That is the instrument Dr. McCoy wielded on “Star Trek.” It has become a generic name for a non-contacting multifunctional medical scanner. That may sound like science fiction, but the X Prize Foundation believes a real tricorder is close enough to reality that it has established a $10 million bounty for the first team to build one.
The remote sensing technology used in tricorders could empower a new generation of medical robots. They could be used to assess and treat people who are injured in natural disasters and battles. They could monitor both bedridden and mobile patients in hospitals, nursing homes, and independent living facilities. They could also monitor crowded airports, train stations, and bus depots for people with signs of communicable diseases.
Scanadu is five to 10 years away from a true non-contacting tricorder, chief operating officer Misha David Chellam said. It plans to have an initial prototype by the end of 2012. It will combine contacting and non-contacting sensors with artificial intelligence software that can assess the data and make recommendations.
The company has lined up $2 million in venture financing. It is led by CEO Walter De Brouwer, who co-founded Starlab, a blue-skies research center with MIT Media Lab founder Nicholas Negroponte. Chief technology officer Ivo Clarysse also came out of Starlab and worked for several Fortune 100 companies.
They have a difficult road ahead. Cramming sophisticated sensors into a durable hand-held device that is economical enough for family use is a challenge. It may preclude using some technologies. Robots, on the other hand, are spacious enough to use a wider range of technologies.
Tricorders
Scanadu?s ideal tricorder would be non-invasive (no needles), non-sampling (no urine or saliva), non-contacting (sensing at a distance), and work without any cooperation on the part of the patient. Its first product will not meet those criteria.
Instead, it will house three contacting elements in a wearable sensor patch. A 3D accelerometer will measure respiration, blood pressure, and heart rate. A thermistor will take temperatures. Microneedles in the patch will draw minute amounts of blood for analysis in a microfluidics lab-on-a-chip. The non-contacting sensor will be a hyperspectral camera that provides spectral data to analyze rashes and throat and ear infections. All three are well-researched technologies, though shrinking them to tricorder size is a challenge.
The tricorder will send data to a smartphone app over Bluetooth. The app will send it to the cloud, where an AI server compares it with previous data and walks the user through any additional steps needed to diagnose the problem.
Several companies have already developed self-diagnosing applications (WebMD Symptom Checker is a popular one). Predictive medical AI programs are just beginning to reach hospitals and consumers. Predictive Medical Technologies, for example, has developed software that can predict a heart attack 24 hours before it occurs. Scanadu will team with partners to build on such software.
Scanadu also plans to compare sensor data with electronic health records. Today, access is strictly limited. As physicians increasingly digitize health data, it grows more likely that patients will have the option of sharing that information.
The combination of cloud-based predictive AI and digital record sharing are preconditions for both tricorders and medical diagnostic robots.
Chellam believes insurance companies will underwrite the device. He noted a study by Wellness Council of America that found 70 percent of doctor’s visits were unnecessary. Another by Excellus discovered that 40 percent of New York emergency room visits were not needed. Insurance companies could save billions of dollars by advising customers when they do not need to seek medical care.
Robots would have different economics. They might serve larger populations and could perform other functions (like ensuring patients take medication), which would spread the cost of expensive diagnostic sensors over more users.
Non-Contact Technologies
Chellam declined to comment on Scanadu’s roadmap, but there are several non-contacting technologies he could tap.
Visualization. Physicians rely on their eyes for diagnosis. Hyperspectral imaging systems go further. They capture hundreds of different spectra for each image. This enables them to evaluate spectroscopically each pixel’s reflectance, emittance, and fluorescence.
NASA developed Hyperspectral imaging for earth observation satellites. In 2008, Themis Vision Systems licensed the technology and developed a compact, 4 lb hyperspectral camera that would be ideal for robots.
The technology could image wounds and burns to assess severity, infection, and healing; scan skin to detect disease; and monitor the eyes for diabetic retinopathy, macular degeneration, and other eye diseases.
More intriguing, researchers have used hyperspectral imaging to measure oxygen saturation in the retina. This could detect diabetes before structural changes occur, according to Dr. Amani Fawzi of University of Southern California, Los Angeles. Similar measurements could characterize skin health. Researchers at George Washington University believe hyperspectral imaging could detect increased heat generated by breast cancer.
Thermography. Thermograms are infrared images that show the heat distribution of an object. They were first used in medical diagnostics in 1957, when Canadian surgeon Ray Lawson found that tumors in women with breast cancer were warmer than the surrounding area.
Over the past decade, as equipment improved, thermography became a mainstream tool for detecting breast cancer, especially in Europe. Cancer has higher metabolic rates than normal tissue and also promotes growth of blood vessels that also heat the body, so it is readily detectable in infrared photos.
Because thermography can visualize small blood vessels through the thin skin of the head, and also the large arteries in the neck, and could provide advance warning of stroke. It has been used by chiropractors and orthopedists to assess spinal injuries and strained and torn muscles. It detects the presence of deep vein thrombosis that could lead to loss of limbs and possibly stroke.
Another application is monitoring respiratory dysfunction, including asthma, allergies, bronchitis, and the influenza. During the SARS epidemic, it was used in airports to scan for passengers with atypical pneumonia. It has helped diagnose appendicitis, irritable bowel syndrome, and colitis, and give advanced warning of urinary disease. It has also been applied to skin wounds and tumors, and ear, nose, and throat problems.
Lasers. Surgeons use lasers as scalpels, but they have a promising future in diagnosis too. Princeton University’s Mid-Infrared Technologies for Heath and Environment (MIRTHE) research center recently showed a laser device that detects blood sugar levels by scanning the skin. The same approach might enable lasers to sample other chemicals in the bloodstream.
MIRTHE is also testing MIR lasers to detect nitric acid (to understand asthma) and ammonia (for kidney data). More frequent monitoring of these gases could allow dynamic adjustment of steroids and better feedback on diet for managing diabetes.
Researchers at Holy Cross University and the Naval Research Laboratory have used a laser Doppler vibrometer to detect the velocity of blood pulsing through the body’s surface arteries. This enables the system to measure heart rate and blood pressure.
Clearly, Scanadu and other Tricorder X Prize competitors have a lot of work ahead. Yet they can take advantage of the growing IT infrastructure needed to apply AI to medical diagnostics, as well as the increasingly rapid development of non-contact sensors.
Fitting those technologies into a small, economical, hand-held device remains a challenge. Building it into a multifunctional mobile robot that serves many people could prove easier technologically and economically.
Those robots may not have the bedside manner of the tricorder’s original owner, Dr. McCoy, but hey, didn’t “Star Trek: The Next Generation” feature an android?
