NMR system and methods for multi-parameter metabolic vulnerability index evaluations

Inventors

Otvos, James D. • Shalaurova, Irina Y.

Assignees

Labcorp Holdings Inc

Abstract

Disclosed are methods and systems to determine a subject's metabolic vulnerability index (MVX) score using at least one defined mathematical model of risk. The methods comprise evaluating various biomarkers to distinguish various health risks. In one embodiment, the method comprises evaluating biomarkers to determine a relative risk of premature all-cause mortality. The model may include NMR-derived measurements of GlycA, S-HDLP, branched chain amino acids (BCAAs), ketone bodies, total serum protein, and citrate in at least one biosample of the subject.

Core Innovation

The disclosure measures metabolic biomarkers in blood plasma/serum using NMR-based measurement. The metabolic biomarkers include GlycA, small HDLP/S-HDLP, branched-chain amino acids (BCAAs), and ketone bodies, with optional inclusion of citrate and serum protein (Protein). The measured NMR signals are used to compute a multivariable Metabolic Vulnerability Index (MVX) for stratifying relative risk of premature all-cause mortality.

MVX is calculated using defined log-linear models, and the model formulation may include interaction terms. The disclosure describes an interaction example using ln(GlycA)*ln(S-HDLP) and another example using ln(Citrate)*ln(Protein). In addition, the index may be decomposed into an inflammation index (INFX) from GlycA and S-HDLP and a metabolic malnutrition index (MMX) from BCAAs and ketone bodies, with optional citrate/protein and interaction.

The disclosure further provides system architectures for acquiring NMR spectra, including flow probe configurations, and for programmatic spectrum deconvolution and MVX computation. Clinical use cases described include monitoring over time and therapy/trial risk stratification. The disclosure also references cohort performance and figures demonstrating mortality stratification and predictive statistics using reported approaches such as Cox proportional hazards model and multinomial logistic regression, including CATHGEN and MESA.

Claims Coverage

Not explicitly described in patent.

Not explicitly described in patent.

Stated Advantages

Documented Applications