Geometric and quantum kernel methods for predicting skeletal muscle outcomes in experimental chronic obstructive pulmonary disease

Abstract Skeletal-muscle dysfunction is an important extrapulmonary feature of chronic obstructive pulmonary disease (COPD), but advanced computational representations require conservative evaluation in small preclinical cohorts. We analysed a cigarette-smoke mouse model of experimental COPD comprising 213 animals with blood and bronchoalveolar-lavage biomarkers to predict tibialis anterior muscle weight, muscle quality, and force. We developed a kernel-geometric quantum hybrid method in which synthetic symmetric positive definite (SPD) references are mapped through a reproducing-kernel Hilbert space, compressed using train-only random projection, normalised, and supplied to low-dimensional simulated quantum regression circuits. We benchmarked this approach against classical Ridge/kernel models, SPD relational representations, and quantum-kernel regression using identical condition-stratified repeated cross-validation folds. Results were endpoint-specific. Biomarker-only Ridge had the lowest RMSE for force, indicating that a compact linear model was sufficient for this endpoint in the present cohort. Synth_ROSE had the numerically lowest RMSE for muscle weight and muscle quality, but paired fold-level testing did not establish statistically significant superiority after Holm adjustment. These findings support leakage-controlled, endpoint-specific benchmarking of SPD and simulated quantum feature maps, not claims of clinical readiness, quantum hardware advantage, or definitive superiority over classical learning.