Non-Invasive Pancreas Ablation Using Histotripsy: Pre-clinical Safety Study in an In Vivo Porcine Model

Abstract

Pancreatic cancer is an aggressive disease that is typically diagnosed late, leading to poor prognosis and overall survival. Even with advances in treatment, the 5-year survival rate for patients is only 12.5%. Histotripsy is a non-invasive, non-thermal and non-ionizing focused ultrasound treatment method that has recently been US FDA approved for treating liver tumors and is currently being investigated for other applications. This study expands on prior work investigating the feasibility of pancreas ablation using ultrasound-guided histotripsy in pigs. Here, we applied histotripsy to healthy pancreas in nine pigs using a therapy system with treatments guided by real-time ultrasound imaging. After treatment, subjects survived for 1 week (n = 3) or 5 weeks (n = 6). Damage to the pancreas and surrounding tissue was characterized using gross morphology, histological analysis, and computed tomography imaging. Treatment zones were visible on immediate post-procedure computed tomography, with lesion dimensions measuring 1.4 ± 0.6 cm, 1.2 ± 0.3 cm, and 1.5 ± 0.4 cm in the anteroposterior, transverse, and craniocaudal planes, aligning with the planned 1.5 cm-diameter spherical treatment volume. At 5 weeks, lesion size reduction was observed in three subjects, decreasing by 49%, 98%, and 100%, respectively. Histotripsy was well-tolerated, with successful pancreas targeting in six of nine subjects. Two cases of potential pancreatitis were noted. Three pigs experienced bowel damage due to poor ultrasound visualization, leading to off-target effects including septic peritonitis and gastrointestinal blockage. These findings suggest histotripsy is a viable approach for pancreatic treatment when the organ is clearly visualized and bubble clouds remain within the tissue.

Introduction

Pancreatic cancer continues to be a fatal disease surpassing breast cancer as the third leading cause of cancer-related death in the United States [1,2]. A main factor contributing to its lethality is that the disease is diagnosed after it has locally advanced or metastasized to other parts of the body—during the localized stage of cancer, patients present asymptomatically or with ambiguous symptoms [3]. Additionally, the anatomical location of the pancreas makes it difficult to access for routine screening [1]. Together, these factors hinder its early diagnosis, leading to 80% of patients presenting with advanced unresectable tumors. For those that are surgical candidates, a pancreaticoduodenectomy, or Whipple procedure [4], is most commonly performed but there is a high morbidity rate of up to 45% associated with this strategy [5]. Chemotherapy is therefore used as a first-line therapy for many patients with unresectable tumors, as well as radiotherapy or a combination of the two, which have shown improved local control rates [6-8]. However, pancreatic tumors have highly dense desmoplastic stroma as a result of cancer-associated fibroblasts in the tumor micro-environment [1,8]. The collagenous stroma produces a physical barrier impeding chemotherapeutic agents from entering the tumor micro-environment while also creating an immunosuppressive environment for targeted therapies [1,7,9]. The limitations associated with these treatment strategies have contributed to the dismal 5-year relative survival rate of ~12.8% for patients [10]. Alternatively, ablative therapies have been investigated for pancreatic cancer including, high-intensity focused ultrasound (HIFU) [11,12], irreversible electroporation (IRE) [13-17], microwave [18-20], radiofrequency [21-26] and cryoablation [27-30]. The latter three have been shown to have serious complications including pancreatic fistulas, pancreatitis and gastrointestinal (GI) bleeding or injury [31], leading to the use of IRE and HIFU as the main ablative strategies [31]. Unlike the other modalities, IRE is non-thermal and relies on a pulsed electric field to induce cell death [31], making it a preferred option near critical structures [27,32]. While it has shown improved survival, IRE is still a minimally invasive technology with probes percutaneously placed into the pancreas [33,34], and there is a risk of generating cardiac arrhythmias with the electric field [32]. Thermal HIFU focuses ultrasound waves to a singular point in the body non-invasively and heats the targeted tissue between 80°C and 100°C to cause coagulative necrosis [35,36]. HIFU has been shown to mitigate pain in pancreatic patients after treatment as well as increase survival [35,37]. However, the heat from this thermal procedure can also cause severe inflammatory changes in the sensitive pancreas and complications include skin burns, acute pancreatitis, and pancreatic leaks/pseudocysts, among others [35]. Therefore, there is potential for a non-invasive, non-thermal treatment strategy to address the limitations of current technologies.

