Artificial Intelligence in Oncology

by Puja Leekha & Rahul Singh

Why is this important to us?

This article discusses the value of using artificial intelligence in healthcare, particularly oncology. Given a large number of patients diagnosed with cancer and a significant amount of data being generated during cancer treatment, there is a specific interest in the application of artificial intelligence to improve oncology care.  

Artificial intelligence can help in better diagnosis of patients by relating the existing case with earlier detected cancer cases of similar nature. Medical data available over past decades and from around the globe can help in finding the best practices to cure this disease. This, in turn, leads to faster diagnosis of patients. 

In this article, we introduce the fundamentals of AI, provide an overview of how it is being currently used in oncology and some of its challenges.

A bit about Cancer

As per WHO, cancer is a generic term for a large group of diseases characterized by the growth of abnormal cells beyond their usual boundaries that can then invade adjoining parts of the body and/or spread to other organs. Other common terms used are malignant tumors and neoplasms. Cancer can affect almost any part of the body and has many anatomic and molecular subtypes that each require specific management strategies.

Cancer is the second leading cause of death globally and is estimated to account for 9.6 million deaths in 2018. Lung, prostate, colorectal, stomach and liver cancer are the most common types of cancer in men, while breast, colorectal, lung, cervix and thyroid cancer are the most common among women.

Cancer Statistics, WHO (2018)

  • Number of deaths – almost – 10 million
  • Number of cases which can be prevented – 30-50%

As per WHO, Cancer Country Profile (2014), there were an estimated 1 million deaths due to cancer-related diseases. Each year, there are new cases, which are being added in this list.

Why is data important in healthcare?

Diagnosing a patient’s condition is one of the most important and challenging tasks in medicine. A humongous amount of data is being generated by various medical tests, scans, etc. done for patients around the world. Historically, this data was stored locally and was not collated & categorized. Thus, this data is often siloed with individuals or within individual institutions.  The cases which were successfully treated were not known to doctor fraternity beyond the doctor treating the patients.

However, healthcare is becoming increasingly connected, at the same time increasingly complex. The industry is trying to address the challenge between encouraging data sharing and maintaining patient’s privacy & trust. For diseases such as cancer, it is of utmost importance to have access to medical data, which can greatly help in the proper diagnosis of the patients.

How Artificial Intelligence is set to change things!

In simple terms, Artificial intelligence (AI) is the simulation of human intelligence processes by machines, especially computer systems. These processes include learning (the acquisition of information and rules for using the information), reasoning (using rules to reach approximate or definite conclusions) and self-correction.

Artificial intelligence is being used in many different fields, specifically more so in healthcare. With the advancements in processing power, intelligent storage and faster connectivity; accessing the data from any part of the globe is easily possible. Programs are now able to interpret different images, compare them and give meaningful information.

AI in medical treatment can process the patient’s history along with current information and suggest doctors about the processes which can be used for treatment. Each doctor in different regions of the world has a specific style of recording patient information. Combining individual styles of writing patient information in different languages across the globe and getting meaningful information which is useful to others can be done using algorithms in AI.

AI algorithms can read, process and make informed decisions based on parameters selected by doctors. This involves the processing of such large scale data & images and giving out relevant data in an output which can be then used by local doctors. The problem of understanding different styles of writing, languages, and images can be programmed in AI. Since AI is also self-learning and evolving, corrections can be applied and then used to display correct information as per the requirements of the doctor. AI can assimilate all research information papers and clinical trial results and reduce physician’s time in analyzing the medical cases.

Al can help doctors in determining, targeting and correlating therapies & drugs for individual patient’s needs. For AI to achieve this, we need more real-world data and consistent information for it to analyze and give outputs in the required format.

Watson (from IBM) for oncology: is a solution which channels information from relevant guidelines, best practices, medical journals, and books. The solution assesses information from patients’ medical records, evaluates medical evidence, and displays potential treatment options ranked by level of confidence; always providing supporting documents. The oncologist can then apply their own expertise to identify the most appropriate treatment options. 

AI applications like Watson can also reduce oncologist’s time in doing the analysis. One can expect better results from analysis when applications have a large amount of data available to compare and correspond to.

In an initial pre-screening test, the Highlands Oncology Group [HOG] from Northeast Arkansas, clinical trial coordinator took 1 hour and 50 minutes to process 90 patients against three breast cancer screenings. Conversely, when Watson’s clinical trial matching platform was used, that job took 24 minutes. “This represents a significant reduction in time of 86 minutes or 78%,” HOG said in a statement.

