This paper presents a concept of automatic needle placement for brachytherapy. For the success of this minimally invasive treatment, a precise and safe placement of needles inside soft tissue is fundamental. The presented concept incorporates information about the needle as well as the tissue to find a suitable needle trajectory. A patientspecific tissue model is derived from different imaging modalities and updated during the insertion. Essential for this concept is that during the robotic placement of the needle, it is continuously verified if proceeding is safe.
Copyright c by the paper’s authors. Copying permitted for private and academic purposes. the outer patient anatomy, it has not reached the desired depth inside the tumor. Second, when the needle is advanced until the tissue layer ruptures, the resulting relaxation may leave the needle to deep inside the tissue. This may lead to unwanted side effects, e.g., puncturing of the bladder during prostate brachytherapy. 2
Conventionally, the needles are inserted manually by trained physicians. Often, a rigid template is used and the needle is primarily advanced be the physician using some sort of image guidance, e.g., ultrasound or, less typical, magnetic resonance imaging (MRI). Either image modality is not free of artifacts and often it is necessary to move the needle back and force multiple times before a final position is reached. Nevertheless, due to the deformations and needle deflection this may not resemble the initially desired placement. Hence, it is common practice to first insert the needles and subsequently identify the position of all needles in images of the target region, and then to optimize the dose delivery. Over the past decade, the idea of robotic brachytherapy has been studied by a number of groups [PBC+14][LHB+12][PNY07]. The promise is a fully optimized and fully automatic treatment, where the needle positions are determined by an optimization algorithm and placed in exactly the same arrangement with respect to the anatomy using an autonomous robotic needle driver. Potential advantages include a further reduction of side effects and the use of few needles to focally treat tumors in an early stage. However, one of the key challenges is the safe insertion of the needles, i.e., in the desired depth within the target while avoiding damage to surrounding structures. The proposed setup resembles an MCPS as illustrated in Figure 5, with the control loop including the robot and its actuators, sensors and imaging, and the tissue inside the patient. Clearly, the robotic insertion is subject to similar difficulties as the manual procedure, i.e., tissue deformation and needle deflection. However, unlike in manual placement the robotic motion of the needle can be more precisely adapted to the sensor readings. For example, we include optical fibers in the needle to obtain high resolution images of the tissue structure in front of the needle tip using optical coherence tomography (OCT). Doppler OCT also allows estimating tissue motion and deformation [OHS12]. We are working on integrating this information into a tissue model, which in turn will be used to estimate and predict the tissue behavior for different needle motion, e.g., to distinguish loading deformation, rupture and cutting through tissue. Hence, the needle driving will be optimized with respect to the tissue properties. Initially, deviations between planned and measured needle trajectory are compared and large errors will lead to stopping and an operator decision whether to proceed or retract. Estimates of the accuracy of the model and impact of failures should ultimately be checked during the procedure. The concept is summarized in Figure 4, Figure 3 shows our experimental setup. We have summarized a concept for automatic needle placement incorporating sensor information, tissue modelling and actuator control to maintain fail safety. Extensions, e.g., online model checking can be implemented, as all sensor data processing, motion planning, and control is realized in software.
Acknowledgments The authors wish to thank Dr. med. Gy¨orgy Kov´acs and Dr. Frank-Andr´e Siebert for helpful discussions regarding HDR brachytherapy and B. Sc. Philipp Koch for providing Figure 1. This work was partially funded by through DFG grant SCHL 1844/2-1. [Koc13]
P. Koch. Experimentelle untersuchung von kra¨ften und deformationen beim einbringen von nadeln in weichgewebe. Bachelor’s thesis, University of Lu¨beck, 2013.