Spectroscopy and Artificial Intelligence to Disrupt the Status Quo in Cervical Cancer Screening

Spectroscopy and Artificial Intelligence to Disrupt the Status Quo in Cervical Cancer

Cervical cancer kills one woman every two minutes, 90% of these women are from low- and middle-income countries. Newly developed testing using biofluids has proven successful in identifying disease markers in, for example, brain cancers and endometrial cancers. Early studies have revealed that this technology is also showing potential for gynaecological cancers using validated human papillomavirus (HPV) test specimens. Urine samples, more easily collected, may make screening more accessible and acceptable to women. Spectroscopy is a portable and relatively simple technology; results are instant, reproducible and reliable. Once we confirm that spectroscopy has the ability to identify potential CIN 2+ by detecting HPV in urine, the test can be miniaturized and adapted to a point of care test. This will be more economical and logistically simpler than what is currently available; no consumables and pre-processing of samples are required. Women with abnormal cervical screening and women with normal screening as controls will be recruited, cervical and urine samples will be obtained. These will be tested for HPV DNA using standard methods and also by spectroscopy for HPV. These spectroscopy signals will be analyzed using artificial intelligence. The results will be compared to tissue samples obtained at colposcopy. This will allow evaluation of the new spectroscopy test. This preliminary study aims to prove the concept the spectroscopy as a simple, affordable screen can be used to radically change cervical cancer screening. Enabling a test that has point of care capabilities has huge implications for women in developed and more significantly in low-and middle-income countries, where cervical cytology and HPV testing have significant logistical problems. A non-invasive test will be preferred by many women. We believe spectroscopy will disrupt the status quo of 'no screening' in the low and middle income countries (LMICs), accelerate elimination of cervical cancer, and thus avert 15 million deaths in next 50 years.

After obtaining informed consent, cervical and urine specimens will be obtained from all enrolled participants. Cervical biopsies in the disease group will be carried out at the providers' discretion per routine colposcopy protocols. Cervical specimens will be collected by the attending gynecologists and urine samples will be collected by the participants. Aliquots of urine and cervical samples will be tested for HPV DNA with cobas 4800 HPV test (Roche Diagnostics) serving as the gold standard at the Newfoundland Public Health Laboratory, St. John's, Newfoundland. The cobas test, however, identifies only two (types 16 and 18) of 14 high-risk HPV genotypes specifically. Therefore, a second set of aliquots will be tested at Seegene laboratory in Kitchener, Ontario to obtain complete genotype profile of all 14 high-risk genotypes as well as quantitative HPV results using Anyplex HPV DNA test (Seegene Inc), a next generation HPV test. Urine samples will also be independently tested by Infrared (IR) Spectroscopy and signals analyzed by Prof. Rehman's lab, Department of Engineering, Lancaster University, Lancaster United Kingdom (UK) to detect HPV. Spectroscopy results will be compared with HPV test results with genotypes to determine concordance between specimen types and between the testing technologies, with a special reference to urine testing. Biopsy confirmed CIN2+ will serve as the clinical endpoint to determine the sensitivity and specificity of urine spectroscopy testing in comparison with urine HPV testing; cervical HPV results will serve as the reference method. We are partnering with NSV-Tech a biotechnology company based in the US and India who have experience with artificial intelligence and cancer screening using tablet-based technology. Spectroscopy signals will be shared with NSV-Tech to develop artificial neural networks (ANN). The spectroscopy outputs linked to the ground-truth (biopsy confirmed abnormal cervical cells (CIN2)+ histology) will be used to develop ANN to improve diagnostic accuracy of this clinically relevant endpoint and remove subjectivity. The study relies on the ability for spectroscopy to detect HPV. The research team is confident this will be feasible. As noted, members of the research team are reviewing reference urine samples previously analyzed using standard HPV detection methods to confirm that spectroscopy is a viable option method. The control specimen collection and analysis uses standard methods. The data collection will take place in an academic colposcopy clinic with adequate patient volume and a proven track record in recruiting patients for such trials. The analysis of the spectroscopy in the UK of the spectroscopy, is in labs that have significant experience in assessing biological samples. The project's international collaboration is significant and brings together world class leaders in the prevention of cervical cancer who all have experience with novel technology and screening trials.

Inclusion Criteria:

• Women undergoing standard HPV screening at the IWK Health Centre and Nova Scotia Health in Halifax, Nova Scotia, Canada

Exclusion Criteria:

-