Fig. 4

a Schematic workflow of nanoparticle-enabled blood (NEB) test for cancer detection. Human plasma is collected from healthy and oncological individuals and incubated with nanoparticles (NPs) to generate personalised NP-protein coronas (PCs) complexes further characterised by direct or indirect analysis. The PC characterization readouts can be paired with clinical blood levels to enhance the diagnostic power of the test. b 1D profiles obtained by SDS-PAGE images derived from direct analysis of personalised graphene oxide (GO)-PCs related to 34 healthy (green) and 34 oncological (red) individuals. Black lines identify the most discriminant molecular weight (MW) region between 20–30 kDa (Area 2). Boxplot of the computed Area 2 for all the processed samples is reported in the inset. ** indicate a Student p-value < 0.001. c Box plots of electrophoretic and clinical blood levels for oncological (red) and healthy (green) sample distributions. Asterisks correspond to Student p-values: * p < 0.05; ** p < 0.001. d AUC obtained by coupling Area 2 and haemoglobin (Hb) as classifiers. e Scatter plot of the Maglev signatures derived from indirect analysis of personalized NP-PCs complexes from 10 healthy and 10 oncological subjects. The black line is the output of linear discriminant analysis (left panel). The output of a blind validation test performed on 5 healthy and 5 oncological samples and superimposed with the distribution of the training test (ellipses) (right panel). f Distributions of Maglev fingerprint and blood levels of 22 healthy and 24 oncological subjects g Receiving operating curve and AUC calculated from the coupling between glycemia blood level and Maglev starting position of the 22 healthy and 24 oncological subjects. Figure adapted from Caputo, D. et al., Cancers 13.1 (2020): 93.; Digiacomo, L. et al. Cancers 13.20 (2021): 5155. and Quagliarini, E. et al. Cancer Nanotechnology 14.1 (2023): 1–12