Thyroid cancer is increasing in prevalence among Indians at a rapid pace. Could new AI technology provide a far more efficient means to diagnose the disease?
Artificial intelligence (AI) — or machine learning — may be the next major leap forward in the field of cancer diagnosis. The capacity for an automated programme to be able to analyse ultrasound images taken from the patient for signs of potential cancer has applications for both thyroid cancer and various other types of the disease — or, indeed, for many other diseases outside the field of cancer.
A study conducted at Thomas Jefferson University in the US has used the approach of ultrasound images being analysed by machine learning programmes as a quick and non-invasive approach to thyroid cancer screening. Such an addition to current screening procedures can rule out more intrusive biopsies in those that display no symptoms. In those patients analysed that the AI programme does find cause for suspicion, further investigation is conducted.
“Machine learning is a low-cost and efficient tool that could help physicians arrive at a quicker decision as to how to approach an indeterminate nodule,” said the study’s lead author John Eisenbrey.
Machine learning works through pattern recognition. The more data that is compiled into the system the more accurate the programme becomes. In the case of recognising cancers, the programme can analyse the expected cellular structure from an ultrasound image. If there is a deviation from the normal cellular structure, these differences are notified. Further procedures can then identify if the differences indicate the suspected presence of cancer.
India desperately needs to incorporate new tools to address its growing burden of cancer in the country. Foremost among the issues plaguing its response to its expanding rates of the condition is its diagnosis rates. In many instances, this is not due to a technological limitation, but rather a lack of access to screening programmes or, for many, a lack of knowledge regarding the symptoms of cancer.
In the case of thyroid cancer, the rates over the last decade have shown a relative increase of 62 percent and 48 percent for women and men respectively. Concerningly, the primary group in which rates have increased is in those aged below 45, indicating an earlier onset age than was previously documented.
AI screening presents a fast and cost-effective method of screening individuals en masse. This could be a major boon to screening programmes for at-risk groups. Earlier diagnosis resulting from these programmes would mean treatment can be administered in the earlier stages of the disease. As such, there is the potential for vastly improved treatment outcomes.