Handheld Near-Infrared Spectroscopy Integrated with Pattern Recognition for Non-Destructive Discrimination of Rice Infected by an Aflatoxigenic Versus a Non-Aflatoxigenic Aspergillus flavus Isolate

Aflatoxin B1 (AFB1), a highly carcinogenic metabolite produced by aflatoxigenic Aspergillus flavus, poses a serious threat to rice safety and public health. This study aimed to establish a rapid, non-destructive method for discriminating between aflatoxigenic and non-aflatoxigenic A. flavus infection and for the exploratory classification of AFB1 contamination levels in rice by integrating a handheld near-infrared (NIR) sensor with pattern recognition analysis. A total of 154 post-harvest rice samples were inoculated with aflatoxigenic A. flavus or non-aflatoxigenic A. flavus, or left untreated as controls. Spectra were collected from each sample, and pattern recognition analysis (PCA, SIMCA, OPLS-DA) was used to discriminate between infection types and to categorize AFB1 content. The uHPLC-MS/MS reference method quantified AFB1 concentrations ranging from 2.70 to 23.59 ppb at early infection stages. PCA revealed distinct separation trends among the three sample groups. The SIMCA model achieved 100.00% accuracy, sensitivity, and specificity for discriminating among the three infection types in internal validation, and hold-out validation on an independent sample set demonstrated effective discrimination, yielding 96.67% overall accuracy. OPLS-DA produced comparable results, with accuracies of 95.90% in internal validation and 96.67% in hold-out validation. Notably, the SIMCA model was also applied in an exploratory manner to classify AFB1 content into low-AFB1 (<4.0 ppb) and high-AFB1 (≥4.0 ppb) groups; under Monte Carlo cross-validation and hold-out validation, this grading model yielded a mean accuracy of 63.10% and 87.50%, which demonstrated the potential feasibility of AFB1-level grading by handheld NIR spectroscopy. This approach shows potential for rapid laboratory-based screening in food supply chains to support timely risk assessment and intervention to protect food safety and public health.