AusCAT Node Setup Summary#

This page provides a summary of the steps for establishing an AusCAT node at a new hospital, clinical centre, or registry, after being approved as a site in the ethics protocol and has received local governance approvals. This guide outlines an overview of the core steps involved in setting up the required infrastructure, deploying the system, and importing data. Detailed instructions for each step are available in separate pages, which can be accessed through the links provided below.

Initial Setup Checklist#

Task

Description

Who is Responsible

Estimated Time of Completion

Approx. Staff Involvement

Ethics Approval

Adding the site investigators to the existing ethics protocol. This step involves collecting required documentation from site investigators and submitting to the HREC for formal inclusion.

Project CPI, AusCAT Technical Centre, Site PI and all Investigators

3–4 weeks (after submission) *

1 day (AusCAT), 1–2 days (Site PI)

Site Governance

Site-level governance approval required by the hospital. This typically involves submitting local governance forms and data custodian agreements. Site PI and investigators will be actively involved.

Site PI and all Investigators, Site Governance Office

6–8 weeks *

1–2 days (Site PI), 1 day (Governance)

Hardware Server Provisioning

Finalising and provisioning server hardware that meets the minimum or optimal specifications required for running the AusCAT infrastructure.

Site Investigators, Site IT Team

2–4 weeks *

1 day (Site investigators), 1–2 days (site IT)

Operating System Setup

Install a supported operating system (e.g., Windows Server 2022) on the provisioned server hardware.

Site Investigators, Site IT Team

1 week **

1 day (Site investigators and Site IT)

VM and Docker Environment

Create two virtual machines and install Docker and Portainer as per the AusCAT infrastructure setup guide.

Site Investigators, AusCAT Technical Team

1–2 weeks **

1 day (AusCAT technical team), 2-3 days (Site investigators)

Deploy Docker Stacks

Deploy the Docker stacks on the two VMs—one for the identified key database and one for the research database.

Site Investigators, AusCAT Technical Team

1 week **

1 day (AusCAT technical team), 1-2 days (Site investigators)

Authentication and Permissions

Obtain the required credentials and configure access permissions for connecting to clinical systems from the data integration tool (e.g., Pentaho).

Site Investigators, Site IT Team

1–2 weeks *

Half a day (Site investigators), 1 day (site IT)

Clinical Data Import

Install and run Pentaho pipelines for importing clinical data into the research database.

Site Investigators, AusCAT Technical Team

1–2 weeks **

2 days (AusCAT technical team), 4-5 days (Site investigators)

DICOM Data Import

Set up the Orthanc server and configure it to receive and store DICOM imaging data exported from the clinical systems.

Site Investigators, AusCAT Technical Team

3–4 weeks (depending on dataset volume) **

3 days (AusCAT technical team), 1-2 weeks (Site investigators)

* The estimated time reflects total processing duration, including passive waiting for approvals or institutional turnaround. Actual staff workload is lower and reflected in the “Approx. Staff Involvement” column.

* *The estimated time may extend due to potential troubleshooting delays such as environment-specific issues, network configurations, permission errors, or hardware/software incompatibilities.

1. Infrastructure Setup#

As a first step, visit the Infrastructure Requirements page to establish the hardware infrastructure. This page provides details on the hardware specifications required for the server, including CPU, RAM, storage, and GPU recommendations for running AusCAT software tools, extracting relevant tabular and image data, executing code, and storing results.

2. Production Deployment#

Once the hardware and server OS (e.g., Windows Server 2022) are ready, the next step is to deploy an AusCAT node in a production environment. This includes the following components:

2a. VM Setup#

  • Install two VMs on the server:

    • KeyDB VM: For hosting the identifiable patient database

    • CatDB VM: For hosting the anonymised research database

  • Detailed VM setup instructions can be found here.

2b. Docker & Portainer Installation#

  • Install Docker and Portainer on each VM running Ubuntu. This is required to manage containerized applications and services.

  • Refer to the Docker and Portainer Installation Guide for step-by-step instructions.

2c. Configure and Deploy KeyVM Docker Stack#

  • Once Docker and Portainer are installed, configure KeyDB secrets and deploy the KeyDB Docker stack on the KeyDB VM.

  • Use the KeyDB stack template provided here.

  • Test the deployment to confirm the stack is running correctly.

2d. Pentaho Installation#

  • Install Pentaho, a data integration tool, on the KeyDB VM.

  • Detailed installation steps can be found in the Pentaho Configuration Guide.

2e. Configure and Deploy CatVM Docker Stack#

  • On the CatDB VM, configure CatDB secrets and deploy the CatDB Docker stack.

  • Use the CatDB stack template provided here.

  • Test the deployment to confirm the stack is running correctly.

For detailed steps, refer to the Production Deployment Guide.

3. Data Import#

Once the infrastructure is in place, the next step is to import clinical and imaging data from the hospital’s clinical systems into the local AusCAT node for research use.

