Artificial Intelligence in Cardiology: Opportunities for Cardio-Oncology

History was made recently with the inaugural and first ever continuing medical education conference on artificial intelligence (#AI) in Cardiology. While most of the presentations were on artificial intelligence or cardiology or both, several sessions also made reference to other fields in which AI has been or is being used, such as Oncology. There was even one study presented on Cardio-Oncology. As study after study was presented, it became clear to me that perhaps several of these techniques and methodologies could potentially be useful to our patients in Cardio-Oncology.

Every single piece of technology started with one single prototype. Every single new piece of software started with one single algorithm. Every single patent started with one single idea. Every single idea started with the impact that disruptive technology could have for at least one single patient – one single case.

As I view various case reports in Cardio-Oncology, I think about how #AI could influence care delivery to potentially improve outcomes and the experience for each patient and their health professionals.

One example that was reiterated in multiple presentations was that of the ECG. Applying #AI to the ECG has been shown in the studies presented to determine the age, sex, and heart condition of the individual. Details were shown for a case of hypertrophic cardiomyopathy (yes, HCM, not just left ventricular hypertrophy) diagnosed via #AI analysis of an ECG that appeared relatively unremarkable to physicians’ eyes. After the septal surgery/procedure, although the ECG then looked remarkably abnormal to physicians’ eyes, the #AI algorithm could identify resolution of the hypertrophic cardiomyopathy.

Another example reiterated throughout the conference was identifying undiagnosed left ventricular systolic dysfunction, in a general community population and also in patients referred to a cardio-oncology practice at a large referral center.

Recently, #AI in Cardiology has been used most frequently for monitoring and detection of arrhythmias, such as atrial fibrillation. Everyone can purchase their own wearable to determine this. Physicians are also now prescribing these wearables for ease-of-use, given their pervasive presence and coupling with smartphones owned by much of the population or provided temporarily by the physician group. Such wearables are transitioning from standalone electrodes, to watches, skin patches, and clothing (e.g., shirts, shorts).

Many direct-to-consumer #AI applications in daily life actually are not wearable, such as Alexa and Siri. One study described the ability of #AI to help diagnose mood disorders and cardiac conditions and risk factors by simply “listening to” and analyzing voice patterns. The timing of a young man’s “voice breaking” can potentially predict his risk for heart disease!

A popular use for #AI in medicine overall is to assist with interpretation of various imaging, such as chest X-rays, MRIs, or CT scans. This applies in Cardiology as well. Further, in Cardiology, #AI is being used to help guide the procurement of echocardiograms. The algorithms provide visual instructions (such as curved arrows) to indicate directions in which the ultrasound probe should be moved to obtain the standard view, to which the algorithm is comparing the image being procured moment-by-moment. The idea is for #AI to help less experienced sonographers or echocardiographers learn and perform echocardiography even more expediently.

The theme of the conference was current advances and future applications of #AI in Cardiology. Accordingly, a historical perspective was given, describing some of the earliest attempts at #AI in various fields. A video of a possible precursor to current automated vacuum cleaners was shown, from archives dating back to the 1960s. In addition to ways in which #AI is now being studied or applied, future opportunities for using #AI were also postulated, for example for coronary artery disease, since stress tests are not 100% sensitive and the gold standard coronary angiography is invasive. #AI could help stratify patients who needed versus did not need the invasive procedure for recurrent convincing symptoms in the absence of a positive stress test. Of course, coronary CT angiography could help fill this gap, but #AI might assist with decision-making sooner.

There have been studies on #AI in Cardiology, and studies on #AI in Oncology, and at least one study in #AI in Cardio-Oncology – a study I predicted; one that is quite intuitive and mentioned above. I propose that we continue to apply #AI in Cardio-Oncology, so that the field can catch up with the rest of Cardiology and Oncology, and help us continue to develop this emergent and burgeoning multidisciplinary subspecialty.

This is an exciting time for me to be alive. I am an early adopter of artificial intelligence. I look forward to seeing more and more the availability of #AI to enhance our use of electrocardiography, echocardiography, wearables, biosensors, voice analysis, and more in Cardiology, and particularly in Cardio-Oncology, with an emphasis on primary and primordial prevention even before secondary and tertiary prevention in the area of Preventive Cardio-Oncology, and especially in women.





Why We Sleep, and Why Don’t We Let Hospitalized Patients Sleep?

