Abstract
Lung cancer is the world's biggest cause of death related to cancer, and its dismal prognosis has been greatly exacerbated by late-stage diagnosis. Even with improvements in treatment strategies, current diagnostic techniques are frequently imprecise, especially when it comes to early-stage detection. A prospective substitute is Raman spectroscopy, which provides a non-invasive, real-time, and extremely sensitive study of biological samples. The objective of this study is to assess the diagnostic efficacy of Raman spectroscopy in the identification and diagnosis of lung cancer across a range of sample types. Nine studies that focused on Raman spectroscopy as a stand-alone diagnostic tool and met strict inclusion criteria were found through a systematic review of the literature published between 2014 and 2024. Statistical methods were used to extract, pool, and show diagnostic measures. The remarkable diagnostic accuracy of Raman spectroscopy was highlighted by its pooled sensitivity and specificity which were 98.68% and 91.81%, respectively. Serum-based research showed the strongest findings, with multivariate models such as PCA-LDA supporting specificity and sensitivity values that, in several cases, reached 100%. Diagnostic accuracy was greatly improved by models such as SVM and CNN, particularly when it came to detecting minute spectral alterations associated with cancer. Raman spectroscopy shows great promise as a lung cancer diagnostic method. However, issues including spectral data standardization, sample preparation heterogeneity and the requirement for bigger, multicentre research needs to be addressed. These results will open the door for the incorporation of Raman spectroscopy into standard clinical procedures.