Ph.D Thesis

Ph.D StudentHakim Meggie
SubjectEarly Detection of Lung Cancer via Breath Samples:
A Clinical Study
DepartmentDepartment of Chemical Engineering
Supervisor PROF. Hossam Haick
Full Thesis textFull thesis text - English Version


Early detection of lung cancer (LC) can improve the 5 years survival by 3-4 folds. However, there is no approved screening protocol for high risk patients. An emerging approach for diagnosis LC relies on the breath analysis of volatile organic compounds (VOCs) that are emitted due to the biological pathways occurring in the human body.

The current thesis aims to tailor a novel chemical nanoarray containing cross reactive sensors in order to compare the signatures and compositions of exhaled VOCs in three clinical trials. In the first clinical trial, breath samples were taken from 72 patients with pulmonary nodules (PNs). The profiles of the VOCs were determined by (I) gas chromatography/mass spectrometry (GC-MS) combined with solid-phase microextraction and (II) the chemical nanoarray. GC-MS analysis identified a significantly higher concentra?tion of 1-octene in the breath of LC, and the nanoarray distinguished significantly between benign and malignant PNs, between adeno- and squamous-cell carcinomas and between early stage and advanced disease. In the second clinical trial, alveolar breath was collected from 87 volunteers: from head and neck cancer patients (HNC), LC patients and healthy volunteers. The choice of HNC patients relates to the fact that these patients often develop a second primary tumor that can affect the entire aero-digestive tract, mostly HNC or LC. The sensors’ signals were analyzed using: (I) principal component analysis (PCA) with analysis of variance (ANOVA) and Student’s t-test and (II) support vector machines (SVM) followed by cross-validation. The results showed clear distinguishment between each pair of the groups. The results were supported by comparative analysis of the chemical composition of the breath through GC-MS. The third clinical trial aimed to explore the difference in the breath print of the four most widespread cancers in the developed world: LC, prostate cancer, and colorectal cancer for men; and breast cancer, colorectal and LC for women. The GC-MS results showed that each cancer has a unique pattern of VOCs, when compared with healthy states. However, GC-MS was not sufficiently accurate to distinguish between the four cancers in a single test.

In summary, this study shows the potential of nanomaterial-based sensors as an inexpensive and por?table tool to further improve the noninvasive-biomarker-based investigation of patients who are not candidates for invasive procedures or in cases where the tissue is hard to sample. To this end, studies involving a larger cohort of patients are needed.