In-Vivo Targeting of Liver Lesions with a Navigation System Based on Fiducial Needles

Abstract

Abstract. We assessed the targeting precision of a previously reported needle-based soft tissue navigation system in-vivo. Eight 2 ml agar nodules were implanted into the livers of two swine as tumor models. One medical expert with experience in punctures and one non-expert used the navigation system to target each nodule twice. The targeting error, which includes the system error and the user error, was defined as the distance between the applicator tip and the center of gravity of the lesions in computed tomography (CT) control scans and was 5.3 ± 2.2 mm (n = 16) for the expert, 3.2 ± 1.3 mm (n = 16) for the non-expert, and 4.2 ± 2.1 mm (n = 32) averaged over both subjects. A robust targeting precision of the order of magnitude of 4 mm could improve the treatment standard for CT-guided minimally invasive interventions in the liver dramatically.

Comparison of freehand-navigated and aiming device-navigated targeting of liver lesions

“…Although this value is close to those obtained with the methods presented in this study, a direct comparison is difficult to make, since our experiment used a static phantom, while the one in (8) was made on live, breathing animals. The main advantage of the approach presented in (8) is the potential to remove the need for respiratory gating, and thus make needle insertions faster. However, the increased Comparison of freehand and aiming device navigation invasiveness, required to precisely track the respiratory motion by inserting needles near the target, represents an additional risk factor for the patient.…”

“…By inserting tracked needles in the liver of live swine, next to the target, in order to precisely track and correct the respiratory motion, a mean TPE of 4.2 AE 2.1 mm was achieved in (8). Although this value is close to those obtained with the methods presented in this study, a direct comparison is difficult to make, since our experiment used a static phantom, while the one in (8) was made on live, breathing animals. The main advantage of the approach presented in (8) is the potential to remove the need for respiratory gating, and thus make needle insertions faster.…”

“…In contrast, fiducial‐based manual registration can be performed in non‐sterile conditions before the intervention. The automated registration method leads to maximum FREs of 0.9 mm, similar to what is achieved in , where motion compensation is achieved by inserting needles around the target. However, the scenario considered here is a static one and this FRE might not reflect the registration accuracy that would be achieved on a breathing patient.…”

“…Conversely, Meier‐Hein et al . and Zhang et al . investigated the feasibility and accuracy of respiratory motion compensation during needle insertion through the insertion of tracked needles near the target.…”

“…None of these approaches take into account patient motion and assume that respiratory gating is used during the intervention, which may be sufficient to compensate for respiratory motion: Widmann et al (7) suggested that applying temporary disconnections of the endotracheal tube reduces the amount of liver deformation to within 4 mm. Conversely, Meier-Hein et al (8) and Zhang et al (9) investigated the feasibility and accuracy of respiratory motion compensation during needle insertion through the insertion of tracked needles near the target. Although these approaches have the potential to remove the need for respiratory gating, and thus make needle insertions faster, their invasiveness represents an additional risk factor for the patient.…”