Michele Astolfi, Martina Masin, Antonio Anfuso, Cesare Malagù, Gabriele Anania, Mascia Benedusi, Giuseppe Valacchi, Giorgio Rispoli
Colorectal cancer represents a global health burden, being the third most frequently diagnosed cancer worldwide, accounting for about 1.9 million new cases annually, and the second leading cause of cancer-related deaths. This highlights the need for innovative and effective methods for cancer detection, improving current diagnostic approaches. In this feasibility study, two nanostructured chemoresistive gas sensors based on tin oxide-titanium oxide composites were selected to detect the metabolic patterns associated with human healthy and colorectal cancer tissues, collected from 26 patients. To further validate the biopsy-derived results, the sensors were tested on two colorectal cancer-derived cell lines, namely Caco-2 and RKO. Both sensors were able to discriminate between healthy and tumor samples by comparing the mean sensor responses, using the paired t-test (p-value = 0.00018), and by using an LDA, achieving a cross-validated accuracy of 0.73. Moreover, the ROC analysis performed on the data projected along the discriminant direction led to an AUC of 0.83. Finally, the sensors exhibited distinct response profiles to Caco-2 and RKO exhalations, as expected given their different biological features. As a proof-of-feasibility study, these findings demonstrate that these sensors could capture a characteristic volatile fingerprint distinguishing the samples, although they cannot identify the specific cancer metabolites.