Ivan Pedrosa
Professor
The University of Texas MD Anderson Cancer Center at Houston
Department of Abdominal Imaging
Genitourinary Translational Imaging Innovation Program - Ivan Pedrosa MD PhD
My research program focuses on the development and validation of innovative imaging technologies that have direct application in clinical practice to improve patient outcomes. The program focuses primarily on kidney cancer although there are opportunities to expand some of the proposed approaches to other genitourinary cancers. The program is organized around three main areas:
1) Develop and validate a non-invasive biomarker of aggressive biology in renal masses with PET/MR
The rapid increase in the incidence of kidney cancer is largely a consequence of the incidental detection of small renal masses (SRMs; cT1a, ≤4 cm) due to the burgeoning use of imaging. However, despite the aggressive treatment of SRMs, kidney-cancer mortality has not declined. There is a critical need for non-invasive methods (‘virtual biopsy’) that enable a paradigm shift emphasizing prompt treatment of aggressive cancers while facilitating safe use of active surveillance in indolent malignant and benign SRMs. We are exploring hybrid examinations with positron emission tomography/magnetic resonance (PET/MR) scanner.
2) Develop an artificial intelligence (AI) pipeline for automatic characterization of renal masses at MRI
We are creating a pipeline with deep learning (DL) models for automated renal mass characterization at MRI to predict tumor aggressiveness and support clinical decision-making, enabling broader use of active surveillance (AS). We will exploit DL models to automatically segment kidneys and renal masses at MRI without human interaction. These models achieve high Dice coefficient scores in about 80% of exams, limiting the number of MRIs that require human interaction.
3) Leveraging metastatic tropism in kidney cancer to develop predictive biomarkers of response
The clinical significance of tumor tropism in metastatic RCC has been recently recognized. RCC tumors metastasizing to the pancreas are characterized by heightened angiogenesis, uninflamed stroma and indolent biology (i.e., frequent PBRM1 mutations, 3p loss, and 5q amplification). Through extensive segmentation of the whole-body tumor burden in about 300 patients with metastatic clear cell renal cell carcinoma (over 5,000 metastases), we confirmed that patients with angiogenic tumors are more likely to develop pancreatic metastasis and respond to antiangiogenic therapy. We are planning to expand these efforts to other histologic subtypes of kidney cancer.
Education & Training
MD, Complutense University of Madrid, 1994
PhD, University of Oviedo, 2017