|
Getting your Trinity Audio player ready...
|
A recent study conducted by researchers at the University of South Australia (UniSA) has unveiled a spectrum of metabolic biomarkers that hold promise in predicting cancer risk. Employing advanced machine learning techniques to analyse data from 459,169 participants enrolled in the UK Biobank, the research identified 84 distinct features that could potentially serve as indicators of heightened cancer susceptibility.
Several of these identified markers were also associated with chronic kidney or liver diseases, raising intriguing questions about potential links between these ailments and cancer. Led by a team of experts including Dr Iqbal Madakkatel, Dr Amanda Lumsden, Dr Anwar Mulugeta, and Professor Elina Hyppönen from the University of South Australia, along with the University of Adelaide’s Professor Ian Olver, this groundbreaking study, titled “Hypothesis-free Discovery of Novel Cancer Predictors Using Machine Learning,” delved deep into the data.
Dr Madakkatel, one of the lead researchers, explained the methodology stated that the team performed an exploratory analysis utilising artificial intelligence and statistical methods to pinpoint factors associated with cancer risk from a pool of over 2800 features.
The study’s outcomes were nothing short of remarkable, with over 40% of the features uncovered by the model proving to be biomarkers – biological molecules that can signify either sound health or underlying health issues, depending on their status. Significantly, some of these biomarkers exhibited dual associations, being linked not only to cancer risk but also to kidney or liver diseases.
Dr Lumsden elaborated on the implications of these findings, noting that the study offers valuable insights into potential mechanisms contributing to cancer risk. She stated that after age, the most significant indicator of cancer risk was identified as elevated levels of urinary microalbumin. Microalbumin, a vital serum protein essential for tissue repair and growth, takes on a dual role when detected in urine, serving as an indicator not only of kidney disease but also as a marker signalling an increased risk of cancer.
The study also identified other indicators of compromised kidney function, such as elevated blood levels of cystatin C, increased urinary creatinine (a waste product eliminated by the kidneys), and an overall reduction in total serum protein, all of which were linked to cancer risk.
Moreover, the research discovered a connection between an elevated red cell distribution width (RDW) – an indicator of the variation in the size of red blood cells – and an increased risk of cancer. Typically, red blood cells are relatively uniform in size, and deviations from this norm can signify higher inflammation and poorer renal function, as well as a heightened risk of cancer.
In addition to these findings, the study highlighted that elevated levels of C-reactive protein, a marker of systemic inflammation, were associated with an increased risk of cancer, along with high levels of the enzyme gamma glutamyl transferase (GGT), a biomarker indicative of liver stress.
Professor Elina Hyppönen, the chief investigator and the director of the Australian Centre for Precision Health at UniSA, emphasised the strength of this study, which lies in the power of machine learning. Professor Hyppönen noted that the model powered by artificial intelligence, has showcased its capability to assimilate and cross-reference a multitude of characteristics, thus uncovering pertinent risk factors that could otherwise remain hidden.
An intriguing aspect of this study was that, despite considering thousands of features, spanning clinical, behavioural, and social factors, a significant proportion were biomarkers reflecting the metabolic state before a cancer diagnosis. While the findings offer promise, Professor Hyppönen stressed the need for further research to confirm causality and clinical relevance.
The implications of this research are profound. With relatively simple blood tests, it might be possible to gain insights into one’s future risk of cancer. This potential early detection could enable proactive measures to be taken at a stage when cancer might still be preventable. The significance of these findings underscores the importance of ongoing research in the field of cancer risk prediction
