# ci2 moonshots session
## Discussion Points
- Sharmila:
- we could propose to the university to help fund models to reduce time of acquisition, increase quality of imaging
- another thing that comes again and again is the push-back from PACS to deploy models because PACS is changing. **It should be priority** for ci2 to make it easier for models to be deployed
- we also need **clinical champions** that keep models being used and making sure they're supported over time (as computers change, etc)
- Researchers should submit intake forms with ci2 so we can evaluate the reasons and costs to make it live.
- **If we can show that time ($$) is being saved, we can ask for resources from the hospital**
- Janine
- That are so many of our techniques that never make it out of the research setting. I think that what we need is the infrastructure or pipelines for recon, automated processing, sending displays to PACS, etc. We've in the past done pieces of it, **but we need support for more robust pipelines**;
- Is there a quick way **to do spectroscopy analysis** that is done by the 3D lab? e.g. Jason put all scripts into sivic and into the scanner. But over time (scanner upgrade?) this stopped working.
- Janine and others should submit as project to ci2
- Peter
- Do we have a way of billing?
- Eugene
- ci2 should also help with interacting with vendors. E.g. GE sometimes is not responsive but maybe the center should take
- Pablo
- Submit your requests to intake (as many as you like!)
- This way we can digest and prioritize based on the whole lab, not just one vocal person
- Maddie
- We have not done a good job at keeping track the impact of our tools. Can we automate the process of measuring impact?
- Sharmila
- What are **useful** KPIs for tracking clinical success of research techniques? Research grants and papers are not helpful for people handling money and taking scans
- This is a good thing to crowd-source
- Stakeholders: Sharmila, Jason, Maddie, (Pablo?)
- Robert Bok
- Metric suggestion ~ # patient scans done on research time/scanners which subtract from **(relieves) the imaging workload on clinical scanners**
- Sharmila says difficult because scanners are now blended between clinical and research
- Bob: relevant to standard of care exams, which benefit patient and reduce patient costs
- Maddie: Time saved by deployed fastMRI sequences? :thumbsup:
- David Vigneron
- Metric suggestion ~ # people who come outside of 50mi radius to be part of research study (people who wouldn't normally go to ucsf)

## Miscellaneous
- How can UCSF/VA collaborate further?
- Introduce imaging protocols to clinical pipelines in both UCSF/VA
- How can UCSF/BIDS/Berkeley collaborate?
- With J&J, anything related to deployment they're highly interested.
- How can we get more involved in hosting challenges
- Peder: What are the painpoints/rewards/... of doing these in the past?
- Information Commons could be a useful resource for this, except for issues with preclinical datasets
- Can prostate be our test bed for the whole university to use the data and do research on? Let's implement Kirti's pipeline on Info Commons as a pilot (Sharmila)

## Call to Action
- Who will lead this initiative in each area?
- Who will spearhead research technique success in clinical? (Segmentation, fastMRI, ...). We need clinicians to make the case that these deployments are useful (Sharmila)
- People who have not volunteered but might be good to involve: Janine, Peter, Jason
- People who have volunteered: Maddie (Pablo)
- This will involve quantifying segmentation
- Who will submit new initatives for deployment/translation?
- EVERYONE (especially MDs!)
- Who is coming the distance to be part of research studies? (Quantifying new patients)
- CRC Core and Steering Committees