Quantifying Post- Laser Ablation Prostate Therapy Changes on MRI via a Domain-Specific Biomechanical Model: Preliminary Findings

Abstract

Focal laser ablation destroys cancerous cells via thermal destruction of tissue by a laser. Heat is absorbed, causing thermal necrosis of the target region. It combines the aggressive benefits of radiation treatment (destroying cancer cells) without the harmful side effects (due to its precise localization). MRI is typically used pre-treatment to determine the targeted area, and post-treatment to determine efficacy by detecting necrotic tissue, or tumor recurrence. However, no system exists to quantitatively evaluate the post-treatment effects on the morphology and structure via MRI. To quantify these changes, the pre- and post-treatment MR images must first be spatially aligned. The goal is to quantify (a) laser-induced shape-based changes, and (b) changes in MRI parameters post-treatment. The shape-based changes may be correlated with treatment efficacy, and the quantitative effects of laser treatment over time is currently poorly understood. This work attempts to model changes in gland morphology following laser treatment due to (1) patient alignment, (2) changes due to surrounding organs such as the bladder and rectum, and (3) changes due to the treatment itself. To isolate the treatment-induced shape-based changes, the changes from (1) and (2) are first modeled and removed using a finite element model (FEM). A FEM models the physical properties of tissue. The use of a physical biomechanical model is important since a stated goal of this work is to determine the physical shape-based changes to the prostate from the treatment, and therefore only physical real deformations are to be allowed. A second FEM is then used to isolate the physical, shape-based, treatment-induced changes. We applied and evaluated our model in capturing the laser induced changes to the prostate morphology on eight patients with 3.0 Tesla, T2-weighted MRI, acquired approximately six months following treatment. Our results suggest the laser treatment causes a decrease in prostate volume, which appears to manifest predominantly at the site of ablation. After spatially aligning the images, changes to MRI intensity values are clearly visible at the site of ablation. Our results suggest that our new methodology is able to capture and quantify the degree of laser-induced changes to the prostate. The quantitative measurements reflecting of the deformation changes can be used to track treatment response over time.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist. Robert Toth, Ph.D., is currently the managing member and majority owner of Toth Technology LLC, and declares no competing interest regarding this publication. This does not alter the authors’ adherence to PLOS ONE policies on sharing data and materials.

Figures

Fig 1. Overview of the registration techniques used to bring the pre-, post-treatment into spatial alignment.

The post-treatment bladder, prostate, and rectum are shown in green, blue, and pink, respectively. The pre-treatment surfaces are shown in grey. The first step is to use a linear registration to account for patient alignment. Secondly, a finite element model (FEM) is used to calculate the deformation due to the bladder and rectum, and removes those deformations. The preceding two steps are necessary in order to remove confounding deformations and isolate the treatment-induced changes to the gland. Finally, a second FEM calculates the remaining deformations on the prostate. These deformations can therefore be assumed to be primarily (if not exclusively) due to shape-based changes from the ablation. The quantitative measurements reflecting such changes can be used to track treatment response over time.

Fig 2. Effect of physical parameters on the FEM used to compensate for the bladder/rectum motion.

The X-axis represents the Young’s modulus (in log-scale). The Y-axis represents the Dice similarity coefficient between the pre-treatment, undeformed prostate (CPreP) and the prostate after inducing, and then removing, a synthetic simulation of the bladder and rectum filling (T^2(C˜PostP)).

Fig 3. Illustration of the ablation related effects on the prostate, bladder, and rectum modeled by the FEM.

((a)) represents the bladder and rectum in pale blue, and prostate in teal. ((b)) represents the synthetic deformation T2. The deformed prostate is shown in yellow, and the arrows in ((b)) represent the direction of the transform (mostly due to the rectum). ((c)) represents the result of the recovered deformation T^2. The deformed prostate is shown in semi-transparent yellow in ((b)) and ((c)). The high level of overlap between the semi-transparent yellow prostate and the teal prostate in ((c)) shows the accuracy of removing the simulated effects of the bladder and rectum motion.

Fig 4. The prostate volume across all eight patients both before and after treatment.

The median change was a 5% decrease in volume following focal laser ablation treatment.

Fig 5. Results for three patients (one per column).

The first row represents the pre-treament MRI scan (IPre). The second row represents the location of the laser during treatment. The third row represents the post-treament MRI scan (IPost). The fourth row represents a heat map of the ablation induced deformations T3. White represents regions of large deformations (2.2 mm), while transparent red represents regions of small deformations (0 mm). Small arrows represent the direction of the deformation (in all cases pointing towards the centroid of the prostate) after removing deformations due to patient alignment (T1) and surrounding tissues (T2). It can be seen that in all patients, the areas with the the largest deformations correspond to the focal laser ablation sites.

Fig 6. Illustration of the focal laser ablation needle locations on T2-w MRI during the procedure for two different patients ((a), (c)).

After registration, the difference between the MRI intensity values (as a relative percent change) are shown as colored values in ((b)) and 3D ((d)), where cool blue colors represent regions corresponding to small differences (0%), while hot red colors represent regions of large differences (11%) in MRI intensity values. Most of the hot colors are correlated with the needle locations (shown as orange arrows).