Cancer Studies at World Molecular Imaging Society Annual Congress Bring New Research to the Forefront

Presentations of AI Tissue Discrimination, Immunotherapy Improvement and Cell-Tracking Technologies Provide Promising Steps

The recent World Molecular Imaging Society (WMIS) annual congress in Montreal provided a plethora of breakthrough research studies including several cancer related presentations. Three of the studies highlighted presentations on Artificial Intelligence (AI) to detect breast cancer, response prediction of cancer immunotherapies and whole body, real-time tracking of single cells. 
     The principle purpose of the study, Sensitivity and Specificity of a Machine Learning Artificial Intelligence System Trained to Detect Breast Cancer, was to investigate the ability of a convolutional neural network (CNN) to discriminate and classify between tissue types. The study was presented by Jeffrey Leal, Ph.D., research associate, Johns Hopkins University. 
         “We hypothesized that a unique tissue signature can be used to classify hypermetabolic tissue types, normal as well as disease, employing AI systems using deep learning based CNN,” said Dr. Leal. “In addition, we believed AI could be used to automate such tissue identification and classification with a high degree of sensitivity and specificity.” 
         Fifty-three baseline PET/CT studies from a multi-institutional study of breast cancer were used to train an AI system based on semantic segmentation. Images were processed using in-house, semi-automated software to generate voxel classification maps of all voxels within the study volume. The trained neural network demonstrated a global accuracy of 98.3 percent detecting 16 tissue class categories. The weighted Intersection of Union (IoU), a ratio of voxels correctly classified, was 97.2%, and the Mean Boundary F1 score (a measure of boundary accuracy) was 85.4percent. For individual tissue types, accuracy rates ranged between 76.7% and 100 percent. Review of the resulting auto-classified image sets allowed visual confirmation of the accuracy of the tissue detection, while also highlighting areas requiring further refinement by the evidence of noise and boundary discordance. 
         “The high-level of performance achieved by this CNN leads us to believe these systems can detect tracer-based tissue signatures, which provide a depth of signal analysis potentially beyond the ability of our classical first-order tools,” said Dr. Leal. “We believe this may provide a whole new dimension in our approach to the use and interpretation of molecular imaging in the future, providing insights into personalized therapy strategies, diagnosis and prognosis which heretofore eluded us.” 
         Another cancer study explored immunotherapies involving checkpoint inhibitors and T-cell transfer. ImmunoPET of the Early Activation Antigen CD69 Enables Response Prediction of Cancer Immunotherapies was presented by Bredi Tako, medical student, Werner Siemens Imaging Center, University of Tübingen. 
         Most cancer immunotherapies, such as immune checkpoint inhibitors (ICI), are based on the reinvigoration of antitumoral T-cell responses and have been implemented as standard of care in various cancers. T-cells express a high amount of the early activation antigen CD69 upon stimulation, which makes it an excellent candidate to monitor immune responses by molecular imaging. The aim of this study was to develop an antibody based immunoPET tracer targeting CD69 and to evaluate its potential use as a novel imaging tool to monitor T-cell activation and responses to ICI therapies. 
         “The purpose of this study was to establish a novel tracer for positron emission tomography (PET) that can be used for imaging of immune cell activation following immunotherapy,” said Mr. Tako. “To this end, we developed an antibody-based imaging method that targets the early activation antigen CD69 and applied it to murine tumor models that respond differently to immune checkpoint inhibitor therapies.” 
     According to Mr. Tako, the research indicated that ImmunoPET imaging of CD69 is a reliable method to distinguish responders from non-responders following immunotherapy for cancer treatment as early as three days after therapy initiation. 
     Early response evaluation by non-invasive imaging allows for a quick therapy modification, which can increase the therapeutic efficiency and outcome of many cancer patients. Since immune cell activation also plays a role in many other diseases, such as autoimmune diseases, non-invasive imaging might also be beneficial in diagnosing and improving treatments for these patients as well. 
         With the progress in cell-based therapies, molecular imaging methods are increasingly being used for in vivo cell-tracking applications. In the research In Vivo Real-Time Tracking of Single Cells in Whole-Body PET/CT, Kyungoh Jung, Ph,D., postdoctoral research fellow, Stanford University, spoke about the sensitivity of PET to track the migration of single cells in mice, in real-time and at the whole-body level. 
         “The purpose of the study was to demonstrate that PET can track much smaller cell populations, down to single cells,” said Dr. Jung. “Using nanoparticles to efficiently ferry radioactivity into cells, hardware optimizations, and a novel cell tracking algorithm, this study demonstrated real-time tracking of single cancer cell metastasizing hematogenously to the lungs.” 
         Dr. Jung believes this system could be used for understanding circulating tumor cells and cell clusters, to investigate biological processes such as cancer development and the early stages of the metastatic cascade. Furthermore, cell-tracking could be used to determine the pattern and kinetics of immune cell mobilization in response to an injury. Finally, the method may find use in clinical trials of cell-based therapies for cancer immunotherapy and regenerative medicine. 
         “In vivo molecular imaging tools are critically important to understand the role of cell trafficking in physiological and pathological processes,” Dr. Jung concluded. 
     “Metastasis, for instance, is the single most important factor for predicting cancer survival, and targeting the metastatic cascade could significantly improve clinical outcomes. This cell-tracking methodology could be applied to develop better therapies against cancer metastasis or to monitor early cell trafficking behavior in clinical trials of cell-based therapies.”

The WMIS is dedicated to developing and promoting translational research through multimodality molecular imaging. The education and abstract-driven WMIC is the annual meeting of the WMIS and provides a unique setting for scientists and clinicians with very diverse backgrounds to interact, present, and follow cutting-edge advances in the rapidly expanding field of molecular imaging that impacts nearly every biomedical discipline. Industry exhibits at the congress included corporations who have created the latest advances in preclinical and clinical imaging approaches and equipment, providing a complete molecular imaging educational technology showcase. For more information:

# # #