This process is divided into two major parts:

  • Clinical data extraction (tabular data)

  • Imaging data export (DICOM files)

3a. Clinical Data (Tabular)#

To extract clinical data from the hospital’s Oncology Information System (OIS) — such as ARIA or Mosaiq, AusCAT uses pre-configured Pentaho pipelines.

  • Visit the AusCAT Centres Extract, Transform, Load (ETL) page to download the appropriate pipeline templates.

  • These templates are customized for both ARIA and Mosaiq, and can be imported into Pentaho Data Integration (Spoon).

  • Once imported, the pipelines can be executed to extract clinical data items (e.g., diagnosis, prescriptions, treatment, site information, etc.) from the reporting server of the clinical SQL database.

Important: Clinical systems often have a dedicated reporting server (read-only) that mirrors the live production server. For security and performance reasons, it is recommended that all extraction should target the reporting server, not the live server.

  • Make sure that Pentaho is properly configured with the correct connection parameters (hostname, port, SQL credentials) to access the reporting server. Without this, data extraction will fail.

For more information, navigate to the Pentaho Configuration Guide for details on running pipelines.

3b. Imaging Data (DICOM)#

Once clinical data is extracted, the next step is to retrieve and import DICOM imaging data (e.g., CT, RTSTRUCT, RTDOSE, RTPLAN) for the same patients.

  • This step involves exporting DICOMs from the clinical imaging system (e.g., Treatment Planning System (TPS) or PACS) to the Orthanc server hosted within the AusCAT node for storage and later processing.

  • Each centre runs a DICOM export script that sends the relevant imaging data to Orthanc via the CTP (Clinical Trial Processor) pipeline, which handles de-identification of DICOM files during transfer.

If your site does not yet have a working DICOM export script, please contact the AusCAT Technical Team. We will provide assistance to configure or adapt scripts suitable for your clinical environment.

External Connection Requirements#

The clinical systems and tools that are involved in the data extraction require external or networked communication between:

  • The Pentaho tool running on the AusCAT node and the OIS SQL reporting server.

  • The clinical imaging systems (TPS/PACS) and the Orthanc/CTP stack for DICOM transfer

Ensure that the appropriate firewall and port access is in place:

  • For SQL queries, port 1433 must be open from the AusCAT VM to the SQL Server.

  • For DICOM transfer, ensure that port 8104 (CTP receive) and 8042/4242 (Orthanc) are properly configured based on the site’s routing preferences.

Usage

Source IP

Source Port

Destination IP

Destination Port

Protocol

2-Way?

SQL Queries

xxx.xxx.xxx.xxx (Reporting Server)

1433

xxx.xxx.xxx.xxx KeyDB (AusCAT)

Dynamic (e.g., 104)

TCP

Yes

DICOM Transfer

DICOM Server

Dynamic (e.g., 105)

CTP on KeyDB (AusCAT)

8104 (or 104)

TCP

No

Notes:#

  • The Pentaho pipelines running on the AusCAT node initiates the connection to the clinical system’s reporting SQL Server (e.g., ARIA, Mosaiq, Epic) to execute data extraction queries. The clinical SQL server then returns the requested data, which is anonymised by CTP before being stored in the AusCAT KeyDB. Since communication is initiated from the AusCAT side and responses are returned from the clinical server, this constitutes a two-way (bi-directional) connection.

  • For DICOM transfers, the clinical system (e.g., TPS or PACS) initiates the send operation using the DICOM protocol (e.g., C-STORE) directly to the CTP node. CTP performs anonymisation and forwards the data to the local Orthanc server. AusCAT does not initiate this communication. Therefore, this setup represents a one-way connection, from the clinical system to the CTP/Orthanc service.

  • Source port for DICOM transfer is dynamic and typically assigned by the OS (e.g., 105), not fixed or required for firewall configuration.

  • The CTP service listens on port 8104 by default. If this is blocked by firewall rules, it can be reconfigured to use port 104, which is a standard DICOM port typically allowed by hospital networks.

  • After receiving and processing DICOM files, CTP forwards them to the internal Orthanc service using port 4242 (this happens locally within the AusCAT Docker environment and usually doesn’t require firewall rules).

  • The Orthanc web interface is accessible on port 8042 and can be used to verify DICOM reception and perform quality checks.

Application Entity Title (AET) Configuration#

  • CTP (on the AusCAT server) does not require or enforce an AE Title. It will accept DICOM files from any AE Title, as long as they are sent to the correct IP and port (usually port 8104 or 104).

  • However, most TPS require the destination AE Title to be defined. In such cases, sites have successfully used a random name like AUSCATSERVER or CTP_RECEIVER when configuring the TPS to send to AusCAT.

  • The AE title is only used internally by the TPS to identify the target — CTP will accept the data regardless of what AET is used.

For additional support or questions, please contact the AusCAT technical team.#

  • Fahim Alam

  • Sasha Barisic

  • Amir Anees