MRI Guided Surgical Planning and 3D Procedure Simulation for Partial Gland Cryoablation of the Prostate


 Purpose This study reports on the development of a novel 3D surgical planning technique to provide pre-ablation treatment planning for partial gland prostate cryoablation (cPGA). Methods Consecutive patients with magnetic resonance imaging (MRI)-visible (PI-RADS v2 score ≥ 3) and biopsy confirmed prostate cancer undergoing partial gland cryoablation (cPGA) (n=20) were prospectively enrolled in an IRB approved study investigating advanced methods of data visualization for treatment planning. All patients enrolled in the study underwent image segmentation of the prostatic capsule, index lesion, urethra, rectum, and neurovascular bundles based upon multi-parametric MRI data. Segmented imaging data were converted to 3D, 3D prostate models were viewed in computer-aided design (CAD) software, and virtual cryotherapy probes with -40°C isotherm volumes were created. Pre-treatment 3D prostate cancer models derived from MRI data were utilized to predict the number of and placement of cryotherapy probes to achieve confluent treatment volume. Treatment efficacy was measured with 6 month post-operative MRI, serum prostate specific antigen (PSA) at 3 and 6 months, and treatment zone biopsy results at 6 months were evaluated. Outcomes from 3D planning were compared to outcomes from a series of 20 patients undergoing cPGA using traditional 2D planning techniques. Results The number of cryotherapy probes utilized matched the 3D plan in 16/20 (80%) of patients. 3D planning predicted a greater number of cryoprobes than 2D planning. Treatment zone biopsy was performed in 13/17 patients in the 3D cohort and was negative in 12 of these (92.9%). For the 3D group, 6 month biopsy was not performed in 4 (20%) due to undetectable PSA, negative MRI, and negative MRI Axumin PET. For the group with traditional 2D planning, treatment zone biopsy was positive in 3/14 (21.4%) of the patients, p = 0.056). Conclusions 3D prostate cancer models derived from mpMRI data provide novel guidance for planning confluent treatment volumes for cPGA. This study prompts further investigation into the use of 3D treatment planning techniques.


Introduction
The relatively recent introduction of multiparametric magnetic resonance imaging (mpMRI) in the 3 diagnostic paradigm for prostate cancer provides a critical improvement in disease localization; mpMRI of the prostate has emerged as the major imaging modality utilized to identify and characterize clinically significant prostate cancer [1][2][3][4]. Coupling mpMRI with targeted prostate biopsy [3,5,6]. Furthermore, this technique improves understanding of disease localization and extent within the gland, thus opening the possibility of targeted treatment via prostate gland ablation (PGA).
Although mpMRI can adequately identify prostate lesions, one major disadvantage of focal therapy for prostate cancer is that it is difficult to accurately determine the "kill" zone. While mpMRI accurately identifies disease location, it may underestimate precise tumor extent by approximately 30% [7].
Working from radical prostatectomy specimens, Le Nobin et al reported that a treatment margin of approximately 13 mm around image visible disease was needed to ensure adequate disease capture [8].
Cryoablation of the prostate utilizes placement of trans-perineal needles (cryoprobes) under ultrasound guidance. Cryoprobes are placed into the prostate tissue using 2D image guidance for localization of prostate tumor. Tissue ablation is achieved through controlled development of an isotherm volume which is based upon the endothermic expansion of argon gas within the cryoprobe.
The 3D size of the isotherm is limited by the length of the cryoprobe and coagulative tissue necrosis is executed through two rapid freeze-thaw cycles. Achieving a temperature of -40°C ensures tissue destruction [16]. Monitoring of the ice ball and treatment temperatures is achieved via 2D ultrasound imaging and select placement of thermocouples. The success of partial gland prostate cryoablation (cPGA) depends upon the development of a 3D ablation volume that entirely encompasses the tumor and its margin within a zone of at least − 40 °C. Utilizing the correct number of cryoprobes in the correct spatial orientation is necessary to achieve this goal [9].
Manual treatment planning typically includes reviewing the pre-operative mpMRI to plan critical steps of the surgical procedure and anticipate potential complications. Surgeons mentally translate twodimensional (2D) images into a three-dimensional (3D) image, but this is dependent on their spatial awareness and experience. This can be especially challenging in cases with irregular shaped lesions, where the treatment can be performed through a series of single probe ablations. If the surgeon's 4 perspective is not consistent with the actual situation, this may lead to treatment failure or major complications. Reported outcomes for prostate cryoablation demonstrate positive biopsy rates from 12-38% [10][11][12][13]. Importantly, most existing data on cryoablation oncologic outcomes stems from treatment approaches that did not use mpMRI for disease localization. The major complication related to prostate gland cryotherapy is post-surgical erectile dysfunction with rates ranging from 45.3-93.0%; however, these complication rates stem from whole gland cryoablation as opposed to partial gland strategies [12,14]. Other complications include urethral sloughing, urinary retention, incontinence, and urethral stricture [11,13,15]. With improved disease localization through mpMRI, limiting treatment to the precise tumor location limits the side effect profile of ablation [16]. The success of this treatment paradigm relies upon accurate disease targeting [17].
Currently, commercial software relies upon 2D images and was developed for whole gland ablation and no commercial tools are available to guide treatment for partial prostate gland cryoablation (cPGA) in 3D. Developing a 3D confluent ablation zone that encompasses the image visible disease as well as the necessary treatment margin is critical to successful cPGA.
To address the inadequacies inherent to 2D mapping techniques, this study reports on the development of a novel 3D surgical planning technique to provide pre-ablation treatment planning for cPGA. Patient-specific 3D models based upon mpMRI are created and the cPGA procedure is simulated using virtual 3D cryoprobes. Prior to cPGA, virtual 3D planning is utilized to confirm the required number and placement of cryoprobes to achieve confluent treatment volume for each unique lesion and margin. Optimization of the treatment plan in 3D by placing a predefined number of cryotherapy probes to best cover the lesion with the estimated − 40°C isotherm surface is expected to save time during the surgical procedure and to ultimately to help improve outcomes following cryotherapy for prostate cancer.

