The study protocol was approved by Tufts University’s Institutional Animal Care and Use Committee (G2021-152). Informed owner consent was required prior to enrollment. Eligibility criteria included dogs of any sex and breed, a minimum of 5 kg in body weight, and between the ages of 1 and 12 years of age with two cytologically confirmed similarly sized lipomas. Target tumor size was a minimum 2 cm and a maximum of 7 cm based on the longest diameter assessment. Additional criteria included a normal physical exam and lack of abnormalities on complete blood count and chemistry panel. Dogs were excluded from the study if they were pregnant, lactating or were deemed ineligible by investigators based on any significant liver, renal or cardiovascular disease.

BXT-786 is a proprietary mix of 0.9% saline and glycerol that permits the formation of an ice slurry amenable to injection through a large gauge needle. Custom-designed syringes containing room temperature, sterile BXT-786, were stored frozen at −22 to −18°C prior to preparation and injection. A custom-designed device (“dock”) was used for generating the BXT-786 coolant slurry immediately prior to injection.

Following screening and enrollment, two similarly sized lipomas on the same patient were designated A and B. Using a standard randomization table, these were assigned to be injected with either BXT-786 coolant or control room temperature 0.9% saline solution. Two blinded, independent observers measured the length, width and height of each lipoma twice, first with calipers and then using ultrasound (Toshiba Apilo 500, Toshiba Viamo PLT-1204 BT Transducer, San Jose, CA, USA) on Day 0. Observers performed their measurements separate from each other and were not present during the injection procedure. Dogs were sedated for injections with intravenous butorphanol alone (0.2 mg/kg), or in combination with dexmedetomidine (2 μg/kg); atipamezole of equal volume was used for intramuscular reversal in those cases. Lipoma sites were clipped and aseptically prepped prior to injection. BXT-786 slurry was administered at a maximum dose of 20 mL per lipoma based on the 7 cm size limit for the lipomas and the presumptive radius of cooling induced by this volume. A similar dose, to the fullest extent possible based on lipoma size, was used for the 0.9% saline injected into the control lipoma (Table 1). This volume was extrapolated from the previous swine study in which a total of 30 mL of BXT-786 was injected into the subcutaneous fat (8). A 17-gauge needle was used for both the BXT-786 slurry and saline injections.

Following the injection procedure, dogs were observed for 1 h to monitor for acute reactions. Dogs were reevaluated at 2, 4, 8, and 12-weeks post-injections to assess for adverse events and to repeat caliper and ultrasound measurements. At 2 weeks post injection, a complete blood count and chemistry profile were repeated. Both lipomas were surgically removed at 12 weeks post injections and submitted for histopathologic analysis. Health related quality of life was assessed via owner completed questionnaire at day 0 and week 8 using a previously established tool (9).

Response to treatment was assessed using the Veterinary Cooperative Oncology Group (VCOG) response evaluation criteria in solid tumors (RECIST v 1.1) (10). Two approaches were used. The first involved single observer baseline and final caliper or ultrasound longest diameter measurements for each tumor (A and B) using the following formula to determine percent change from baseline: [final length-baseline length]/baseline length. To account for variability in assessor measurement that might affect response assessment, the average baseline longest diameter measurement (caliper or ultrasound) across the two observers for each target lesion at baseline and at the final time point was used to determine a “final” response assessment: [final average length-baseline average length]/baseline average length. A complete response (CR) was defined as complete resolution of the tumor, either by measurement or as documented by histopathology (10). A partial response (PR) was defined as a ≥30% reduction in the target lesion, taking as a reference the baseline measurement (10). Progressive Disease (PD) was defined as a 20% or greater increase in the longest diameter of the target lesion, taking as a reference the baseline measurement (9). Stable Disease (SD) was defined as neither sufficient shrinkage to qualify as CR/PR, nor sufficient increase to qualify for PD (10).

Adverse events were reported and graded in accordance with the standard Veterinary Cooperative Oncology Group – Common Terminology Criteria for Adverse Events (VCOG-CTCAE) (11). Patients experiencing adverse events received appropriate medical therapy. Principal investigators graded the severity of adverse events and assigned attribution of the adverse event as “unrelated,” “unlikely,” “possible,” “probable” or “definite” in relation to coolant injection (11).

The reliability testing on tumor volume was performed by computing the concordance correlation coefficient (CCC) along with its 95% confidence interval. Reliability analyses were computed to estimate the reliability testing for repeated measures on tumor volume for ultrasound and caliper separately between the two observers and for the reliability testing for repeated measures on tumor volume irrespective of the measurement device used. A mixed effects model was used to compute the CCC, evaluating the observer effect on the change in tumor volume across visits, a subject specific random intercept was evaluated to account for the within-subject correlation due to repeated measurements using the epiR packages in RStudio (12). For the model that examines reliability testing irrespective of the measurement device used, a device measurement effect on the change in tumor volume across visit is included in the mixed effects model. A Bland–Altman plot was used to visualize the agreement with the limits of agreement (bias ± 1.96 std) calculated using Bland and Altman’s formula for repeated measures using the BlandAltmanLeh package in RStudio (13, 14). A paired t-test was used to assess changes in quality of life comparing baseline assessment to that at 8 weeks post injection.

