(A) the material level
From the physical level, PACS can be divided into four layers: network user layer, access layer and core layer.
Schematic diagram of PACS application hierarchy
Schematic diagram of PACS application hierarchy
The core layer and the resource provider layer form a "pyramid" structure from bottom to top. Among them, the network user layer is a large number of terminals or workstations in the network; Access layer refers to the collection of network devices (such as secondary switches and hubs). ) connected to the terminal or workstation in the network user layer for network interconnection; The core layer refers to the collection of network devices (such as servers, routers, firewalls, etc.). ) the network devices of the access layer are brought together to form the whole network interconnection; The resource providing layer refers to many medical equipment terminals in PACS network, such as (CT, US, DR, etc. ).
(2) Application level
From the application level, PACS can be divided into three layers: small PACS and departments.
Schematic diagram of PACS application hierarchy
Schematic diagram of PACS application hierarchy
Class PACS and hospital-level PACS form an "embedded" structure from the inside out. Among them, MINI-PACS is a system for planning small medical institutions or individual departments, and the MINI-PACS system must also include professional image reporting systems such as ultrasound and endoscope. Department-level PACS refers to the department structure proposed for medium-sized hospitals, which closely integrates the hospital's existing HIS/RIS system and establishes a patient-centered department imaging center. PACS in the whole hospital is mainly aimed at the whole hospital architecture proposed by large hospitals, which fully realizes the workflow of digital film reading and diagnosis in the imaging department of the whole hospital and realizes the electronic management of the imaging department of the whole hospital.
workflow
The existing mainstream PACS manufacturers have followed the following standard processes at the beginning of developing PACS systems.
PACS business flow chart
PACS business flow chart
(A) check the information registration input
The registration workstation at the front desk can input the patient's basic information and examination application information, or automatically input the patient's information by searching the HIS system (if there is HIS and integrated with PACS/RIS), and make triage registration, follow-up registration, application form scanning, application form printing and triage arrangement for patients.
(2) Worklist service
Once the patient information is input, other workstations can be automatically called directly from the main database of PACS system without manual input again; Medical imaging equipment with worklist service can directly extract the basic information list of related patients from the server, while imaging equipment without worklist function can input patient information through the console of medical imaging equipment or extract registration information through the triage desk.
(3) Image acquisition
For standard DICOM equipment, the acquisition workstation can automatically (or manually) forward the image to the PACS main server after the inspection is completed or during the inspection.
(d) Non-DICOM conversion
For non-DICOM equipment, after receiving the registration information, the collection workstation can use the MiVideo DICOM gateway to collect the images in the inspection process, and the collected images are automatically (or manually forwarded by the equipment operation technician) forwarded to the PACS main server.
(5) Image access
After the patient completes the image examination in the examination room, the doctor can read, browse and process the image through the network of the film reading room, and can print the film and deliver it to the patient.
When the image needs to be accessed, the PACS system is automatically called from the main server disk array or the pre-server connected with it according to the path set in the background.
In the image display interface, doctors can generally measure the length, angle, area and other image post-processing. In mainstream PACS, in addition to the measurement function, they also provide image post-processing functions, such as scaling, moving, mirroring, inversion, rotation, filtering, sharpening, pseudo-color, playing, window width and window level adjustment.
(6) Report editing
After the patient completes the image examination, professionals will evaluate and analyze the image quality. After completing the image quality evaluation and control, the diagnostician can edit the image diagnosis report, and according to the authority of the diagnostician, conduct the initial diagnosis report and report review respectively. In the process of writing reports, templates of commonly used diagnostic words can be used to reduce the workload of doctors' keyboard input. In the process of reviewing the diagnosis report, the revised contents can be modified to obtain information such as clinical diagnosis, detailed medical history and historical diagnosis, and the report can be saved as a typical case for other similar diagnoses for general practice to learn and improve.
The audited report is output by the printer, signed by the doctor and submitted, and the diagnosis report is uploaded to the main server for storage and backup. The printed report cannot be modified, but it can only be read for reference.
6 mode data
Storage technology architecture
The most important thing that distinguishes PACS from other medical information systems such as HIS and LIS is mass data storage. Reasonable design of data storage structure of PACS is the key to the success of PACS construction. A large hospital has a large number of modern large-scale medical imaging equipment, and the amount of data generated by image examination every day is about 4 GB (uncompressed raw data), and the total amount of data in one year is about (1200GB). With the rapid development of hospital business and the introduction of new imaging equipment, this data volume may further increase. In addition, how to improve the efficiency of online data random access is also a very critical issue.
Therefore, most of the existing PACS medical image information system providers adopt the strategy of hierarchical storage (HSM), which divides PACS storage into two levels: online storage and offline storage. Two kinds of storage media with different performances are used to meet the requirements of large capacity and high efficiency respectively, and low-speed ultra-large capacity storage equipment (offline storage server) is used for permanent storage; High-speed storage devices (SAN) are used for online data storage, ensuring extremely efficient access to online data. For historical data of more than 2 years, it is stored in offline storage devices, while online storage devices only store data of the last 3 years.
Storage form
DICOM file refers to the medical file stored according to DICOM standard.
DICOM files are composed of multiple data sets. The data set shows the related properties of real-world information objects, such as the patient's name, gender, height and weight. The data set consists of data elements, which contain the values of the information object attributes to be encoded and are uniquely identified by the data element tags. There are three structures of data elements, two of which are typed to represent VR (whether it appears or not depends on the transmission syntax), the difference lies in its long expression, and the other does not include typed representation. Type indicates the data type in the data element, which is a string of length 2. For example, the VR of a data element is FL, which means that the data type stored by the data element is floating point. All data elements contain labels, value lengths and data value bodies.
The label is an unsigned integer pair of 16 bits, arranged in order including group number and element number. The data elements in the dataset should be arranged in ascending order according to the tag number of the data elements, and they should appear at most once in the dataset.
The length of the value is 16 or an unsigned integer of 32 bits (depending on explicit VR or implicit VR), which indicates the exact length of the data value and is recorded in an even number of bytes. This length does not include data element tags, VR and value length fields.
The data value body represents the value of the data element, and its length is even bytes. The data type of this field is clearly defined by the VR of the data element. The data element field consists of three public fields and an optional field.
data structure
Take the mainstream super PACS system in Guangdong market as an example.
At present, there are 36 tables in the database of SUPER PACS system, which are divided into public tables, special tables for digital cinema, special tables for radiation, special tables for ultrasound and special tables for remote use. Among them, the four main tables of patients, examinations, series and images play a key role.
Patient table is used to store the basic information of patients, and its application scope involves all subsystems of super PACS. The research table is used to store the examination information of patients, and its application scope involves all subsystems of SUPER PACS. Sequence table is used to generate image sequence table, and its application scope involves SUPERPACSR DICOM radiation system. The image table is used to save system image records.