Year 3. March 4. Science Monday – Research Extravaganza.
I am pleased to headline this edition of Science Monday with some recent good news. Our department of medicine (DoM) has been ranked the second highest NIH (National Institutes of Health)-funded department of medicine in the country! This is the fourth year in a row that we have ranked either number 1 or 2. These data are from the federal 2023 fiscal year according to the Blue Ridge Institute for Medical Research. This dataset summarizes NIH grants awarded to principal investigators (PIs) of record (corresponding PI) for each award. It should be noted that there are many grants that are awarded to multiple PIs, sometimes in the same, or different institutions, but Blue Ridge credits the award to the department and institution of the corresponding PI. This achievement is a testament to our faculty and staff’s commitment to lead in innovation, transform care, and advance health for all through the studies that they are leading across our divisions.
Below you will find the list of top 10 NIH funded corresponding PI’s in our department as listed by the Blue Ridge Institute. It is important to note, that several of these awards include multiple PI’s from the DoM and we may not see their names listed here. In fact, 232 DoM faculty secured NIH federal funding in 2023 which contributed towards our exceptional standing. As you can see from the infographic below, there are 14 investigators in the DoM with greater than $5M in NIH funding and 30 investigators in the DoM with $2-5M in NIH funding, which includes funds awarded as co-principal investigators or as principal investigators on parent grants that could be administered by other departments or institutions. As explained above, their names might not be listed by Blue Ridge, which only captures corresponding PI's. You may view a complete list of DoM PI’s who are listed in Blue Ridge as having secured NIH funding in FY23 HERE.
Our NIH grants provide one snapshot of our funding portfolio. When we include all sources of funding, federal (including NIH, VA and CDC) and other (Including sponsored clinical trials and foundation awards), the DoM secured over $450 million in support of your groundbreaking scientific efforts. I applaud and extend gratitude to every PI, fund manager, grant writer and administrative staff who contributed to these efforts, which solidify our standing as a top research department.
With that, I am thrilled to introduce you to a selection of studies that illustrate the depth, strength and diversity of research studies in the department of medicine. Enjoy this research extravaganza!
DoM IM Resident Amber Tang, MD, Publishes QI Research that Informs Improvement of Guideline-Directed Medical Therapy
Amber Tang, MD, is a second-year internal medicine resident, who joined us following completion of medical school -- also as a Bruin. During her first year of residency, Dr. Tang began to garner an interest in quality improvement (QI) with regards to cardiology care. As a result, she got involved with the DoM's Quality Improvement Program, which offers residents robust exposure to QI in both the inpatient and outpatient settings through QI training, QI research support, and QI project development.
As part of her involvement with the QI program, Dr. Tang set out to conduct a systematic review of guideline-directed medical therapy (GDMT) -- the backbone of pharmacological intervention for patients with heart failure with reduced ejection fraction (HFrEF), which consists of four foundational drugs that target different disease pathways of heart failure.
The paper, “Interventions for Optimization of Guideline-Directed Medical Therapy: A Systematic Review” recently published in JAMA Cardiology, empirically reviews the success of different intervention types for this type of therapy.
It can be difficult to get patients onto the correct doses of all four drugs, because it requires overcoming a range of patient and system barriers to achieve proper titration and sustainability of high doses for patients. "There have been many interventions to improve GDMT," Dr. Tang shared. "Our goal was to understand what kinds of interventions have been tested, and which are the most effective." To do so, Dr. Tang and team categorized interventions into four categories: 1) interdisciplinary interventions, 2) electronic health record alerts, 3) clinician education, 4) patient education. What they found is that the most successful interventions were the interdisciplinary interventions that set up medication titration clinics either by phone or in-clinic, often following a set algorithm for patients to get them up to higher doses of GDMT. There were positive signals for the other intervention types as well, yet data on those interventions are limited.
Dr. Tang shared that she designed this systematic review because it's the type of summary she wishes she had when designing new heart failure QI initiatives, and that ideally, is the starting point of further research on this topic. "My hope is that this research can serve as a building block for other care providers in my position who are making decisions on what GDMT interventions to implement and the impact it might have on patients," she stated. Dr. Tang also expressed deep gratitude for UCLA's novel QI program that has supported her work as a QI scholar and shaped the trajectory of her career to be oriented towards quality improvement as the bottom line.
