Publications by authors named "Thushika Mahendiran"
Lancet Digit Health
October 2019
Article Synopsis
- Deep learning algorithms show significant potential in medical diagnostics, prompting a systematic review to evaluate their accuracy against healthcare professionals in classifying diseases through medical imaging.
- From 31,587 studies screened, 82 met the criteria for inclusion, where sensitivity of deep learning methods ranged from 9.7% to 100% (mean 79.1%), and specificity ranged from 38.9% to 100% (mean 88.3%).
- In a subset of studies comparing deep learning with healthcare professionals, the pooled sensitivity for deep learning models was 87% compared to 86.4% for professionals, indicating comparable diagnostic performance.