Ten client-owned dogs were enrolled in the study following written owner consent. All medical histories and physical examinations obtained on enrolled dogs were clinically normal and revealed no overt signs of illness. Mean patient age was 10.18 years. The patient population consisted of eight castrated males and two spayed females. Breeds included mixed breed dogs (n = 5), Labrador Retrievers (n = 2) and one each of the following breeds: American Staffordshire Terrier, Belgian Malinois, and Cocker Spaniel.

While every effort was made to identify dogs with similarly sized lipomas, this was not always possible. Consequently, the volume of injection of the 0.9% saline and the BXT-786 coolant was sometimes variable (Table 1, Case 1, 6, and 9). Additionally, due to the constraints of lipoma size, there were several instances where less than 20 mL total volume of coolant or saline was injected into the lipoma to prevent leakage of material out from the injection site secondary to increased interstitial pressures.

Response was assessed using RECIST criteria by measuring the longest diameter of each lipoma using calipers or ultrasound (Table 2). Based on caliper assessments, the majority of the tumors were determined to be stable in size prior to surgical excision by both observers. Discordant response assessments occurred for 2/10 of the placebo treated tumors and 3/10 of the BXT-786 injected tumors. However, based on the average response assessment across the two observers, 9/10 placebo and 7/10 BXT-786 were deemed to have achieved SD. Only two tumors were determined to have a PR by caliper assessment (one placebo, one BXT-786).

Baseline visit ultrasound longest diameter measurements are not available for patient 1 as the initial approach to ultrasound assessment using the Butterfly IQ+ device was found to be inadequate; the device was not suitable for capturing accurate length of the tumors due to limitations with the probe. Consequently, the Toshiba Apilo 300 system was used for all subsequent patients to provide more accurate assessments that could be readily stored in the Tufts PACS server for reference. Greater variability in response assessment was observed when using ultrasound as the measurement tool. In this setting, discordant response assessments occurred for 3/9 of the placebo treated tumors and 4/9 of the BXT-786 injected tumors. However, based on the average response assessment across the two observers, 6/9 placebo and 6/9 BXT-786 treated lipomas were deemed to have achieved SD using ultrasound. No tumors were determined to be PR by ultrasound. Importantly, no CRs were noted for any lipomas, regardless of the method used for assessment.

A total of six adverse events were reported in four patients (Table 3), although most of these were deemed unrelated to the BXT-786 treatment. One dog developed cellulitis and infection at the coolant lipoma injection site necessitating early removal of the lipoma. Health related quality of life assessment aggregate data scores at baseline and week 8 post injection were compared. No statistically significant change was found over this time (95% CI −2.06, 7.3%, p-value 0.2405) indicating that the injections did not negatively impact quality of life (Figure 1).

Four measurements (ultrasound assessments of both lipomas by two independent observers, caliper measurements of both lipomas by two independent observers) were taken prior to treatment and at weeks 2, 4, 8, 12 of the study. Using the volumetric formula length × width × height = volume, the size of the tumor volume measured by ultrasound ranged from 1,297 to 76,319 mm³ with a mean of 2,030 mm³. The same tumors measured by caliper ranged from 3,240 to 223,015 mm³ with a mean of 41,284 mm³. This discrepancy is not unexpected given that calipers also include skin in the lipoma measurement.

Repeated measures of tumor size (volume calculations) were tested for reliability within each measurement modality. Relative concordance was achieved between observers for caliper (0.76, 95% CI: 0.66–0.83) while the concordance correlation coefficient was slightly larger for ultrasound (0.82, 95% CI: 0.74–0.88). Reliability testing for repeated measure on tumor volume irrespective of the measurement device measurement used, resulted in a relative concordance (0.79, 95% CI: 0.73–0.84). Concordance between ultrasound and caliper was also visually assessed using a Bland–Altman plot (Figure 2).

Histological evaluation of lipomas post excision demonstrated evidence of fibrosis, inflammation and necrosis in 8 of the 10 lipomas injected with BXT-786 coolant (Table 4). These changes are consistent with the proposed mechanism of action of the coolant. Conversely, inflammatory histopathologic changes were only reported in one lipoma injected with control room temperature sterile 0.9% saline. It is possible that in this instance, the large bore needle induced tissue injury that contributed to the observed changes.