Using Disruptive Technology In Healthcare – Can Machine Learning Be Employed To Diagnose Breast Cancer

Breast ultrasound elastography is an emerging imaging technique used by doctors to help diagse breast cancer, w researchers identified critical role machine learning can play in making this technique more efficient and accurate in diagsis. 

Breast cancer is leing cause of cancer-related death among women. It is also difficult to diagse. Nearly one in 10 cancers are misdiagsed as t cancerous, meaning that a patient can lose critical treatment time. 

On or hand, more mammograms a woman has, more likely it is she will see a false-positive result. After 10 years of annual mammograms, roughly two out of three patients who do t have cancer will be told that y do and be subjected to an invasive intervention, most likely a biopsy. 

Using more precise information about characteristics of a cancerous versus n-cancerous breast lesion, this methodology has demonstrated more accuracy compared to tritional modes of imaging. In  case of breast ultrasound elastography, once an im of affected area is taken, im is analysed to determine displacements inside tissue. Using this data and physical laws of mechanics, spatial distribution of mechanical properties like its stiffness is determined. After this, one has to identify and quantify appropriate features from distribution, ultimately leing to a classification of  tumour as malignant or benign. problem is final two steps are computationally complex and inherently challenging. 

In research published in Journal of Computer Methods in Applied Mechanics and Engineering one of researchers of study, Ass Oberai sought to determine if y could skip most complicated steps of this workflow entirely. 

Cancerous breast tissue has two key properties: heterogeneity, which means some areas are soft and some are firm and n-linear elasticity, which means  fibres offer a lot of resistance when pulled inste of initial give associated with benign tumours. 

Kwing this, Oberai created physics-based models that showed varying levels of se key properties. He n used thousands of data inputs derived from se models in order to train machine learning algorithm. 

"If you h eugh data available, you wouldn't. But in case of medical imaging, you're lucky if you have 1,000 ims. In situations like this where data is scarce, se kinds of techniques become important," said Oberai. Through eugh examples, algorithm is able to glean different features inherent to a benign tumour versus a malignant tumour and make correct determination. Oberai and his team achieved nearly 100% classification accuracy on or syntic ims. Once algorithm was trained, y tested it on real-world ims to determine how accurate it could be in providing a diagsis, measuring se results against biopsy-confirmed diagses associated with se ims."We h about an 80 percent accuracy rate. Next, we continue to refine algorithm by using more real-world ims as inputs," Oberai said. 

Also Re: Artificial Intelligence Could Be Used To Find Out If re Is Discrimination At Various Levels