Histotripsy is an emerging focused ultrasound ablation method with the potential to precisely target and ablate tumors non-invasively [38]. The method utilizes high amplitude focused ultrasound pulses at a low duty cycle (<1%) that are focused within a targeted tissue [38-40]. Due to this focusing, high peak negative pressures are achieved and when the pressure exceeds a threshold intrinsic to the targeted medium, cavitation forms. The cavitation “bubble cloud” is comprised of microbubbles that rapidly expand and collapse, imposing high stress and strain on the surrounding tissue, which leads to cell rupture [41]. Using a short burst duration to mitigate heat build-up, histotripsy supplies a non-thermal mechanical breakdown of the target tissue, differing from thermal ablation techniques such as HIFU, microwave, radiofrequency and cryoablation [27]. The mechanical aspect of histotripsy overcomes limitations associated with thermal modalities such as incomplete ablation near major vessels [42], lack of treatment precision [43] and damage to critical structures [38]. Prior work has shown that histotripsy is capable of generating vessel-sparing and duct-sparing ablation due to these tissues being more resistant to histotripsy-induced tissue damage [38,44-48], and tissue selectivity may vary depending on the target tissue and critical structure properties [39,44,47]. The feasibility and safety of histotripsy have been established for treating liver tumors in both pre-clinical [27,49-54] and clinical studies [55,56]. Results have shown that histotripsy treatment is well-tolerated, with the ablated volume reduced by 90% at 8 weeks after treatment, allowing clinicians to immediately assess treatment effectiveness on follow-up imaging [52,53,55,56]. The THERESA trial was the first clinical trial of histotripsy treatment for malignant liver tumors to establish the technology’s safety and feasibility in humans [55], with no major adverse events reported. Furthermore, a multicenter clinical trial (HOPE4LIVER) using histotripsy to treat liver tumors demonstrated the therapy’s technical success for clinical adoption [56]. The promising results from these clinical trials supported the recent approval of the histotripsy Edison platform by the US FDA for the treatment of liver tumors.

Unlike the liver, the pancreas is encased in a delicate fibrous capsule making it more sensitive [57]. Due to the delicate nature of the pancreas, there are major concerns about causing injury to critical structures within the tissue, such as the pancreatic duct, as well as inducing pancreatitis. Clinically, 20%–30% of acute pancreatitis patients will experience recurrent attacks, with 10% of these developing chronic pancreatitis [58]. While there is less risk around acute pancreatitis, the development of severe or chronic pancreatitis has serious concerns as the disease significantly reduces quality of life and life expectancy [59]. There are also risks of off-target injury to the overlying stomach and bowel given the anatomical location of the pancreas. While histotripsy appears promising for abdominal targets, it is imperative that these concerns are evaluated for the pancreas. Previously, our group investigated the proof-of-concept and feasibility of applying histotripsy to healthy pancreas in 11 pigs using a prototype system with a 500 kHz therapy transducer [60]. The first 3 subjects in the study were fasted prior to histotripsy treatment to minimize gas contents in the bowel, but results showed a poor acoustic window as gas blockage was still present, resulting in difficulties in targeting the pancreas and off-target bruising in the overlying bowel after treatment. The remaining subjects were administered a special pre-treatment diet consisting of custard that was laced with simethicone and bisacodyl to further reduce overlying bowel gas. The pre-treatment preparation improved ultrasound visualization of the pancreas and enabled the generation of histotripsy bubble clouds to be formed inside the pancreas with a reduction in off-target injury. The results showed no major adverse events after histotripsy treatment of the pancreas, including no major intra-abdominal bleeding or hemorrhage, main vessel thrombosis, pancreatitis or adjacent organ injury/toxicity including intestinal perforation or fistulas. Post-treatment enzyme levels showed transient non-clinically significant elevation, and subjects did not exhibit any signs of pancreatitis or fistula formation during the study. No clinically significant change in pancreatic function was noted as assessed via enzyme panels, with no signs on computed tomography (CT) imaging or at necropsy of pancreatic fistulas. Limitations of this study included small subjects (5–8 kg) with relatively small ablation volumes (as small as 0.5 cm³). While ablation was noted on US imaging, the pathologist did not confirm any ablation zones with histopathology, likely due to the small size of the generated ablation zones and tissue deformation/lesion resorption after treatment.

The goal of the current study was to expand on our prior investigation by applying histotripsy to healthy pancreas in larger pigs with longer survival times (up to 5 weeks) to evaluate the safety profile of histotripsy. There were three main hypotheses: (i) histotripsy can safely generate ablation in the pancreas without inducing pancreatitis; (ii) histotripsy ablation will not cause significant injury to critical structures within the pancreas such as major vessels and ducts; and (iii) histotripsy can safely be applied through overlying tissues and will not induce clinically relevant injury to the bowel when the bubble cloud is confined within the pancreas.

Materials and methods

Animal model

A total of nine healthy pigs (four male and five female, 6–8 weeks old, 17–24 kg; Virginia Tech Swine Farm, Blacksburg, VA, USA) were treated with histotripsy in this study. After arriving at the facility, the pigs were closely monitored three times in their initial week to allow them to adjust to their new environment in adherence to the protocols set by the Virginia Tech Institutional Animal Care and Use Committee. The primary aim of this research was to assess the viability of using histotripsy to target and ablate the pancreas in vivo under ultrasound guidance. In this series of experiments, histotripsy was targeted on the pancreas of all subjects, each of which was supplied a specialized diet laced with a laxative designed to reduce bowel gas. This dietary approach mirrored previous studies that evaluated HIFU in healthy pig and human pancreas [61,62]. The distinctive diet included sweetened custard infused with simethicone (2 mL/pig) and bisacodyl (5 mg/kg) and was administered once daily for the 4 days preceding treatment. On these days, subjects had free access to water and their regular diet, and the pigs were fasted overnight (~12 h) 1 day prior to treatment. On the morning of treatment the subjects were only given the specialty diet (with a double dose of simethicone and bisacodyl) roughly 1 h prior to histotripsy application to allow for digestion.

