Skip to main content

Artificial intelligence, search technologies, and cancer care

by Futures Centre, Oct 2
6 minutes read

Artificial Intelligence (AI) refers to dynamic technology that can learn on its own apart from prefabricated algorithms or patterns of user behavior. This means that a machine powered by AI should be able to improve its efficiency and expand its abilities over time as it collects new information and “learns” new skills. While people might be more familiar with “pseudo-intelligent” digital assistants like Apple’s Siri and Amazon’s Alexa, AI is beginning to initiate change in various industries. One area that could reap huge benefits from AI is the healthcare industry.

One of the immediate assumptions about how AI could aid healthcare tends to be in reference to electronic health records (EHRs). Record-keeping and documentation can take up massive amounts of time in a medical office. Many positions in the medical field could benefit from intelligent machines that can categorize information, organize data, and communicate pertinent information to patients and providers.

However, AI is positioned to do much more than just handle EHRs, especially with the help of search technologies. A recent algorithm created by Google proved the potential of neural networks (a form of AI software that relies on data to improve) when it reached a more accurate patient prognosis than doctors working in a hospital. Experts were especially impressed with the algorithm’s ability to “sift through data previously out of reach: notes buried in PDFs or scribbled on old charts.” In short, this AI system was able to collect, sort, and synthesize a large amount of information in order to predict patient prognosis. Rather than simply filing and categorizing information, a combination of AI and search technologies could make it possible to turn aggregates of information into systems that can draw significant conclusions about patient prognosis and the care they might need.

Similarly, significant progress has been made in recent years in areas that seek to apply the powers of AI to cancer diagnosis, patient care, and even finding cures. In fact, Microsoft announced in 2016 that it aimed to develop a cure for cancer within 10 years. This plan includes the development of a computer made from DNA that could live inside human cells, identify affected cells, and allow scientists to “reprogram” those cells in order to fight the disease.

Though a definitive cure for cancer might be a ways off still, AI is currently proving its utility in the world of cancer care by helping doctors diagnose different forms of cancer. Technology that can identify cancer early on could be invaluable, especially for deadly cancers with especially poor survival rates. AI could prove to be especially useful for radiologists because AI systems could expedite the process for reviewing scans produced by medical imaging systems, making it possible to treat a patient more quickly.

AI Systems Diagnosing Cancer

For particularly aggressive cancers, an early diagnosis is crucial to improving prognosis. Though AI systems have already been used by radiologists to help doctors diagnose things like broken wrists and strokes, professors and researchers at Johns Hopkins Medicine in Baltimore are embarking on a journey to train computers to diagnose pancreatic cancer.

Pancreatic cancer is treatable, but the disease is often deadly because patients usually receive a diagnosis when it is too late to surgically remove malignant tumors. The team at Johns Hopkins hopes to train computers to recognize pancreatic cancer early by working to develop a tumor-detecting algorithm that could be built into CT scanner software. Ideally, this technology would make it possible to diagnose pancreatic cancer before patients have any symptoms, which could greatly improve prognosis and broaden options for  treatment.

Another form of aggressive cancer that presents significant challenges when it comes to reaching an early diagnosis is mesothelioma. Mesothelioma is a cancer that develops in the lining of the lungs, heart, and abdomen. With only 3,000 cases diagnosed each year, many people are unaware of this cancer and its risk factors. Along with this, the cancer is often misdiagnosed early on because many of the symptoms are mistaken for signs related to more common illnesses, including asthma and COPD. These combined factors have resulted in mesothelioma having an especially poor prognosis and a very short life expectancy for patients.

AI has the power to help reach a mesothelioma diagnosis more quickly, which could lead to vast improvements in survival rates and expand curative and palliative care options. In addition to identifying malignant mesothelioma tumors with the help of medical imaging, AI could also compare symptoms to aggregates of patient data or flag relevant factors in a patient’s medical history. This latter point could prove to be extremely important, as mesothelioma symptoms can take anywhere from 10 to 50 years to appear. Therefore, identifying a pattern of symptoms or relevant risk factors in a person’s history (such a clear link to asbestos exposure) could be a crucial step in determining the right care. Search technologies could potentially improve the implementation of this technology through taking in large amounts of data, organizing it, and flagging relevant information.

Improving Care with AI and Clinical Trials

For many forms of cancer, clinical trials are essential for advancing treatments and finding cures. In addition to finding ways to detect and prevent cancer, clinical trials can improve quality of life for patients during and after treatment, and AI is assisting in these capacities as well.

For example, researchers at MIT are relying on unique machine learning to improve the quality of life for patients diagnosed with glioblastoma, the most common and aggressive form of brain cancer. The goal of this research is to reduce toxic chemotherapy and radiotherapy dosing. Using current treatment regimens, an AI system uses “self-learning” machine learning techniques to adjust doses in the hopes of finding an “optimal treatment plan with the lowest possible potency and frequency of doses that should still reduce tumor sizes to a degree comparable to that of traditional regimens.”

While the goal in the trial is reducing tumor sizes, this research simultaneously places a conscious focus on improving quality of life and attempting to lessen illness and harmful side effects. Most notably, the researchers were also able to design the model to treat patients individually, which is not something typically done during traditional clinical trials. potentially contributing to decreased efficiency and effectiveness with certain tactics.

The introduction of AI is exciting for the future of the healthcare industry. For cancer care specifically, a combination of AI systems and search technologies could unlock an array of possibilities. Even though the true potentials of these developments are still largely unknown, as many of them are in research or trial stages, it is clear they hold the potential to improve everything from administrative operations to patient care and cures.

Resource links:


by Futures Centre Spotted 1994 signals

Have you spotted a signal of change?

Register to receive the latest from the Futures Centre.
Sign up

  • 0
  • Share

Join discussion

Related signals

Our use of cookies

We use necessary cookies to make our site work. We'd also like to set optional analytics cookies to help us improve it. We won't set optional cookies unless you enable them. Using this tool will set a cookie on your device to remember your preferences.

For more detailed information about the cookies we use, see our Cookies page.

Necessary cookies

Necessary cookies enable core functionality such as security, network management, and accessibility. You may disable these by changing your browser settings, but this may affect how the website functions.

Analytics cookies

We'd like to set Google Analytics cookies to help us to improve our website by collecting and reporting information on how you use it. The cookies collect information in a way that does not directly identify anyone. For more information on how these cookies work, please see our 'Cookies page'.