Urine-based SERS and multivariate statistical analysis for identification of non-muscle-invasive bladder cancer and muscle-invasive bladder cancer
Qingshan Zhong,Lei Shao,Yudong Yao,Shuo Chen,Xiuyi Lv,Zhihan Liu,Shanshan Zhu,Zejun Yan
DOI: https://doi.org/10.1007/s00216-024-05595-0
2024-10-18
Analytical and Bioanalytical Chemistry
Abstract:Bladder cancer (BC) is an epidemiological urologic malignancy that continues to increase each year. Early diagnosis and prognosis monitoring is always significant in clinical practice, especially in distinguishing non-muscle-invasive bladder cancer (NMIBC) from muscle-invasive bladder cancer (MIBC), due to the various depths of tumor invasion related to different therapeutic schedules and recurrence rates. Common diagnostic approaches are too invasive or generally inefficient in accuracy and specificity. In this work, a totally non-invasive and cost-effective method is established by investigating urine samples using surface-enhanced Raman spectroscopy (SERS) and multivariate statistical analysis. The comparison of urine SERS spectra shows the intensities of characteristic peaks for DNA/RNA, hypoxanthine, albumin, D-( +)-galactosamine, fatty acids, and some amino acids are distinguishable in BC occurrence and invasion progression. A PLS-LDA-based two-step binary classification scheme is performed on urine SERS spectra and the diagnostic accuracies were 97.7% and 96.3% for healthy individuals versus BC patients and NMIBC versus MIBC patients, respectively. Moreover, the impact of urine SERS spectral lengths in reaching high-precision recognition of BC is investigated. The results show that the Raman peaks at 803, 893, 1139, 1375, and 1466 cm −1 play an essential role in correctly categorizing healthy control, NMIBC, and MIBC patients, and SERS spectra ranges from 400 to 1600 cm −1 are enough for this identification task. These findings provide a sensitive, label-free, rapid, and totally non-invasive way for assessment of invasion depth of BC to its early diagnosis and prognosis monitoring, as well as valuable insights for selecting reasonable spectral range to enhance the measurement efficiency especially in large-scale sample datasets.
chemistry, analytical,biochemical research methods