Clinical trials Quality Assurance platform

Project description

Aim

Denmark has a strong tradition of performing national clinical trials. Collaborations between centers require well-functioning radiotherapy (RT) quality assurance (QA) procedures. The aim of this work package (WP) is to generate a multi-center radiotherapy QA facility which allows for a supple audit of the radiotherapeutic part of the clinical trials with tailored software for efficient and high-quality radiotherapy QA.

Background

In clinical trials, the potential improvement of standard treatment is investigated. However, the quality of RT can vary considerably among different treatment centers or from patient to patient and therefore hinder a clinical trial in producing the correct result. Either by failure to identify a new effective intervention or by incorrectly being unable to reject the experimental arm. Just by having varying quality, the statistical power of a study may be reduced. More and more focus has been addressed to these issues in recent years and now well-conducted clinical RT trials require some form of RT QA. The QA procedure should act as a positive addition to the clinical trial and should not be a hindering and resource-intensive process. The whole process needs to be a well-integrated part of the patient information and treatment preparation process. This requires software with tailored possibilities and automation processes for each specific clinical trial.  

Methods

Using the infrastructure of the national radiotherapy image bank (WP 12), an online platform for RTQA will be developed. Data from the image bank will be acquired and processed via internal and external software-modules. The platform should enable the dynamic development of new modules with features for specific trials.

The basis for RT QA has already been created through the national RT image bank. Here RT dose plans and medical images can be submitted and dose metrics can be extracted through mapping of organs at risk (OARs) and paring of dose information. The current process is relatively labour intensive and thus not suited for online and large-scale QA. Since radiotherapy clinical trials have many similarities but also unique features, large parts of the QA requirements are often common, but most trials will have specific needs that require QA attention.

WP13 will create an RT QA platform for interfacing with the RT image bank. This interface will allow small software programs to extract information from the treatment plans and validate that this information adheres to the specific clinical trial protocol.

The following modules are planned:

Clinical trial audit of imaging contours. From any given secure browser, a pre-approved user should be able to visualize a patient with the contours of targets and OARs. If possible, the co-registered images should be available for validation of contours. The visualization should be observed in transverse slice views and the contours should be able to be toggled on and off. The contours that from the protocol need a review should be flagged and should be evaluated according to “per protocol”, “minor deviation”, “minor violation“ (take action) or “major deviation”. Comments for individual contours should be possible. The latter two gratings should automatically generate a report to the patient center when the review is finalized.

Clinical trial audit of radiotherapy treatment plans. From any given secure browser, a preapproved user should be able to visualize a patient with the contours and isodoses. The visualization should be seen in a transverse slice view and the contours and doses should be able to be toggled on and off. The protocol specified dose requirements should automatically be validated and let the reviewer change or comment the evaluation according to “per protocol”, “minor deviation”, “minor violation” (take action) or “major deviation”. The reviewer should be able to review the DVH and isodoses easily.

Dummy run contouring audit. For quality assurance of contouring, the standard setup should support a module that can distribute dummy run patients, collect contours from the different participating centers and analyze the contouring differences using metrics like dice similarity index and mean surface distance.

Dummy run treatment planning audit. For QA of treatment planning, the standard setup should support a module that can distribute dummy run patients, collect treatment plans from the participating centers and analyze the dose metric differences using metrics like dose-volume histograms (DVHs) and normal tissue complication probabilities (NTCPs).

Facility interface with external software for auto-segmentation. For patients in the image plan bank, it should be possible to re-contour with an external auto-segmentation program, for large patient cohorts. The module should be able to communicate with an external program for the export of anonymous patients and the import of new contours to the correct patient. 

Automatic radiotherapy quality assurance of specific clinical trial contours. The contours of an incoming patient to the image plan bank should be validated by comparing it to auto-segmented contours. If the mass mid-point, volume or other metric does not match the segmentation, then the contour should be flagged and manual expert validation would be needed. This process should be able to validate both the nomenclature and roughly the contours. 

Automatic radiotherapy quality assurance of specific clinical trial treatment plans. Plan entering the image plan bank should be validated by comparing it to a fully automatically generated plan. If the mean or specific dose metrics important for the specific protocol are significantly improved in the automatic plan, then the plan and dose metrics should be flagged and a report should automatically be sent to the appropriate persons.     

Each software addition will be validated before it will have access to the image plan bank data.      

Expected results

An RTQA platform for interfacing with the radiotherapy image bank with the following modules:

1. Online browser-based audit of contours. The audit will be conducted via online visualization and review of the medical images including the contours.
2. Online browser-based audit of RT plans. The audit will be conducted via online display and browsing capabilities offering easy validation of the isodoses covering/missing targets.
3. Automatic RTQA of incoming contours and RT plans (> 3yrs).

Impact/Relevance/Ethics

There is a need for a multi-center RT QA platform to make sure that the quality of the radiotherapeutic part of the treatment is equally high among centers participating in the national clinical trials. Since the aim of the national center is to stimulate the national collaboration within radiotherapy for patient benefit, an RT QA platform is highly relevant to ensure high-quality treatment. Ethics will be described in relevant clinical trials.