying the right set of priors, a query to PACS on PatientID and PatientName may not be efficient, as sometimes, the MiddleName, Prefix and suffixes are not consistent. Implementations should also consider the demographic changes in the Prior studies based on IHE PIR.
-Not all PACS support matching/returning optional parameters such as Modalities in Study, Study Description, BodyPart etc. In such cases, the rules configured may not work efficiently. A study on such support in PACS at connectathon revealed that out of 15 PACS, 40% of the PACS did not support returning ProtocolName as a return key in DICOM Query.
-For editing free text fields like Study Description, the hospitals may not have a common guideline followed by all users. There can be various key words which mean the same thing. Systems have to cope with handling such multiple values per field.
-Some PACS open only single Association per Node. So when a Prefetch is in progress, it might block the user from doing Manual Q/R with the PACS.
-Not all prior studies are required in all cases. In some Oncology cases the baseline, nadir and previous two studies are sufficient.
-Based on the study throughput in different hospitals, systems need different intervals for doing the Prefetch based on MWL. It can range from once every minute to once everyday.
-Sometimes due to the presence of the unsupported DICOM IOD appearing in a study the prior may not be retrieved. In such cases, it is required to let the user know about thissituation.