NON-INVASIVE DIAGNOSIS OF LUNG CANCER BASED ON CFD MODELING AND FRACTAL ANALYSIS

Due to a great majority of lung cancer patients dying within one year after being diagnosed with apparent symptoms, developing a diagnostic/monitoring technique for early-stage lung cancer is in critical demand. Conventionally, lung cancer diagnostic approaches are costly, and they increase the health risks caused by invasiveness and radiation hazards. In this work, a new diagnostic technique using aerosol fingerprints in the breath test is explored based on computational fluid dynamics (CFD) modeling and fractal analysis. At first, the three-dimensional symmetrical human lung model is constructed by Solidworks. Then, the large eddy simulation-discrete-phase model (LES-DPM) approach is used in CFD modeling to model the airflow pattern and aerosol behaviors in the human lung model. After that, the box-counting method is employed in fractal analysis to calculate the fractal dimension of exhaled aerosol patterns in the human lung model. Finally, the fractal distributions of exhaled aerosols in the breath test at different respiratory intensities and aerosol-releasing positions are specifically investigated. The results show that the coupled CFD modeling and fractal analysis is a reliable method for deciphering the complexity of exhaled fingerprints which is shown to quantify and differentiate the exhaled aerosol patterns adequately. Meanwhile, the exhaled aerosol fingerprints are found to be relevant to the respiratory intensity, and the distribution of exhaled aerosols exhibits a unique pattern at different respiratory intensities. Therefore, the abnormal respiration of patients, which hints at the severity of lung cancer, can be judged by aerosol fingerprints and fractal dimensions. In addition, the exhaled aerosol fingerprints are associated with the aerosol-releasing positions. Different aerosol-releasing positions lead to different aerosol distribution patterns. It is feasible to locate the site of lung cancer by judging the aerosol fingerprints and fractal dimension. This study is helpful in determining the respiratory abnormalities caused by lung cancer and diagnosing the location where the lung cancer occurs.

Automated summary

Due to the high mortality rate of lung cancer patients, there is a critical demand for early-stage diagnostic/monitoring techniques. This study explores a new diagnostic technique using aerosol fingerprints in the breath test based on computational fluid dynamics (CFD) modeling and fractal analysis. The results show that the coupled CFD modeling and fractal analysis method is reliable for quantifying and differentiating exhaled aerosol patterns. The abnormal respiration and location of lung cancer can be judged by analyzing the aerosol fingerprints and fractal dimensions.