Feasibility of MRI-Guided Transperineal Implantation of Microdevices for Drug Delivery and Response Assessment in Prostate Cancer

Abstract

Background: Prostate cancer (PCa) treatment often involves systemic therapies with varying mechanisms of action, affecting individuals differently. Implantable microdevices (IMDs) are designed to test multiple drugs within a patient's tumor, but the feasibility of MRI-guided placement in PCa has not been evaluated.

Purpose: To provide proof of concept for placing IMDs into lesions with MRI guidance to predict patient-specific responses to therapies.

Study type: Prospective.

Population: Fifteen participants undergoing prostatectomy for PCa.

Field strength/sequence: 3T MRI With T2-weighted (T2W).

Assessment: In-bore MRI-targeted placement of IMDs was performed. Intra-procedural MRI scans were reviewed by a radiologist, using needle artifacts on T2W images to guide IMD placement. A genitourinary pathologist performed Gleason scoring around the IMDs. Drug response analysis included Enzalutamide + Nivolumab, Enzalutamide + Docetaxel, and single-agent Enzalutamide.

Statistical tests: Mann-Whitney U test for continuous variables, p < 0.05 for significance.

Results: Of 53 IMDs implanted into suspicious lesions in 14 participants, 48 (90%) were successfully placed within the lesions. The average distance from the needle tip to the tumor was 8.32 ± 4.02 mm. Larger lesion size (p = 0.009) and lower prostate imaging-reporting and data system score (p = 0.031) were significantly associated with successful IMD placement. Of the 53 IMDs, 49 (92.4%) were retrieved for histopathology and drug response analysis. In four participants, Gleason scores around the device were lower than preplacement biopsy in two and equal in two. Additionally, drug analysis in one patient demonstrated the feasibility of drug response analysis, revealing differences in apoptotic index, lymphocyte infiltration, dysplastic cell composition, and cellular profiles for each treatment. No complications or adverse events occurred.

Conclusion: IMDs can be effectively and safely placed in prostate lesions using MRI guidance, with feasible histological and drug response analyses.

Evidence level: 2. Technical Efficacy: Stage 1.

Keywords: MRI; interventional radiology; microdevice; prostate cancer.

Figure 1.

A 2.2 cm lesion in the left peripheral zone affects the mid to base and posteromedial/posterolateral regions (indicated by white arrowheads) in Case #14. On (A) T2WI, the lesion was a well-defined, homogeneous mass that is moderately hypointense, confined to the prostate. On (B) Diffusion-Weighted Imaging (DWI), the lesion shows significant hyperintensity. On (C) Apparent Diffusion Coefficient (ADC) map, the lesion appears notably hypointense. (D) Dynamic Contrast-Enhanced (DCE) MRI was also positive. There is probable spread beyond the prostate into the surrounding fat, posteriorly and anteriorly near the apex. This lesion is classified as PI-RADS 5. Imaging was performed using a 3T MRI Scanner (MAGNETOM Prisma, Siemens Healthineers, Erlangen, Germany) with the following parameters: T2WI had a field of view (FOV) of 220×220 mm, a matrix of 320×256, a repetition time (TR) of 5060 ms, an echo time (TE) of 100 ms, a flip angle of 130°, a thickness of 3 mm, and a resolution of 0.68×0.68×3.00 mm. DWI had an FOV of 112×220 mm, a matrix of 90×46, a TR of 4800 ms, a TE of 64 ms, a flip angle of 90°, a thickness of 4 mm, and a resolution of 1.22×1.2 mm with a b-value of 1400 sec/mm². The ADC used five evenly spaced b-values (0–750 sec/mm²). DCE had an FOV of 260×260 mm, a matrix of 192×54, a TR of 2.99 ms, a TE of 1.09 ms, a flip angle of 15°, a thickness of 3 mm, and a resolution of 1.35×1.35×3.00 mm.

Figure 2.

[Left] T2-weighted MR image at the time of IMD deployment. [Right] Magnified view of the white rectangle area overlaid in the left image. The approximate device location (arrows) was marked at the site of IMD placement.

Figure 3.

Axial T2-weighted (T2W) MR images showing the marked device location (circle D) and tumor coordinates (circle T). (A) The distance between the two coordinates was measured using the 3D Slicer. (B) A segmentation (indicated by the green area) was performed in the 3D Slicer, based on correlation with pre-procedure mpMRI images.

Figure 4.

Axial images following microdevice placement. (A) Three microdevices are clearly visible on CT as areas of hyperdensity (red arrows) near the targeted lesion. (B) A PI-RADS 5 lesion located in the right transition zone, anterior midgland, is identified on MRI (yellow arrowheads).

Figure 5.

The hematoxylin and eosin (H&E) slide from Case #11 shows the device’s location (indicated by double-headed arrows) within tissue sections. At approximately 2.1 mm from the top and cranial side of the device (left), the device is positioned within Gleason score 3+3 and benign tissue. At approximately in the middle of the device, 1.8 mm below, it is located within Gleason score 4+3 tissue.

Figure 6.

Tumor tissue analysis with implanted microdevice illustrating drug-specific effects in an index tumor, demonstrating feasibility (Case #22). [Left Column] H&E-stained images highlight the tumor tissue’s microdevice area and residual cavity. The first image (Enzalutamide + Nivolumab ROI) shows diffuse prostatic adenocarcinoma with Gleason 4+3 morphology. The second image (Enzalutamide + Docetaxel ROI) reveals fibrotic regions and smooth muscle sheets with cribriform tumor glands (Gleason 4+4 morphology). The third image (Enzalutamide ROI) displays smooth muscle and fibrotic sheets with tumor glands exhibiting Gleason 3+4 morphology. [Middle Column] Immunohistochemistry (IHC) staining for cleaved caspase-3 (CC3), where red signals indicate CC3-positive cells and blue signals represent CC3-negative cells. The Enzalutamide + Nivolumab ROI shows a higher amount of positive signal compared to the other two drug treatments. IHC signal quantification in regions of interest (ROIs) was performed using Mikaia 1.6 software. [Right Column] Cell-type recognition and density analysis using Mikaia 2.0 AI. The AI identifies dysplastic cells (red), lymphocytes (yellow), fibroblasts (turquoise), and epithelial cells (orange). The Enzalutamide + Nivolumab ROI shows a higher number of lymphocytes compared to the other regions, where fibroblasts dominate.

Figure 7.

[Left] Apoptotic index (%) of the three examined drug areas – The percentage of apoptotic cells calculated as the ratio of apoptotic cells to total cells, multiplied by 100. Total cell count and apoptotic cells were identified based on immunohistochemistry (IHC) staining for cleaved caspase-3 (CC3) and quantified within defined regions of interest with Mikaia 1.6 image analysis software. Enzalutamide + Docetaxel drug combinations showed a higher apoptotic index compared to the other two agents. [Right] Cell density analysis expressed as cells/mm² within the ROIs, providing a detailed evaluation of the tumor microenvironment. The first three columns represent the lymphocyte, the second three the fibroblast, the third the dysplastic cells, and the fourth the epithelial cell percentage within the ROI per drug combinations (every first column per category belongs to Enzalutamide + Nivolumab, every second to Enzalutamide + Docetaxel and every third to single agent Enzalutamide).