Program for Innovation in Imaging Informatics - P(I)³

Yale Radiology is a preeminent department for the advancement of radiologic technologies and has always been at the forefront of these endeavors. The Program for Innovation in Imaging Informatics sits at the nexus of imaging innovation, operations, and quality patient care. In collaboration with Yale New Haven Health System (YNHH), this group drives initiatives across the Yale University-YNHH continuum with the goal of creating an ecosystem of technology drive change management.

Our mission is to drive a culture of continuous improvement and steward the missions of the radiology department, university, and YNHH delivery networks across the state. With this focus the P(I)3 team has driven advancements in many areas including being at the leading edge of artificial intelligence systems implementations and designing the reading room paradigm of the future.

Strategic goals for our team include standardizing key informatics processes, improving the experience of all in our community, and continuing to lead cutting edge initiatives and research that put Yale Radiology at the forefront of innovation across the world.

Yale Radiology is committed to sustainable innovation throughout our department. The YDR RadICL Program is designed to bring faculty lead ideas into fruition. We do this by creating an environment of innovation within our department and streamlining the processes to bring those ideas into reality. Through this initiative our faculty are partnered with resources across the university and health system to create sustainable tools that support our operational and patient care needs. We welcome industry partnerships within this endeavor.

If you are a faculty member with an idea submit here: RadICL Innovation Hub Idea Submission

Actively engage with members of our department to identify their interests and discuss how our informatics personnel and relationships can assist.

In collaboration with YDR residency program directors and the center for virtual simulation, develop IR, US, GI, and other training modules following the format of the successful contrast simulation program.

Expand and support remote reading to include trainees. Pilot home reading of Breast Imaging remote reading, and if successful, negotiate contracts and support remote reading for additional section members.

Develop an interactive program to assist radiologists with skills and confidence to interact with patients in a variety of clinical scenarios: Discussing reports, incidental findings, complications, radiation, contrast, pregnancy and image tests. Educational Program

1) Development of a workflow efficient PACS based automated brain tumor segmentation and radiomic feature extraction for clinical implementation

• Authors: Tej Verma

2) PACS-Integrated Lesion Tracking Tool Enables Efficient Lesion Measurement And Reveals High Proportion Of Metastatic Lesions Showing Mixed Response To Radiosurgery

• Authors: Gabriel Cassinelli, Tej Verma, Khaled Bousabarah, PhD, Leon Jekel, Sara Merkaj, Ryan Bahar, BS, Sandra Abi Fadel, MD, Ichiro Ikuta, MD, MingDe Lin, PhD, Antonio Omuro, MD, Mariam Aboian, MD/PhD

3) Applying a Glioma-Trained Deep Learning Automatic 3D Segmentation Algorithm for Primary CNS Lymphoma Segmentation - Preliminary Results

• Authors: Gabriel Cassinelli, Sara Merkaj, Khaled Bousabarah PhD, Leon Jekel, Sandra Abi Fadel MD, Ichiro Ikuta MD, MingDe Lin PhD, Anita Huttner MD, Antonio Omuro, MD, Mariam Aboian, MD/PhD

4) Are Deep Learning or Conventional Machine Learning Algorithms Best For Differentiating Gliomas from Primary CNS Lymphomas? - A Systematic Review

• Authors: Gabriel Cassinelli Petersen, Julia Shatalov, Tej Verma, Waverly Brim BS, Sara Merkaj, Ryan Bahar BS, Harry Subramanian, MD, Jin Cui PhD, Michele Johnson MD, Ajay Malhotra MD, Antonio Omuro MD, Mariam Aboian MD/PhD

5) A fully PACS integrated machine learning pipeline for automatic glioma segmentation and grade prediction

• Authors:Sara Merkaj, Tal Zeevi, Khaled Bousabarah, Eve Kazarian, MingDe Lin, Andrej Pala, Lawrence Staib, Leon Jekel, Ryan Bahar, Gabriel Cassinelli, Niklas Tillmanns, Jin Cui, Ichiro Ikuta, Sandra Abi Fadel, Jitendra Bhawnani, Irena Tocino, Mariam Aboian

6) Minimal studies needed to train novel PACS-based glioma auto-segmentation tool to translate to clinical practice

• Authors: Sara Merkaj, Khaled Bousabarah, MingDe Lin, Andrej Pala, Gabriel Cassinelli, Leon Jekel, Ryan Bahar, Niklas Tillmanns, Jin Cui, Ajay Malhotra, Jitendra Bhawnani, Irena Tocino, Mariam Aboian

7) TRIPOD analysis of machine learning studies in glioma segmentation

• Authors: Niklas J. Tillmanns, Avery E. Lum, Gabriel Cassinelli, Sara Merkaj, Tej Verma, Harry Subramanian MD, Ryan C. Bahar, Waverly Brim, Jan Lost, Leon Jekel, Sam Payabvash MD, Ichiro Ikuta MD, MingDe Lin PhD, Khaled Bousabarah M.Sc, Michele H. Johnson MD, Jin Cui, Ajay Malhotra MD, Bernd Turowski Prof. Dr., Mariam S. Aboian MD/PhD

8) Why is AI based glioma segmentation not in clinical practice? - A systematic review of the current literature

• Authors: Niklas J. Tillmanns, Avery E. Lum, Gabriel Cassinelli, Sara Merkaj, Tej Verma, Harry Subramanian MD, Ryan C. Bahar, Waverly Brim, Jan Lost, Leon Jekel, Sam Payabvash MD, Ichiro Ikuta MD, MingDe Lin PhD, Khaled Bousabarah M.Sc, Michele H. Johnson MD, Jin Cui, Ajay Malhotra MD, Bernd Turowski Prof. Dr., Mariam S. Aboian MD/PhD