Localization accuracy of two electromagnetic tracking systems in prostate cancer radiotherapy: A comparison with fiducial marker based kilovoltage imaging

Highlights

• RayPilot transmitters may migrate within the prostate between treatment fractions.

• Localization accuracy of the RayPilot suffers from instability of the transmitters.

• Localization accuracy of Calypso is comparable to kV imaging of fiducials.

Abstract

The aim of this study was to evaluate the localization accuracy of electromagnetic (EM) tracking systems RayPilot (Micropos Medical AB) and Calypso (Varian Medical Systems) in prostate cancer radiotherapy. The accuracy was assessed by comparing couch shifts obtained with the EM methods to the couch shifts determined by simultaneous fiducial marker (FM) based orthogonal kilovoltage (kV) imaging. Agreement between the methods was compared using Bland-Altman analysis. Interfractional positional stability of the FMs, RayPilot transmitters and Calypso transponders was investigated. 582 fractions from 22 RayPilot patients and 335 fractions from 26 Calypso patients were analyzed. Mean (± standard deviation (SD)) differences between RayPilot and kV imaging were 0.3 ± 2.2, −2.2 ± 2.4 and −0.0 ± 1.0 mm in anterior-posterior (AP), superior-inferior (SI) and left-right (LR) directions, respectively. Corresponding 95% limits of agreement (LOA) were ±4.3, ±4.7 and ±2.1 mm around the mean. Mean (±SD) differences between Calypso and kV imaging were −0.2 ± 0.6, 0.1 ± 0.5 and −0.1 ± 0.4 mm in AP, SI and LR directions, respectively, and corresponding LOAs were ±1.3, ±1.0 and ±0.8 mm around the mean. FMs and transponders were stable: SD of intermarker and intertransponder distances was 0.5 mm. Transmitters were unstable: mean caudal transmitter shift of 1.8 ± 2.0 mm was observed. Results indicate that the localization accuracy of the Calypso is comparable to kV imaging of fiducials and the methods could be used interchangeably. The localization accuracy of the RayPilot is affected by transmitter instability and the positioning of the patient should be verified by other setup techniques. The study is part of clinical trial NCT02319239.

Introduction

Accurate treatment localization is crucial part of modern radiotherapy. Its importance is emphasized in prostate radiotherapy as the prostate position is known to change relative to bony structures and skin surface between the fractions (interfraction motion) [1], [2], [3] and during the treatment (intrafraction motion) [4], [5], [6]. The motion is mainly caused by variations in rectum and bladder filling and it can range up to 1–2 cm during the treatment course [7], intrafraction motion being generally smaller in extent, typically less than 5 mm [5], [6]. Inaccuracies in the localization may deteriorate the target dose distribution [8]. Uncertainties in treatment localization, target delineation and treatment delivery are covered by planning target volume (PTV) margins around the clinical target volume (CTV). However, PTV margins expand the irradiated volume thus exposing larger volumes of nearby critical organs to high doses. High doses in rectum and bladder (especially trigone region) has been associated with late rectal bleeding [9] and incidence of acute and late urinary toxicity [9], [10], respectively. Uncertainties in localization, and thus PTV margins, could be reduced by accurate and frequently performed localization methods.

Several localization methods, such as ultra-sound (US), kilovoltage (kV) and megavoltage (MV) imaging, cone beam computed tomography (CBCT) and the use of fiducial markers (FM) have been applied in image-guided radiotherapy (IGRT) of the prostate. US methods increase the localization accuracy when compared to positioning based only on skin marks [11] but clinically unacceptable accuracy have also been reported [12]. US probe pressure can cause prostate displacement if handled without care, which have been seen both for transabdmominal [13] and transperineal probes [14]. Several studies have compared the accuracy of US localization methods to kV and MV imaging using FMs and found latter methods outperforming the US [11], [15], [16]. MV and kV imaging as such have poor soft tissue contrast and positioning of the patient is based on alignment of bony structures. CBCT imaging provides three-dimensional (3D) information of the imaging volume but the poor soft tissue contrast limits the accuracy of identifying the prostate from surrounding tissues. FMs implanted into the prostate provide a surrogate for prostate position, as they are made of radiopaque material and are discerned well in kV or MV images. Implantation of the FMs is made prior to treatment planning and the localization is based on alignment of the markers in reference images and kV or MV images (2D match) or markers in planning CT and CBCT images (3D match). FMs have been in use since the 1990s and the feasibility of FMs in the IGRT of the prostate has been shown in many studies [2], [17], [18]. Imaging of FMs has proven to be reliable localization method having accuracy on the order of ≤1 mm [19] which reduces the systematic positioning error [20] and enables the margin reduction [21], [22]. The imaging of FMs together with 3D soft tissue analysis is currently the most effective and widely available localization technique in the IGRT of the prostate [23] and is established practice in many radiotherapy clinics.

To monitor continuously the organ intrafraction motion a real-time electromagnetic (EM) localization system, Calypso (Varian Medical Systems), was introduced [24], [25]. Calypso consists of three transponders implanted into the prostate and an EM source/receiver array which is setup above the patient during the treatment. Calypso provides real-time 3D position information of the three transponders and thus the prostate and can be used for localization and intrafraction motion tracking. Another EM tracking system currently in the market is RayPilot (Micropos Medical AB), which consists of a wired transmitter, implanted into the prostate for the treatment course and removed afterwards, and a detector array which is setup on the treatment couch [26]. RayPilot provides the 3D position of the transmitter in real-time and can also be used in localization and tracking of the prostate. EM systems provide a non-ionizing alternative to kV- and MV-imaging systems and high sampling rate of the systems guarantees that positional information of the prostate is not missed during the treatment.

The localization accuracy of Calypso have been investigated in many studies both in laboratory conditions [19], [25], [27], [28], [29] and in clinical situations [4], [5], [30], [31], [32], [33] whereas there is a lack of information about the localization accuracy of RayPilot in clinical environment. Kindblom et al. [26] reported submillimeter accuracy of RayPilot in phantom fixture and compared 3D positional difference between RayPilot and orthogonal X-ray imaging when the transmitter was implanted in a urethral catheter. However, urethral catheter implantation does not represent the intended application method of implanting the transmitter into the prostate gland.

The aim of this study is to investigate the localization accuracy of the RayPilot and Calypso systems in comparison with fiducial marker based localization for a group of clinically treated prostate cancer patients. This is accomplished by comparing the translational couch shifts, proposed by the EM systems and couch shifts proposed by FM based kV-image alignment acquired simultaneously with the EM systems. Couch shifts represent the isocenter offsets recorded by the localization method in question. The stability and migration of the RayPilot transmitter, the gold seed fiducials and Calypso transponders are evaluated as well. To our knowledge, this is the first large study investigating the accuracy of the RayPilot system in clinical environment. The study is part of a clinical trial (ClinicalTrials.gov NCT02319239) which aims at developing extremely hypofractionated treatment protocol for prostate cancer.