Spectral delayed luminescence system for human saliva screening

Abstract

Delayed luminescence (DL) is a measurement method utilizing the decay of photon counts recorded, after one second or more, after a sample is exposed to a stimulating light source. DL has been studied on various human samples including blood serum, lung cells, cancerous and tumor cells, and the skin, except saliva. Previous studies have cross-correlate the DL of respective human samples with a range of diseases comprising of diabetes, leukemia, lung cancer, and tumor. However, recent studies have shown that DL is not cell type specific due to the unknown mechanism of the photon emission. Hence, this method is not viable for the diagnosis of complex diseases but it is proposed for non-invasive disease screening. Saliva, which can be obtained non-invasively, was anticipated for the screening of diseases using the DL method. Therefore, the aim of this study is to identify the potential application for the screening of diseases using the spectral delayed luminescence (SDL) of saliva. In order to achieve this, a prototype DL system was developed. The prototype DL system was then tested with and without the cooling of the photomultiplier tube (PMT). Illumination of the sample in the DL system was then tested with the ultraviolet (UV) light emitting diode (LED) which was then compared against the white LED. DL without the PMT cooling system, shows lower photon count deviations when compared to the PMT cooled at -8 ˚C for stimulation time ranging from 50 to 950 ms. In addition, the UV LED stimulation showed higher DL photon counts compared to white LED stimulation. Optimal stimulation time was then iterated for the SDL measurements of saliva and tongue swab, and it was found that 0.5 s is the optimal stimulation time. The first set of SDL measurements of saliva and tongue swabs were model fit into eight classes of mouth conditions. The classification performances of the respective models were then tested against the second data set of SDL measurements from the same respective sample. Results shows that the most significant application of the SDL system is in detecting conditions related to mouth sores. This significance is dependent on both the SDL measurements of saliva and tongue swab with detection performance of 100 % sensitivity, 85 % specificity, and 93 % non-error rate.