Feasibility of Alpha-DaRT source deployment according to a treatment plan

A total of 206 sources were implanted in four phases, with two swine in each phase (see Table S1, Fig. 1). In the first phase, the sources used were not radioactive (inert sources), while in the subsequent three phases, all sources were active. The radial applicator was easy to use and placed multiple sources in a relatively short time (minutes) according to the predetermined configuration (Fig. 2). The duration of the entire procedure did not exceed one hour per animal, including anchor placement and implantation of all sources in both hemispheres (up to 42 sources per animal).

Fig. 2

Representative images of Alpha-DaRT sources implantation to the swine brain. (A) After assembling the template-frame on the swine ‘s head, holes were drilled to access the brain. (B) The radial applicator was assembled on the template-frame and a biopsy needle was fixed on the rotation mechanism. (C) The Radial Cartridge loaded with sources was inserted into the biopsy needle and the sources were deployed one by one at the planned location in an umbrella-like configuration using the handle. (D) the umbrella-like configuration of Alpha-DaRT sources is presented under the entry point (drilling hole). (E) Intraoperative fluoroscopy of anesthetized swine for applicator needle localization verification prior to source implantation. Red arrow – applicator’s needle tip. Green arrow – base of the skull. White asterisk - three (3) metal screws of the customized stereotactic frame applied to the animal skull

Following the insertion of the sources in 1–3 layers of 7 sources (unless mentioned otherwise) per layer to the left or right hemispheres, animals were monitored for a period of up to 90 days. During the follow-up period, CT and MRI scans (Fig. 3), blood, CSF, urine, and feces samples were collected, and clinical and neurological evaluations were performed.

Fig. 3

CT and MR scans of Alpha-DaRT sources in the swine’s brain. (A) Two-planar CT images. In each image, the virtual sources (i.e., the sources as represented in the treatment plan software) are shown as green marks, while the estimated region subject to an alpha dose of > 20 Gy they create is shown as a green-filled contour. The brain contour is represented in magenta. (B) Three-planar MR images for the same animal as A. The sources are shown as black stains and are marked by red circles. (C) Three-planar MR Image data sets with various sequences commonly utilized for brain tumors (T1W ± Gadolinium, FLAIR, T2W, DWI, and SWI) of animal DP13253 brain. Implant location-bilateral parietal. Mild intra-implant microhemorrhages left greater than right. No evidence of infection or inflammation. Minimal intra-implant tissue edema bilaterally. White asterisk – near right side implant location

Due to the close proximity of the biopsy needle to the skull during the deployment of the sources in one animal (DP13577), one source did not come out of the cartridge (see Table S1).

Movement of sources in the follow-up period

Analysis of source movement was done relative to the 3D source enclosure, defined as the minimal convex surface that encapsulates all sources together with a 3.6-mm margin, corresponding to registration errors and errors in identifying the source coordinates (see more details in the Supplementary Information). Out of a total of 206 sources, only one source partially protruded out of the 3D source enclosure with the maximal distance of 0.6 mm (Table 1, Fig. S3). In the first animal operated on in the study (swine DP13176), one source was unintentionally inserted into the middle part of the right ventricle due to the pig ventricles not being clearly visible on CT. This source moved through the CSF-filled ventricle anteriorly to the frontal horn between days 14 and 21. The target location was refined for subsequent insertions. This source was excluded from the movement analysis. Notably, by comparing days 0–60 and days 0–66 of animal DP13176 it can be shown that the MRI performed on day 64 did not affect the movement level.

Table 1 Spatial-temporal localization analysisClinical and neurological evaluation during the follow-up period

All clinical evaluations including neurological evaluations were normal except for one animal (DP13254) that exhibited temporary neurological deficiencies, starting from day 7. The neurological symptoms were mild front legs ataxia, mild head tilt and circling to the right, and possibly vision impairment. These symptoms, except for the suspected vision impairment, gradually improved and were resolved by day 23. These clinical signs were mild and did not prevent the animal from eating. Therefore, it was not prematurely euthanized and completed the entire trial. Notably, in this animal, the sources were placed at a deeper location than planned in the left hemisphere as can be seen in Fig. S4. No treatment-related findings were observed in follow-up measurements.

CSF, hematology, and chemistry

No abnormal findings were found in the blood tests. CSF, glucose, and lactate values were normal. Total protein values were elevated in all animals on day 14, as may be expected from the tissue damage following source insertion. In animal DP13253, the high values (179 mg/dl) increased steadily throughout the trial. Animal DP13332 had elevated levels of total protein in CSF that normalized by day 60. In one animal (DP13254) the total protein values were very high (416 mg/dl, peak on day 14), and on day 60 the levels were still high but lower than on day 14. This animal had the highest number of sources implanted. For the other two swine, no CSF samples were available on termination day. Of note, there was no correlation between clinical symptoms and total protein levels.