Last month, I wrote about my newfound fascination with wearables and the physiological parameters I now measure on myself. One of these metrics is my “recovery score,” (Figure 1) which is heavily determined by the quantity, quality, and consistency (the regularity of the times at which I go to sleep each night and wake up each morning) of my sleep. Now that I wake up each morning being graded by my wearable on how well I slept, I have implemented strategies to improve my scores: wearing an eye mask, using blue light-blocking glasses at night, using my phone in bed, going to sleep at a regular time. These indeed have improved my recovery scores, and I honestly feel I sleep more efficiently as a result – I feel just as rested after 6 hours of sleep now as I previously did only with 7-8 hours of sleep.

Figure 1: Sample recovery score from my wearable device (Whoop strap).

Figure 1: Sample recovery score from my wearable device (Whoop strap).

My focus on my own sleeping habits led me to start reading the book, “Why We Sleep” by Dr. Matthew Walker. It is a fascinating review of the physiological role of sleep (or at least, the best knowledge we have of it), and I recommend it to anyone interested in the topic. While sleep has been studied intensely for decades, there still is much we do not know about how it benefits our bodies and minds. Some recent high-profile studies on sleep caught my attention though, including one published in Science that implicates the sleep-wake cycle in the regulation of the Alzheimer’s Disease-related tau protein in the brain.1,2 Another recent study suggests that when we cannot repay accrued “sleep debt” over the course of a week, at least with regards to its effects on metabolic dysregulation.3 However, as I continued to read about the importance of sleep and its health benefits, I could not help but think about some of the most sleep-deprived people we encounter regularly – our own patients.

Anyone who has worked in the inpatient hospital setting knows that admitted patients are regularly disrupted throughout the evenings in the hospital, and it is a rare evening they can get a full, restful night’s sleep. Indeed, one study showed that inpatients on average get 83 fewer minutes of sleep compared to their sleep duration at home.4 Some factors preventing sleep are related to the nature of their hospitalization and are difficult to circumvent – severe pain, hemodynamic instability, infections that produce significant discomfort (e.g., fever, cough).

Yet many factors are systematic and iatrogenic. Often, nurses receive orders to check vital signs every 4 hours, and thus are required to check in on patients in the middle of the night to measure blood pressures. Intravenous medication infusion pumps alarm seemingly every 5 minutes. Daily lab work is ordered to be performed early in the mornings, so that results will be ready by the time physician teams begin their rounds. And in teaching institutions, medical students and housestaff “pre-round” in the early hours (sometimes as early as 5:00am), waking their patients to briefly interview and examine them (Figure 2).

elderly man in hospital

Figure 2: Depiction of early morning “pre-rounding.”

While clearly many hospitalized patients are in the hospital precisely because they need close monitoring, there is a large subset of admitted patients who, despite the need to be hospitalized, could be provided an environment in which they can get a full night’s rest. Many hospital systems are attuned to these issues, and I am proud to say that mine is as well, but many of the systematic factors that interfere with sleep are unfortunately just part of the rounding culture at academic hospital centers (early labs, pre-rounding, etc.).

A recent Twitter thread from a former co-resident of mine, Dr. Dan Wheeler (@WheelerMed), made me reflect on the purpose of morning rounds. Dr. Wheeler’s thread highlighted the heterogeneity in how different teams rounded and the challenges this may pose to trainees. Yet I could not help but think that the major problem with rounds is that, in my opinion at least, it is not patient-centered. In fact, given that many features of rounds interfere with an inpatient’s ability to obtain adequate sleep, rounds actually impede patient recovery.

Isn’t patient recovery our ultimate goal? When I spend so much time trying to improve my own “recovery score,” shouldn’t I also focus energy to improve the recovery scores of those who need it much more than I do?

As trainees, we arguably do not have much control of the rounding schedule and culture. But, my goal in writing this post is to challenge us to be mindful of these factors when we eventually do have the capacity to restructure the inpatient hospital routine, to focus appropriate attention and efforts to allow for adequate patient sleep and recovery. This field may actually be an exciting avenue for research, particularly in wearables that can appropriately monitor inpatients’ physiological parameters during sleep without disturbing them.

So before you go entering your patients’ rooms at 5:30am in the morning to pre-round, I encourage you to take an extra moment to ask yourself whether it’s truly necessary to wake them, or whether you can allow them the chance to improve their recovery score just a bit more.



