A framework for the correction of slow physiological drifts during MR-guided HIFU therapies: Proof of concept

Cornel Zachiu (1) , Baudouin Denis de Senneville (1, 2) , Chrit T. W. Moonen (1) , Mario Ries (1)

1 University Medical Center [Utrecht]
2 IMB - Institut de Mathématiques de Bordeaux

Purpose: While respiratory motion compensation for MR-guided HIFU interventions has been extensively studied, the influence of slow physiological motion due to, for example, peristaltic activity, has so far been largely neglected. During lengthy interventions, the magnitude of the latter can exceed acceptable therapeutic margins. The goal of the present study is to exploit the episodic work-flow of these therapies to implement a motion correction strategy for slowly varying drifts of the target area and organs at risk over the entire duration of the intervention. Methods: The therapeutic work-flow of an MR-guided HIFU intervention is in practice often episodic: Bursts of energy delivery are interleaved with periods of inactivity, allowing the effects of the beam on healthy tissues to recede and/or during which the plan of the intervention is re-optimized. These periods usually last for at least several minutes. It is at this time scale that organ drifts due to slow physiological motion become significant. In order to capture these drifts we propose the integration of 3D MR scans in the therapy work-flow during the inactivity intervals. Displacements were estimated using an optical flow algorithm applied on the 3D acquired images. A preliminary study was conducted on 10 healthy volunteers. For each volunteer 3D MR images of the abdomen were acquired at regular intervals of 10min over a total duration of 80min. Motion analysis was restricted to the liver and kidneys. For validating the compatibility of the proposed motion correction strategy with the work-flow of an MR-guided HIFU therapy, an in-vivo experiment on a porcine liver was conducted. A volumetric HIFU ablation was completed over a time span of 2h. A 3D image was acquired before the first sonication, as well as after each sonication. Results: Following the volunteer study, drifts larger than 8mm for the liver and 5mm for the kidneys prove that slow physiological motion can exceed acceptable therapeutic margins. In the animal experiment motion tracking revealed an initial shift of up to 4mm during the first 10min and a subsequent continuous shift of ∼2mm/h until the end of the intervention. This leads to a continuously increasing mismatch of the initial shot planning, the thermal dose measurements and the true underlying anatomy. The estimated displacements allowed correcting the planned sonication cell cluster positions to the true target position, as well as the thermal dose estimates during the entire intervention and to correct the NPV-measurement. A spatial coherence of all three is particularly important to assure a confluent ablation volume and to prevent remaining islets of viable malignant tissue. Conclusions: This study proposes a motion correction strategy for displacements resulting from slowly varying physiological motion that might occur during an MR-guided HIFU intervention. We have shown that such drifts can lead to a misalignment between interventional planning, energy delivery and therapeutic validation. The presented volunteer study and in-vivo experiment demonstrates both the relevance of the problem for HIFU therapies and the compatibility of the proposed motion compensation framework with the work-flow of a HIFU intervention under clinical conditions.

Domaines: Traitement du signal et de l'image [eess.SP]

Cornel Zachiu, Baudouin Denis de Senneville, Chrit T. W. Moonen, Mario Ries. A framework for the correction of slow physiological drifts during MR-guided HIFU therapies: Proof of concept. Medical Physics, 2015, 42 (7), pp.4137-4148. ⟨hal-01159816⟩