A Retrospective Multicentric Study of Electrochemotherapy in the Treatment of Feline Nasal Planum Squamous Cell Carcinoma

Feline squamous cell carcinoma (SCC) is currently treated with surgery, radiation therapy and electrochemotherapy (ECT). Both the efficacy and/or safety of ECT were evaluated as a sole therapy with bleomycin to treat feline nasal planum SCC (npSCC). Sixty-one cats were enrolled. Local treatment response was evaluated as complete remission (CR), partial remission (PR) or stable disease (SD). Recurrence rate (RR), disease-free interval (DFI) and progression free survival (PFS) were calculated. A six-point scale was used for ECT toxicity. The median tumor size was 1.5 cm. CR was achieved in 65.6% of cases, PR in 31.1% and SD in 3.3%. The overall response rate was 96.7%, RR was 22.5%, median DFI was 136 days, and median PFS was 65.5 days. ECT toxicity was ≤2 in 51% of cats. Tumor recurrence/progression (p = 0.014) and local treatment response (PR: p < 0.001; SD: p < 0.001) influenced survival time. Cats with toxicity >2 showed a higher probability of tumor recurrence/progression. Tumor-related death was higher in cats with PR (p < 0.001) and recurrence/progression (p = 0.002), in ECT treatment with 1 Hz (p = 0.035) and 1200 V/cm (p = 0.011) or 1300 V/cm (p = 0.016). Tumor size influenced local treatment response (p = 0.008) and toxicity (p < 0.001). ECT is an effective treatment for feline npSCCs and should be considered as the first-line procedure for low-stage tumors.


Introduction
Squamous cell carcinoma (SCC) is a malignant tumor arising from epidermal cells with differentiation to keratinocytes (squamous cells). In the skin, squamous cells can be found in the epidermis, the dermis, as well as nail beds and foot pads, along with the oral tissue sarcomas, perianal neoplasms, oral squamous cell carcinomas, melanomas and nasal tumors in dogs. The feline literature is mostly restricted to cutaneous SCCs and soft-tissue sarcomas [29][30][31][32][33][34][35].
ECT is minimally invasive, with minor or no side effects. It can be repeated several times, requires general anesthesia and the costs are lower than other treatment approaches, such as radiation therapy [21]. ECT can also be used in combination with several other therapeutic approaches, such as surgery, radiation therapy and gene therapy [30,36,37]. ECT has been used as a single-treatment approach [9,38] or combined with the current standard of care approaches for the treatment of nasal planum SCC (npSCC), such as surgery [33,39]. One report also described a case in which it was combined with radiation therapy, with some local complications [40].
The aim of this multicentric and retrospective study was to evaluate the efficacy and safety of ECT in the treatment of feline npSCC in a case series.
From the clinical records available at each facility, details on the study population (breed, age and sex) and information on tumor size (caliper measurements) were collected.
ECT was offered to the owners as an option alongside standard treatment approaches, such as surgery and/or radiation therapy. Informed consent illustrating the entire ECT procedure was signed by each owner. ECT as the sole therapy was required for inclusion in this study.

Staging
The patients were diagnosed with npSCC using cytology (17/61) or histology (34/61) and, in some cases, with both (10/61). Patient staging was performed in cats that already had medical records with complete information regarding clinical examination, fine needle aspirates of the regional lymph nodes if enlarged and three view thoracic radiographs. Skull radiographs were performed depending on the clinician in charge at the time of the diagnostic work-up.
Only cats without metastatic disease in the regional lymph nodes or lungs were included in the study.
Since ECT was performed under general anesthesia, a complete blood count, coagulation profile and biochemistry profile were performed to rule out underlying diseases at the clinician's discretion. All the cats, except two showing unrelated chronic kidney disease (CKD, IRIS II), had unremarkable blood chemistry parameters.

Electroporators, Electrical Parameters and Anesthesia Protocol
The equipment and their electrical characteristics (frequency and amplitude to electrode distance ratio) used for each case treated are listed in Table 1. Although different pulse generators were used in this multicentric study, all delivered 8 monophasic square pulses of 100 µs each. The different protocols used for general anesthesia depended on the veterinary facility and the clinician in charge (see Table 1).

