Development of early noninvasive methods for lung cancer diagnosis is among the most promising technologies, especially using exhaled breath as an object of analysis. Simple sample collection combined with easy and quick sample preparation, as well as the long-term stability of the samples, make it an ideal choice for routine analysis. The conditions of exhaled breath analysis by preconcentrating volatile organic compounds (VOCs) in sorbent tubes, two-stage thermal desorption and gas-chromatographic determination with flame-ionization detection have been optimized. These conditions were applied to estimate differences in exhaled breath VOC profiles of lung cancer patients and healthy volunteers. The combination of statistical methods was used to evaluate the ability of VOCs and their ratios to classify lung cancer patients and healthy volunteers. The performance of diagnostic models on the test data set was greater than 90 % for both VOC peak areas and their ratios. Some of the exhaled breath samples were analyzed using gas chromatography coupled with mass spectrometry (GC-MS) to identify VOCs present in exhaled breath at lower concentration levels. To confirm the endogenous origin of VOCs found in exhaled breath, GC-MS analysis of tumor tissues was conducted. Some of the VOCs identified in exhaled breath were found in tumor tissues, but their frequency of occurrence was significantly lower than in the case of exhaled breath.
Applying GC-MS with the use of VOC peak area ratios instead of peak areas allows to distinguish the samples of lung cancer patients and healthy volunteers with higher performance.
Non-invasive, simple, and fast tests for lung cancer diagnostics are one of the urgent needs for clinical practice. The work describes the results of exhaled breath analysis of 112 lung cancer patients and 120 healthy individuals using gas chromatography-mass spectrometry (GC-MS). Volatile organic compound (VOC) peak areas and their ratios were considered for data analysis. VOC profiles of patients with various histological types, tumor localization, TNM stage, and treatment status were considered. The effect of non-pulmonary comorbidities (chronic heart failure, hypertension, anemia, acute cerebrovascular accident, obesity, diabetes) on exhaled breath composition of lung cancer patients was studied for the first time. Significant correlations between some VOC peak areas and their ratios and these factors were found. Diagnostic models were created using gradient boosted decision trees (GBDT) and artificial neural network (ANN). The performance of developed models was compared. ANN model was the most accurate: 82–88% sensitivity and 80–86% specificity on the test data.
Development of simple, fast, and non-invasive tests for lung cancer diagnostics is essential for clinical practice. In this paper, exhaled breath and skin were studied as potential objects to diagnose lung cancer. The influence of age on the performance of diagnostic models was studied. Gas chromatography in combination with mass spectrometry (MS) was used to analyze the exhaled breath of 110 lung cancer patients and 212 healthy individuals of various ages. Peak area ratios of volatile organic compounds (VOCs) were used for data analysis instead of VOC peak areas. Various machine learning algorithms were applied to create diagnostic models, and their performance was compared. The best results on the test data set were achieved using artificial neural networks (ANNs): classification of patients with lung cancer and young healthy volunteers: 88 ± 4% sensitivity and 83 ± 3% specificity; classification of patients with lung cancer and old healthy individuals: 81 ± 3% sensitivity and 85 ± 1% specificity. The difference between performance of models based on young and old healthy groups was minor. The results obtained have shown that metabolic dysregulation driven by the disease biology is too high, which significantly overlaps the age effect. The influence of tumor localization and histological type on exhaled breath samples of lung cancer patients was studied. Statistically significant differences between some parameters in these samples were observed. A possibility of assessing the disease status by skin analysis in the Zakharyin-Ged zones using an electronic nose based on the quartz crystal microbalance sensor system was evaluated. Diagnostic models created using ANNs allow us to classify the skin composition of patients with lung cancer and healthy subjects of different ages with a sensitivity of 69 ± 2% and a specificity of 68 ± 8% for the young healthy group and a sensitivity of 74 ± 7% and a specificity of 66 ± 6% for the old healthy group. Primary results of skin analysis in the Zakharyin-Ged zones for the lung cancer diagnosis have shown its utility, but further investigation is required to confirm the results obtained.
19 September 2024 in Molecular Diagnosis & Therapy
We present a review of the main results achieved by various research groups in the studies on identifying biomarkers of lung cancer among volatile organic compounds (VOCs) in exhaled breath. Known methods for determining VOCs in exhaled breath, sampling and storing samples, preconcentrating VOCs, and statistical processing of the obtained data are considered; their advantages and disadvantages are discussed. A list of the main biomarkers of lung cancer is summarized. The main problems hindering the introduction of the analysis of exhaled breath into clinical practice are discussed.