“Alveolar macrophage lipid burden correlates with clinical improvement in patients with Pulmonary Alveolar Proteinosis”
Pulmonary alveolar proteinosis (PAP) is a life-threatening, rare lung syndrome for which there is no cure and no approved therapies. Elizabeth Tarling, PhD, FAHA notes that the disease is one of lipid build up, characterized by something called alveolar macrophage foam cell formation -- in essence, a lack of homeostasis of alveolar macrophages, which are an important type of immune cells that are one of your body's first line of defense against outside pathogens and pollutants. They also help to maintain the right amount of lipid in your air sacs so that you can breathe -- not too much, or too little. When lipid builds up in the lungs in patients with PAP, because alveolar macrophages have too much lipid, it becomes difficult to breathe.
Dr. Tarling shared that while there is a decent amount of literature on the clinical presentation of PAP, research is lacking regarding how the lipid in alveolar macrophages changes throughout the course of disease, from diagnosis to treatment.
Dr. Tarling, who is an associate professor in the UCLA Division of Cardiology, took the scientific community a step further in closing this gap in her recent co-authored paper “Alveolar macrophage lipid burden correlates with clinical improvement in patients with Pulmonary Alveolar Proteinosis.”
The study is a long-time collaboration in partnership with DoM Senior Executive Clinical Vice Chair Tisha Wang, MD, who works with many patients with PAP and other lung conditions, and therefore was able to collect their fluids for Dr. Tarling to study. The old saying, one person's trash is another's treasure couldn't be truer than in this context! Dr. Tarling took these fluid samples from patients with and without PAP and studied how the lipid buildup changed over time in response to a range of treatments -- including the standard of care treatment whole-lung lavage (also called lung washing) and more experimental treatment such as inhaled GM-CSF.
Their team found that in comparison to alveolar macrophages from non-PAP patients, the alveolar macrophages of patients with PAP showed increased levels of more types of lipids than previously understood. They also found that the lipid composition of these cells varies greatly by patient, showing the diverse nature of the disease. Finally, the study indicated that lipid levels in alveolar macrophages could be used as a biomarker, but that collecting it from these samples isn't feasible in practice because of the invasiveness of whole-lung lavage. Dr. Tarling and her team next hope to conduct a similar study using blood samples from patients that do and don't have disease, to see if we can identify a blood biomarker for lipid buildup in alveolar macrophages. Overall, this study is a triumph both with regards to scientific research and long-term multidisciplinary collaboration.
"This paper is 10 or 12 years in the making and required collecting patient samples over a long period of time and dedication to this project," Dr. Tarling shared. "It's one of the projects I'm most proud of because it crosses from basic science to medical translation through a significant effort of all people involved. It's a really big deal."
“Benchmarking Open-Source Large Language Models, GPT-4 and Claude 2 on Multiple-Choice Questions in Nephrology”
Ira Kurtz, MD, chief of the UCLA Division of Nephrology, is confident that Artificial Intelligence (AI) will one day replace many physician responsibilities, with implications for reducing health costs, expanding care access, and improving medical education. In response, he's dedicated to understanding and improving the knowledge base of a type of advanced AI called large language models (LLMs), especially when it comes to complex subspecialties like nephrology. This quest is the basis for one of his recent studies, “Benchmarking Open-Source Large Language Models, GPT-4 and Claude 2 on Multiple-Choice Questions in Nephrology," which was published in the inaugural edition of NEJM AI.
The paper, coauthored and championed by Dr. Kurtz, presents the first comprehensive study comparing the performance of several LLMs in the field of nephrology. To do so, the study evaluates how well seven different models perform in answering multiple-choice questions from the Nephrology Self-Assessment Program (nephSAP) -- an advanced self-assessment tool consisting of 858 questions primarily used for nephrology board and recertification exams. While Dr. Kurtz and team don't have access to the distribution of human scores on nephSAP, they do know that a "passing" score is 75%. Their methodology was therefore relatively simple: ask each model all 858 questions and assess their scores.
Importantly, Dr. Kurtz and team weren't just comparing each model's individual success, they were also comparing the success of closed source LLMs versus open source LLMs. Open source LLMs are models with publicly available source code, meaning that anyone can access and train the model. The authors assessed five open source LLMs: Llama2-70B, Koala 7B, Falcon 7B, Stable-Vicuna 13B, and Orca-Mini 13B. Closed source LLMs are the opposite: they're proprietary models with a source code that is not publicly available. Closed source LLMs are developed and maintained typically by private companies that can buy large amounts of training data and have vast computational resources. The authors assessed two closed source LLMs: GPT-4 and Claude 2.
