Detection of nasopharyngeal cancer using confocal Raman spectroscopy and genetic algorithm technique

Abstract

Raman spectroscopy (RS) and a genetic algorithm (GA) were applied to distinguish nasopharyngeal cancer (NPC) from normal nasopharyngeal tissue. A total of 225 Raman spectra are acquired from 120 tissue sites of 63 nasopharyngeal patients, 56 Raman spectra from normal tissue and 169 Raman spectra from NPC tissue. The GA integrated with linear discriminant analysis (LDA) is developed to differentiate NPC and normal tissue according to spectral variables in the selected regions of 792-805, 867-880, 996-1009, 1086-1099, 1288-1304, 1663-1670, and 1742-1752 cm-1 related to proteins, nucleic acids and lipids of tissue. The GA-LDA algorithms with the leave-one-out cross-validation method provide a sensitivity of 69.2% and specificity of 100%. The results are better than that of principal component analysis which is applied to the same Raman dataset of nasopharyngeal tissue with a sensitivity of 63.3% and specificity of 94.6%. This demonstrates that Raman spectroscopy associated with GA-LDA diagnostic algorithm has enormous potential to detect and diagnose nasopharyngeal cancer.

Characterization and noninvasive diagnosis of bladder cancer with serum surface enhanced Raman spectroscopy and genetic algorithms

“…Our groups previously have applied GA-LDA algorithms to search significant Raman bands for noninvasive detection of nasopharyngeal tumor from the normal and neoplastic nasopharyngeal tissues Raman spectra. A diagnostic sensitivity of 69.2% and specificity of 100% for distinction of nasopharyngeal neoplasia is provided by the diagnostic models developed with identified seven Raman bands 38 .…”

Raman Spectroscopy Analysis of the Biochemical Characteristics of Molecules Associated with the Malignant Transformation of Gastric Mucosa

“…The Raman peaks at 1658 cm -1 , 1033 cm -1 , 1266 cm -1 and 1127 cm -1 represent proteins [4]-[6], [13], [20]. Compared with normal tissue, the position of 1658 cm -1 ,1127 cm -1 , 1033 cm -1 and 1266 cm -1 were shifted in cancer tissue to various degrees, suggesting that the interactions between chemical bonds of amino acids are weakened in cancer cells.…”

“…It has been reported that the sensitivity and specificity of using Raman spectrometry to diagnose gastric mucosal lesions in vivo are 85%–95% and 90%–98%, respectively [13]. Scientists now mainly focus on the differential comparison of Raman spectra, the establishment of diagnostic models and principles by combining Raman spectrometry and multivariate statistics, and distinguishing malignant versus benign tumors, pathological subtypes, degree of differentiation, and lymph node metastasis [1], [4]–[7], [10]. Raman spectrometry has not been utilized to its full potential to analyze the microstructure of molecules and the mechanisms and principles associated with malignancy of tissue and cells.…”

Surface-enhanced Raman spectroscopy + support vector machine: a new noninvasive method for prostate cancer screening?

“…Perhaps it will be helpful to resolve the problem with the aid of feature selection techniques, such as genetic algorithms. Huang's group and our group have conducted similar studies by feature selection [16][17][18].…”