  1. Holth JK, Fritschi SK, Wang C, Pedersen NP, Cirrito JR, Mahan TE, Finn MB, Manis M, Geerling JC, Fuller PM, Lucey BP, Holtzman DM. The sleep-wake cycle regulates brain interstitial fluid tau in mice and CSF tau in humans. Science. 2019;363:880–884.
  2. Noble W, Spires-Jones TL. Sleep well to slow Alzheimer’s progression? Science. 2019;363:813–814.
  3. Depner CM, Melanson EL, Eckel RH, Snell-Bergeon JK, Perreault L, Bergman BC, Higgins JA, Guerin MK, Stothard ER, Morton SJ, Wright KP. Ad libitum Weekend Recovery Sleep Fails to Prevent Metabolic Dysregulation during a Repeating Pattern of Insufficient Sleep and Weekend Recovery Sleep. Curr Biol. 2019;29:957-967.e4.
  4. Wesselius HM, van den Ende ES, Alsma J, Ter Maaten JC, Schuit SCE, Stassen PM, de Vries OJ, Kaasjager KHAH, Haak HR, van Doormaal FF, Hoogerwerf JJ, Terwee CB, van de Ven PM, Bosch FH, van Someren EJW, Nanayakkara PWB, “Onderzoeks Consortium Acute Geneeskunde” Acute Medicine Research Consortium. Quality and Quantity of Sleep and Factors Associated With Sleep Disturbance in Hospitalized Patients. JAMA Intern Med. 2018;178:1201–1208.



Apple Watch, Fitbit or RESPeRATE – Can They Assist in Lowering Blood Pressure?

which device should I choose?

We see people walking around with the wearable devices everyday without regard to whether they really make a difference with metabolic parameters. These devices all have the capability of prompting wearers to take steps, stand up when sitting, but they can also alert wearer to slow breathing. A poster presented at Experimental Biology in Orlando by Evan D. Jette, a student from the Usselman’s lab at McGill University in Montreal, QC, Canada made the argument these wearable devices can potentially lower blood pressure (BP). I was interested in whether there was more research surrounding these wearable devices, especially Fitbit and Apple because I noticed they were prevalent among conference goers.

Evan’s research suggested there was a trend with blood pressure among clinical populations (high BP or T2DM) indicating that slower breathing (~15 breaths per min) can potentially have a positive impact on diastolic BP. He further indicated that the Fitbit may have been optimal in lowering BP via respiratory rate due to the ability of the Fitbit to customize breathing to the wearer rather than generating a standardized rate for all subjects. These data leave me to wonder, since the RESPeRATE is marketed to lower blood pressure by controlling breathing, and most people own wearable devices such as commercially available Fitbits or Apple Watches, would these devices really assist in BP modulation?

There is a significant amount of literature surrounding the heart rate capability of the wearable devices, but negligible data referencing changes in blood pressure. The exception is RESPeRATE taking the stand that their product is “clinically proven to lower blood pressure”. In the study by Jette, participants that undergo low to moderate exercise exhibited no significant difference in heart rate with either the Fitbit or the Apple Watch. However, under extreme conditions such as high impact training, accuracy was reduced across both devices. The Fitbit provided heart rates that were equivalent to the Polar monitor (a heart rate monitor strapped to the subject’s chest). I wonder under these experimental conditions will RESPeRATE have similar outcomes.

Nevertheless, I did not find any data that supporting these wearable devices playing a role in reduced BP. I think the Usselman’s group is on to something with exploring the use of these wearable devices to modulate BP. However, a healthy lifestyle that incorporate the AHA Life’s Simple 7 will assist in blood pressure reduction. So, keep wearing your fitness devices to maintain an enhanced level of motivation and stay connected to a community of people that will support your BP reduction goal.


Wearables in Medicine: Try It Before You Prescribe It?

Much of the “buzz” in the air among ACC19 attendees revolved around the Apple Heart Study. There was a wide variety in reactions to the study results – from underwhelmed to measured to overzealous.  After some reflection, my personal reaction is that I’m just glad this study was performed – now we have some data for one of the most widely used wearable devices by our patients. Patients will, and already have, come to me with questions about the Apple Watch and its heart rhythm monitoring capabilities, and now I have some numbers available to help me address their concerns.

As the Apple Heart Study is likely just the beginning of an impending flood of wearable and virtual enrollment studies, physicians will undoubtedly be asked more and more questions about data collected by our patients’ devices. Just the other week in clinic, I had one new patient present with concerns about his cycling performance over the past few months. I fortunately noticed he was wearing an Oura ring device, and I asked him if he wore the ring during his cycling rides. He was shocked – he had yet to encounter a physician who knew what the Oura ring was, let alone be comfortable with analyzing the variety of data it measured. Fortunately, I had just chatted at length with a colleague who uses the Oura ring, as I was in the market for a wearable fitness monitor at the time. Yet even my cursory knowledge of the device seemed to deepen the patient-physician relationship in that first clinic visit.

