Karthik Murugadoss

Karthik Murugadoss

Auto-generated profile, based on publicly available information

Karthik Murugadoss

Nference, 1 Main Street, East Arcade building, Cambridge, Massachusetts, USA

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 May 2025 in JCO Oncology Practice
Verified
Journal Article
Open access
TSTyler B. Sandahl
Tyler B. Sandahl
SSScott A A Soefje
Scott A A Soefje
RFRafael Fonseca
Rafael Fonseca
Sikander AilawadhiSikander Ailawadhi
Sikander Ailawadhi
RPRicardo Parrondo
Ricardo Parrondo
SKShaji Kumar
Shaji Kumar
PURPOSE: Teclistamab is initiated with a step-up dosing (SUD) schedule to mitigate the risk of cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS). Early teclistamab users commonly received SUD in a hospital setting. This study aimed to evaluate safety and health care resource utilization (HRU) in real-world patients with multiple myeloma who initiated teclistamab SUD in an outpatient setting. METHODS: This was a retrospective study using Mayo Clinic's electronic medical records from October 26, 2022, to October 31, 2023. Patient characteristics were summarized for all patients treated with teclistamab and separately for patients who started SUD outpatient. SUD pattern, safety, HRU, and post-SUD dosing schedule were described in patients with complete SUD. RESULTS: At data cutoff, 65 patients received ≥1 teclistamab dose, including 58 patients who initiated SUD outpatient (median age, 69.2 years; male, 63.8%; White, 89.7%). Among 57 patients who completed SUD in an outpatient setting, all received premedications on the days of teclistamab administrations per label recommendation; 18 (31.6%) developed CRS (13 grade 1, four grade 2, and one grade 4) and two developed ICANS (one each with grade 2 and 4). All CRS and ICANS resolved with supportive care and all patients continued treatment. Eighteen patients were admitted to the hospital for CRS treatment, with a median CRS-related hospital stay of 2 days per admission. Most (60%) doses during SUD required <1 hour clinic time between administration and checkout. Post-SUD, clinic time for treatment doses decreased to <30 minutes for most doses (82%). CONCLUSION: Outcomes of this study support outpatient administration as a safe and feasible option for teclistamab SUD to potentially reduce HRU and improve patient experiences.

20 June 2024 in medRxiv
Verified
Preprint
Open access
PDF file
KMKarthik Murugadoss
Karthik Murugadoss
SKSaivikas Killamsetty
Saivikas Killamsetty
DDDeeksha Doddahonnaiah
Deeksha Doddahonnaiah
NINakul Iyer
Nakul Iyer
MPMichael Pencina
Michael Pencina
SASankar Ardhanari
Sankar Ardhanari
The natural language text in electronic health records (EHRs), such as clinical notes, often contains information that is not captured elsewhere (e.g., degree of disease progression and responsiveness to treatment) and, thus, is invaluable for downstream clinical analysis. However, to make such data available for broader research purposes, in the United States, personally identifiable information (PII) is typically removed from the EHR in accordance with the Privacy Rule of the Health Insurance Portability and Accountability Act (HIPAA). Automated de-identification systems that mimic human accuracy in identifier detection can enable access, at scale, to more diverse de-identified data sets thereby fostering robust findings in medical research to advance patient care. The best performing of such systems employ language models that require time and effort for retraining or fine tuning for newer datasets to achieve consistent results and revalidation on older datasets. Hence, there is a need to adapt text de-identification methods to datasets across health institutions. Given the success of foundational large language models (LLMs), such as ChatGPT, in a wide array of natural language processing (NLP) tasks, they seem a natural fit for identifying PII across varied datasets. In this paper, we introduce locally augmented ensembles, which adapt an existing PII detection ensemble method trained at one health institution to others by using institution-specific dictionaries to capture location specific PII and recover medically relevant information that was previously misclassified as PII. We augment an ensemble model created at Mayo Clinic and test it on a dataset of 15,716 clinical notes at Duke University Health System. We further compare the task specific fine tuned ensemble against LLM based prompt engineering solutions on the 2014 i2b2 and 2003 CoNLL NER datasets for prediction accuracy, speed and cost. On the Duke notes, our approach achieves increased recall and precision of 0.996 and 0.982 respectively compared to 0.989 and 0.979 respectively without the augmentation. Our results indicate that LLMs may require significant prompt engineering effort to reach the levels attained by ensemble approaches. Further, given the current state of technology, they are at least 3 times slower and 5 times more expensive to operate than the ensemble approach.