The team found that the overall success of open-source LLMs in answering the nephSAP questions correctly was 17.1% to 30.6%. In contrast, Claude 2 and GPT-4 (the closed source LLMs) scored 54.4% and 73.3%, respectively. These findings indicate that some models are already performing exceptionally well, with closed LLMs leading the way. This is likely in part due to a couple of factors: 1) an increased number of model parameters in closed models, and 2) training on higher quality curated data that is not available publicly. Dr. Kurtz shared that ideally, in the future, physicians and healthcare systems will either get access to closed source LLMs (for a fee) and fine-tune them on specific medical datasets, or the capability of open source LLM models will improve to the point that they will also be fine-tuned in this context. Regardless, both approaches would lead to the development of more sophisticated and accurate models in the medical field in the future. Doing so, Kurtz says, will pave way for the future of medicine by allowing for adaptive individualized medical education, improving medical decision-making, enhancing medical educational opportunities for patients, and moving us closer as a society to personalized medical care. I would also add that these findings have significant implications for how tests will be administered in the future!
"Respiratory distress in SARS-CoV-2 exposed uninfected neonates followed in the COVID Outcomes in Mother-Infant Pairs (COMP) Study"
Since the onset of the COVID-19 pandemic and the experimental trials for COVID-19 vaccines that followed shortly after, researchers have been working to answer questions about the impact of the SARS-CoV-2 virus on infants who were born to individuals infected while pregnant and whether maternal vaccination protects against adverse infant outcomes. This study, led by Karin A. Nielsen, MD, professor of pediatrics in the UCLA Division of Infectious Diseases, with assistance from one of her mentees, Mary Catherine Cambou, MD, newly appointed assistant clinical professor of medicine in UCLA Division of Infectious Diseases, is a first-of-its-kind study that sheds light on both of those questions.
"We knew pretty early on that SARS-CoV-2 was associated with adverse reproductive and neonatal outcomes," shares Dr. Nielsen, including higher rates of respiratory distress (RD) among neonates exposed to SARS-CoV-2 in utero, even though the risk of vertical (mother-to-child) transmission is low. The team was surprised to find out, in following their cohort of babies exposed to maternal COVID-19 in pregnancy, that some of them developed respiratory distress in the newborn period. The mechanism behind RD in infants exposed to maternal illness in utero was unclear, as was the frequency with which this happened. At first, the team proposed a hypothesis that there were higher rates of RD among newborns exposed to in utero SARS-CoV-2 because they were more likely to be premature. However, the data didn't quite add up, as there ended up being cases of RD among infants who were not premature, as most babies with RD were late preterm or term.
To dig in further, Dr. Nielsen, Dr. Cambou, and their team carried out an analysis to investigate potential causes of RD in this newborn cohort, as part of the COVID-19 Outcomes in Mother-Infant Pairs (COMP) study, which is a longitudinal cohort study to evaluate SARS-CoV-2 infection in pregnancy. While none of the newborns had COVID-19, they identified increased levels of cytokines in infants with RD who had maternal COVID-19 exposure during pregnancy. These findings suggested that prenatal exposure to SARS-CoV-2 may activate an inflammatory cascade in the newborn airway, leading to ciliary beat dysregulation of the airway epithelium and higher production of inflammatory proteins and immunoglobulins associated with reactive airway disease, which can result in RD. This discovery led them to wonder if it was possible that vaccination was attenuating cytokine responses in pregnant individuals and lessening this inflammatory cascade for newborns. In addition to these empirical hints was a clinical one: a neonatologist helping care for infants in their study population remarked that babies from their study population were frequently admitted to the NICU with respiratory distress.
Leveraging the existing COMP study, the team set out to evaluate the relationship between maternal COVID-19 vaccination and neonatal RD among 199 infants in the study. What they found was that neonates born to unvaccinated individuals had a 3-fold higher risk of RD compared to neonates born to vaccinated individuals, even when accounting for preterm delivery. In addition, a proteomic analysis in a subset of newborns born to unvaccinated mothers revealed a robust inflammatory response, with the main pathways associated with ciliary function, macrophage activation, and enhanced IgE production. The results shed light on both the impact that prenatal exposure to SARS-CoV-2 can have on newborns (and by what mechanisms), and the crucial role that COVID-19 vaccination plays not just in protecting a pregnant person, but also their newborn. “We never stop learning about COVID," Dr. Nielsen shared. "Respiratory distress in newborns is a big problem and can be really significant, especially in areas with low resources. If there's no NICU it could be disastrous. The evidence we found is one more reason that pregnant people should get the vaccine." Dr. Cambou agrees, and shares that there's yet more to uncover about the impact of the virus on infants and the role of vaccines, and that their research continues.
"The Combined Effect of Immune Checkpoint Inhibitors and Tyrosine Kinase Inhibitors on Thyroid Function"
Melissa Lechner, MD, PhD, runs the onco-endocrinology group at UCLA, which serves cancer patients who develop endocrine complications as a result of their treatment. Her work sits at an interesting intersection in medicine. As advanced cancer therapies get more precise and effective, the more potential endocrine complications there are to study and address. Two of these recent, successful advanced cancer therapies are immune checkpoint inhibitors (ICIs) and tyrosine kinase inhibitors (TKIs), which treat solid malignancies.
Researchers have identified thyroid dysfunction as a common side effect of each, with rates estimated at 30–40% for TKI and 10–20% for ICI. However, little is known about the effect of combined ICI+TKI therapy on thyroid function, as a new era of combined therapy emerges. Dr. Lechner and her team, including former UCLA endocrinology fellow Karen Tsai, MD, aimed to fill this gap in their study "The Combined Effect of Immune Checkpoint Inhibitors and Tyrosine Kinase Inhibitors on Thyroid Function," which evaluated the incidence, clinical features, and risk factors for developing thyroid abnormalities during ICI+TKI therapy and the relationship to cancer outcomes.
The study found the rate of thyroid dysfunction to be 63.2% -- far higher than the rates of thyroid dysfunction associated with each of the treatments individually. Findings also suggested a more rapid onset of thyroid dysfunction than previously seen with ICI or TKI therapy alone. Implications are significant, with study authors recommending "close monitoring of thyroid function during initial therapy and multidisciplinary care with endocrinology." Ideally, with close monitoring and cross-disciplinary care, providers will be able to achieve early detection of thyroid abnormalities and appropriate thyroid hormone replacement therapy.
"Our study identifies thyroid dysfunction from combined ICI+TKI therapy as an emerging disease group that should be evaluated more," Dr. Lechner explains. " Even with this pilot study the rate of thyroid disease is higher than we thought." In terms of future work, Dr. Lechner hopes they can identify the mechanisms leading to higher rates of thyroid dysfunction in this combined treatment, and then translate these findings to patient care so that patients can safely receive the most effective cancer treatments.
“Machine learning-directed electrical impedance tomography to predict metabolically vulnerable plaques”
Kaidong Wang, PhD, a post-doctoral researcher within the Cardiovascular Engineering Research Laboratory, directed by Tzung K. Hsiai, PhD, has recently made significant strides in the field of bioengineering and translational medicine. His latest contribution, an article titled "Machine Learning-Directed Electrical Impedance Tomography to Predict Metabolically Vulnerable Plaques," was recently published in the Journal of Bioengineering & Translational Medicine. This groundbreaking study aims to revolutionize the way atherosclerotic plaques are characterized, offering a novel approach to predict their vulnerability to rupture. This advancement is particularly pivotal in addressing the diagnostic challenges associated with plaque rupture, a leading cause of acute coronary syndromes, and holds promise for enhancing heart health management, including for patients affected by long covid.
Atherosclerosis, characterized by the buildup of plaque within the arterial lining, leads to arterial wall thickening and stiffening, significantly reducing blood flow. Rupture of these plaques is a critical event leading to most acute coronary syndromes. Dr. Wang and his team's research paper elaborates on the composition of atherosclerotic lesions, highlighting the presence of a necrotic core, a thin fibrous cap, macrophages, and calcification, with the necrotic core predominantly comprising metabolically active oxidized low-density lipoprotein (oxLDL) crystals. The team's objective was to enhance the detection of these metabolically active, oxLDL-rich plaques.
Employing a novel approach, they developed a machine learning-directed electrochemical impedance spectroscopy (EIS) platform to meticulously analyze oxLDL-rich plaques, utilizing immunohistology as the benchmark for accuracy. Demonstrating the clinical applicability of the EIS sensor, the research team successfully deployed it in the coronary arteries of an explanted human heart from a transplant patient, enabling the reconstruction of 3D EIS profiles of oxLDL-rich atherosclerotic plaques in both the right coronary and left descending coronary arteries. This process was further validated in the common carotid arteries of an un-embalmed human cadaver.
