The explosion of imaging technologies that evaluate structural, physiologic and molecular properties of cancer, has generated large and diverse amounts of data, ranging from diverse molecular imaging probes to structural and physiologic imaging at the organ level. This transition of oncologic imaging from its industrial era to it is information era has necessitated the development of analytical methods that 1) extract from this data information that is clinically and biologically relevant; 2) integrate imaging, clinical nd genomic data via rigorous statistical and computational methodologies in order to derive models valuable for understanding cancer mechanisms, but also for diagnosis, prognostic assessment, response evaluation, and personalized (or precision) treatment management; 3) are available to the biomedical community for easy use and application, with the aim of understanding, diagnosing, and treating cancer. Building and disseminating advanced information technology for computational cancer phenomics, largely captured by diverse imaging technologies, is the emphasis of the proposed work. In particular, we propose to develop and widely distribute the cancer phenomics toolkit (CapTk), a software suite integrating advanced oncologic image computing and analytics tools that have been developed by our groups here at Penn and offer sophisticated quantitative analytics of oncologic images well beyond currently used methods. Importantly, these informatics tools have been developed in the context of active clinical studies and collaborations and have therefore being inspired and tested by real clinical needs. Although the main focus of our work is CapTk, which we view as an advanced computational suite that can be incorporated into, and further enable, various commercial (e.g. BrainLab, Hologic's, GE's, see letters) and non-commercial (the open source Slicer is our research platform) workstations, we will also use this software suite to develop two focused research prototype workstations leveraging upon the unique strengths of our work and clinical studies at Penn: an Advanced Neuro-Oncologic Computational Imaging Workstation (ANCI) and the Advanced Breast Computational Imaging Workstation (ABCI). In particular, we will pursue the following specific aims:
Aim 1) To refine, extensively document and integrate our image analysis algorithms and software into CapTk, which will have 3 major components pertaining to imaging measurements and decision support: Image Registration Suite; Imaging Feature Extraction Suite; Suite for Imaging Analytics and Predictive Modeling.
Aim 2) To develop and test two focused research prototype workstations, ANCI and ABCI, aiming to provide diagnostic and treatment decision support mechanisms for brain and breast cancer management.
Aim 3) Disseminate software and knowledge, via a) deploying our software to the clinic at Penn as well as to selected collaborating institutions, in order to test and further refine the software accordig to feedback; b) freely distributing our software to the research community; c) organizing various educational activities for both clinicians and informaticians, and setting up and maintaining support forums.

Public Health Relevance

This project will develop advanced computer analysis methodology for interpretation of radiologic images of cancer, emphasizing brain and breast cancer. The functionality of the software will substantially transcend limitations of current analysis of cancer images, and will open the way for more precise and effective surgical planning as well as for more specific diagnosis of cancer based on its imaging characteristics, eventually leading to individualized medicine.

National Institute of Health (NIH)
National Cancer Institute (NCI)
Resource-Related Research Projects--Cooperative Agreements (U24)
Project #
Application #
Study Section
Special Emphasis Panel (ZCA1-TCRB-W (M1))
Program Officer
Redmond, George O
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Pennsylvania
Schools of Medicine
United States
Zip Code
Zhao, Xiaomei; Wu, Yihong; Song, Guidong et al. (2018) A deep learning model integrating FCNNs and CRFs for brain tumor segmentation. Med Image Anal 43:98-111
Li, Hongming; Galperin-Aizenberg, Maya; Pryma, Daniel et al. (2018) Unsupervised machine learning of radiomic features for predicting treatment response and overall survival of early stage non-small cell lung cancer patients treated with stereotactic body radiation therapy. Radiother Oncol 129:218-226
Binder, Zev A; Thorne, Amy Haseley; Bakas, Spyridon et al. (2018) Epidermal Growth Factor Receptor Extracellular Domain Mutations in Glioblastoma Present Opportunities for Clinical Imaging and Therapeutic Development. Cancer Cell 34:163-177.e7
Pisapia, Jared M; Akbari, Hamed; Rozycki, Martin et al. (2018) Use of Fetal Magnetic Resonance Image Analysis and Machine Learning to Predict the Need for Postnatal Cerebrospinal Fluid Diversion in Fetal Ventriculomegaly. JAMA Pediatr 172:128-135
Rathore, Saima; Bakas, Spyridon; Pati, Sarthak et al. (2018) Brain Cancer Imaging Phenomics Toolkit (brain-CaPTk): An Interactive Platform for Quantitative Analysis of Glioblastoma. Brainlesion (2017) 10670:133-145
Rathore, Saima; Akbari, Hamed; Doshi, Jimit et al. (2018) Radiomic signature of infiltration in peritumoral edema predicts subsequent recurrence in glioblastoma: implications for personalized radiotherapy planning. J Med Imaging (Bellingham) 5:021219
Davatzikos, Christos; Rathore, Saima; Bakas, Spyridon et al. (2018) Cancer imaging phenomics toolkit: quantitative imaging analytics for precision diagnostics and predictive modeling of clinical outcome. J Med Imaging (Bellingham) 5:011018
Han, Xu; Kwitt, Roland; Aylward, Stephen et al. (2018) Brain extraction from normal and pathological images: A joint PCA/Image-Reconstruction approach. Neuroimage 176:431-445
Rathore, Saima; Akbari, Hamed; Rozycki, Martin et al. (2018) Radiomic MRI signature reveals three distinct subtypes of glioblastoma with different clinical and molecular characteristics, offering prognostic value beyond IDH1. Sci Rep 8:5087
Zhu, Xiaofeng; Li, Hongming; Fan, Yong (2018) Parameter-Free Centralized Multi-Task Learning for Characterizing Developmental Sex Differences in Resting State Functional Connectivity. Proc Conf AAAI Artif Intell 2018:2660-2667

Showing the most recent 10 out of 35 publications