A single drop of unprocessed whole blood, placed at the inlet of a 3D-printed cartridge, yields a quantitative cancer biomarker result in under 15 min—without centrifugation, pipettes, cold-chain reagents, or any laboratory instrument. This is made possible by a silica nanoparticle–chitosan (SiNP–chitosan) nanocomposite that transforms a hydrophobic stereolithography-printed substrate into a near-superhydrophilic, antifouling interface (contact angle 74.3° → 25.3°; Laplace pressure amplified 5.2-fold; fibrinogen adsorption reduced 12-fold), driving autonomous capillary plasma separation directly from finger-prick blood without any external actuation. Anti-GPC3-functionalized gold nanoparticles (AuNPs, ∼30 nm) embedded in chitosan hydrogel beads within the same cartridge transduce antigen binding via localized surface plasmon resonance (LSPR) into two simultaneous, independent optical signals, namely, instrument-free RGB colorimetry (TCS3200/Arduino) for point-of-care use and laboratory-grade LSPR ratiometry for cross-validation—providing intrinsic dual-channel confirmation that no single-transducer platform can match. Glypican-3 (GPC3), an HCC-specific biomarker superior to α-fetoprotein, is detected at LOD = 0.24 ng mL−1 directly from unprocessed whole blood in the same step that separates the plasma. A stereolithography-fabricated device (45 × 18 × 6 mm) incorporates a geometry-programmed micropillar array coated with a silica nanoparticle–chitosan (SiNP–chitosan) nanocomposite. Anti-GPC3-functionalized AuNPs (∼30 nm; TEM core 30.2 ± 2.8 nm) transduce antigen binding into LSPR ratiometric shifts (A₅₈₀/A₅₂₀) and a correlated RGB colorimetric signal, validated in a DEN-induced murine HCC model. The results are as follows: contact angle reduced from 74.3° to 25.3°; Laplace pressure amplified 5.2-fold (86 to 451 Pa); plasma recovery 35–42% of total blood input volume within 30 s; LOD = 0.24 ng mL−1 (LSPR channel)/0.31 ng mL−1 (RGB channel); CV < 10% (both channels); spike recovery 97.3%; ELISA agreement R2 ≈ 0.99; tumor burden correlation R2 ≈ 0.98; diagnostic AUC = 1.00 (95% CI: 0.993–1.000; per timepoint: week 9, AUC = 0.98; week 11, 0.99; weeks 13–16, 1.00) in a controlled inbred murine DEN-HCC model (n = 10 per group). This AUC = 1.00 result must be interpreted in context: it was obtained in a tightly controlled, single-strain, single-sex, small-sample (n = 10 per group) murine cohort and reflects ceiling-level discrimination in a maximally homogeneous preclinical setting. It is not a projection of human diagnostic performance. Such ceiling-level AUC values in small homogeneous animal cohorts are well-documented in biosensor proof-of-concept literature Hanley and McNeal. Radiology 1982, 143, 29–36 and carry a recognized risk of optimistic bias: the controlled model eliminates the inter-individual biological heterogeneity, comorbidity, and preanalytical variability that will reduce AUC substantially in human populations. The AUC is reported to characterize analytical discrimination power of the platform under maximally favorable conditions, not to claim clinical diagnostic performance. This preclinical proof-of-concept platform demonstrates, in a murine DEN-HCC model, a broadly applicable framework for decentralized nanoscale cancer biomarker diagnostics. The RGB channel enables fully instrument-free point-of-care operation; the LSPR channel provides laboratory-grade cross-validation. Prospective human clinical validation is planned before clinical deployment.
Hassan et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: