Setup

This section will walk you through setting up Clara DICOM Adapter using the Clara Deploy AI COVID-19 Classification Pipeline as an example.

Install Clara DICOM Adapter Helm Chart

To install the Clara DICOM Adapter Helm Chart, please follow these steps:

  1. If the system doesn't already have Docker and Kubernetes installed, first run the Bootstrap script.
  2. Install Clara Deploy CLI.
  3. Install the Clara DICOM Adapter Helm Chart using the following command:
$ clara pull dicom

Next, we'll configure the Clara DICOM Adapter.

Configure Storage Mount

DICOM Adapter, by default, stores all received DICOM instances temporarily under /clara-io/clara-core/payloads which is mounted to /payloads inside the DICOM Adapter container.

Disk storage size required for the temporary space depends on many factors: size per job, time to process each pipeline per job and the total number of jobs per day, etc. It is advised to allocate enough space for the temporary space or perhaps mount /clara-io on an network attached storage (NAS) devices.

Note

To change the default storage location on the host machine , find and modify the hostPath property in ~/.clara/charts/dicom-adapter/values.yaml and restart the DICOM Adapter.

Note

The default size of the persistent volume claim created for the mount is 50Gi. To increase or decrease the size of the volume claim, find and modify the volumeSize property in ~/.clara/charts/dicom-adapter/values.yaml and restart the DICOM Adapter.

Configuring Clara DICOM Adapter

The DICOM Adapter configuration is stored as JSON in ~/.clara/charts/dicom-adapter/files/appsettings.json. The default settings enable DICOM C-STORE SCP and C-STORE-SCU and set listening on port 104.

Note

If Clara DICOM Adapter is deployed via helm chart, find and modify dicomPort in ~/.clara/charts/dicom-adapter/templates/values.yaml instead of the appsettings.json file.

Default Settings (appsettings.json)

{
  "ConnectionStrings": {
    "DicomAdapterDatabase": "Data Source=dicomadapter.db"
  },
  "DicomAdapter": {
    "dicom": {
      "scp": {
        "port": 104,
        "logDimseDatasets": false,
        "rejectUnknownSources": true
      },
      "scu": {
        "aeTitle": "ClaraSCU",
        "logDimseDatasets": false,
        "logDataPDUs": false
      }
    },
    "storage" : {
      "temporary" : "/payloads",
      "watermarkPercent": 75,
      "reserveSpaceGB": 5
    }
  },
  "Logging": {
    "LogLevel": {
      "Default": "Information",
      "System": "Warning",
      "Microsoft": "Warning",
      "Microsoft.Hosting.Lifetime": "Warning",
      "Microsoft.AspNetCore.Mvc.Infrastructure.ControllerActionInvoker": "Error"
    }
  },
  "AllowedHosts": "*"
}

Please refer to Configuration Schema for a complete reference.

Note

Before running DICOM Adapter, adjust the values of watermarkPercent and reserveSpaceGB based on the expected number of studies and size of each study. Suggested value for reserveSpaceGB is 2x to 3x the size of a single study multiply by the number of configured Clara AE Titles.

Note

If DICOM Adapter is restarted before a C-STORE-RQ completes, associate is properly released and before it was able to create a job, the received DICOM instances are dropped upon restart.

Note

If DICOM Adapter ran out of available stroage space while receiving instances, the instances that were received and stored may trigger a new job after timeout depending on the logic of the Job Processor.

Starting Clara DICOM Adapter

Once you have configured the Clara DICOM Adapter, run the following command to start the service:

clara dicom start

Enable Incoming Associations

Before setting up a new Clara AE Title, you must first register the COVID-19 pipeline: Refer to the COVID-19 Setup Guide to register the pipeline.

Once you have registered the pipeline with Clara Platform and have the Pipeline ID, you can now set up a new Clara AE Title to accept DICOM associations and store DICOM instances.

First, create a new Clara AE Title:

$ clara dicom create aetitle -a COVIDAET pipeline-covid=<PIPELINE-ID>

.. Note:: Per the DICOM standard, the length of the aeTitle value should not exceed 16 characters.

Next, create a DICOM Source to allow that DICOM device to communicate with the DICOM Adapter:

$ clara dicom create src -a MYPACS -i 10.20.30.1

Now you have a DICOM device with the AE Title MYPACS registered at IP address 10.20.30.1, and Clara DICOM Adapter will now accept DICOM associations from this device.

.. Note:: If you would like DICOM Adapter to accept any incoming DICOM association without verifying the source, you may set DicomAdapter>dicom>scp>rejectUnknownSources to false in the configuration file.

Exporting Processed Results

If the pipeline (in this case, the COVID-19 pipeline) generates results in DICOM format and needs to export the results back to a DICOM device (e.g. PACS), you will need to create a DICOM destination:

$ clara dicom create dest -n MYPACS -a MYPACSAET -i 10.20.30.2 -p 1104

With the command and arguments above, we have created a new DICOM Destination named MYPACS with the AE Title MYPACSAET, at IP address 10.20.30.2 and port 1104. Please note that the name argument (-n) must match the arguments defined in the pipeline. Please refer to Register Results Operator for a complete reference.

Summary

With all the steps completed above, the Clara DICOM Adapter is now ready to receive DICOM associations with the AE Title COVIDAET. To summarize, you've done the following:

  1. Install the Clara DICOM Adapter Helm Chart.
  2. Register the COVID-19 pipeline with Clara.
  3. Set up the Clara AE Title to run the COVID-19 pipeline.
  4. Set up the DICOM Source to allow incoming association requests.
  5. Set up the DICOM Destination for exporting processed results.