How AI helps to predict breast cancer: MIT researchers answer

AI has advanced in all fields of lives. Now, it has developed to the era where several types of diseases, be it curable or uncurable can be predicted easily with AI power. Very recently, MIT has found a way to Use AI to Predict Breast Cancer Up to 5 Years in Advance. With this deep-learning image classification model, Rather than taking a one-size-fits-all approach, we can personalize screening around a woman’s risk of developing cancer using Artificial intelligence.

Modern medication is round-faced with the challenge of getting, analyzing and applying a large amount of knowledge necessary to solve complex clinical problems. The development of medical artificial intelligence has been related to the development of AI programs meant to assist the practician within the formulation of identification, the making of therapeutic decisions and the prediction of outcome.

They are designed to support health care employees in their everyday duties, assisting with tasks that rely on the manipulation of data and knowledge. Such systems include Artificial neural networks (ANNs), fuzzy expert systems, evolutionary computation, and hybrid intelligent systems.

During the past years, there has been a significant increase in the level of interest regarding content-based data processing and case-based reasoning using AI. One major area of accelerated growth in the field of medical informatics where a lot of research is done towards the development of diagnostic tools, designed to support the work of medical professionals.

Since the first breast-cancer risk model from 1989, development has largely been driven by human knowledge and intuition of what major risk factors might be, such as age, family history of breast and ovarian cancer, hormonal and reproductive factors, and breast density as reported by MIT

However, most of these markers are only weakly correlated with breast cancer. As a result, such models still aren’t very accurate at the individual level, and many organizations continue to feel risk-based screening programs are not possible, given those limitations.

Rather than manually identifying the patterns in a mammogram that drive future cancer, the MIT/MGH team trained a deep-learning model to deduce the patterns directly from the data. Using information from more than 90,000 mammograms, the model detected patterns too subtle for the human eye to detect.

Each AI technique has its own strengths and weaknesses. Neural networks are mainly concerned with learning, fuzzy logic with imprecision and evolutionary computation with search and optimization. The advantages of these technologies can be combined together to produce hybrid intelligent systems which can work in a complementary manner. Their action permits a hybrid system to accommodate sensibly, extract knowledge from raw data, use human-like reasoning mechanisms, deal with uncertainty and imprecision, and learn to adapt to a rapidly changing and unknown environments.

There are many different hybrid systems available and the popular ones are ANNs for designing fuzzy systems, fuzzy systems for designing ANNs, and Genetic Algorithms for automatically training and generating neural network architectures. Once again, the applying of hybrid intelligent systems has been explored in several various clinical scenarios. Breast cancer is one of these applications of AI.

There is compelling proof that medical AI will play an important role in aiding the practician to deliver health care expeditiously within the twenty-first century. There is very little doubt that these techniques can serve to reinforce and complement the‘medical intelligence’ of the longer term practician.

However, It is widely acknowledged that neural network modeling requires large amounts of data, moreover, they have not fulfilled the expectation of some proponents that it would eclipse more conventional statistical techniques, moreover, several issues associated with neural network derivation demand that developers apply rigorous engineering practices in their studies. While parallel studies have identified neural network methods among the most prevalent non-traditional methodologies for data analysis, in realizing their potential application needs to overcome obstacles including the need for expanded databases and the need to establish multidisciplinary teams and lack of appropriate gold standards.