ATU-10 Targeting histopathology workload crisis using attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy

Introduction For the past 200 years histopathology has been the gold-standard for identification of many diseases. Biopsy procedures typically generate 1–0 sample sections for analysis, the whole process is time consuming and expensive. The Royal College of Pathologist recently published a workforce census showing that only 3% of NHS histopathology departments have enough staff to meet clinical demand [1] as consequence the reported waiting time for exam results is currently up to 14 weeks. This leaves us in desperate need of a change [2]. DynamX is developing a system that will allow to identify healthy samples within 15 seconds, removing unnecessary histopathological analysis, creating a faster route to treatment, when necessary, together with the possibility of more screening programs. Methods We combine infrared spectroscopy, biological information, machine learning and artificial intelligence to extract information from a tissue sample. A schematic representation of the process is given in the picture below. Each sample has a its own FTIR fingerprint that once acquired can be used to distinguish between heathy and abnormal cells. Our algorithm has the potential to be coupled with any commercially available hardware. The biopsy is scanned immediately after resection, no sample preparation is required so the sample can re-enter the routine diagnostic pathway if necessary. Results A 400 patient multicentre phase 2 IVD clinical trial (ISRCTN 33779, IRAS196045), focused on oesophageal and colon cancer screening has been recently completed. Primary objective of the trial was to evaluate the validity of the algorithm in discriminating among heathy and abnormal cells. Secondary objective consisted in testing and implementing our user interface. Over 1600 biopsies/polyps were analysed and unique spectral differences were identified. The algorithm achieved 96% sample sensitivity and 56% sample specificity. Human Factors analysis of the software was performed following study completion, along with feedback from a focus group with study nurses from the clinical trial, allowing for a rigorous assessment of the common problems. Conclusions Results of the completed clinical trial were extremely positive. DynamX technology showed the potential to be particularly beneficial to the time and price allocated for each histopathological exam and to assist in the assessment of the disease. The study allowed us to collect a broad variety of samples to build up our internal library. In order to expand our technology to diagnose a larger variety of diseases a clinical trial involving 450 patients on detection of presence of infrared biomarkers for cancer in cheek cells is about to start. References R.C. of Pathology Meeting pathology demand histopathology workforce census. 2018. Maung R. Pathologists workload and patient safety. 2016.