Get the Flash Player to see this rotator.


Robotics     UMD

Search robotics news archives

Bookmark and Share


Robotic Haptic Feedback System for Bx/RFA of Breast Tumor
under Continuous MRI

Main Participants

J. P. Desai, R. Gullapalli, U. Tan, B. Yang, and A. McMillan


This project is sponsored by NIH.


MRI-Compatible robotics, Breast biopsy (Bx), Radiofrequency ablation (RFA), Pneumatic actuation, Teleoperation, Haptic feedback, MRI-compatible force sensor


In 2009, The American Cancer Society estimated that 192,370 women would be diagnosed with invasive breast cancer and in addition 62,280 new cases of in situ breast cancer would be detected. One in 8 women born today are likely to be diagnosed with breast cancer during their lifetime. Although these statistics are discouraging, positive trends are evident as a result of innovations in diagnosis and treatment over the past decade.

Recent, large-scale studies reported in The Lancet and The New England of Journal of Medicine demonstrate the value of MRI as an effective tool in the diagnosis of breast cancer. Hence, coupling diagnosis with MRI based biopsy (Bx) will lead to better delineation of the tumor margin. For breast cancer treatment, radiofrequency ablation (RFA) has emerged as a promising approach for early stage breast cancer with maximum effectiveness and conservation of healthy breast tissue without full surgical intervention.

While Ultrasound (US) and Computed tomography (CT) is the imaging modality for RFA, they are not without their drawbacks. In US, there is limitation on the precision with which a needle placement can be performed and the image distortion due to the formation of microbubbles during RFA obscures the visualization of the tumor. CT lacks the ability of providing soft tissue contrast as good as MRI and also exposes the patients to radiation dose. MRI on the other hand provides excellent soft-tissue contrast, no ionizing radiation, and accurate thermographic maps at the ablation site.

Although the utility of RFA has been demonstrated in treating several types of lesion, the recurrence rate of tumors has not been as low as one would desire (8-47%). The recurrence of cancer after RFA or misdiagnosis in a breast biopsy using US, CT, or MRI could be a result of planning the trajectory towards the tumor based on images obtained earlier and not under continuous MRI guidance. Although a subjective evaluation is performed after the placement of the needle, even small errors in the trajectory could lead to sub-optimal results for biopsy and RFA. Hence, we are working towards developing a novel robotic approach with haptic feedback under continuous MRI for performing Bx/RFA of breast tumor. This work is in collaboration with the University Of Maryland School of Medicine.


The objectives of this project are:

1. Development of a controller for pneumatic actuation. The use of traditional electrical motor is prohibited in strong magnetic field and hence MRI-compatible actuation has to be pursued. Pneumatic actuation is chosen over hydraulic actuation because hydraulic systems can suffer from cavitation and fluid leakage.

2. Development of a 3-DoF MRI-compatible force sensor. A number of conventional force sensors are not compatible for use in MRI since they affect the quality of the image. One popular solution is to develop an optical force sensor. An MRI-compatible force sensor is crucial for providing haptic (sense of touch) feedback to the physician. We believe that the integration of haptic interface will improve the success rate of the teleoperated procedure by enabling the physician to characterize the tissue type (normal, fat, or cancerous, for example) during Bx/RFA of breast tumor.

3. Development of multi-DoF MRI-compatible compact robotic system. We envision building a MRI compatible robotic system that is compact enough to fit in the bore of the MRI and perform interventional procedures on the breast under teleoperated guidance from the physician located outside the MRI room.

Overview of Approach

fig1Pneumatic Controller: The strong magnetic field limits the actuation techniques that can be used in MRI. Though ultrosonic/Piezo motors, hydraulic actuators, and pneumatic actuators can be used in MRI, hydraulic actuators and pneumatic actuators normally do not produce significant image artifact in MRI. Since hydraulic actuation suffers from cavitation and fluid leakage, we are currently pursuing pneumatic actuation for our robotic system in MRI. Precise position control of such a pneumatic system is very challenging. Commercially available pneumatic valves are not MRI-compatible and hence they have to be placed a significant distance away from the bore of the magnet. On the other hand, significant pressure has to be transmitted along long transmission lines to actuate the pneumatic actuator in the MRI. This induces time-delay and finite time is required for pressure build up in the cylinder chamber. The problem is further complicated by the presence of highly non-uniform friction during the range of motion of the device. In our preliminary work, we have evaluated adaptive friction compensation controllers and sliding mode control techniques. Fig. 1 shows our current 1-DoF pneumatically actuated needle driver prototype for use in MRI. Through careful modeling of the various components in the pneumatic system, we have demonstrated that the device has excellent MRI-compatibility and produces negligible image artifacts during precise position control tasks.

fig23-DoF Optical Force Sensor: Optical approach is pursued here for MRI-compatibility and an overview of a typical optical force sensor is shown in Fig. 2. The light source emits light via the optical cable attached to the force sensor. At the same time, there is an optical cable at the receiving end to transmit the light back into the control room where all the non-MRI compatible equipment is stored. Through this approach, electrical wires, which act as antennas and pick up RF signals and corrupt the image quality, are thereby eliminated. In our current prototype, a force on the loading point will cause a deformation in the elastic frame structure, resulting in a displacement in the reflector. This displacement will change the intensity of the reflected light. Hence, by monitoring the reflected light intensity, the force acting on the loading point can be computed. A topology optimization technique is used to design the elastic frame structures for maximizing the sensitivity of the force sensor during loading in all three principal directions.

fig3A prototype of the 3-DoF force sensor has been built and is shown in Fig. 3.

Multi-DoF compact robotic system for Bx/RFA of breast tumor: A multi-DoF pneumatically actuated device is currently being designed for breast biopsy/ RFA of breast tumor under continuous MRI. The device will be built with high strength polymer materials to achieve MRI-compatibility. Taking into account the limited space inside the MR bore, a parallel mechanism is chosen as the kinematic design for the device. Parallel mechanism is also known for its rigidity and is preferred over serial mechanism because strong metals like steel cannot be used in this project. This robotic device will be teleoperated by a master device placed inside the control room.

Relevant publications

U-Xuan Tan, Bo Yang, and Jaydev P. Desai, “Design and Development of a 3-Axis MRI-compatible Force Sensor,” IEEE International Conference on Robotics and Automation, Alaska, USA, May 2010.

Rebecca Kokes, Kevin Lister, Rao Gullapalli, Bao Zhang, Alan McMillan, Howard Richard, and Jaydev P. Desai, “Towards a teleoperated needle driver robot with haptic feedback for RFA of breast tumors under continuous MRI,” Medical Image Analysis, vol. 13, no. 3, pp. 445-455, June, 2009.

Rebecca Kokes, Kevin Lister, Rao Gullapalli, Bao Zhang, Howard Richard, and Jaydev P. Desai, “Towards a Needle Driver Robot for Radiofrequency Ablation of Tumors under Continuous MRI,” IEEE International Conference on Robotics and Automation, Pasadena, CA, USA, 2008.


Prof. Jaydev P. Desai
Director—Robotics, Automation, and Medical Systems Laboratory
Department of Mechanical Engineering
0160 Glenn L. Martin Hall
University of Maryland College Park, MD-20742
Phone: 301-405-4427
Email: jaydev (at)


Back to top      
Clark School Home UMD Home ISR Home