Treatment (ECT) Protocol
ECT was combined with the standard dose of bleomycin (15,000 UI/m 2 , European Farmacopeia) except for one cat, treated at GVS, for which a lower dosage was used (10,000 UI/m 2 ). Based on recent publications, lower doses of bleomycin (10,000 IU/m 2 rather than 15,000 UI/m 2 ) were shown to be equally effective in clinical response. One researcher (JAI) used the lower dosing regimen as a standard approach with ECT [41,42]. In all cases, bleomycin was administered intravenously. Eight minutes after drug administration, the clinician proceeded with the application of electric pulses.
All the cats were treated with ECT alone. A single ECT treatment was performed in 39/61 (63.9%) cats. The remaining cats, 22/61 (36.1%), had more than one ECT due to recurrence, presence of stable disease (SD) or to potentiate treatment efficacy ( Table 2).
After ECT, cats were administered additional support therapy if needed. The posttreatment plan differed at each facility ( Table 2).

Follow-Up and Treatment Outcome
ECT patients were monitored once or twice a week for the first four weeks, followed by a monthly checkup for an additional three months and then every 3-4 months. The followup visit was performed by the referral veterinarian. Some clinicians also photographed the healing process and the progression of the patients.
The local response to the treatment was evaluated as: (CR) complete remission and no sign of the primary tumor; (PR) partial remission, i.e., at least 30% reduction in the tumor but not the disappearance of the mass, and (SD) stable disease where the size of the tumor showed no reduction or enlargement (less than 30% reduction in size (PR) or less than 20% increase in size (PD)-progressive disease) [43].
Treatment response rate, survival time (time from first treatment to the day of the last checkup or death) and median survival time (MST) were calculated. In cats that experienced CR and subsequent local recurrence, the recurrence rate (RR) and diseasefree interval (DFI) were calculated. DFI was defined as the time from first treatment to the day of local recurrence or evidence of metastasis. In cats that experienced PR or SD, progression free survival (PFS) was defined as the time from first treatment to the day of local progression of the disease or evidence of metastasis. Finally, the treatment outcome of the patients was evaluated: alive without tumor, alive with tumor, dead without tumor and dead with tumor. MST was calculated for the deceased cats and median follow-up was calculated for the cats that were still alive at the conclusion of the study period.

Statistics
The data were initially analyzed as descriptive statistics. Size, age, DFI, PFS and survival times were analyzed as non-parametric, continuous variables and presented as median and range. Categorical variables (gender, breed and number of treatments) and continuous variables (pulse frequency, pulse to electric distance ratio, local treatment response and toxicity score) were presented as absolute and relative frequencies. Local treatment response was divided at the beginning into three categories as follows: CR, PR and SD. However, due to only a few cases with SD, PR and SD were considered as a single category when local treatment response to ECT was investigated. The toxicity score for each cat was subtyped into two groups: toxicity score ≤2 and >2. Alive cats with neither tumor progression/recurrence nor metastasis and cats that died due to tumor-unrelated causes were censored at the last follow-up. The only cat treated at GVS was not included in the following statistical analysis, since a specific ECT protocol was applied (Table 1).  OVG/Electrovet EZ (10) 0.5-6 Meloxicam 0.05 mg/kg, PO and Amoxicillin and clavulanic acid 20 mg/kg, PO for 5-7 days. For wound cleaning and crust removal: saline and local ointments: VEA cream PF ® (antioxidant cream with Vit. E and Polyphenol), One Vet spray ® (neem oil, St John's wort oil and olive oil), Iruxol ® (collagenase and chloramphenicol) or Hypermix ® (neem oil and St John's wort oil) The data were evaluated by inferential statistics. Kaplan-Meier curves and two-sided log-rank tests were used as the univariate analysis to compare survival times and times of recurrence/progression occurrence among categorical and continuous variables. Cox regression (proportional hazard regression) analysis was used for continuous variables (age and size), which, if significant, with a p < 0.05, were entered into multivariate analysis (Cox regression, proportional hazard regression).
A univariate analysis (binary logistic regression analysis) was used to establish the relationship between gender, pulse frequency, amplitude to electrode distance ratio, presence of recurrence/progression, age and tumor size and the tumor specific survival, local treatment response and the toxicity score. Variables with a p < 0.05 on univariate analysis were entered into multivariate analysis (binary logistic regression analysis).
For each continuous variable significantly associated with survival, local treatment response and toxicity, a receiver operating characteristic (ROC) analysis was performed and the area under the curve (AUC) was calculated. Diagnostic cut-offs and their sensitivity and specificity were determined according to the maximum Youden index.
All statistics were analyzed with IBM SPSS ® Statistics.

