Dr. Patel has worked on developement and clinical/preclinical evaluation of various robotic systems for MRI-guided cancer diagnosis and treatment and safety in robot-assisted retinal surgeries using machine learning. INSPIRE Lab will be exploring similar research problems in X-Ray/CT/Ultrasound guided robotic systems and their evaluation in real cilnical environment and developement of next generation technologies in minimally invasive robotic surgeries. If this sounds like a place where you would like to be at, please contact Dr. Patel for more information.
This project aims at development of a skull-mounted, four-degree-of-freedom (4-DOF) sliding arch mechanism robot for Computed Tomography (CT)-guided burr hole neurosurgery procedures such as biopsy and ventriculostomy. The robot’s design is optimized to be compact and lightweight, allowing it to be directly attached to the patient’s skull without needing a bulky stereotactic frame. Though the mounting mechanism for the presented robot is optimized for neurosurgery procedures, it can be adapted to assist various image-guided, needle-based percutaneous interventions by replacing the mounting mechanism. Only CT-compatible components are utilized for the robot mechanism and the control system. This work presents the mechanical design, robot kinematics, calibration procedure, and preliminary bench-top accuracy assessment of the 4 DOF robotic manipulator. The manipulator was evaluated for targeting accuracy in a benchtop environment, showing an average translational error of 2.66 mm and 3.85 mm in the X and Y axes, respectively. Achieved targeting accuracy is sufficient for performing biopsies for tumors larger than 5 mm and ventriculostomy.
Duration: Dec 2021 : May 2023
Funding Agency: BIRAC
Amount: 44.57 Lakh
Reinforcement Learning for Surgical Planning
Path planning algorithms for minimally invasive neurosurgery involve avoiding critical structures such as blood vessels and ventricles while following needle kinematics. The majority of planning solutions proposed in the literature use sampling-based algorithms. This paper introduces a Flexible Needle Path Generation framework with Non-Holonomic constraints (FNPG-NH), an extension of our FNPG framework. FNPG-NH uses deep Reinforcement Learning (RL) based methods such as Deep Deterministic Policy Gradient (DDPG), Proximal Policy Optimization~(PPO), and Soft Actor-Critic (SAC) to obtain a kinematically feasible path for a bevel-tipped flexible needle using a nonholonomic model. RL algorithms presented in this work generate the control input for needle rotation based on the rewards generated by the environment. The deep RL algorithms are trained on an environment that consists of (1) ventricles segmented from T1 images of the healthy volunteers using atlas-based segmentation, (2) blood vessels segmented from MRA volumes of the same volunteer using thresholding, and (3) tumor volume from labeled BraTS 2020 dataset and placed at an anatomically relevant location. The paths generated by the reinforcement learning algorithm and the traditional sampling-based algorithm RRT are compared for various performance metrics. The reinforcement learning model was trained on 20 volumes and validated on 68 volumes, and RRT was evaluated on the same 68 validation volumes. The results show that the trajectories generated by the FNPG-NH framework are safer, shorter, and take less time than RRT while avoiding critical structures such as ventricles and blood vessels.
Duration: Dec 2021 : Dec 2023
Funding Agency: IIT Madras
Amount: 5.00 Lakh
Ultrasound Guided Liver Biopsy Robot
Cancers of abdominal organs such as liver, lung, stomach, and kidney are among the leading cause of death due to cancer in India. In most cases, these are identified by performing a biopsy procedure under image (Ultrasound or Computed Tomography) guidance. While CT imaging provides larger field of view, it results into exposure to radiation and requires a dedicated imaging facility to perform the biopsy procedure under CT-guidance. On the other hand, ultrasound imaging systems are portable, do not require large operating room and provides real-time imaging capabilities. Performing biopsy of abdominal organs under ultrasound guidance could extend the benefits of image-guided biopsy procedure to larger population. However, ultrasound-guided biopsy procedures are challenging and requires extensive experience in interventional radiology and often requires multiple needle insertions before a good quality tissue sample could be acquired. A robotic system that could precisely align the biopsy needle to desired target lesion identified in a real-time ultrasound image could alleviate these challenges and result in shorter procedure time, a smaller number of needle insertion attempts and make it available to larger population. The overarching aim of this project is to develop and evaluate an ultrasound-guided robotic system for minimally invasive percutaneous interventions such as biopsy, brachytherapy, and ablation. The ultrasound-guided robotic system will provide precise placement of diagnostic/therapeutic instruments into the target tumor/tissue in the abdominal regions (liver, lung and kidney etc) and enable a streamlined workflow for minimally invasive procedures under real-time ultrasound guidance. Following are the objectives of the proposal. 1:Develop an affordable ultrasoundguided robotic device with precise targeting in desired region of the target organ. 2: Develop an embedded control system for precise motion control of the robot. 3:Create an intuitive surgical planning and navigation application that would eliminate manual guesswork and human error. 4:Provide 3D visualization of the surgical scene and clinical workflow management for shorter procedure times. Successful completion of the proposed project will result in better cure for the patient, shorter procedure time for the physician/hospital and make minimally invasive US guided percutaneous procedures more accessible to lesser privileged population.