During each histotripsy session the animals underwent general anesthesia, maintaining the oxygen flow rate (L/min) with approximately 1.5%–2.5% isoflurane. Once anesthetized, the subjects were placed in the supine position on a surgical table with their feet loosely restrained to ensure stability. The abdominal area was shaved and then a depilatory cream (Naircare, Ewing, NJ, USA) was applied for 5–6 min before using a wet towel for removal. Throughout the treatment, vital signs including heart rate, blood oxygen levels, respiratory rate and temperature were monitored closely using the Propaq Encore Vital Signs Monitor (Welch Allyn, Beaverton, OR, USA) by trained personnel. Blood samples (1–2 mL) were collected before treatment, immediately after, and weekly after histotripsy treatment for complete blood cell and pancreatic enzyme panel analysis. Following treatment, while still anesthetized, the pigs underwent an abdominal helical CT evaluation covering non-enhanced and contrast-enhanced triple-phase CT through the abdomen. After the procedure the subjects were allowed to recover and then returned to their housing for either 1 week or 1 month, adhering to the previously outlined protocol. Pigs that successfully recovered after treatment were visually monitored in recovery kennels for 2–3 h until fully recovered, after which they were returned to their standard housing. Daily monitoring for general health, behavior, diet, and skin lesions was conducted post-recovery. After the study the animals were sedated and euthanized through an intracardiac injection of pentobarbital sodium and phenytoin sodium (Euthasol, MWI Animal Health, ID, USA). The timeline for the study is shown in Figure 1. All procedures used in this work were reviewed and approved by the Virginia Tech Institutional Animal Care and Use Committee (protocol #22-022 and #25-038).

Figure 1.

Timeline. All subjects underwent pre-treatment preparation prior to treatment. Blood was collected before and immediately after treatment. Immediate post-computed tomography (CT) data were collected for all subjects and then subjects were placed in either an acute or chronic group for follow-up blood sampling and CT scans.

Histotripsy treatment procedure

Histotripsy was applied to the porcine pancreas using a clinical system previously described that used a workflow developed in prior treatments for liver cancer [52] and employed in our first healthy pig study (Fig. 2a) [60]. Prior to treatment a veterinary radiologist resident (M.E.) performed freehand ultrasound imaging with a curvilinear probe (Model C6-1, GE Healthcare, USA) to identify the pancreas and determine the desired treatment location (Fig. 2b). A reservoir filled with ultrasound medium (degassed water) was placed over the subject’s abdomen to aid ultrasound propagation to the targeted pancreas (Fig. 2c). In this study, an improved “clinical reservoir” was utilized where an acoustically transparent elastic membrane contacted the subject’s abdomen as the base of the reservoir to fill with degassed water instead of our previous method using hand-cut drapes. Before placing the membrane on the skin, a thin layer of castor oil was rubbed over the entire abdominal area. A coupling band was then used to swipe air bubbles between the membrane and patient abdomen by slowly dragging the band from one side of the patient, across the abdominal area, to the other side. This action was repeated twice or until there were no visible air bubbles. After that, the reservoir was filled with degassed water (12–14 L) for acoustic coupling. A support platform was used to prevent excess water weight from directly loading the pig’s body. The treatment head (700 kHz therapy transducer and imaging probe), mounted onto a programmable robotic arm with 6 degrees of freedom (UR5e, Universal Robots, Novi, MI, USA), was then manually positioned to align with a clinically relevant acoustic window identified during freehand imaging. The body to head of the pancreas was targeted in all subjects as this region had the most ideal acoustic window for imaging and therapy application.

Figure 2.

Treatment workflow. (a) A histotripsy system (HistoSonics) was used in this study. (b) A veterinary radiology resident first performed freehand ultrasound (US) imaging to identify the pancreas (red dashes) and optimal acoustic pathway. (c) The “reservoir” was placed over the subject’s abdomen, filled with degassed water and the treatment head was submerged. Coax US imaging was collected (d) before, (e) during and (f) after histotripsy treatment. (g, h) Freehand US imaging and CT were collected immediately post-treatment. All subjects underwent immediate post-treatment CT with three subjects from the chronic group undergoing a second scan 5 weeks post-treatment.

Prior to treatment, the focus of the treatment head was aligned to the targeted location in the pancreas using coaxial imaging (Fig. 2d). Unlike our previous feasibility study [60], this work used a clinical system that had a coaxial imaging probe capable of telescoping. The feature allowed the radiologist to telescope the probe to its fullest extent on subjects with significant gas so that the probe was touching, or almost touching, the subject’s abdomen. In doing so, this minimized the stand-off distance and allowed for some visualization of the targeted pancreas. Then the radiologist slowly retracted the imaging probe in an attempt to anatomically align the focus of the therapy transducer in the targeted pancreas region as there was less stand-off when the imager was fully extended. Histotripsy treatment was still applied to these subjects to assess the safety profile in a “worst-case” scenario.