Results
The study population of 61 cats was mostly represented by the domestic short hair (DSH) (56/61, 91.8%) and only five cats (8.2%) were domestic long hair (DLH). The median age of the study population was 11 years old and ranged from 5 to 17 years. All cats included had been neutered, and they were almost equally distributed between females (31/61, 50.8%) and males (30/61, 49.2%). The median tumor size was 1.5 cm (range 0.2-6.0 cm).
The survival time of the whole study population ranged from 13 to 2929 days, with an MST of 286 days. Overall, the RR for cats with CR (40/61, 65.6%) was 22.5% (9/40). Cats with local tumor recurrence after CR had a median DFI of 136 days (range 29-302 days). Ten out of 61 cats (16.4%) that achieved PR (9/10) and SD (1/10) had local progression with a median PFS of 65.5 days (range 16-264 days).
At the end of the study period, 14/61 (23%) cats were still alive and 47/61 (77%) cats died. Among the cats that were still alive at the end of the study period, twelve (12/14, 19.7%) cats were in CR and 2/14 (3.3%) still had a tumor after treatment. In the group of cats that died during the study period (47/61.8%), twenty-four (24/47, 39.3%) died without a tumor with CR, and twenty-three (23/47, 37.7%) died with a tumor with PR or SD. The MST for cats that died without tumor was 872 days (22-2,929 days) and 193 days (13-362 days) for those that died with a tumor. The median follow-up for cats still alive at the end of the study period was 394 days (range 29-2798 days).
The Kaplan-Meier graphs for survival time (days) are reported in Figure 1a   The Kaplan-Meier graphs for recurrence/progression (DFI/PFS) (days) are reported in Figure 2a  The Cox regression analysis for continuous variables showed significant results only for tumor size (p = 0.004, OR 1.47, 95%CI 1.13-1.92). The multivariate analysis for DFI/PFS was significant only for toxicity score (toxicity ≤2 as reference category: p = 0.025; toxicity >2 OR 3.03, 95%CI 1.15-7.99).
Data from the univariate analysis and multivariate analysis used to establish the relationship between variables are presented in Table 3 for tumor-specific survival, Table 4 for local treatment response and Table 5 for toxicity. Local treatment response and recurrence/progression were independently associated with tumor-specific survival. Tumor size was associated with local treatment response. Tumor size >1.7 cm was associated with higher risk of PR/SD (AUC 0.79, 95%CI 0.67-0.91, p < 0.001, sensitivity 76%, specificity 72%). Tumor size >1.1 cm was associated with higher risk of toxicity score >2 (AUC 0.88, 95%CI 0.78-0.98, p < 0.001, sensitivity 93%, specificity 83%).  Data from the univariate analysis and multivariate analysis used to establish the relationship between variables are presented in Table 3 for tumor-specific survival, Table 4 for local treatment response and Table 5 for toxicity. Local treatment response and recurrence/progression were independently associated with tumor-specific survival. Tumor size was associated with local treatment response. Tumor size >1.7 cm was associated with higher risk of PR/SD (AUC 0.79, 95%CI 0.67-0.91, p < 0.001, sensitivity 76%, specificity 72%). Tumor size >1.1 cm was associated with higher risk of toxicity score >2 (AUC 0.88, 95%CI 0.78-0.98, p < 0.001, sensitivity 93%, specificity 83%). Table 3. Univariate and multivariate analyses for the relationship between variables and tumor-specific survival.