The primary objective of the Apple Heart Study was to test the ability of the Apple Watch and its rhythm analysis algorithm to accurately detect atrial fibrillation. While atrial fibrillation detection is clearly an important tool, as recently described by AHA blogger, Dr. Christa Trexler, there are a variety of data being collected by wearables that may have tremendous value for our ability to optimally care for our patients, as these measurements lend insight into the 99+% of the time our patients spend outside of our clinic room with us. These include routinely measured factors (such as heart rate, step counts, and even blood pressure), but they are also measuring parameters and providing assessments of factors we do not routinely use in clinical practice, such as heart rate variability (HRV), “sleep quality,” and recovery/readiness indices. So, should we start incorporating these latter measurements in our patient care?

Figure: (Left) Apple Watch Series 4. (Right) Whoop Strap 2.0, demonstrating my Recovery Score for the day (based on my recent sleep patterns and recent cardiovascular workloads). (Both) Demonstration of probably wearing too many wearable devices.


As I mentioned, I was in the market for a wearable device, and somehow I now have two: the Apple Watch Series 4 and the Whoop Strap 2.0 (Figure 1). Overkill? Absolutely. But in experimenting with these devices, I’ve become incredibly fascinated with the HRV, sleep quality scores, and recovery. These are metrics that counterbalance our typical recommendations of increased physical activity with adequate rest and recovery. However, while parameters such as HRV were heavily studied in the 1990s and remain very much present in the current literature (search for “heart rate variability” on Pubmed yields 1665 publications), we have not routinely interfaced with these parameters in modern cardiology practice.

Yet I’ve found myself poring over my own device-measured physiological data and have already used it to plan my days. For instance, when my Whoop strap notifies me that I’m “in the green” and adequately recovered (as in Figure 1), I plan a more intense workout. Conversely, when I haven’t had adequate sleep for consecutive nights, I will be reminded by the Whoop app that my body is not primed for significant strain, and I will focus my efforts on restorative exercises, such as stretching and an early bedtime. While it seems silly to rely on a device to tell me how my body should feel, it has at the very least strengthened my own practice of reflecting on my health daily, a practice that can easily be forgotten amid busy clinical and research training.

With the increasing popularity and use of these wearables, and now that Apple Heart Study has paved the way for massive amounts of patients to be enrolled in studies using wearables, more of our patients will be using the Apple Watch, the Whoop strap, and other similar devices. To better prepare for this inevitable influx of personalized data, I feel that it is useful for clinicians to have our own experiences with these devices and apps. In my own experience, I’ve seen it enhance clinic encounters with patients, and by learning more about how devices monitor my own physiology, I believe it can help me better counsel my patients on how to monitor theirs.

What wearables do you use? Has your own use of wearables already impacted your management of patients? Would love to hear your thoughts via Twitter @JeffHsuMD.


The Future of Wearable Technology & Detecting Atrial Fibrillation – An Update!

Last November at AHA18, I was lucky enough to catch a talk from one of the investigators, Dr. Marco Perez, working on the Apple Heart Study, where he described the goals of the project. I even wrote about it for the blog I wrote at AHA18 in Chicago, which you can find here.

To quickly recap, this unprecedented collaboration between Apple and Stanford is a progressive clinical trial that uses data from Apple Watch devices from over 400,000 participants. The main purpose of this study was to examine if atrial fibrillation (AFib) can be reliably diagnosed from irregular pulse notification data from wearable devices. In November, Dr. Perez mentioned that they wouldn’t have data until early 2019, and this past March, they released some results that are really exciting.

Highlights from the findings include:

  • Around 0.5% of participants received irregular pulse notifications, which was particularly important since people were concerned that these devices would potentially over-notify people.
  • The pulse detection algorithm of the Apple Watch has a 71% positive predictive value – this was compared to simultaneous electrocardiography patch recordings.
  • The majority of the time (84%) when participants received irregular pulse notifications, they were found to be in AFib at the time of the notification.
  • 1/3 of the participants who received irregular pulse notifications and were then followed up by using an ECG patch over a week later were found to have AFib. This isn’t entirely surprising since AFib is an intermittent condition, so it’s not uncommon for it to go undetected in subsequent monitoring.
  • 57% of people who received irregular pulse notifications sought medical attention afterwards.
  • A clinical trial could be conducted in this large-scale virtual manner

The last point is particularly important because this is the first study of its kind. It was almost completely virtual, at least from the standpoint that the researchers analyzing the data never once had any contact with the participants. Additionally, the scope of the number of people who were analyzed is impressive – 400,000 participants is quite the sample size, especially for someone like me who studies heart disease in mice!