Methods
Patients with MRI-visible, biopsy proven prostate cancer (PI-RADS v2 score ³ 3) scheduled to undergo cPGA (n=20) were enrolled in our Institutional Review Board approved prospective study investigating 5 advanced methods of data visualization for patients with prostate cancer. Patient-specific 3D prostate cancer models were developed as described below. A comparison group (n=20) composed of men undergoing cPGA using 2D planning techniques was retrospectively evaluated.
The patient demographics for the 2D and 3D planning groups are shown in Table 1.

Patient-Specific 3D Prostate Cancer Models
Patient-specific 3D anatomical prostate cancer models that highlight the prostate, prostate tumor, urethra, neurovascular bundles, and rectal wall were created from the mpMRI data [23]. Regions of interest were segmented using Mimics 21.0 (Materialise, Leuven, BE) and visualized in 3D format with computer aided design (CAD) software (3-matic, Materialise, Leuven, BE).

3D Surgical Planning/Simulations
Virtual treatment simulation was performed in the 3-matic software pre-treatment for all patients in the 3D planning group and retrospectively post-treatment for the 2D planning group. The 3D prostate model was oriented in a supine position allowing the simulation to be performed in the same alignment as the cPGA operating procedure and a 1cm margin was created around each tumor.
Virtual cryotherapy probes were then selected and manually placed into the software in a spatial orientation to ensure confluent -40 º C isotherm encompassing both the tumor and the margin. This model was assessed in multiple views to ensure treatment confluence. The distances between the center of each probe were measured in order to reproduce the plan during the operation. In addition, contours of the anatomy and selected cryotherapy probes were generated on the 2D MR images.

Operating Procedure
All cryoablation procedures were performed under general anesthesia in a dorsal lithotomy position. A BK Flexfocus 800 biplanar ultrasound probe (model # 8808) attached to a Civco brachytherapy stand and stepper was utilized to visualize the prostate. Healthtronicsä cryoablation equipment was utilized to perform all ablations procedures.
6 2D Planning Method: For patients undergoing treatment with 2D planning, the Healthtronics TM software package was utilized to plan probe location. This software utilizes a 2D rigid registration of the prostate in an axial view on ultrasound. Probe placement is then guided by the 2D software in order to optimize probe-toprobe distance, probe-to-capsule distance, and probe -to urethra distance. This software does not utilize any MR-US fusion technology. MR tumor location is targeted using visual estimation. Visual estimation is performed preoperatively using image measurements on axial and sagittal MR images.
These measurements are translated to real-time US imaging to achieve visual estimation in lesion targeting. Cryotherapy probes are then placed under axial and sagittal ultrasound guidance. Each needle is placed via a 16 gauge brachytherapy grid with 2.5 mm distance between each grid location.

3D Planning Method:
The same software and equipment as described above is utilized for 3D planning with the exception of the pre-treatment planning as described above. The location of the pre-planned cryoprobes are then placed according to the 3D treatment planning, also using visual estimation. Again, no fusion software was available on the ultrasound for these ablation procedures. Cryoablation Procedure: After

Evaluation of Treatment
In order to measure treatment efficacy, treatment zone biopsy results at 6 months were evaluated.
Post-operative MRI and PSA at 3 and 6 months were also performed. The Kruskal-Wallis H Test was performed to determine if there was a difference in positive biopsy rates for the 2D and 3D planning groups. Statistical evaluation was carried out in SPSS Software (IBM, Armonk, NY).