Discussion
The present retrospective study evaluated the efficacy and safety of ECT in the treatment of feline npSCC in a group of 61 cats. To date, only a few studies have been published on this topic, with a limited study population [9,32,38,44].
As with all facial SCCs, npSCC can also be locally invasive, with the outcome of the treatment approaches influenced by the stage and invasiveness of the tumor, with an improved outcome in lower-stage lesions (such as Tis and T1) [9]. Because of the major cosmetic and functional defects after surgery, nosectomy, which was considered the standard care for the treatment of canine npSCCs, should not be considered as the first-choice treatment for these tumors in cats when ECT is available [2]. Studies describing the use of cryosurgery have reported CR rates of up to 84% with fairly long response durations [10]. Other reports, however, have reported high recurrence rates (up to 73%) even for low-stage tumors, due to the impossibility of the evaluation of surgical margins after cryosurgery [4,10,11]. Another major limitation of cryosurgery for feline npSCCs is the multiple treatments needed in most cases [4,10,11].
Radiation therapy has been used to treat npSCCs in cats using many different protocols, including orthovoltage, megavoltage and proton beam irradiation. The main limitations of radiation therapy include the multiple anesthesias needed for treatment planning and delivery, the occurrence of early and late side effects after treatment, and the high cost [2,4,11,12,15,21].
Brachytherapy with strontium-90 has shown very good results, with favorable CR rates in two studies [13,14]. However, when the treatment was evaluated in a larger population, the occurrence of new lesions outside the radiation field penumbra was reported in 33% of patients [14]. In a recent study, the response to treatment was 100%, with a median DFI of 916 days [21].
Photodynamic therapy has also been used in feline npSCCs with excellent results. Although it can be safely repeated, when applied to deep seated tumors, the recurrence rate is high [16,17]. With this treatment, the photosensitizer can be administered both topically and systemically. When administered topically in feline SCCs, 85% of cats in the study obtained CR; however, 51% of them had tumor recurrence after a median time of 157 days [16]. Conversely, when the photosensitizer was administered intravenously, the response rate was 100%, with a 20% recurrence rate. One year after the treatment, 75% of the population maintained local control [17].
ECT, the topic of the present paper, is a novel treatment that combines both anti-tumor and favorable cosmetic effects at a reasonable cost, and it is increasingly used in veterinary oncology [9,22].
In a report by Spugnini et al. including nine cats with mostly npSCC treated with ECT combined with intratumoral administration of bleomycin, complete local control was obtained in 77.7% cats, which lasted for three years. Only two cats in the study died with local tumor recurrence [38]. Tozon et al. described a single ECT session coupled with systemic bleomycin for the treatment of SCCs located on the head in a population of 11 cats. The authors reported that CR was achieved in 81.8% of the population, with only two cats experiencing tumor recurrence, two and eight months, respectively, after achieving CR [9].
In a second study, Spugnini et al. reported the use of ECT coupled with systemic bleomycin for the treatment of feline periocular SCC and advanced SCC of the head. In this report, an overall response rate of 89% (21/47 CR and 2/47 PR) with a median PFS of 30.5 months was achieved [32].
Dos Anjos et al. compared two ECT protocols, using the standard dose of bleomycin (15,000 UI/m 2 ) and a reduced dose (10,000 UI/m 2 ) in cats with cutaneous SCC, treated with ECT alone. A group of 56 cats was included in the study, divided into two groups: 22 treated with the standard dose and 34 with the lower dose of bleomycin. The response rate and median DFI were 87.4%, 240 days and 100%, 210 days, respectively [44].
To the best of our knowledge, ours is the first study on the use of ECT alone on a large population of cats affected by SCC located exclusively on the nasal planum.
We combined ECT with the systemic administration of bleomycin, achieving a very good response rate with low RR or progression of the disease. The population consisted of 61 cats with non-metastatic npSCC, with a median tumor size of 1.5 cm. Similar data on tumor size have been reported in other studies of ECT in feline SCCs [9,32,38].
ECT was used as the sole therapy in all 61 cats, 96.7% of which responded to the treatment and 3.3% remained in SD. CR was achieved by 65.5% and PR by 31.1% of the cats included in the study. Our data on the treatment response rate are similar to, if not slightly better than, previous studies that describe ECT in feline SCCs, in which the RR ranged from 77 to 100%. The degree and completeness of the response also varies among studies and greatly depends on the clinical stage and tumor size [9,32,38,44].  obtained a CR rate of 44.6% but they included periocular and advanced SCCs of the head. However, when lower-stage tumors were included, CRs up to 81.8% were reported. Caution is needed when comparing our data with the previous literature, as this is the first study that only considers feline SCCs located on nasal planum treated with ECT. In our study, most of the patients (63.9%) required only one ECT session, which is in accordance with previously published studies [9,32,38,44]. However, 22 cats (36.1%) were treated more than once (Table 2). Multiple ECT sessions were mostly performed in cats that responded to the first ECT with PR or SD, and in cats that resulted in CR after one treatment but then had a recurrence. Among those cats treated with multiple ECT sessions, five patients received more than one ECT even if CR had been obtained after the first session, in order to potentiate the CR. This decision was taken by the clinicians in charge when a possible recurrence of the disease was suspected, based on the appearance of the tumor. In this study, the overall MST was 286 days (range 13-2929 days) and the overall RR for cats in CR was quite low (22.5%), with a median DFI of 136 days (range 29-302 days). Cats with local disease progression (10/61) had a median PFS of 65.5 days (range 16-264 days).
