User Survey: Feature Prioritization (Spring 2019)
In April 2019, we sent our current ProKnow DS users an electronic survey with which they could, if they desired, cast their votes on which major features they want to be released next. We presented 11 features and allowed the user to set a priority level for each, selecting from the following choices:
- Priority 1 (As soon as possible)
- Priority 2 (Important but not immediate)
- Priority 3 (Nice to have eventually)
- Priority 4 (Not important)
We plan to implement all 11 of the mentioned features eventually, so the purpose of this survey was not to determine “if” but rather “when,” i.e., to allow our users to help us build our near-term road map of major features.
Below is a summary of the results, followed by a legend of feature descriptions.
Table 1. Summary of results, sorted by mean priority rank. The lower the mean priority, the more important the feature to the responding users (N = 18).
|Integrated Image Fusion||11||3||2||1||1.59|
|DICOM: Push to TPS/other||8||6||3||0||1.71|
|Oblique Image Sets||8||4||4||1||1.88|
|Contouring: 3D Auto-margin||7||6||5||0||1.89|
|Advanced Task Management||5||4||6||1||2.19|
|Automated Actions (Based on Triggers)||5||5||5||2||2.24|
|Advanced Patient Search||3||7||5||2||2.35|
- Integrated Image Fusion. Interactive tools to allow the registration of two image sets using user-controlled translations and rotations (rigid body, not deformable). The secondary image set will be resampled to the frame of reference of the primary with coincident axial slice positions.
- Contouring: 3D Auto-margin. Create new anatomical structures via expansion (positive auto-margin) or contraction (negative auto-margin) of an existing structure.
- Contouring: Auto-by-threshold. Create new anatomical structures by image pixel values and edge detection.
- Advanced Patient Search. Improve the patient search tools to allow filters based on existing metadata, computed metrics, and custom metrics (e.g., search by modality, ROI name, max dose) and allow multiple criteria. This feature will be specifically useful for building collections/cohorts from a large number of uploaded but yet-to-be-organized patient datasets.
- Oblique Image Sets. Upload and archive oblique slices (e.g., MRI) plus DICOM registration objects (output from other software systems), then perform image resampling for the registered image sets to enable them to be viewed together in ProKnow’s interactive viewer for contouring, dose review, etc.
- DICOM: Query/Retrieve. Allow the direct pull of DICOM data stored by ProKnow, as initiated by another medical device (e.g., TPS)
- DICOM: Push to TPS/Other. Allow the direct push of DICOM data from ProKnow DS to another medical device (e.g., TPS)
- Interactive DVH. Allow sampling DVH curves (vol-at-dose or dose-at-vol) via mouse and/or keyboard entry of sampled dose or volume. Also all zooming of DVH curves to show certain dose ranges.
- Automated Actions (Based on Triggers). Define trigger criteria (e.g., specific DICOM values) that will initiate automated actions (e.g., create scorecard from template, create checklist from template, add patient to collection) based on specific events (e.g., “Dose/Plan/Structure Set Uploaded”, “Dose/Plan Uploaded”, “Structure Set Uploaded”, “Dose added to Collection”).
- Anonymization Advancements. This would allow users to have a local tool/database with which they can optionally anonymize (de-identify) data prior to storage on ProKnow, then re-identify when downloading locally or pushing to TPS through DICOM DS. A followup feature would be to allow for on-the-fly de-identification (e.g., of Patient ID and Name) for display in the ProKnow UI, enabled by real-time queries of the local anonymization/de-anonymization lookup.
- Advanced Task Management. Allow a user to view and filter tasks assigned to them. This is an easy way to see “To Do” tasks as well as completed tasks to make workflow more efficient.