Because AFib is an extremely common condition, affecting between 2-6 million people in the United States, that often goes undiagnosed, understanding that wearable technology can aid patients in detecting their condition is huge. Also, with the increase in technology in our health care sphere (specifically in cardiology), something that was expertly discussed by Dr. John Chen earlier this year on his blog is that this is just the beginning in understanding how wearable technology can help us treat disease. We now have another tool in our kit, and this one looks promising.

In Stanford’s press release, Dr. Perez said, “The performance and accuracy we observed in this study provides important information as we seek to understand the potential impact of wearable technology on the health system. Further research will help people make more informed health decisions.”

Interestingly, Johnson & Johnson and Apple recently announced their plans to build off these preliminary results by partnering together in a new project called HEARTLINE. This study will focus on an older population (~ 65 years) of around 150,000 participants who, due to their age, are at a higher risk of AFib. It’s a really exciting time to be in the cardiology field since this is just the beginning of this type of research, which is full of therapeutic potential.

It’s also really thrilling to be able to follow-up with this study, especially since it all started with my father-in-law, who has AFib, being nervous about using wearable technology to detect his condition. I’m excited to share this data with him as well since maybe he’ll feel a little more comfortable using it now.

Examples of the notifications that participants in the Apple Heart Study receive. Courtesy of Apple

Examples of the notifications that participants in the Apple Heart Study receive. Courtesy of Apple


Tech in Cardiology

Tech in Cardiology

On a recent flight from San Francisco, I found myself sitting in a dreaded middle seat.  To my left was a programmer typing way in Python, and to my right was an oncologist flipping through a slide set on chemotherapy trials.  While this may sound like the beginning of a bad joke, I remember this moment because it got me thinking about the influence of tech on medicine.  The purpose of my trip, by the way, was to interview for a fellowship position in cardiology, a specialty with arguably some of the most impressive tech.



Not to discount advances in medical devices (e.g. leadless pacemakers, bioprosthetic valves), the emergence of consumer-facing wearable devices is as trendy as ever.  Google recently collaborated with AHA to build its fitness app (Google Fit), which uses algorithms to quantify physical activity in terms of “heart points.”1  The Apple Health app now incorporates EKG capabilities, allowing patients to record episodes of arrhythmias—something I have certainly witnessed in cardiology clinic.2


Big data

Big data is an increasingly prominent component of clinical research, and a number of joint ventures with medical and tech leaders have emerged.  One Brave Idea3 is a research collaboration between AHA and Verily (Alphabet’s life sciences division) which uses genomics to study coronary artery disease.  Meanwhile, Verily’s Project Baseline4 is a massive longitudinal observational study—a modern version of the Framingham Heart Study.


Artificial intelligence

AI could eventually play a prominent role in medical diagnosis and decision-making.  The Stanford Machine Learning Group5 has developed a neural network that outperforms cardiologists in diagnosing arrhythmias on EKG—a significant improvement on existing algorithms which are often unreliable.  AI also carries vast potential in radiologic interpretation.  Already, Veril is using machine learning to interpret retinal images not only to detect diabetic retinopathy and macular edema but also to extrapolate information about cardiovascular risk.6



Electronic medical records represent an obvious space for tech innovation.  Fast Healthcare Interoperability Resources (FHIR) are making it easier to share health information across our disjointed EMR systems.  Providers are now able to push health data directly to patients’ iPhones using Apple Health Records.7  One can only speculate whether we will see a legacy software giant compete directly in the EMR space.


Cardiology and the rest of medicine has long excelled at basic science and translational research, but digital tech is increasingly creeping in.  We are in a tech zeitgeist, and this is good for both patients and providers.



  1. https://www.heart.org/en/news/2018/08/21/google-just-launched-heart-points-here-are-5-things-you-need-to-know
  2. https://www.apple.com/healthcare/site/docs/Apple_Watch_Arrhythmia_Detection.pdf
  3. https://www.onebraveidea.org/
  4. https://verily.com/projects/precision-medicine/baseline-study/
  5. https://stanfordmlgroup.github.io/projects/ecg/
  6. https://blog.verily.com/2018/02/eyes-window-into-heart-health.htm
  7. https://www.apple.com/healthcare/health-records/