23 April 2020 in medRxiv
Verified
Preprint
Open access
PDF file
TWTyler Wagner
Tyler Wagner
FSFNU Shweta
FNU Shweta
KMKarthik Murugadoss
Karthik Murugadoss
SASamir Awasthi
Samir Awasthi
AVAj Venkatakrishnan
Aj Venkatakrishnan
VSVenky Soundararajan
Venky Soundararajan
Understanding temporal dynamics of COVID-19 patient symptoms could provide fine-grained resolution to guide clinical decision-making. Here, we use deep neural networks over an institution-wide platform for the augmented curation of clinical notes from 77,167 patients subjected to COVID-19 PCR testing. By contrasting Electronic Health Record (EHR)-derived symptoms of COVID-19-positive (COVID pos ; n=2,317) versus COVID-19-negative (COVID neg ; n=74,850) patients for the week preceding the PCR testing date, we identify anosmia/dysgeusia (27.1-fold), fever/chills (2.6-fold), respiratory difficulty (2.2-fold), cough (2.2-fold), myalgia/arthralgia (2-fold), and diarrhea (1.4-fold) as significantly amplified in COVID pos over COVID neg patients. The combination of cough and fever/chills has 4.2-fold amplification in COVID pos patients during the week prior to PCR testing, and along with anosmia/dysgeusia, constitutes the earliest EHR-derived signature of COVID-19. This study introduces an Augmented Intelligence platform for the real-time synthesis of institutional biomedical knowledge. The platform holds tremendous potential for scaling up curation throughput, thus enabling EHR-powered early disease diagnosis.

2 December 2025 in medRxiv
Verified
Preprint
Open access
PDF file
AVAj Venkatakrishnan
Aj Venkatakrishnan
KMKarthik Murugadoss
Karthik Murugadoss
VSVenky Soundararajan
Venky Soundararajan
Background: Glucagon-like peptide-1 receptor agonists (GLP-1RAs) are widely used for obesity and type 2 diabetes, yet substantial variability in weight-loss response remains poorly understood. Methods: We conducted a retrospective cohort study using de-identified electronic health records and performed 1:1 propensity matching on age, sex, type 2 diabetes status, baseline BMI and weight, index year, and follow-up duration. The matched cohorts included 10,339 tirzepatide-treated and 10,339 semaglutide-treated patients. Patients were categorized by maximum weight loss over two years into five response groups. Adverse events were identified through AI-enabled curation of clinical notes. Weight-loss trajectories, demographic patterns, adverse-event profiles, and pre- to post-treatment disease-prevalence changes were compared across drugs. Results: Patients treated with tirzepatide lost more weight than those treated with semaglutide (mean reduction, 14.7% vs. 10.8%; p<0.001). High-response rates (≥15% weight loss in year 1) were nearly doubled with tirzepatide (42.6% vs. 21.6%; p<0.001), accompanied by faster monthly weight-loss velocity (2.54% vs. 2.18%). AI-enabled curation showed that tirzepatide was associated with lower prevalence of gastrointestinal and systemic adverse events. For both tirzepatide and semaglutide, women were more represented among high responders than the minimal weight-loss group (<5% weight loss) and White patients were more represented among high responders, whereas Black and Hispanic patients were more represented among the minimal weight loss group. Conclusions: In this large, propensity-matched real-world cohort, tirzepatide was associated with greater and faster weight loss than semaglutide. Marked demographic disparities highlight the need for precision approaches to obesity treatment.

