Joseph W. Carlson, M.D., Ph.D.

Dr. Carlson's research program addresses problems in the diagnosis, molecular classification, and biomarker discovery of gynecologic malignancies.

A central theme is the application of molecular classification to diagnostic problems that cannot be resolved by morphology and immunohistochemistry alone. High-grade endometrial and ovarian serous carcinomas present particular classification challenges: these tumors are histologically identical across anatomic sites yet are staged and treated under different clinical frameworks, and the criteria used to distinguish them are subjective and poorly reproducible. Dr. Carlson's laboratory uses DNA methylation profiling, RNA sequencing, and somatic mutation data to develop machine learning classifiers for objective, reproducible molecular diagnosis in these scenarios. A related program investigates interobserver variability in gynecologic pathology, rigorously quantifying where and why pathologists disagree and measuring the downstream clinical consequences of that disagreement on staging, treatment selection, and trial eligibility.

A second major area is the molecular pathology of uterine mesenchymal tumors, including undifferentiated uterine sarcoma and uterine leiomyosarcoma. During his tenure at the Karolinska Institute, Dr. Carlson led a research program supported by the Swedish Cancer Society, Region Stockholm, and Radiumhemmets Research Funds that identified molecularly distinct prognostic subgroups within undifferentiated uterine sarcoma, demonstrating for the first time that a subset of patients has substantially better outcomes than previously recognized, and characterized extracellular matrix remodeling pathways and tumor immune microenvironment features as drivers of leiomyosarcoma aggressiveness. This program produced more than 40 peer-reviewed publications and supported the training of multiple graduate students and postdoctoral fellows.

A third area is computational pathology and artificial intelligence applied to gynecologic cancer. In collaboration with Professor Yan Liu's laboratory in computer science at the University of Southern California, Dr. Carlson has developed and published multimodal deep learning algorithms that integrate digital whole-slide images, blood-based molecular features, and clinical variables to predict ovarian cancer treatment response. This work has progressed to a pending US patent application (Application No. 19/381,256, University of Southern California, filed November 2025).

A fourth area is the identification and clinical translation of therapeutic biomarkers in gynecologic cancers. This work focuses on tumor-expressed targets relevant to antibody-drug conjugates, CAR-T cell therapies, and targeted agents. Of particular interest are folate pathway targets, including folate receptor alpha (FOLR1), which are highly expressed in a subset of gynecologic malignancies. Standardized, validated biomarker testing algorithms for these targets are not yet established in clinical practice. Dr. Carlson's laboratory is developing the pathologic and molecular framework needed for clinical translation, including assessment of expression patterns across histotypes and disease stages, correlation with molecular subgroups, and evaluation of the analytical performance of companion diagnostic assays.

Selected Grants

• 2021–2023, Swedish Cancer Society (20 0842 PjF), PI
• 2022–2023, Ming Hsieh Institute, University of Southern California, Co-PI
• 2020–2023, Swedish Foundation for International Cooperation in Research and Higher Education (BR2019-8511), PI
• 2018–2020, Swedish Cancer Society (CAN 2017/473), PI
• 2016–2019, Region Stockholm Senior Clinical Scientist Award, PI
• 2018–2020, Radiumhemmets Research Funds (174053), PI