Once the focus was aligned in the desired region, energy delivery from the therapy transducer was initiated. Histotripsy was applied non-invasively in this pre-clinical proof-of-concept study using the clinical pulsing parameters designed for liver tumor destruction on the Histo-Sonics Edison System, as outlined in the DeNovo Grant application (DEN220087) and user guide [63,64]. This system was characterized in accordance with IEC 60601-2-62 using proprietary test methods and special controls. As in previous work, the focal pressure was slowly increased to identify the cavitation threshold at seven points within the planned treatment volume until a robust bubble cloud was achieved. A 1.5 cm-diameter spherical volumetric ablation (1.8 cm³) was applied to the targeted region in the pancreas. This size was selected in an attempt to target the largest spherical volume while keeping the ablation zone fully encapsulated by the pancreatic tissue. Subjects were treated at treatment voltage settings ranging from 44% to 68% of the maximum output. The entire procedure used ultrasound imaging to monitor the treatment in real time (Fig. 2e). Post-histotripsy treatment, coaxial US imaging was obtained and the therapy transducer was removed (Fig 2f). Freehand ultrasound and CT imaging followed to also assess for any tissue damage (Fig. 2g, 2h), and superficial thermal injuries on the skin were observed in two subjects. In both cases, the pancreas was located at a shallow depth (<5 cm from the skin surface), which is shallower than the allowable clinical treatment range for the system [63,64]. While histotripsy itself is a non-thermal, mechanical ablation modality, the HistoSonics clinical system includes thermal safety features designed to minimize heat accumulation in superficial tissues when delivering treatments at standard depths (>5 cm). However, to enable treatment at these unusually shallow depths, the system’s thermal safety constraints were bypassed to allow for histotripsy delivery in this pre-clinical setting. This adjustment, while necessary to reach the target anatomy, likely contributed to the observed superficial thermal injuries. In both pigs these thermal injuries (grade I) resolved over the course of 2–3 weeks (Supplementary Fig. 1). The animals recovered and survived for the necessary length based on their group, as described above. A full necropsy was conducted on all subjects by a board-certified veterinary pathologist (S.C.O. or K.E.) at the study’s conclusion.

Contrast-enhanced CT imaging

Subjects underwent an abdominal helical CT evaluation that included non-enhanced and contrast-enhanced triple-phase CT through the abdomen immediately post-procedure, with three subjects having a second scan 5 weeks post-procedure (Fig. 1). CT images were acquired with a multi-slice helical CT unit (Aquilion 64, Toshiba Medical Systems Corp., Shimoishigami, Otawara-Shi, Tochigi, Japan). For imaging, each subject was placed in a supine position and kept under general anesthesia for the entirety of the scan. The tube rotation speed was 0.5 s, the slice thickness was 1 mm and the pitch factor was 0.828. All phases of scanning were initiated at the cranial aspect of the diaphragm and extended caudally to the level of the pelvic inlet. Iohexol (350 mg I/mL, Omnipaque 350, GE Healthcare) was used as the contrast medium at a dose of 0.86 mL/lb and was injected with a power injector (Medrad Stellant, Indianola, PA, USA). Scan delay was triggered from the abdominal aorta at the level of the diaphragm for the arterial phase, measuring the mean of Hounsfield units (HU) in the non-enhanced aorta and adding 30 HU to determine the HU to trigger the system to begin scanning. The pancreatic phase was scanned immediately after the arterial scan was completed and the table was translated back to the area of the diaphragm. The equilibrium phase was scanned after 3 min from the time the contrast medium injection started. Images were evaluated by a veterinary radiologist (M.E.) and a human radiologist (T.Z.).

Necropsy and histological analysis

A full necropsy and tissue harvest was performed by a board-certified veterinary pathologist (S.C.O. or K.E.) at the time of each subject’s sacrifice, and the procedure followed the same protocol used previously [60]. Gross damage to internal organs was determined, as previously described, with inspection in situ and ex vivo with serial sections spaced approximately 1 cm apart. Serial slicing of pancreatic tissue was performed after fixation to preserve tissue architecture. Pancreatic tissue and any tissue with noted gross changes were visually inspected and fixed in 10% formalin for at least 24 h before sectioning, embedding, and staining. All tissues were stained using hematoxylin and eosin to assess for histotripsy damage. Additionally, any other changes in the structure and density of collagen and other tissue structures (i.e., vasculature and bile ducts) within the ablation volume were also noted.

Results

The goal of this study was to evaluate the safety profile of histotripsy pancreas ablation. The treatment results are summarized in Table 1 for all subjects. As observed in prior work, the radiologist was able to visualize the pancreas in all subjects with freehand ultrasound imaging after applying pressure to the abdomen with the probe, thus dislodging the overlying gas-filled tissues out of the beam path (Fig. 3a, 3b). The targeted pancreatic region remained identifiable on coaxial US imaging in six out of nine subjects (Fig. 3c) and was unable to be identified in the remaining three subjects due to overlying gas blockage (Fig. 3d). In subjects that had minimal gas, the histotripsy bubble cloud was visualized on coaxial US imaging and maintained throughout treatment (Fig. 3e), corresponding to the situation that would be applied in a clinical setting. In subjects where the pancreas was only visualized after telescoping the coaxial US probe, the bubble cloud was not visualized during treatment (Fig. 3f). While the inability to visualize the pancreas and bubble cloud would preclude the application of histotripsy in a human clinical setting, therapy was applied using anatomical alignment in order to assess the safety of continuing therapy under these conditions. The results showed a positive safety profile during and immediately after treatment, with no changes in vital signs and animal behavior. There were also no acute signs of damage or injury on immediate post-treatment CT scans to critical structures such as the stomach, portal vein, inferior vena cava, aorta, and pancreatic duct for all subjects.

Table 1.

