Geometric and Quantum Kernel Methods for Predicting Skeletal Muscle Outcomes in chronic obstructive pulmonary disease

Chronic obstructive pulmonary disease (COPD) affects hundreds of millions of people worldwide, and skeletal-muscle dysfunction is clinically important. Quantum machine learning is increasingly explored for biomedical prediction, but its value in small biomarker cohorts requires benchmarking against strong classical baselines. We analysed a cigarette-smoke COPD cohort of 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 quantum regression circuits. We benchmarked this approach against classical ridge/kernel models, SPD relational representations, and quantum-kernel regression (QKR). All methods were evaluated using condition-stratified repeated cross-validation. The largest numerical improvement was observed for muscle weight, where the proposed method had the numerically lowest mean root mean squared error (RMSE), approximately 1.8% below the best classical comparator; paired fold-level testing did not establish statistically significant superiority after Holm adjustment, but the endpoint is biologically meaningful. The method also had the numerically lowest mean RMSE for muscle quality. For force, biomarker-only Ridge performed best, suggesting a more linear endpoint structure.