Q & A: XPRIZE Tricorder

During the XPRIZE Tricorder Competition, we were asked by various press outlets to answer questions regarding our experiences and challenges. I wrote up our responses – and for posterity sake – am placing it here so people can read our answers. So here it is:

_____________________________________________________________________________________

Q1. Can you tell us a little about your Tricorder version. What makes it unique and what advantages do you feel your technology offer?

A1: The Tricorder unit designed by our team consists of 3 main components: (1) the scope set, (2) blood/urine/breath tests, and (3) vital signs monitoring set. The scope set utilizes a Bluetooth enabled magnifying camera to obtain high-resolution images of the skin and tympanic (ear) membrane; the blood/urine/breath tests are, as expected, employed to analyze fluids or breath dynamics to diagnose conditions such as urinary tract infection, diabetes, and COPD; and the vital signs monitoring set enables both calibration and acquisition of continuous vital sign parameters (temperature, heart rate, respiration, blood pressure, and oxygen saturation).

We believe that the unique and advantageous features of our Tricorder unit can be attributable to several factors. First, its user-friendliness: thanks to the experience and expertise of HTC, remarkably extensive time and thought was placed on the user experience – from form factor to presentation to video instructions, large amount of effort was made to ensure that the user had a seamless and effortless experience with our portable device after numerous test cases and iterations.

Second, its analytical methodology: we employed state-of-the-art mathematical algorithms to analyze physiological signals and incorporated machine-based learning techniques for analyzing acquired images. Cloud computing was also used to perform some of the computationally heavy steps. Fortunately, due to our past and ongoing research in biological signal analyses, we felt well-prepared to incorporate novel, and importantly validated, methodology into our Tricorder device.

And third, its AI Diagnostic Algorithm: we created a diagnostic algorithm that simulated the strategic approach used by clinicians to arrive at a specific diagnosis/diagnoses. This algorithm helped determine what question should be asked next and how to calculate risks for each condition at each interval step. This algorithm notably integrated all the information made available – including age, gender, BMI, vital signs, reported symptoms, and biological test results.

_____________________________________________________________________________________

Q2. What were the biggest challenges when trying to develop this technology? How difficult was it to model your technology after something that is essentially science fiction?

A2: Probably, the biggest challenge in developing this technology was the restrictive criteria set by the Qualcomm XPRIZE competition. The whole Tricorder set could not weigh more than 5 pounds and yet the unit had to be reliable and user-friendly. In fact, we had to eliminate some heavier and more sophisticated components during the process of the competition as a result. However, the limitations were ironically liberating as well. We had to think outside of the box, and having to operate within the confines of the competition rules forced us to think in more unconventional ways to establish a diagnosis.

Using COPD (chronic obstructive pulmonary disease) as an example, we thought of assessing exercise tolerance (change of heart rate or oxygenation with activity), ratio of inspiratory time to expiratory time, analyses of chest wall movement dynamics, acoustic characteristics of the chest wall, electrical impedance of the chest, and so on. Identifying which of these approaches should be pursued without adequate clinical data and potentially at the expense of other tests (given weight restrictions) was also one of the big challenges of this competition.

The most difficult thing about emulating a science fiction, imaginary concept is knowing where to draw the line between what is ambitious yet potentially achievable and what is clearly beyond the capabilities of present technologies. An added factor to this challenge is not fully knowing what our modern capabilities really are. In this competition, so many disciplines are necessarily invoked – from device/hardware technologies, physiology, clinical medicine, software/app programming, biophysics, material science, user psychology, etc. – and with each added discipline, it becomes truly difficult to keep abreast of where the cutting edge lies in each discipline and importantly where they can intersect.

A novel electromagnetic detector is great but becomes pointless if there is no physiological basis for such measurements. Likewise, a novel, promising physiological test would be impossible to implement if it requires invasive and maybe biologically incompatible material. And then one begins to ask, “Maybe there is physiological basis for electromagnetic measurements?” or “Perhaps there is a biologically compatible material out there that we are completely unaware of?”. It was interesting to see how many times promising ideas were shot down simply because it was not feasible in one or more specific disciplines or because we lacked the adequate expertise in relevant area of interest. The science fiction, nevertheless, provides the critical vision and lofty goals that we frequently neglect in our daily mundaneness of life and work.

_____________________________________________________________________________

Q3. How long of a process has it been from concept to development? What steps remain to help bring this device to market?

A3: The process took approximately 3 years from concept to generation. Prior to this competition, we had been doing preliminary work in the portable diagnosis of conditions such as atrial fibrillation and sleep apnea, but the idea of assembling a device capable of diagnosing up to 15 medical conditions – and diverse ones at that – did not get initiated until the Qualcomm XPRIZE Tricorder Competition announcement in 2012. This competition pushed us to think broadly and ambitiously, and is a real tribute to what XPRIZE and Qualcomm are doing to push the boundaries of science.

In regards to bringing this device to market, a lot of work remains to be done in optimizing this device for clinical use. Certainly, this device needs to be validated and proven reliable in clinical studies. But equally important is the utility and usefulness of this device from the perspective of both user and health care provider. From our perspective, there is little point in introducing this product to the market if it merely becomes a burden to the daily lives of patient or provider. Our goal is to put out a product that meaningfully improves people’s health and enhances one’s quality of life.

__________________________________________________________________________________

Q4. In what ways do you think this device could have the biggest impact?

A4:  In the near future, we foresee this technology having its biggest impact in rural, remote areas where access to medical care is limited. We look forward to working with governments and NGO’s to make this a possibility.

In the long-term, this technology will likely be most transformative by introducing continuous, long-term data into the healthcare system. Our medical system is largely predicated on one-time, episodic tests, and for the first time, we are entering an era where continuous, highly granular, and multimodal data are increasingly available. If our belief that dynamics of physiological signals contain useful, often unintuitive information turns out to be correct, then the medical system may be transformed by having new understanding and approaches to both health and disease.

In clinical medicine, it is not infrequent to have multidisciplinary rounds. For instance, for pancreatic cancer, you have the gastroenterologist, surgical oncologist, radiation oncologist, oncologist, pathologist, and radiologist who spend time discussing about the appropriate treatment course for a specific patient. With the availability of rich, continuous data for each patient, can you now imagine how a computational scientist or mathematical physiologists would be a needed team member to help analyze acquired time series and determine how a certain medication or lifestyle intervention affected one’s day-to-day body function?

Then consider the additional availability of diet, exercise, sleep, and even emotional data – you would have the need for additional expertise (e.g., nutritionist, sleep specialist, psychologist, etc.), the data infrastructure to securely store and transmit data, tools for data visualization and analyses, and even a systems-specialist capable of integrating the information in a unifying manner. In the process, medicine would become increasingly personalized and precise. There would be a transformative effect not only in how medicine is practiced but also in how the whole medical infrastructure (from data management to specialized personnel to coordination of information) is organized and managed. Such is the future these technologies are potentially taking us.

____________________________________________________________________________________________

Q5. Finally, are there any plans to further enhance this technology, or evolve its capabilities to better serve patients?

A5: We plan to develop a 2nd generation Tricorder device system and hope to apply this technology to remote areas where medical access is limited. In the immediate future, we hope to apply this to rural areas of China and then to extend the implementation to other developing countries/areas such as India, South America, and Africa. As we obtain more data, we anticipate that the AI algorithm and analytical methodologies will be improved further.