An integrated platform for decoding hydrophilic peptide fingerprints of hepatocellular carcinoma using artificial intelligence and two-dimensional nanosheets.

Hydrophilic peptides (HPs) play a critical role in the pathogenesis of hepatocellular carcinoma (HCC). However, the comprehensive and in-depth high-throughput analysis of specific changes in HPs associated with HCC remains unrealized, due to the complex nature of biological fluids and the challenges of mining complex patterns in large data sets. The clinical diagnosis of HCC still lacks a non-destructive and accurate classification method, given the limited specificity of widely used biomarkers. To address these challenges, we have established a multifunctional platform that integrates artificial intelligence computation, hydrophilic interaction extraction of HPs, and MALDI-MS testing. This platform aims to achieve highly sensitive HP fingerprinting for accurate diagnosis of HCC. The method not only facilitates efficient detection of HPs, but also achieves a remarkable 100.00% diagnostic accuracy for HCC in a test cohort, supported by machine learning algorithms. By constructing a panel of HPs with 10 characteristic features, we achieved 98% accuracy in the test cohort for rapid diagnosis and identified 62 HPs deeply involved in pathways related to liver diseases. This integrated strategy provides new research directions for future biomarker studies as well as early diagnosis and individualized treatment of HCC.

Multidimensional Interactive Cascading Nanochips for Detection of Multiple Liver Diseases via Precise Metabolite Profiling.

It is challenging to detect and differentiate multiple diseases with high complexity/similarity from the same organ. Metabolic analysis based on nanomatrix-assisted laser desorption/ionization mass spectrometry (NMALDI-MS) is a promising platform for disease diagnosis, while the enhanced property of its core nanomatrix materials has plenty of room for improvement. Herein, a multidimensional interactive cascade nanochip composed of iron oxide nanoparticles (FeNPs)/MXene/gold nanoparticles (AuNPs), IMG, is reported for serum metabolic profiling to achieve high-throughput detection of multiple liver diseases. MXene serves as a multi-binding site and an electron-hole source for ionization during NMALDI-MS analysis. Introduction of AuNPs with surface plasmon resonance (SPR) properties facilitates surface charge accumulation and rapid energy conversion. FeNPs are integrated into the MXene/Au nanocomposite to sharply reduce the thermal conductivity of the nanochip with negligible heat loss for strong thermally-driven desorption, and construct a multi-interaction proton transport pathway with MXene and AuNPs for strong ionization. Analysis of these enhanced serum fingerprint signals detected from the IMG nanochip through a neural network model results in differentiation of multiple liver diseases via a single pass and revelation of potential metabolic biomarkers. The promising method can rapidly and accurately screen various liver diseases, thus allowing timely treatment of liver diseases.