2023 Welcome from the MnRI Director (Nikos)

First, we wish you all a healthy and prosperous 2023. As a new year starts, with hopes that new computational tools will revolutionize healthcare, this column focuses on robotics and automation for cancer treatment. Computer vision, robotics, and machine learning continue to revolutionize cancer treatment, with pertinent activities ranging from assisting in surgeries to defining tumors and blood vessels and enabling minimally invasive procedures.

One area in which MnRI is very active is Cancerous Tissue Recognition (CTR). Segmentation methodologies are continuously integrating advancements at the forefront of machine learning and computer vision, providing a variety of inference schemes for data such as images (e.g., CT scans, histopathological data). Histopathological data, in most cases, come in the form of high-resolution images, making methodologies operating at the patch level more computationally attractive. Such methodologies capitalize on pixel-level annotations (tissue delineations) from expert pathologists, which are then used to derive labels at the patch level. The ultimate objective is a digitally connected health system that augments clinician capabilities by providing powerful feature descriptors that may describe malignant regions. The team working on this problem consists of Panagiotis Stanitsas, Anoop Cherian, Vassilios Morellas, Resha Tejpaul, Nikos Papanikolopoulos, and Alexander Truskinovsky.

In work related to kidney cancer, Nick Heller, Resha Tejpaul, and Chris Weight are looking to accelerate the development of reliable kidney and kidney tumor semantic segmentation methodologies, potentially fueling the development of other AI-powered classification tools. According to the American Cancer Society, kidney cancer is among the 10 most common cancers. The same organization estimates that there were 79,000 kidney cancer cases in the United States in 2022 and around 14,000 deaths. One usual form of kidney cancer is Renal Cell Carcinoma (RCC). The research team uses data from hundreds of kidney cancer patients to train a deep neural network to assist in the segmentation of malignancies in the CT scans of the patients. The potential of this work also lies in the creation and use of new RENAL nephrometry scores that can assist clinicians in determining proper treatment.

One company in this area with which we are fortunate to collaborate is Histosonics, which focuses on histotripsy delivered by a robotic mechanism to destroy diseased tissue. The method is non-thermal and operates at the sub-cellular level. Histosonics currently employs several of our MS in Robotics graduates and students. MnRI is looking for more partners in this area and will continue investing in this broad research line.

Have a great academic year,
Nikos Papanikolopoulos
Minnesota Robotics Institute Director