17 November 2025 in medRxiv
Verified
Preprint
Open access
PDF file
AVAj Venkatakrishnan
Aj Venkatakrishnan
KMKarthik Murugadoss
Karthik Murugadoss
VSVenky Soundararajan
Venky Soundararajan
Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have reshaped obesity treatment, yet weight-loss outcomes remain highly uneven in real-world care. Using a federated biomedical platform integrating 23 million de-identified U.S. patient records, we analyzed 135,349 individuals treated with GLP-1RAs and stratified them as “super responders” (>15% weight loss), “moderate responders” (5–15% weight loss), “minimal weight-loss group” (<5% weight loss), and “weight regainers”. super responders reversed nearly two decades of age-associated weight gain in one year, representing approximately a decade more weight reversal than moderate responders. Compared with Wegovy (semaglutide), Zepbound (tirzepatide) showed 47% higher odds (CI: 33–61%) of super-response and 30% lower odds (CI: 23–37%) of minimal weight-loss. Likewise, relative to Ozempic (semaglutide), Mounjaro (tirzepatide) showed 284% (CI: 265–304%) higher odds of super-response and 48% (CI: 46–51%) lower odds of minimal weight-loss. AI-enabled curation processed more than 14 million clinical notes and 15 million structured records covering 1,426 disease terms across the year before and after GLP-1RA initiation. Wegovy and Ozempic super responders showed marked post-treatment increases in vomiting compared with pre-treatment baselines, as reflected by pre-to-post rate ratios (RR 0.37, p=0.014 and RR 0.09, p<0.001). In contrast, Zepbound super responders showed significantly lower post-treatment vomiting relative to baseline (RR 2.34, p<0.001), indicating brand-specific gastrointestinal tolerability profiles. Ozempic (RR 0.24, p<0.001) and Mounjaro (RR 0.17, p<0.001) super responders each showed significant post-treatment increases in diagnoses of protein–energy malnutrition, suggesting a need for whole-body compositional imaging to distinguish beneficial fat loss from unintended lean-mass loss. Novel signals for therapeutic expansion also emerged. Compared with pre-treatment baselines, Zepbound showed significantly reduced post-treatment encounters for recurrent major depressive disorder (pre-to-post RR 12.6, p<0.001) and asthma (pre-to-post RR 2.6, p<0.001). Patient stratification prior to therapy initiation revealed pre-treatment signatures that can guide GLP-1RA choice, with Zepbound super responders showing lower sleep apnea prevalence (baseline RR 0.42, p<0.001) and higher muscle stiffness prevalence (baseline RR 2.4, p=0.037). This study pinpoints actionable physiological signatures and GLP-1RA brand-specific opportunities that emerge from heterogeneous real-world responses, outlining a map for guided precision obesity interventions.

6 April 2026 in medRxiv
Verified
Preprint
Open access
PDF file
KMKarthik Murugadoss
Karthik Murugadoss
AVAj Venkatakrishnan
Aj Venkatakrishnan
CGChristopher Gregg
Christopher Gregg
VSVenky Soundararajan
Venky Soundararajan
Semaglutide is often optimized for weight loss, but whether longer-term cardiovascular benefit tracks achieved weight loss or therapeutic exposure levels remains unclear. Using a federated deidentified U.S. electronic health record network of 29 million patients, including 505,874 semaglutide-treated individuals, we leveraged multimodal AI technologies to analyze 47,199 patients with baseline cardiovascular disease. We quantified dose escalation and weight change during the 0–2-year period after semaglutide initiation (landmark period) and assessed cardiovascular outcomes during the 2–4-year period (post-landmark). In propensity-matched comparisons during the landmark period, semaglutide was associated with lower cardiovascular events than metformin, DPP-4 and SGLT2 inhibitors. Higher maximum semaglutide dose was associated with greater weight loss during the landmark period (3.15% additional weight loss per 1 mg increase; r=−0.97, P<0.001), and lower post-landmark risk of all-cause mortality (RR 0.42, p<0.001), composite cardiovascular events (death, myocardial infarction, or stroke; RR 0.51, p<0.001), cerebrovascular disease (RR 0.50, p<0.001), heart failure (RR 0.55, p<0.001), and valvular/rheumatic heart disease (RR 0.71, p=0.025). In contrast, greater achieved weight loss during the landmark period did not show a consistent monotonic association with lower post-landmark cardiovascular risk (All-cause mortality p-value=0.14, composite cardiovascular endpoint p-value=0.55). Integrating insights from a single cell GLP1R expression atlas was used to infer how semaglutide pharmacology may tie into heart-specific signaling, beyond what is reflected by body-weight reduction alone. The strongest prevalence-weighted GLP1R signal was observed in the pancreas, followed by the heart, where GLP1R engagement potential (GEP) was considerable across cardiomyocyte, cardiac endothelial, and rarer immune cell populations. Together, semaglutide cardiovascular benefit appears organized more by maximum dose attained than by achieved weight-loss magnitude, setting the stage for beyond-obesity trial designs that integrate whole-body spatial intelligence.