Embedded Files
Generic PACS Network Diagram
What you need to setup a Picture Archiving & Communication System (PACS) at your facility
What you need to setup a Picture Archiving & Communication System (PACS) at your facility
This section will guide you on how to setup PACS. I will try to keep it as less technichal as possible.
Medical images move from modality to storage archive via Local Area Network (LAN). This is a push process. From the storage archive, images are pulled (on demand) to remote viewing stations (via the same LAN). For this process to be successful, a number of conditions must be met.
Modalities (Radiology equipment, ECG machines, Endoscopy machines, etc.) must have the same network configuration (i.e. they must be on the same network). This sounds simple but it's not. In Malawi, most of the equipment is procured through development partners. Therefore, it is not uncommon to have multiple brands in one facility. Five different brands means you have to engage with five engineers from all the vendors for you to successfully connect all the modalities to the hospital LAN. To mitigate this, it is highly encouraged to have at least one radiographer with PACS Admin privilleges on all the modalities. This should be a priority during installation of all medical imaging equipment.
Licensing. Ultrasound machines rarely come with DICOM abilities, unless specified during procurement. Medical images are transmitted using the DICOM (Digital Imaging Communication in Medicine) standard. Consult technical team whenever you plan to procure imaging equipment.
ICT Staff. To successfully setup PACS, you either need a service engineer (who in most cases will try to sell you their company's own PACS system) or an ICT person with knowledge of computer networking. PACS is mostly about networking (general computer networking and special DICOM networking).
Technical Part
Technical Part
Any application/software on a PACS network (aquisition console, storage server, remote viewing station) is called an Application Entity (AE). It is assigned an Application Entity Tittle (AET). AETs contain alpha-numeric characters without spaces. AETs are unique that is, no two AEs can have the same AET on the network (this would cause conflicts). It is advisable to have all AETs in capitals (Some AEs would view AET CR01 as different from cr01, thereby breaking the DICOM network). If you have two Fuji CR consoles on the network, one should be assigned FCR01 and the other one FCR02 (Certain special characters are also allowed: CSL-FCR).
Remote viewing stations are only assigned an AET if they request images from the storage server via C-MOVE protocol. This means the storage server must also have the information of that particular viewing station on it's AET list. C-MOVE protocal is like begging. "I am viewing station XYZ in radiology. Would you let me view images belonging to John Doe, please?" The storage server then verifies "who is station XYZ, do I have him on the list of known/allowed viewing stations?" If all conditions are met, XYZ retrieves the images. Otherwise, an "ASSOCIATION" error is displayed on XYZ. On the other hand, if C-GET protocol is used (on demand - "I want these images"), remote viewing stations don't need to have unique AETs. However, it's still good practice to have all AEs populated with unique AETs.
Apart from AETs, all workstations on the PACS network need an IP address and Port number. Two computers on the network cannot have the same IP address - IP conflict would occur and break the whole network. This is why it is essential to have networking staff on site so that all modalities and storage server are assigned fixed IP addresses. In addition, these fixed IP addresses must be reserved in the DHCP server (this is where you need a networking person the most. If IPs are not reserved, they will be assigned to another computer on the network leading to conflicts). Special viewing stations (Radiologist, Neurosurgeon, Orthopedic surgeon) also need fixed IP addresses. It is common practice for radiographers to push post-processed images (3Ds, MIPs, etc) to these workstations as such, they need to be identified on the PACS network and configured on acquisition consoles. The rest of the viewing stations can have DHCP (the network router automatically assigns IPs to clients).
If a computer has multiple viewers or storage applications, make sure they have unique port numbers. Say you have RadiAnt and Weasis on the same computer, they should be registred in storage server as: WEASIS-ER01; 192.168.10.17:11112 and RADIANT-ER01; 192.168.10.17:11113 (same IP, different AETs, different port numbers). "ER01" at the end of the AET helps you identify the station's location (e.g. Emergency Room 01).
Association negotiation between Storage Servce Class User (SCU) - Modality and Service Class Provider (SCP) - Storage Server
More Technical: DICOM Conformance Statement
More Technical: DICOM Conformance Statement
Say you have everything right but you still can't send or retrieve images. In certain instances, it has to do with the DICOM Storage Subject-Object Pair (SOP) Class (see attached link from dicom.nema.org). For this particular reason, technical team should always ask for the DICOM Comformance Statement for all modalities as well as DICOM applications. A DICOM Conformance Statement is a document that details which features of the DICOM standard a specific medical imaging system supports, helping determine interoperability between different systems.
Page updated
Google Sites
Report abuse