Speakers
Ed Wilkes
Kamaljit Chatha
Chelsey Walsh
Parallel session
Chair: Ed Wilkes
This parallel session will cover a variety of topics surrounding the practical applications and ethical implications of using AI/ML in our future practice.
11.30 - 12.00 - The complex pathway to AI implementation in the NHS - Kamaljit Chatha
There is real drive nationally to investigate and implement exciting developments that Artificial Intelligence (AI) has the potential to deliver. At the same time there have been several parallel developments in setting out the governance and evidence required to ensure that the roll out of machine learning and AI software in the healthcare system is implemented in a safe manner and does not cause harm to patients. This presentation will use an ongoing AI project being developed for Bowel Cancer Screening, ColonSys, to highlight the various legal, scientific stages and regulatory approvals needed to clinically validate and implement AI in the NHS setting.
12.00 - 12.30 - Can artificial intelligence replace biochemists? - Chelsey Walsh
From grocery shopping to essay writing, artificial intelligence (AI) tools are becoming a staple part of modern life. Media publications focus heavily on the positive impact this technology, and in particular, chatbots can have in improving productivity and efficiency in a busy, technological world. So, with patients able to access their own laboratory results through the NHS app, and increasing wait times for GP appointments, the public may feel comfortable asking the internet for help in understanding their own test results. However, the safety and accuracy of responses given by chatbots is unknown compared to the standard of interpretation provided by a qualified, state-registered clinical scientist or chemical pathologist.
We therefore undertook a small-scale study that aimed to assess the accuracy and safety of two freely available AI tools providing interpretation of thyroid function test results as if posed by laboratory scientists or patients. A range of scenarios were generated to mimic a typical selection of cases and combination of blood test results seen when working in the laboratory, and presented to a group of practicing Biochemists and Chemical Pathologists as well as the two AI chatbots. The main aims were to see if AI in its current form could replace human clinical knowledge? Most importantly, are these widely used and freely available tools safe for use in this way? We looked into the appropriateness of the advice offered by these chatbots and tried to decide whether or not they could one day take our jobs!
12.30 - 1.00 - SPECTR: Automating the detection of monoclonal gammopathy using serum protein electrophoresis and deep learning - Xavier Dieu
Serum protein electrophoresis (SPEP) is a routine analysis in medical laboratories. Its main indication is screening, diagnosis, and follow-up of monoclonal gammopathies. Human interpretation is to this day still mandatory for highlighting relevant pathological patterns and avoiding pitfalls that may alter trace interpretation. This human step hinders throughput, harmonization, and security of results.
Ed completed his PhD and post-doctoral work at Barts Cancer Institute, where he developed computational methods for predicting drug response to small kinase inhibitors in cancer models. He went on to train as a clinical scientist in the NHS and is now the Head of Research at Kinomica. His interests primarily lie in predictive modelling, statistics education, systems biology, and (phospho-)proteomics.
Pronouns: He/him/his
Dr Kamaljit Chatha is a Consultant Clinical Biochemist and Director of the Bowel Cancer Screening Midlands and North West Hub at the University Hospitals Coventry and Warwickshire (UHCW). Kam started her career in Laboratory Medicine as a Biomedical Scientist after completing a BSc in Biochemistry and Pharmacology. As a Biomedical Scientist she then went on to obtain an MSc in Biomedical Sciences and switched to the Clinical Scientist training program.
Kamaljit has a keen interest in research and development and has several publications that covers the novel and innovative use of biomarkers to detect chronic disease such as cardiovascular disease and renal disease. She also has an interest in screening, as highlighted by her work in Bowel Cancer Screening, which includes two Innovate UK grants to combine FIT with artificial intelligence systems to improve bowel cancer detection in the screening programme, ColonSys and the symptomatic FIT service, BowelSys. Kamaljit has been actively involved with regional and national Association for Laboratory Medicine (LabMed) committees and is currently Director of Publications and Communications for LabMed, promoting the profession and the excellent work it delivers. Kamaljit is also an Honorary Clinical Lecturer for the University of Warwick and is part of the Institute of Cardiometabolic Medicine (ICMM) at UHCW.
Dr Xavier Dieu holds a Medical Degree with a specialty in Laboratory Medicine, Biochemistry and Molecular Biology. He defended his PhD thesis in 2022, in the University of Angers, after developing several new laboratory medicine tools relying on the use of innovative data science methods, including machine learning and deep learning. Besides his clinical activity focused on the screening and diagnosis of rare diseases through molecular biology, his research projects aim at improving harmonization, throughput, and the patient’s security in the clinical laboratory, through artificial intelligence.