Postoperative MRI analysis

All MR Image data sets were reviewed by a senior neuro-radiologist (JMG). MRI scans did not show any evidence of infection or inflammation. All active animals had mild peri- or intra-implant-zone edema. The local damage to the tissue surrounding the sources was demonstrated (minimal peri- /intra-implant tissue reaction) without any non-adjacent tissue reaction or other clinically significant abnormal findings. All active animals had mild intra-implant microhemorrhage. One animal (DP13578) had one macroscopic peri-implant hemorrhage on day 53, which had no clinical significance based on the animal’s neurological examination. It was not observed on MRI day 81 as it was probably absorbed. The clinical follow-up of this animal was extended to day 90 with no relevant findings.

212Pb and geiger counter measurements

212Pb specific activity (corrected from the measurement time to the sample collection time) on days 7–8 was 1.5-5.0 Bq/ml in blood, 0.7–14.8 Bq/ml in CSF, 1.1–2.9 Bq/ml in urine, and 0.1-3.0 Bq/g in feces. The ratio between the blood, CSF, and urine-specific activities on day 14–15 and day 7–8 was consistent, within error, with the expected decay of 224Ra (with which 212Pb is in secular equilibrium from ~ 2 days post-treatment and onward). No feces samples were taken on days 14–15. When normalized by the total 224Ra activity on the Alpha-DaRT sources (calculated for the time of sample collection), the specific 212Pb activity per µCi 224Ra on day 7–8 was (mean ± std) 0.14 ± 0.04 Bq/ml/µCi for blood, 0.16 ± 0.16 Bq/ml/µCi for CSF, 0.11 ± 0.05 Bq/ml/µCi for urine, and 0.04 ± 0.04 Bq/g/µCi for feces. Notably, the normalized specific activity recorded in blood is > 10 times smaller than observed in Alpha-DaRT-treated patients (with skin and head and neck tumors) [24]. This may indicate possible reduced clearance of 212Pb by the blood due to the blood-brain barrier. A gross estimate for the alpha-particle absorbed dose to the swine brain (outside of the treatment region) is described in the Supplementary Information. For the largest treatment in this study, consisting of 126 µCi 224Ra, it is ~ 4 × 10− 4 Gy, ~ 30,000 timessmaller than the estimated tolerance alpha dose of 12 Gy [28].

The external dose rate was measured at 0–30 cm away from the swine’s head in two swine (along with a comparative background measurement) during the first 5 days from sources insertion. The values (e.g., ~ 0.01 mGy/h at 30 cm for 126 µCi 224Ra) were similar to those observed in skin patients at 30 cm and consistent with exposure estimates using the MicroShield code [13].

Pathology and histopathology

Macroscopically, in all swine, the Alpha-DaRT sources were located within the cortex and subcortex cerebral regions (Fig. 4A). None of the brains presented evidence of gross tissue necrosis. In animal DP13176 the histopathological analyses showed that one Alpha-DaRT source was located at the frontal horn of the right ventricle. In swine DP13177 several Alpha-DaRT sources penetrated the caudate nucleus.

Fig. 4

Representative images of the swine’s brain pathology at the end of the follow-up period, post Alpha-DaRT insertion. A. Gross pathology of a cluster of Alpha-DaRT sources in cerebral cortex, swine. B. Five optically empty spaces in the cerebrum adjacent to the lateral ventricle. C. Optically empty spaces (circular & longitudinal) surrounded by mineralization, fibrin, foreign body material (bone dust and fragments & keratin), fibrinoid necrosis of the parenchyma and the wall of blood vessels and intravascular fibrinous thrombi. D. Mononuclear perivascular cuffs in cerebral parenchyma. E. At the margins of the optically empty space the histopathologic findings include neutrophils and eosinophils inflammatory infiltrate, reactive endothelium, giant cells, liquefactive necrosis, edema in the neuropil and Gitter cells. F. Meningeal and parenchymal cerebral hemorrhages

The tissue findings were consistent with moderate to severe, focal-extensive, meningoencephalitis with liquefactive necrosis. In all the sections, the histopathological findings were locally confined adjacent or around the Alpha-DaRT sources or cluster of Alpha-DaRT sources. Namely, no abnormal findings were observed at distant sites (not beyond approximately 3–8 mm from the source, according to gross pathology observations) in active sources compared to inert sources.

In all animals, the histopathological findings were localized with varying degrees of severity (Fig. 4B and F). In a few sections, histopathological findings included minimal vacuolization (possible artifact), hemosiderin-laden macrophages, reactive meningeal cells, and “red neurons” (suspected as neuronal necrosis).