EUVIMEDOpdag EUVIMED
Open access · CC BY 4.0

Pilot evaluation of multimodal optical pulse signals for lipid-related analysis

Suchin Trirongjitmoah, Kantapon Keawtip, Phalakorn Suebsamran, Parinya Chamnan, Kriengsak Treeprapin

Array · 2026-12 · 32 · 101270

DOI: 10.1016/j.array.2026.101270 ↗

At a glance

Journal article in Array, 2026-12. Written by Suchin Trirongjitmoah and 4 co-authors. 33 references. The full text is freely available (open access, CC BY 4.0). Funded by: Thailand Science Research and Innovation.

AbstractBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Optical pulse waveforms contain information related to vascular and cardiometabolic state, but favorable joint model performance does not establish an independent optical biomarker contribution. This pilot study evaluated synchronized contact photoplethysmography (cPPG), near-infrared imaging photoplethysmography (NIR-iPPG), and speckle plethysmography (SPG) for nine lipid-related outcomes in 20 fasting adults. Three 30-s recordings were acquired per participant and summarized to participant-level features. In the primary fully nested leave-one-subject-out analysis, age, sex, body mass index, heart rate, and optical features competed for a fixed Top-5 predictor set, with linear regression, partial least-squares regression (PLSR), and random forest regression evaluated separately. Triglycerides (TG) showed the strongest performance, with PLSR achieving R 2 = 0.483 (95% bootstrap CI 0.103–0.646). Permutation testing supported TG for linear regression and PLSR after false discovery rate correction (both p = 0.005, q = 0.045). Optical configuration selection was outcome-dependent, with cPPG, NIR-iPPG, and SPG showing different selection patterns across lipid outcomes. However, optical-only TG-PLSR performance was negative ( R 2 = −0.697), and residual optical augmentation did not improve the clinical reference ( R 2 = 0.378 vs. 0.334). No positive incremental optical R 2 had a paired 95% confidence interval entirely above zero. Sensitivity analyses also showed variation with analytical choices. Thus, joint TG models contained predictive structure, but a stable independent optical contribution to lipid prediction was not demonstrated. The point estimates should be considered provisional, and larger independent cohorts are required to determine whether optical pulse signals provide reproducible incremental information beyond basic participant characteristics.

Funding

  • Thailand Science Research and Innovation

Cite

APA

Trirongjitmoah, S., Keawtip, K., Suebsamran, P., Chamnan, P., & Treeprapin, K. (2026). Pilot evaluation of multimodal optical pulse signals for lipid-related analysis. Array, 32, 101270. https://doi.org/10.1016/j.array.2026.101270

Vancouver

Trirongjitmoah S, Keawtip K, Suebsamran P, Chamnan P, Treeprapin K. Pilot evaluation of multimodal optical pulse signals for lipid-related analysis. Array. 2026;32:101270. doi:10.1016/j.array.2026.101270

BibTeX
@article{Trirongjitmoah2026,
  author = {Trirongjitmoah, Suchin and Keawtip, Kantapon and Suebsamran, Phalakorn and Chamnan, Parinya and Treeprapin, Kriengsak},
  title = {Pilot evaluation of multimodal optical pulse signals for lipid-related analysis},
  journal = {Array},
  year = {2026},
  volume = {32},
  pages = {101270},
  doi = {10.1016/j.array.2026.101270}
}
Start for free →