Answers are grounded in this paper and cite their sources.
Demonstrates improved colorectal cancer detection in patients using novel serum Raman technology, suggesting effective monitoring post-surgery.
Serum Raman spectroscopy offers a promising minimally invasive approach for colorectal cancer (CRC) detection, but clinical translation has been limited by a fundamental sample-preparation problem: when a droplet dries on a conventional flat glass substrate, capillary flow and competitive protein adsorption drive suspended biomolecules toward the periphery, producing a ring-like enrichment pattern often described as a coffee-ring or Vroman-type effect and resulting in large spatial spectral variance. Here, we employed a hierarchical nanoporous layer (HNL) glass substrate whose superhydrophilic, three-dimensional porous architecture enables rapid and uniform biomolecular concentration, and integrated it with 1064 nm Raman spectroscopy and support vector machine (SVM) classification to develop a reproducible serum Raman platform for CRC detection and postoperative monitoring. Serum samples prospectively collected from 20 CRC patients preoperatively and postoperatively and from 20 age- and sex-matched healthy controls were analyzed. HNL glass reduced spatial spectral variance at 1450 cm–1 by approximately 4.6-fold compared with conventional glass while preserving spectral profiles, and the SVM achieved a mean test area under the ROC curve (AUC) of 0.9923 (standard deviation, 0.0102; range, 0.966–1.000) across 30 independent patient-level stratified train-validation-test splits. In a representative split, the patient-level ROC curve based on preoperative scores yielded an AUC of 1.000 for all 40 participants, although this point estimate should be interpreted cautiously given the small sample size. Principal component analysis (PCA) of patient-level mean spectra showed partial but visible separation between preoperative CRC and control groups, with PC1 and PC2 accounting for 23.2% and 12.8% of total variance, respectively. In an additional reproducibility experiment, three independently fabricated T-4800 HNL substrate batches showed an interbatch relative standard deviation (RSD) of 2.5% at 1450 cm–1 and pairwise spectral correlations of 0.9989–0.9992, whereas conventional soda-lime glass showed a spatial RSD of 10.9% under matched conditions. Longitudinally, CRC prediction scores decreased significantly after curative resection, whereas the score remained persistently elevated in a noncurative stage IV case. These proof-of-concept findings demonstrate that HNL-based serum Raman spectroscopy combined with SVM provides a reproducible sensing platform for CRC diagnosis and postoperative monitoring and warrant validation in larger multicenter cohorts.
No takes yet. Share an insight, caveat, or question.
Ito et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: