Read on Twitter
Our team @Inflammatix_Inc (plus some great collaborators) put out three manuscripts this week, and I think a thread is in order, as they form a powerful trio demonstrating the power of what we’re doing.
First, we published the proceedings of a clinical advisory panel on InSep™. InSep is a 29-mRNA rapid test to diagnose and risk-stratify patients with suspected acute #infections. This meeting was held in 2018 with KOLs across clinical lab, emergency medicine, and ICU. /2
Their input in understand the clinical relevance of host-response testing and in things like threshold setting, target performance, etc., became instrumental in the later customer requirements of InSep. /3
Some key advisory panel learnings were understanding sensitivity / specificity targets for diagnostic scores, and how best to support modern ED workflows.
https://www.mdpi.com/2075-4426/10/4/266 /4
https://www.mdpi.com/2075-4426/10/4/266 /4
Second, a preprint of a prospective validation of our diagnostic algorithm for InSep (BVN-2), this time with our colleagues from Greece. We show excellent diagnostic performance for identifying bacterial and viral infections (substantially better than PCT). /5
We also show that BVN2 matches a lot of the requirements laid out in manuscript #1 above, in particular obtaining very high sens/spec at the outer bands of the bacterial score. In addition, we show that the score appears to be well-suited for #COVID-19 #diagnostic workflows. /6
Many thanks to our collaborators from Greece, especially Prof. Giamarellos-Bourbolis. Next comes peer review: https://medrxiv.org/cgi/content/short/2020.12.01.20242321v1 .
But wait, there's more! /7
But wait, there's more! /7
Third, we unveil a new 6-mRNA score for risk stratification of viral infections, with validation in #COVID. Training performance (mortality prediction in viral infections) is almost identical to the test performance across multiple independent cohorts (both COVID and non). /8
Preprint for the 6-mRNA viral severity score here. Again, thanks to external collaborators!
https://www.medrxiv.org/content/10.1101/2020.12.07.20230235v1
So what is the common theme? /9
https://www.medrxiv.org/content/10.1101/2020.12.07.20230235v1
So what is the common theme? /9
First, they both demonstrate that our #ML team (led by @ljbuturovic) creates algorithms that basically don’t lose discriminatory power in external cohorts. /10
Second, our team @Inflammatix_Inc is working on tests that are likely to have a major impact by meeting clinical requirements for changing workflows – rapidly and accurately. /11
Incredibly, there are several more manuscripts in progress that should come out in the coming months. Always looking for more collaborators! More here: https://inflammatix.com
/fin
/fin