Results
The 3D surgical plan was successfully simulated in all 20 patients selected to undergo pre-procedural 7 3D planning. 3D planning for a representative patient is shown in Fig. 2  For the 2D planning group, 13 returned for follow-up biopsy. Of these, ten (76.9%) had negative 6 month post ablation biopsy. Three patients (23.1%) were positive in the ablation zone, one patient with Gleason score 3 + 3 in the medial margin and two with Gleason score 3 + 4 in the treatment zone. Of these patients, the predicted plan using 3D modeling matched the actual plan in once case (4 cryoprobes for both) and predicted more cyroprobes in two cases. For these cases, three 8 cryoprobes were utilized, but one 3D plan predicted 5 cryoprobes but only 3 were utilized and the other predicted that 4 cryoprobes would be necessary to cover the target volume. Although 3 patients had positive findings post-operatively in the 2D planning group as compared to one patient in the 3D planning group, this did not reach statistical significance (p = 0.056). No post-surgical complications were reported for either group.

Discussion And Conclusions
Prostate cancer is the most common cancer in American men, accounting for almost 1 in 5 new cancer diagnoses annually [24]. The majority of prostate cancer is diagnosed following a screening evaluation of serum PSA followed by prostate biopsy [25]. The widespread use of this diagnostic paradigm most often identifies prostate cancer at an early stage when it is localized to the prostate gland [25][26][27]. Early identification of the disease allows for multiple management strategies, including surveillance for low grade disease and whole gland radiation and radical prostatectomy for intermediate and high-risk disease. Due to significant treatment toxicities associated with both radiation and radical prostatectomy, partial gland ablation (PGA) for prostate cancer aims to achieve oncologic control while mitigating side effects by limiting treatment to only regions of known cancer and preserving normal surrounding tissue.
Multiple technologies have been employed for focal therapy including high-intensity focused ultrasound (HIFU), cryotherapy, electroporation, radiofrequency ablation, and photodynamic therapy [28][29][30][31][32]. While an organ sparing strategy is widely employed in multiple oncologic treatments including kidney and breast cancer, employing this approach for prostate cancer has been limited by challenges in precise determination of tumor location and volume within the prostate gland [33].
Multi-parametric MRI is increasingly utilized for detection, localization, and staging of prostate cancer and offers a potential tool for image guided PGA of prostate cancer [34,35]. Despite this significant advance, achieving a confluent "kill zone" for MRI-guided PGA remains a significant challenge. In this study, we report the use of 3D prostate cancer models used in conjunction with mpMRI and advanced 3D visualization software methods to plan and simulate a theoretic zone of cryoablation for imageguided cryotherapy ablation of prostate cancer. 9 The pre-operative 3D prostate cancer models are helpful in planning the cryotherapy procedure.
These 3D models easily conceptualize the location of the tumor within the prostate as well as provide guidance on the extent of the necessary treatment margin (in this study a 1 cm margin was utilized) to predict an adequate "kill" zone. The 3D models also provide a comprehensive understanding of the 3D surgical anatomy including an understanding of the relationship of surrounding critical structures to the proposed treatment zone.
The virtual cryotherapy probes also allowed the exact "kill zone" to be predicted pre-operatively, thereby facilitating the operating procedure. The procedure was successfully carried out in all patients following the 3D virtual surgical planning procedure. In regards to the cryotherapy probe selection, there was a strong correlation between the planned number and the actual number used in the surgical procedure (80%), which suggests that the 3D surgical plan helped to guide the procedure.
Although there was no difference in operating times between groups, less variation was seen in the 3D planning group. In addition, in this small cohort, a greater number of patients in the 3D planning group were negative for cancer post-operatively as compared to those in the 2D planning group, with 1/17 (5.9%) and 3/13 (23.1%) positive for cancer at follow-up biopsy for the 3D and 2D groups respectively. The fact that 3D planning predicted a greater number of cryoprobes than 2D planning and that there was a higher success rate in the 3D cohort suggests that 3D planning allows for a more comprehensive assessment of the coverage area needed for successful tumor ablation. Future studies with more patients will assess how this method of procedure simulation compares to traditional 2D planning with mpMRI and how it impacts long-term treatment efficacy.

Conclusions
This study represents a preliminary exploration of a novel 3D treatment planning approach to cPGA of the prostate. Meaningful differences between 3D planning and traditional 2D planning were not possible in this study due to the small cohort and retrospective nature of the evaluation; however, the results encourage additional study in a larger cohort.

Ethics approval and consent to participate
This study was approved by the NYU Langone Health Institutional Review Board (IRB). All patients signed written informed consent to participate.

Consent for publication
All authors provided consent for publication

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.