Weird idea/question of the day - would it be possible to use anthropometric data to predict cancer risk?
Is it possible that machine learning-assisted anthropometric analysis could provide better risk assessment than genetic screening for cancer or other conditions (e.g., connective tissue disorders, heart failure, etc.)? Relative
proportions are reflective of relative growth rates of organs and
tissues which in turn are the product of both genes and environment and
are dependent on factors such as the balance of growth hormones and
nutrient intakes. It seems that if a large number of individuals were
scanned photographically - or much better with MRI (since more interesting measurements could be taken, e.g. width of the aorta) and the scans were provided as input, along
with data from their health records, to an appropriately designed machine learning algorithm, it may be possible to predict risk of various diseases based on anthropometric parameters alone.
This occurred to me just prior to learning that a new company, Ezra, is offering whole body MRI scans for early cancer detection for just under $2,000. Seems like a great thing and I hope they expand to more cities soon. Eventually they may amass a very large dataset (though more than likely it will be unfortunately skewed in terms of socioeconomic status) and the possibility of utilizing it in this way is intriguing. Prevention is even better than early detection, and one would hope prediction would enable better prevention.
[Correction: a previous version of this post referred to anthropometrics as biometrics ; Wikipedia has an explanation of the difference in meaning.]
Bibliography [will be expanded]
Bandera et al. (2016) 'The use and interpretation of anthropometric measures in cancer epidemiology: A perspective from the world cancer research fund international continuous update project' - https://pubmed.ncbi.nlm.nih.gov/27352197/
Giovanucci (2019) 'A growing link—what is the role of height in cancer risk?' - https://www.nature.com/articles/s41416-018-0370-9
Rahman et al. (2011) 'Hand pattern indicates prostate cancer risk' - https://pubmed.ncbi.nlm.nih.gov/21119657/
Scutt et al. (2006) 'Breast asymmetry and predisposition to breast cancer' - https://breast-cancer-research.biomedcentral.com/articles/10.1186/bcr1388
Comments
Post a Comment