Using Machine Learning and Centrifugal Microfluidics at the Point-of-Need to Predict Clinical Deterioration of Patients With Suspected Sepsis Within the First 24 H
Bibliographic record
Abstract
Abstract Introduction: Sepsis is the body's dysfunctional response to infection associated with organ failure. Delays in diagnosis have a substantial impact on survival. Rationale: Lab-on-a-chip (LOC) technologies have the potential to advance care. We have previously shown extraction and detection of molecular markers using a centrifugal-based LOC system. In parallel, a 99 gene signature able to predict clinical deterioration based on the emergence of a cellular reprogramming profile associated with the inability of cells to respond to pathogens has been published. This signature was pathogen agnostic, predictive of immunological endotypes and severity outcomes, but too large for point of care (POC) testing.Methods: Here, we hypothesized that a reduced signature, would retain the ability to predict clinical deterioration in patients with suspected sepsis. Prospective samples from 586 patients were used in conjunction with machine learning and cross-validation to narrow the 99-gene expression signature to just six genes (Sepset). The ability of the signature to identify patients with sepsis was validated in nine published studies. The ability of the signature to predict clinical deterioration in patients with suspected sepsis within the first 24 hours (h) of clinical presentation was determined by routine semi-quantitative PCR (N=248). A compact prototype instrument, the PREcision meDIcine for CriTical care (PREDICT) device, that detects the Sepset classifier using digital droplet PCR in less than 3 h using 50 µL of whole blood was developed and its analytical performance tested in an independent cohort of label-free patients with suspected sepsis (N=30).Results: The accuracy of the Sepset signature (∼90% in early intensive care unit (ICU) and 70% in emergency room patients) was validated in 3,178 patients from existing independent cohorts. RT-PCR showed 98% sensitivity to predict worsening of the sequential organ failure scores or admission to the ICU within the first 24 h following Sepset detection in 248 independent patients. The PREDICT prototype was developed as a stand-alone centrifugal microfluidic instrument that integrates the automated workflow for detection of the Sepset classifier in whole blood using digital droplet PCR. PREDICT had a high sensitivity of 92%, specificity of 89%, and an overall accuracy of 88% in identifying the risk of imminent clinical deterioration in patients with suspected sepsis.Conclusions: The current study describes the development of a molecular risk classifier of clinical deterioration and onset of sepsis, and a novel POC device to measure this at the bedside, as well as the proof-of-concept demonstration using real-world patient samples.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".