The study's findings are particularly noteworthy, with the DenseNet model achieving an impressive 92.59% accuracy in predicting plaques at risk of rupture after 100 epochs of training. This research marks a significant leap towards the clinical adoption of intravascular EIS as a proactive diagnostic tool, offering a real-time, early warning system to identify patients at elevated risk of cardiovascular disease and potentially save lives by preventing plaque rupture.
The keto diet protects against epileptic seizures. Scientists are uncovering why
The following are excerpts from a UCLA Health News piece.
"The high-fat, low-carbohydrate ketogenic diet is more than just a trendy weight-loss tactic. It has also been known to help control seizures in children with epilepsy, particularly those who don’t respond to first-line anti-seizure medications. In a new UCLA study published in the journal Cell Reports, researchers demonstrate that the changes the diet causes in the human gut microbiome — the trillions of bacteria and other microorganisms that live in the digestive tract — can confer protection against seizures in mice. In the hope of finding new ways to more effectively treat seizures in the approximately one-third of people with refractory epilepsy who don’t respond to existing anti-seizure medications, Lum sought to understand the underlying molecular mechanisms behind the diet’s alteration of the human gut microbiome.
Previous research conducted by Hsiao’s lab had found that in a mouse model bred to mimic epilepsy, mice fed a ketogenic diet had significantly fewer seizures than mice fed a standard diet. Lum took that research step further, studying how the gut microbiome is beneficially altered in children with epilepsy who start ketogenic diet therapy. To that end, he transplanted fecal samples from pediatric epilepsy patients on the diet into mice to gauge whether the diet-associated gut microbiota would protect the mice against seizures. The study found that the mice that received fecal transplants from patients collected after a month on the diet were more resistant to seizures than mice that received pre–ketogenic diet fecal transplants.
Importantly, the study also found that in the pediatric patients, the ketogenic diet altered key gut microbiome functions related to fatty acid oxidation and amino acid metabolism — and that these changes were preserved when the fecal matter was transplanted into the mice.
While more research on these changes is needed, Lum said, the study holds promise as a step toward finding new microbiome-based therapies for pediatric epilepsy patients who do not respond to standard anti-seizure medications."
Please congratulate your research colleagues today for the work that they are doing to advance new knowledge, often quietly behind the scenes.
Dale
P.S.
I had lunch this week with DoM trainees and alums from the National Clinical Scholars Program (NCSP). An impressive group who are advancing health services research in the DoM while building a bench of future researchers. At lunch I learned of a collaboration between a fellow Dr. Jessica Zhang and faculty member Dr. Sae Takada that was triggered by something that they read in one of my emails. Here is the note that Jessica sent me.
“ Hi Dr. Abel,
Thank you for having lunch with some members of the UCLA National Clinician Scholars Program (NCSP) community on 2/27.
We wanted to share the connection that was made possible from your newsletter. Please let me know if we can provide any other information.
Newsletter Connections: Recently, Dr. Sae Takada’s work on the impact of social determinants of health (SDOH) on healthcare utilization among UCLA patients was highlighted in Dr. Abel’s 2/5 weekly newsletter and piqued the interest of Department of Medicine Quality Program post-discharge workgroup members Brian Le (Program Manager), Dr. Wendy Simon and Dr. Erin Dowling (Hospital Medicine, Medical Directors of Inpatient Quality for the DOM), and Dr. Jessica Zhang (UCLA National Clinician Scholars Program fellow, of which Dr. Takada is an alum). They connected with Dr. Takada to learn more about her insights and hope to incorporate a better understanding of SDOH into reducing post-discharge follow-up appointment no-shows and unplanned readmissions and improving transitions of care. Thank you Dr. Abel for creating a platform to facilitate collaboration within the UCLA community
Thank you,
Jessica
Jessica Zhang, MD
Postdoctoral Fellow | Internal Medicine
National Clinician Scholars Program | UCLA
------------------------------------------------------
The National Clinician Scholars Program
A two-year fellowship advancing health and health care through scholarship and action.
Related Posts
This week is a special edition of “Reflections from the Chair” as I have invited our regional ambulatory medical directors to share with us the[...]
Evan M. Shannon, MD, MPH Named 40 Under 40 Leader in Minority Health Join us in applauding Evan M. Shannon, MD, MPH for being named[...]
Our faculty who serve at the VA Greater Los Angeles Health Care (VA GLA), are an integral part of the department of medicine (DoM). As[...]