I’ve been thinking about whether some of the challenges in rheumatology (heterogeneous diagnoses, variable treatment response, and “seronegative” disease) might partly reflect a mismatch between the dimensionality of the immune system and what we measure clinically.
In practice, a lot of decision-making still relies on relatively low-dimensional inputs: autoantibodies (RF, anti-CCP, ANA subsets) and physician perception of symptoms (swollen/tender joints).
But from a systems immunology perspective, disease-relevant states seem much higher-dimensional: cell-state distributions, transcriptional programs, tissue-specific immune niches (e.g., synovium), etc. Tools like bulk and single-cell RNA-seq, CyTOF, and spatial methods can capture some of this structure, at least in research settings.
I’m curious how people here think about the bottleneck:
– Is it primarily a measurement problem (we’re not capturing the right features / resolution)?
– Or more of an interpretation/actionability problem (we can measure high-dimensional states, but don’t yet know how to map them to treatment decisions)?
Would also be interested in whether people think tissue-level profiling (e.g., synovial transcriptomics) is likely to be clinically tractable? I really admire the STRAP trial and as a patient in the USA would love if there were something similar here that I could join (that included a synovial biopsy and analysis using spatial transcriptomics to predict treatment response). But I'm not aware of similar trials here.
I wrote a short piece exploring this idea (I compared rheumatology to endocrinology where measurement reshaped the field) but mainly posting to hear how others think about this:
| https://mrsbot.substack.com/p/what-rheumatology-can-learn-from |
|---|