Machines Eradicating Cancer
At the SxSw Interactive Conference in Austin this week, Former Vice President Joe Biden challenged all innovators to think BIGGER. Biden’s Cancer Moonshot Task Force , established last January, brings together 20 government agencies and more than 70 private-sector companies with one aim — “eliminate cancer as we know it.” Even the most unorthodox medical collaborators are bringing their domain expertise to the table. Amazon, for example, is offering free cloud hosting for researchers globally whom are now accessing valuable data over 80 million times.
To eradicate one of the most brutal killers on the planet requires more than just American ingenuity; it means working across disciplines and borders with novel solutions battling tumors at the microscopic level. The latest weapon in the arsenal could be a swarm of microrobots that could invasively attack cancer cells. Researchers last month at Phillips in Hamburg, Germany showcased their latest discovery of controlling single and multiple microrobots with external magnetics. Imagine injecting microrobots within the body and then controlling their movements through external magnetic fields to perform targeted functions within organs. Such micromachines could be used to deliver a therapeutic agent to a specific location in a minimally invasive manner or control adaptive implants. In addition, a “swarm” of micromachines could effectively deliver targeted chemotherapeutic drugs into the complex architecture of a solid tumor.
Precisely controlling a swarm of bots in the body is a serious challenge. Controlling each micromachine in a swarm individually is important to perform complex tasks in the body. However, applying a uniform magnetic stimulus to the whole swarm and hoping for the best does not facilitate individual control of each microrobot. Making individual robots perform specific tasks is not possible using a one-size-fits-all stimulus. The challenge lies in limiting a magnetic field to a highly specific area to control only the micromachines within that area. The new Phillips research uses a novel approach by generating magnetic fields through a focal point containing a very low magnetic field gradient. The result is that every microrobot outside the focal point can be “locked” in place, while the “unlocked” microrobot in the focal point can be controlled through another magnetic field to perform specific tasks. In addition, the focal point can be moved, allowing for machines to be locked and unlocked at will.
“Our method may enable complex manipulations inside the human body,” says lead researcher Jürgen Rahmer. He imagines that a robotic swarm could eventually deliver cancer-killing radioactive “seeds” specifically inside tumors in the body, treating cancer with pinpoint accuracy and sparing healthy tissue and side effects.
In addition to microrobots, researchers are also using magnetic fields with nanorobots for targeted treatments. Nanomedicine is one of the fastest growing areas of the cross-section of biology and robotics (see previous post). Today, scientists can package powerful chemotherapy drugs on nanobots 200 times smaller than a red blood cell. Typically nanobots are made of DNA particles that trick the body to thinking they are natural entities. Today, nanotechnology is deploying new control methods that act more like autonomous cars than biologic compounds.
Researchers in Canada from McGill University, Université de Montréal and Polytechnique Montréal demonstrated last August how drugs can be delivered into tumors using bacteria guided by computer-aided magnetic fields. Unlike microrobots, the bacteria can be self-directed into the most affected cells by detecting areas with depleted levels of oxygen. Novel drug delivery systems are running the gamut from magnetic field to artificial intelligence control systems.
Earlier this year, scientists from Israel’s Bar-Ilan University began human trials using another type of drug-delivering nanobots to treat cancer. Professor Gal Kaminka of Bar-Ilan’s Robotics and Artificial Intelligence Lab explained his research is taking more of a software approach than typical medical practices. “We realized that the way the people were building the nanorobots slowed their progress. They develop a specific robot for every drug and disease. We wanted the doctor, who is an expert in medicine, not to need to understand nanorobotics and not to deal with the technical aspects of how the robot moves from point A to point B. When I write software in Java or any other computer language, I don’t need to think about which telephone or microprocessor it will run on. In most cases, I write the code and put it into software called a ‘compiler’ that adapts it for me to a specific machine,” explains Kaminka.
Through his “compiler-approach” Kaminka translates medications into nanobot swarm behavior to deliver triggered-payloads directly to the source of the tumor. It can even deliver multiple medicines in a cycling of nanobots that are correctly sequenced. In a sense, he is building an autonomous medical transportation network that will be capable of leveraging the highways of the blood stream to respond to infection and disease.
While nanobot platforms are still in development, questions remain whether computers could more accurately detect cancer than trained physicians. Google, a leader in algorithms, recently tested this premise on breast cancer images with its Camelyon16 Project. The goal of of the project is “to evaluate new and existing algorithms for automated detection of metastases in hematoxylin and eosin stained whole-slide images of lymph node sections.” The project aims to reduce the workload of pathologists in diagnosing cancers more efficiently.
According to the Google blog, “the prediction heatmaps (shown above) produced by the algorithm had improved so much that the localization score (FROC) for the algorithm reached 89%, which significantly exceeded the score of 73% for a pathologist with no time constraint. We were not the only ones to see promising results, as other groups were getting scores as high as 81% with the same dataset. Even more exciting for us was that our model generalized very well, even to images that were acquired from a different hospital using different scanners.”
Google’s aim is to augment the experienced pathologist versus replace them with computers. According to the blog, “we envision that algorithms such as ours could improve the efficiency and consistency of pathologists. For example, pathologists could reduce their false negative rates (percentage of undetected tumors).”
These bleeding edge initiatives would not be possible without the support and funding by the National Institute of Health (NIH). We are on the cusp of ending the plague of cancer in our lifetime, unfortunately the new 2018 Trump budget is planning an 20% reduction in NIH funds. Now more than every the health of our society depends on the private innovation partnership between medicine and mechatronics.