A compact superlattice as a label-free surface-enhanced Raman scattering substrate for noninvasive urine testing for the diagnosis of lung cancer.

The development of noninvasive cancer diagnosis approaches may provide convenient, remote, painless diagnosis, instant healthcare, and postoperative follow-up. Label-free surface-enhanced Raman spectroscopy (SERS) is becoming a powerful approach for the detection of various potential biomarkers in cancer diagnosis; it avoids focusing on one or several specific targets, thus it may achieve the comprehensive and accurate diagnosis of cancer. In this work, a compact nano-superlattice is constructed as a label-free SERS substrate for the Raman analysis of urine samples from healthy people and lung cancer (LC) patients before and after surgery for the noninvasive diagnosis and postoperative monitoring of LC. Multiple chemometric methods, including principal component analysis (PCA), linear discriminant analysis (LDA), and orthogonal partial least-squares discriminant analysis (OPLS-DA) are applied for the classification of SERS spectra obtained from different samples. LDA and OPLS-DA outperform PCA to discriminate lung cancer preoperative patients from postoperative and healthy persons with higher efficiency. The high accuracy based on LDA and OPLS-DA is evaluated by calculating the area under the curve (AUC) of receiver operating characteristic (ROC) curves, demonstrating AUC values of 0.926 and 0.986 for LDA and OPLS-DA, respectively. In addition, the noninvasive SERS analysis of urine samples shows significant superiority for LC diagnosis over typical clinical biomarker tests such as carcinoembryonic antigen (CEA) testing in serum. These results demonstrate that label-free SERS associated with multivariate analysis is a promising tool for the noninvasive diagnosis and postoperative monitoring of LC patients.