Summary of all pig treatments

	Pig ID	Weight (kg)	Sex	US pancreas visualization	Bubble cloud visualized	Average voltage (%)	CT zone (anteroposterior x transverse x craniocaudal) (cm)	CT results	Side effects
1 week	A-1	22	Female	Body to head	Yes	42	1.3 × 1.1 × 1.4	Treatment zone in the pancreas	None
	A-2	23	Female	Body to head	No	57	1.0 × 1.0 × 1.2	Treatment zone in the pancreas	Mild pancreatitis on histology
	A-3	24	Male	None	No	68	2.5 × 1.4 × 1.4	Treatment zone in the pancreas w/extension into adjacent small bowel loop	Septic peritonitis; subject had to be culled 2 days after treatment
5 weeks	B-1	24	Male	None	No	66	Immediate post-CT: 1.1 × 1.1 × 1.5 5-week post-CT: not identifiable	Treatment zone in the pancreas 5-week imaging shows no clear treatment zone and is likely microscopic in size	Skin burn
	B-2	22	Female	Body to head	Yes	44	1.8 × 1.2 × 1.7	Treatment zone in the bowel loops in the right abdomen	None
	B-3	20	Male	None	No	67	Immediate post-CT: 0.9 × 0.9 × 1.4 5-week post-CT: 0.8 × 0.8 × 0.9	Treatment zone identified within a lymph node between the head of the pancreas and the PV	Skin burn
	B-4	19	Female	Body to head	Yes	49	Immediate post-CT: 1.9× 1.7 ×2.4 5-week post-CT: 0.5 × 0.5 × 0.5	Treatment zone in the pancreas 5-week imaging shows zone has involuted with development of mild dilation of the pancreatic duct	Mild to moderate pancreatitis; elevated amylase and lipase; mild dilation of pancreatic duct
	B-5	17	Male	Body to head	Yes	45	0.7 × 0.6 × 0.9	Treatment zone identified at the periphery of a bowel loop	GI blockage; subject had to be culled ~3 weeks after treatment
	B-6	17	Female	Body to head	Yes	47	1.6 × 1.5 × 1.8	Treatment zone in the pancreas with preservation of a patent gastroduodenal artery. Non-occlusive thrombus of PV	None

AE, adverse event; GI, gastrointestinal; LN, lymph node; PV, portal vein; TX, treatment; US, ultrasound; PV, portal vein.

Figure 3.

Histotripsy treatment results. (a, b) The pancreas was identified on freehand US imaging in all subjects (red dashed line). (c) In subjects with minimal gas blockage, the pancreas continued to be visualized on coaxial US imaging pre-treatment (red dashed line), but not in subjects with significant gas blockage (d). (e) Bubble clouds were identified and maintained throughout treatment in subjects with minimal gas blockage but not in subjects with significant gas content (f). Treatment videos can be found with this supplemental videos. (g) A hypoechoic region was observed immediately post-treatment in subjects with minimal gas blockage but not in subjects with gas blockage (h).

Immediately after treatment, hypoechoic regions were identified on US imaging in the regions of the tissue volume that were targeted in five out of nine subjects (Fig. 3g). Hypoechoic lesions were not observed in other subjects due to limitations in the ultrasound imaging, corresponding to the same patients in which the bubble cloud could not be visualized during treatment (Fig. 3h). CT results showed that histotripsy lesions were identified in all pigs immediately post-treatment, including those where lesions were not observed on ultrasound imaging. Lesions were identified to be confined within the pancreas in the six out of nine subjects who had clear acoustic windows for therapy (Fig. 4a). Treatment zones identified on immediate post-procedure CT were 1.4 ± 0.6 cm by 1.2 ± 0.3 cm by 1.5 ± 0.4 cm (mean ± standard deviation) for the anteroposterior, transverse, and craniocaudal planes, respectively, which were in good agreement with the planned 1.5 cm-diameter spherical treatment volume. For the three subjects that had a second CT scan at 5 weeks, Pig B-3 had a 49% decrease in the treatment zone within the peripancreatic lymph node, Pig B-4 had a 98% decrease in the treatment zone within the pancreas and Pig B-1 had a treatment zone most likely microscopic in size at the end of the 5-week period (Fig. 4b), suggesting involution and resorption of the ablated tissue. Both CT and necropsy showed no unexpected or off-target effects in these animals (Fig. 4a-d). Histological analysis did not identify lesions in the pancreas in these subjects at the final study time point, suggesting that treatments were well-tolerated over the course of their survival and the resulting lesions had been almost completely resorbed by the time of necropsy (Fig. 4e).

Figure 4.

Pancreas targeting computed tomography (CT) and necropsy results. (a) Immediate post-procedure CT with treatment zone between green arrows at the head of the pancreas in one subject. (b) 5-week post-treatment CT: location of the treatment zone with normal pancreas noted at the site (green arrow). No unexpected or off-target effects were identified. There were no notable lesions or damage outside of the planned treatment volume observed upon gross necropsy; representative images from the same subject as CT scans showing (c) the peritoneal cavity, (d) normal pancreas and (e) histological staining of the pancreas. No cellular destruction within the pancreas was identified by staining in these subjects, as represented by (e).

One subject (Pig B-6) had a treatment zone within the pancreas with preservation of a critical structure post-treatment. The treatment zone was seen as a well-defined area of hypo-enhancement in the pancreatic head on immediate post-procedure CT, with a patent gastroduodenal artery traversing the treatment zone (Fig. 5a-c), consistent with prior histotripsy studies where there is maintenance of blood vessels in the treatment zone [54]. There was free fluid and mesenteric edema adjacent to the treatment zone, consistent with acute inflammatory changes. Additionally, there was a non-occlusive thrombus at the periphery of the main portal vein with extension into the liver. Necropsy and histological results for this subject were consistent with the findings in Figure 4 (c-e).

