About

Justin Ko is a Clinical Professor in the Department of Dermatology at Stanford University. He is affiliated with the Center for Artificial Intelligence in Medicine & Imaging (AIMI) at Stanford. His work focuses on the application of artificial intelligence and imaging techniques in the field of medicine, particularly dermatology. As a faculty member at Stanford, he contributes to research and education in AI-driven healthcare solutions, advancing the integration of innovative technologies into clinical practice.

Research topics

• Medicine
• Artificial Intelligence
• Dermatology
• Internal medicine
• Political Science
• Computer Science
• Engineering ethics
• Family medicine
• Law
• Psychiatry
• Nursing
• Engineering
• Immunology
• Pathology
• Physical therapy
• Emergency medicine

Selected publications

• 62096 Association Between Governmental Financial Support Programs and Decreased Medical Care Affordability Delays in Pediatric Atopic Dermatitis

  Journal of the American Academy of Dermatology · 2025-09-01

  article
  PublisherDOI

• Dynamical mean-field analysis of adaptive Langevin diffusions: Replica-symmetric fixed point and empirical Bayes

  ArXiv.org · 2025-04-22

  preprintOpen access

  In many applications of statistical estimation via sampling, one may wish to sample from a high-dimensional target distribution that is adaptively evolving to the samples already seen. We study an example of such dynamics, given by a Langevin diffusion for posterior sampling in a Bayesian linear regression model with i.i.d. regression design, whose prior continuously adapts to the Langevin trajectory via a maximum marginal-likelihood scheme. Results of dynamical mean-field theory (DMFT) developed in our companion paper establish a precise high-dimensional asymptotic limit for the joint evolution of the prior parameter and law of the Langevin sample. In this work, we carry out an analysis of the equations that describe this DMFT limit, under conditions of approximate time-translation-invariance which include, in particular, settings where the posterior law satisfies a log-Sobolev inequality. In such settings, we show that this adaptive Langevin trajectory converges on a dimension-independent time horizon to an equilibrium state that is characterized by a system of scalar fixed-point equations, and the associated prior parameter converges to a critical point of a replica-symmetric limit for the model free energy. As a by-product of our analyses, we obtain a new dynamical proof that this replica-symmetric limit for the free energy is exact, in models having a possibly misspecified prior and where a log-Sobolev inequality holds for the posterior law.

  PublisherOA PDFDOI

• 0428 Correlates to barrier function before and after dupilumab treatment

  Journal of Investigative Dermatology · 2025-07-21

  articleOpen access
  PublisherOA PDFDOI

• Low-rank matrix estimation with inhomogeneous noise

  Information and Inference A Journal of the IMA · 2025-03-26 · 2 citations

  article

  Abstract We study low-rank matrix estimation for a generic inhomogeneous output channel through which the matrix is observed. This generalizes the commonly considered spiked matrix model with homogeneous noise to include for instance the dense degree-corrected stochastic block model. We adapt techniques used to study multi-species spin glasses to derive and rigorously prove an expression for the free energy of the problem in the large size limit, providing a framework to study the signal detection thresholds. We discuss an application of this framework to the degree corrected stochastic block models.

  PublisherDOI

• Baricitinib Effectiveness and Patient Impressions in Routine Clinical Practice: 12-month Results from a Prospective Observational Real-world Study

  SKIN The Journal of Cutaneous Medicine · 2025-11-10

  articleOpen access1st authorCorresponding
  PublisherOA PDFDOI

• Exploring implementation of interventions to facilitate integration in fragmented healthcare systems

  Learning Health Systems · 2025-01-15 · 1 citations

  articleOpen access

  Introduction: Stanford Medicine is working to better coordinate care across the Stanford healthcare system, as well as improve patient and provider experiences in seeking and receiving care. This study aimed to explore the complexities of moving from a fragmented to an integrated academic healthcare system and to identify and explain factors (e.g., facilitators and barriers) of the implementation of three interventions meant to improve patient experience, reduce staff burden, and integrate health care systems across faculty and community settings. Methods: We conducted qualitative semi-structured interviews via Zoom with faculty and community physicians. Interviews were audio-recorded, professionally transcribed, and analyzed using the Consolidated Framework for Implementation Research (CFIR) and open coding. Using consensus coding approaches, researchers met regularly to discuss themes and adaptations to CFIR. Results: = 26). Factors impacting integration included the following: (1) physicians supported the interventions, promoting mission alignment; (2) physicians were motivated for change, reporting the existing system was intolerable; (3) physicians reported different priorities between clinics: faculty versus community and primary care versus specialty; (4) physicians prioritized interpersonal versus system solutions; (5) specialists were wary of unintended consequences of integration, specifically inappropriate bookings or patients being redirected to other clinics. Broadly speaking, facilitator factors 1-2 focused on the openness to, and tension for, change; and barrier factors 3-5 promoted or sustained variation across specialties and faculty/community clinics. Conclusions: Our results illustrate the challenges and opportunities of moving from a fragmented to an integrated healthcare system and emphasize the importance of building shared culture, collaboration, and coordinated actions across and within an integrated healthcare network.

