Sensing Device for Colorectal Cancer Preventive Screening through Fecal Odor Analysis:Clinical Validation Outcomes

Abstract

Colorectal cancer (CRC) is a worldwide diffused pathology for both men and women and it is the third most common cancer for both genders, in the United States[1] However, if promptly diagnosed, CRC is also one of the most curable tumor types (90% curability rate at stage I), therefore prevention is fundamental to avoid degeneration. Here the clinical validation outcomes of a preventive screening method based on a patented device, SCENT A1, composed by a core of nanostructured MOX sensors[2]. SCENT A1 is capable of detecting the presence of CRC from fecal odor with an in-vitro, inexpensive and non-invasive methodology. Fecal odor can be affected by the presence of CRC-biomarkers produced by membrane peroxidation and metabolic alterations[3,4]. The screening analysis adopted by NHS of Italy and of other countries is fecal occult blood test (FOBT). In Italy, FOBT is performed every two years on subjects from 50 to 69 years old and shows a huge percentage of false positives (about 65% according to data). Positives are then invited to undergo colonoscopy. SCENT A1 is capable of reducing FOBT false positives and so the total number of non-operative colonoscopies with related risks (e.g. bowel perforation). The device is composed by a core of five chemoresistive MOX sensors, chosen and calibrated for this specific aim [5-10]. Sensors (Figure 1) have been synthesized at Sensor Laboratory of the University of Ferrara and are composed by a nanostructured semiconductor film screen-printed onto an alumina substrate and a platinum heater to modulate operating temperature. In the ending protocol phase, the chosen sensors have been limited to two, one composed by Iron and Samarium oxides and the other by Tin and Titanium oxides, to simplify the device, facilitating a future large scale reproduction, without any information loss. The data collection and analysis software employs support vector machine (SVM) [9]. The clinical validation protocol of this device started in May 2016, in collaboration with Hospital St. Anna of Ferrara, Delta Hospital of Lagosanto and Department of Public Health of Ferrara. A total of 398 fecal samples of FOBT-positive subjects have been analyzed by SCENT A1 system with the k-fold cross validation method. Samples have been grouped into two macrocategories, depending on the gold-standard (colonoscopy) outcomes: 260 healthy subjects (negative to colonoscopy); 138 colorectal adenomas and carcinomas (positive to colonoscopy), 54 low-risk adenomas and 84 high-risk adenomas and carcinomas. The system has been capable of distinguishing among these two cathegories with a sensitivity (TPR: true positive rate) and specificity (TNR: true negative rate) respectively of of 84,1% and 82,4%, defined as follows: TPR=TP/P=TP/(FN+TP);TN/N=TN/(FP+TN); where TP, TN, FP and FN are the abbreviations that indicate respectively true positives, true negatives, false positives and false negatives. The choice to group carcinomas, high- and low-risk adenomas into a single class and not into two diverse classes as before [10], has been done in order to simplify the test and to ensure greater doctor protection, despite a small loss in specificity. Low-risk adenomas are, in fact, often similar to healthy tissues and they have a risk of evolving into cancer comparable to the risk of colonoscopy complications. As already emerging from a previous study [10], in fact, the major error of classification (only 57%) concerned just low-risk tumors that emit gases halfway between healthy and tumoral samples. What emerges is that, if SCENT A1 test is carried out on FOBT-positives subjects as a second check, it would guarantee a reduction of about two thirds of the colonoscopies performed on healthy patients. A new multicenter protocol, involving also Ospedale Maggiore of Bologna (Italy) and the private clinic Quisisana will start at the end of 2020, in order to test diverse devices of the same type also to FOBT and colonoscopy negatives, in to identify interval cancers. Moreover, if employed by the NHS of countries without screening system, this method will significantly reduce CRC-mortality rate.[1] [https://www.cancer.org/cancer/colon-rectal-cancer/about/key-statistics.html[2] Italian #: RM2014A000595, European #: 3210013 (Germany, UK);[3] B. Szachowicz-Petelska et al., NMR in Biomedicine V5,226–233, 1992;[4] T. G. de Meij et al., International Journal of Cancer (2014), 134, 1132–1138;[5] C. Malagù et al., Sensors, 14, 18982-18992, 2014;[6] G. Zonta et al., Sensors and Actuators B, 218, 289-295, 2015;[7] N. Landini et al., Scholars' Press, ISBN-13: 978-3-639-76538-0;[8] G. Zonta et al., Sensors and Actuators B, 238, 1098–110, 2016;[9] G. Zonta et al., Sensors and Actuators B 262 (2018) 884–891;[10] G. Zonta et al., Sensors and Actuators B: Chemical, Volume 301, 12 December 2019, 127062;[11] B. Scholkopf, A.J. Smola, 2001 (ISBN: 9780262253437); Figure 1