Figure 5.

Case of vessel-sparing. Immediate post-procedure computed tomography with a well-defined treatment zone between the green arrows. The yellow arrow denotes edema in the mesentery adjacent to the pancreas. (a) The orange arrow identifies preservation of the patent gastroduodenal artery traversing the treatment zone. (b) Yellow arrows denote edema in the mesentery between the pancreas (caudal to the image) and the stomach. (c) Non-occlusive thrombus in the main portal vein (dark blue arrow) with extension into the liver (purple arrow).

While the majority of subjects responded well to histotripsy treatments when the pancreas was clearly visualized, one subject (Pig B-4) presented with mild pancreatitis post-histotripsy that remained stable throughout the study without requiring interventional treatment or causing signs of distress. In this case, the treatment zone was identified in the pancreas on immediate post-treatment CT with no unexpected or off-target effects (Fig. 6a). At the 5-week CT scan, there was mild dilation of the pancreatic duct (Fig. 6b). Necropsy showed no off-target damage, but a subtle lesion was noted on the capsule of the pancreas (Fig. 6c, 3d). Histological analysis showed fibrosis of the pancreas indicating chronic pancreatitis, which was also confirmed by the elevated amylase and lipase levels at 1 and 2 weeks post-treatment (Fig. 6e). A second subject (Pig A-2) was observed to have a similar focus of inflammation and fibrosis in the pancreas noted on histology at 1 week post-histotripsy (Fig. 6f). However, this was not associated with increased levels of amylase or lipase within the serum. This, combined with the lack of gross lesions within the pancreas at necropsy or evidence of clinical abnormalities following histotripsy, supports this histological finding as likely to be sub-clinical with regard to the development of pancreatitis. In the other seven out of nine subjects, there were no other signs of pancreatitis on histology, with amylase and lipase levels also remaining within their reference ranges (Supplementary Table).

Figure 6.

Evidence of pancreatitis in two subjects. (a) One subject showed the treatment zone between the green arrows at the head of the pancreas on immediate post-procedure computed tomography (CT). (b) A 5-week CT scan showed an involuted treatment zone at the green arrow with a mildly dilated pancreatic duct in the pancreatic tail at the yellow arrow. (c, d) Necropsy results show a subtle lesion on the inflamed pancreas (blue arrow) with histological analysis confirming significant chronic pancreatitis (e). (f) A second subject showed similar lesions of pancreatitis histologically but with no associated gross lesions.

Two off-target events were observed in subjects that had bowel damage due to unclear visualization of the pancreas. These subjects had to be euthanized early as they showed deterioration in their health. In one subject (Pig A-3), immediate post-procedure CT revealed the treatment zone as a well-defined area of hypo-enhancement at the right lateral aspect of the pancreas with apparent extension of the treatment zone to involve an adjacent small bowel loop (Fig. 7a). There were no signs of bowel wall compromise, such as air in the wall of the bowel, air in the peritoneal cavity, extraluminal fluid or inflammatory changes. Necropsy results revealed septic peritonitis (Fig. 7b) most likely due to GI perforation, although the perforation could not be identified during necropsy. Off-target damage was not observed in any of the other tissues nor was damage to the pancreas (Fig. 7c). Histological results revealed the pancreas to be autolyzed (Fig. 7d) due to septic peritonitis making it present differently than the normal pancreas. The duodenum presented with inflammation and ingesta free material (brown, honeycombed material indicated by black arrows, Fig. 7e). The presence of this material within the abdominal inflammation confirms a GI perforation even though it could not be identified during necropsy. For Pig B-5, a treatment zone was noted at the periphery of a bowel loop, extending into the mesentery on immediate post-procedure CT (Fig. 7f, 7g). There was edema of the adjacent bowel wall, although no pneumatosis or pneumoperitoneum was noted. Upon necropsy, the pig had a notable GI blockage of an adhesion between the duodenum and jejunum with autolyzed overlying tissue and pancreas (Fig. 7h, 7i). Histological results did not show ablative damage to the pancreas but sporadic groups of acinar degeneration and necrosis attributable to sepsis (Fig. 7j). These results may be treatment-related because the subject’s planned treatment zone extended ~1–4 mm beyond the pancreas, encompassing part of the small bowel. The same volume was used to keep all treatment volumes uniform throughout the study, although the volume could not be contained inside the pancreas of this subject due to the pancreas appearing smaller in size on US imaging. A third subject also had mis-targeting of the bowel but did not show any changes in vital signs or behavior, nor any adverse events. This subject reached its 5-week survival, suggesting the lesion resolved over the course of the study.

Figure 7.

Off-target effects. (a) One subject had an immediate post-procedure computed tomography (CT)with the treatment zone identified within the pancreas (green arrows) and extension into the small bowel (light blue arrows). (b, c) Necropsy identified septic peritonitis with no damage to the pancreas, but tissue was autolyzed compared with the normal pancreas. (d) Histological analysis shows an autolyzed pancreas and the duodenum covered in a thick layer of inflammation, which is mostly neutrophils as well as fibrin and necrotic cells (e). (f) Another subject’s immediate axial CT image showed a treatment zone involving the bowel wall and mesentery between green arrows. Light blue arrows denote an edematous bowel wall, which is seen as thickened and less dense than the normal bowel wall (orange arrow). The yellow arrow denotes gas within the lumen of the edematous bowel loop. (g) Coronal CT image showing an edematous bowel wall at the light blue arrows compared with a normal-appearing bowel wall at the orange arrows. The yellow arrow denotes gas within the bowel lumen. (h) Adhesion between the duodenum and jejunum is noted as a gastrointestinal blockage. (i) The pancreas appears to be autolyzed. (j) Histological analysis shows sporadic individual to small groups of acinar degeneration and necrosis attributable to sepsis.