  PublisherOA PDFDOI

• Most Baricitinib Responders Achieved Full Scalp Hair Regrowth: Findings from Adult and Paediatric BRAVE-AA Trials

  SKIN The Journal of Cutaneous Medicine · 2025-11-10

  articleOpen access

  Background Achievement of full scalp hair regrowth is an important outcome for patients with severe AA. Baricitinib is a selective JAK inhibitor that is approved in adults to treat severe AA and has now been studied in 257 adolescents (ages 12 to <18 years). In both populations, significant scalp hair regrowth (SALT score <20) was achieved in clinical trials. This analysis evaluates complete response thresholds of SALT score <10 and <5 among SALT score <20 responders, through Weeks (Wk) 36 and 52 in adolescent and adult populations, respectively. Methods Among BRAVE-AA1/2 (NCT03570749, NCT03899259) and BRAVE-AA-PEDS (NCT05723198) trials, baricitinib 4mg and 2mg responders (SALT score ≤20) were evaluated for achievement of SALT score ≤10 or ≤5 at Wk4, 8, 12, 16, 24, 36, and 52*. Results Among those patients treated with baricitinib 4mg, the proportion of adult responders who achieved SALT score <10 at Wk36 and Wk52 was 74.6% (150/201) and 75.0% (171/228); and the proportion in adolescents at Wk36 was 86.1% (31/36). Among those patients treated with baricitinib 4mg, the proportion of adults who achieved SALT score <5 at Wk36 and Wk52 was 55.2% (111/201) and 62.7% (143/228); and the proportion in adolescents at Wk36 was 61.1% (22/36). Some patients achieved these responses as early as Wk8 in both adult and adolescent populations.Among those patients treated with baricitinib 2mg, the proportion of adults who achieved SALT score <10 at W36 and W52 was 62.2% (51/82) and 71.0% (66/93); and the proportion in adolescents at W36 was 78.3% (18/23). Among those patients treated with baricitinib 2mg, the proportion of adults who achieved a SALT score <5 at W36 and W52 was 42.7% (35/82) and 47.3% (44/93); and the proportion in adolescents at W36 was 60.9% (14/23). Some patients achieved these responses as early as W8 in both adult and adolescent populations. Conclusion Baricitinib treated adult and adolescent patients who were responders (SALT score ≤20) had increasing depth of response over time with continued therapy, and the majority achieved full scalp hair regrowth. Footnote *Week 52 data is not yet available for the BRAVE-AA-PEDS trial.

  PublisherOA PDFDOI

• Augmented Intelligence and Dermatology – Part I: Core Concepts and Applications

  Journal of the American Academy of Dermatology · 2025-03-01 · 1 citations

  review
  PublisherDOI

• 297 Trajectory AD Severity Modeling to Better Define Dupilumab Response: A Framework for Precision Biomarker Discovery

  Journal of Investigative Dermatology · 2025-11-01

  article
  PublisherDOI

• Trends in level of service provided by advanced practice clinicians and dermatologists in outpatient setting, 2012–2020

  Archives of Dermatological Research · 2025-04-29

  letter
  PublisherDOI

Frequent coauthors

• Dilraj Kalsi

  104 shared

• Daniel Mullarkey

  Stanford University

  104 shared

• Lucy Thomas

  University of Queensland

  102 shared

• Roxana Daneshjou

  Stanford Medicine

  101 shared

• Jack Greenhalgh

  Instituto Pirenaico de Ecología

  88 shared

• Vijaytha Muralidharan

  University of Alberta

  85 shared

• Haiwen Gui

  Stanford University

  84 shared

• Jesutofunmi A. Omiye

  83 shared