But there are a few challenges …

1.Interpretability

It is also referred to as explain-ability. This is a big challenge. It refers to the ability to be able to understand the results or inferences that come out of the AI models. For example, looking at the patient’s data, past history, an AI model could predict that a patient can develop skin cancer in 3 years’ time. However, it is challenging to explain how the AI model arrived at that outcome.

2. Skills gap

To successfully merge AI into oncology we need to address the knowledge gaps. Currently, oncologists are given little or no training on AI/DL kind of technologies, limiting their ability to understand & adopt AI algorithms. Similarly, data scientists have limited knowledge in oncology which limits their ability to understand all use-cases.

3. Data sharing & privacy

One Vital consideration and a real challenge for many oncology doctors around the world is the collation and collaboration of relevant information and medical practice around the globe. Due to various local rules and regulations, one method of data collation may be valid in a country but the same thing may be illegal in another country and different regulatory framework.

It is suggested that a global framework under the aegis of WHO can be framed which contains all the treatment and scans of all cancer patients. These scans and the treatment outcome can be used to establish data and rules for teaching AI. This vast network of information can be then used with artificial intelligence helping doctors to diagnose the patients.

“There has to be a human-agency-first kind of principle that lets people feel empowered about how to make decisions and how to use AI systems to support their decision-making,” – Soumitra Dutta, a professor from Cornell SC Johnson College of Business, states a vital aspect of data collection and collaboration for making this kind of diagnosis a success.

Take home messages

1.Early Detection: Accurate diagnosis of cancer can be life-saving for patients particularly in cases where symptoms are detected at a later stage. AI helping in early detection is another compelling use-case. 

2. Transforming healthcare: Over the past decade, AI has evolved a lot and is of tremendous help in multiple healthcare use-cases.  Due to an explosion of electronic data, advances in technological infrastructure, and groundbreaking research in DL neural networks, AI is poised to make life-changing impacts on the medical field and oncologic care.

3. In Oncology: At present, AI has shown promise in improving cancer imaging diagnostics and treatment response evaluation, predicting clinical outcomes, and catalyzing drug development and translational oncology.

4. Challenges: However, challenges still remain—such as validation and proving generalizability, concerns over interpretation, and the widening knowledge gap between clinical and data science experts.

5. Transforming the future: As researchers, data scientists help address these challenges, AI has the potential to transform oncology, harnessing the power of big data to drive cancer care into the 21st century and beyond.


For more details on the topic please contact us here.

Puja Leekha

As the daughter of a defence officer, she has had the tremendous good fortune of traveling & living across the length & breadth of India. The multi-dimensional Indian culture has enriched and shaped her personality during her growing-up years. An excellent student, her love for computers started early-on in life and she followed her passion to its logical conclusion. An accomplished IT engineer at IBM, she has diverse experience in various leadership roles at IBM. A family person at heart, over the weekend you will find her cooking delicious cuisine for her two sons.
Linkedin Profile – https://www.linkedin.com/in/pujaleekha/

Rahul Singh

A keen learner, an observer of life, he uses his fine intellect and unique vision in various dimensions of life that interest him. His skill and expertise in nature photography are appreciated by an ever-increasing fan base. His dedication to his profession as a mechanical engineer has gained him professional acclaim from business leaders both in India and across the globe. If he is not busy capturing the Himalayas on his camera or building intricate designs for his next engineering project, you will find him enjoying the company of his family or close friends.


References:

WHO: https://www.who.int/cancer/en/
IBM: https://www.ibm.com/in-en/marketplace/clinical-decision-support-oncology
Wharton at UPenn: https://knowledge.wharton.upenn.edu/article/people-first-artificial-intelligence-strategy/

Watson, IBM, and other names are trademark brands of respective companies.


Attribution for images: Pixabay. com

5 thoughts on “Artificial Intelligence in Oncology”

  1. great job Puja and Rahul. IBM Research and IBM POWER systems have been in the forefront of this domain. “Due to an explosion of electronic data, advances in technological infrastructure, and groundbreaking research in DL neural networks, AI is poised to make life-changing impacts on the medical field and oncologic care!”

  2. I beⅼieve everything ѕaid made a lot of sense. But, what
    about this? what if you compoѕed a catchier post title?

    I ain’t sayіng your content isn’t good., however what if you adⅾеd somethіng
    tһat makes people want more? I mean Artificial Intelligence in Oncology –
    WRITE & DESIGⲚ is a little vanilla. You should glance
    at Yahoo’s home page and see how they write newѕ headlines to get viewers to opеn the links.
    You might add a vidеo or a picture or two to get readers
    excited about еverything’ve got to say. In my opinion, it would make your posts a little livelier.

Leave a Reply

Your email address will not be published. Required fields are marked *