Discussion

In this study, we investigated the safety profile of histotripsy pancreas treatments in a healthy swine model. This study was designed to evaluate the response to histotripsy pancreas ablation with 1- and 5-week survival groups. Overall, the results from this study support our first hypothesis that histotripsy can be safely applied to the pancreas without inducing off-target injury as long as the histotripsy bubble cloud is precisely confined to a targeted region within the pancreas. Bubble clouds were successfully generated inside the head or body of the pancreas where the targeted tissue could be clearly visualized on coaxial US imaging, with suspected bubble clouds in all other subjects based on the auditory sounds during treatment as well as the post-treatment effects observed on US and CT. As seen in previous liver studies, a benefit of a histotripsy system is the ability to monitor the treatments in real time with ultrasound image guidance as the bubble cloud appears hyper-echoic. In this study, the bubble cloud was visualized in five out of nine subjects, suggesting the need for continued improvements in bowel preparation to further reduce gas blockage during treatment. Hitotripsy lesions were seen to reduce in size or fully involute at 5 weeks post-treatment, which has been observed in previous pre-clinical [52,53] and clinical [55] studies of histotripsy in the liver, suggesting that histotripsy lesions generated in the pancreas will also be well-tolerated by patients.

Inflammation and fibrosis were observed in one 1-week subject (Pig A-2) and one 5-week subject (Pig B-4). However, only the 5-week pig exhibited functional consequences, as evidenced by concurrent serum elevations in amylase and lipase levels, and dilation of the pancreatic duct on CT. The 1-week pig showed no additional changes in the morphology of the pancreas on CT or via gross or histological examination, nor were there changes in serum amylase or lipase. Collectively, this suggests that, at least at the 1-week timepoint, the inflammation and fibrosis were of minimal functional consequence. Additionally, for the 1-week subject, it is not entirely certain whether the observed inflammation was treatment-related or possibly pre-existing. As there was no identifiable treatment lesion observed on gross necropsy or histology, it is difficult to confirm whether this sample was taken from the exact site of treatment. Therefore, it is difficult to confirm whether inflammation was directly caused by the treatment or represents an unrelated finding. For the 5-week subject, it is hypothesized that pancreatitis may have been caused in this subject due to the treatment zone going through the epithelial casing of the pancreas and not being fully encased by pancreatic tissue, which was observed for this lesion during necropsy. It is worth noting that the treatment zone only extended through the epithelial casing of the pancreas for this one subject with elevated serum levels which was not observed in other subjects, supporting this hypothesis. Additional studies are warranted to further explore this possibility. In this subject, amylase and lipase levels were elevated three times the upper reference limit 1 week post-treatment. While the amylase returned to normal at 2 weeks, the lipase level remained elevated for 2 weeks before returning to normal. At the 5-week survival endpoint, the amylase dropped below the lower reference limit (Supplementary Table); these results agree with what is seen clinically for pancreatitis [65-67]. Furthermore, one study investigating the development of postoperative pancreatitis in 52 human patients observed pancreatitis within 7 days of the operation in 26 of the patients [68]. The results from this study and our current findings suggest that monitoring after histotripsy treatment in the pancreas will be critical, especially at the 1-week mark, to identify signs of pancreatitis for mitigation. The acute pancreatitis observed 1 week after histotripsy treatment is not expected to be life threatening if the disease presents as mild edematous pancreatitis, which has a mortality rate of less than 1% [58,68]. It is also worth noting that the current study applied histotripsy to healthy pancreas rather than pancreatic tumors. Applying histotripsy to a pancreatic tumor ranging from 1 to 4 cm in diameter (without ablating surrounding healthy pancreas) is expected to further reduce the risk of inducing pancreatitis in the clinical setting. Our group has successfully demonstrated targeting and treating pancreatic tumors in vivo [69-71]. Together, these findings suggest that histotripsy ablation is likely to be safe, with a similar or reduced risk of inducing pancreatitis compared with other minimally invasive ablation methods as long as histotripsy is confined to the targeted pancreatic tumor.

The results from this study also supported our second hypothesis that critical vessels and ducts can be spared during histotripsy treatment, which agrees with previous reports [44,48]. Treatments were targeted adjacent to the portal vein, which was preserved in all cases. Additionally, one subject showed preservation of the patent gastroduodenal artery that traversed the treatment zone. Of note, one subject had a nonocclusive thrombus in the main portal vein with extension into the liver, which is frequently seen post-histotripsy in swine [52] and human liver [55] treatments in the vicinity of the portal venous branches. Non-occlusive thrombus did not lead to any long-term complications, consistent with prior liver studies showing acute thrombus resolving within 1–4 weeks [52,72]. Overall, these results further highlight how histotripsy’s tissue selectivity offers a potential advantage over thermal ablation methods, which frequently prove ineffective at ablating tissue near major vessels. These studies also demonstrate the potential of generating effective ablation of borderline resectable or non-resectable pancreatic tumors that are located in regions surrounding major blood vessels such as the celiac artery, common hepatic artery, and superior mesenteric artery [73,74].

In general, the results from this study support our third hypothesis that histotripsy can safely be applied through overlying tissues without injuring the bowel when the bubble cloud is maintained within the pancreas. However, this work also highlights the importance of pre-treatment dietary preparation to mitigate overlying bowel gas for improved imaging and targeting. Two off-target events were observed as consequences of mis-targeting the bowel when there was a poor acoustic window, further demonstrating the need for a clinical treatment workflow that prevents treatments from proceeding under these conditions. Histotripsy was still applied in these subjects, as the goal of this study was to evaluate the safety profile of histotripsy even under these adverse conditions that could occur in patients. Although early ex vivo work has suggested safe histotripsy application through partial gas blockage of the bowel [75], the results of our study warrant caution for histotripsy application in the pancreas when treating tissue immediately adjacent to gas-filled bowel. Strategies should be developed to protect the bowel when the pancreas and/or bubble cloud cannot be visualized on B-mode US imaging. These strategies could include advanced passive cavitation detection techniques that employ transmit-and-receive therapy transducers that supply histotripsy pulses and detect cavitation signals [76]. Passive cavitation detection methods would allow pancreatic tissue breakdown monitoring through the cavitation collapse time and could spatially locate the bubble cloud to monitor for off-target cavitation in the bowel [77].

There were also superficial skin injuries observed on two subjects although minor, additionally emphasizing the need for pancreas-specific treatment workflows that are designed to avoid all of these treatment side effects. It is important to highlight that the pig model represents a more clinically challenging target than the human case due to shallower treatment depths (<5 cm), which the clinical histotripsy system was not optimized for. This system maximized histotripsy treatment efficiency rates while being within the allowable thermal limits for human liver treatments and it will be imperative to follow the thermal safety limits for the pancreas clinically. These superficial skin burns are consistent with prior studies demonstrating that thermal injury to the body wall or limb can occur during histotripsy treatments when overly aggressive pulsing parameters are used [51,52,78]. In the future, clinical treatment parameters will need to be expanded to enable safe and non-thermal histotripsy to be applied closer to the skin. Pre-focal cavitation is another concern that must be considered when treating at shallow depths. Although such issues are unlikely to present the same challenges in human pancreatic cancer patients, where the pancreas is typically located deeper than in this pre-clinical study, they remain important considerations for other applications involving superficial targets such as soft tissue sarcoma, osteosarcoma, and others [79-82].

The results from this study suggest that histotripsy can be used for safe and effective pancreas ablation when using proper treatment planning and targeting strategies. However, it is important to note some limitations of this study. First, histological evidence of histotripsy-induced damage was not observed in the 1-week treatment group, which may be attributed to lesion resorption during this time frame. Although posttreatment CT imaging successfully identified lesions in all subjects, the inclusion of pre-treatment CT scans would have provided a more robust comparison and improved lesion characterization over time. Next, for clinical treatment of a pancreatic tumor, therapy would not be applied without confirmed visualization of the target. Overlying bowel gas still affected visualization of the target and bubble cloud even with the pretreatment diet. These findings further underscore the significance of improving image guidance for histotripsy in the pancreas, which is surrounded by critical organs such as the intestine. In order for clinicians to avoid unwanted mis-targeting of bowel or other structures surrounding the pancreas, it is imperative that the targeted tissue is clearly visualized and the focal bubble cloud is monitored within the pancreas. While ultrasound imaging was used in this study for targeting and monitoring of the cloud, other imaging modalities, or a combination, may provide better confirmation and visualization. While current histotripsy systems are not guided by magnetic resonance imaging or CT, fusion imaging has recently been integrated into the clinical systems in order to utilize pre-treatment CT/magnetic resonance images to help with target confirmation and other work is investigating cone beam CT as an alternative or complement to US imaging [83-85]. Our group is also exploring the development of endoscopic histotripsy systems for image-guided histotripsy ablation of the pancreas as an alternative approach to deliver histotripsy precisely in patients that prove difficult to target using the transabdominal approach.

Conclusion

Overall, this study indicated the safety of histotripsy for pancreas ablation and illustrated some of the benefits of histotripsy, including the ability to destroy pancreatic tissue near large vessels, the reduction in size of lesions post-treatment, and the ability to guide and monitor treatments in real time using US imaging. The results of this pre-clinical study suggest that histotripsy can safely and effectively be applied in the pancreas when the target is clearly visualized and a focal bubble cloud is maintained within the pancreas on US imaging. This study included the evaluation of histotripsy under adverse conditions where overlying bowel gas obstructed pancreas visualization and targeting, with results supporting the recommendation that histotripsy is not applied under these conditions in a clinical setting for pancreatic cancer patients. The results highlight the adverse effects of bowel damage when treatments are performed without an effective acoustic window. The results also assessed the risk of inducing pancreatitis, with potential pancreatitis only observed in two out of nine subjects, with only one subject showing functional consequences with elevated serum levels, corresponding to a rate below what has been reported for other therapies. This finding further highlights the potential of histotripsy to be developed as a safe and effective approach for non-invasively ablating pancreatic tumors.

Supplementary Material

Supplementary material associated with this article can be found in the online version at doi:10.1016/j.ultrasmedbio.2025.07.026.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.