Local treatment toxicity was assessed using a previously published, subjective sixpoint scale [29]. This is the first study where this grading system has been applied in cats for the evaluation of local treatment toxicity due to ECT, as, until now, it has only been reported in canine tumors treated with ECT [29][30][31]. In the present study, local treatment toxicity after the treatment was absent or mild (score ≤ 2) in 51% of patients. However, the remaining 49% of our study population had a toxicity score of >2. Although an objective comparison with the literature is difficult due to the subjective nature of this toxicity scoring system, our data regarding local treatment toxicity are consistent with findings previously described in other studies on feline SCCs [9,32,38].
From the multivariate analysis, the two variables with the highest influence on the survival time and with an increased probability of cats dying from tumor were the presence of tumor recurrence or progression of the disease (p = 0.014) and the local treatment response. Cats with PR (p < 0.001) or SD (p < 0.001) had a 23.97-and 29.42-times higher probability of dying of tumor-related causes, respectively, compared to the cats in CR. This is not surprising since cats with recurrence or progression of the disease tended to demonstrate a poorer prognosis in most of the published studies [44].
Recurrence or progression of the disease was significantly influenced by the local treatment response (p = 0.024) (Figure 2a), tumor size (p = 0.004) and toxicity score (p = 0.018) (Figure 2b). While a correlation between local treatment response and tumor size with recurrence or progression of the disease appears logical, and has been confirmed in the literature also considering other treatment approaches, the link between treatment outcome and local toxicity is surprising. In fact, the multivariate statistical analysis showed that cats with a toxicity score of >2 had a three-times higher probability of having a tumor recurrence or progression of the disease than cats with a toxicity score of ≤2 [9,32,[44][45][46]. To the best of our knowledge, to date, no association between a higher number of local side effects and the recurrence or progression of the disease in tumors treated with ECT has been described. These data highlight the possibility that higher toxicity could influence the efficacy of the treatment and should be considered in future studies.
Death due to the presence of the tumor was significantly higher in cats with PR (p < 0.001) compared to the cats that achieved CR. Moreover, cats treated with 1 Hz (p = 0.035) had a lower chance (OR 0.07) of dying from a tumor compared to the cats treated with 5000 Hz. Similarly, cats treated with 1200 V/cm (p = 0.011) and 1300 V/cm (p = 0.016), compared to the cats treated with 1000 V/cm, had a higher probability of tumor-related death. Lastly, cats with recurrence or progression (p = 0.002) were logically more susceptible to tumor-related death. The impact of the local treatment response and of the recurrence or progression was also confirmed in the multivariate logistic binary regression (Table 3).
Local treatment response was influenced by pulse frequency (p = 0.035), meaning that cats treated with 1 Hz had a higher probability of achieving CR than cats treated with 5000 Hz. This is quite surprising as studies on both human and veterinary patients have proven that the application of 1-Hz or 5000-Hz pulses does not influence the clinical outcome. In light of our findings, data should be carefully considered and confirmedideally with a prospective, comparative study [26,47]. Cats with a toxicity score of ≤2 had a higher probability of achieving CR compared to cats with a toxicity score of >2. Local treatment response was also influenced by size (p = 0.001). The multivariate analysis confirmed that local treatment response was highly influenced by tumor size (p = 0.008) ( Table 4).
Local treatment toxicity was influenced by pulse frequency for the frequency of 500 Hz leading to lower toxicity compared to 5000 Hz (p = 0.008). This finding has also never been reported in the literature and is possibly influenced by the multicentric nature of our study. In preclinical models, pulse frequency has been reported to influence the number and intensity of muscular contractions and therefore possible treatment-associated pain, which would be difficult to evaluate in veterinary patients [48,49]. In addition, the amplitude to electrode distance ratio also had an impact on the toxicity of the treatment, showing that tumors treated with 1300 V/cm showed higher toxicity than those treated with 1000 V/cm (p = 0.01). This result was also found by Torrigiani et al. (2019); however, it has not been reported elsewhere. Tumor size also significantly influenced the toxicity score (p < 0.001) ( Table 5), which is in agreement with the literature on humans. In fact, one study describing the risk factors of pain associated with ECT treatment of cutaneous metastasis of various cancers found that the size of metastasis was associated with a higher pain score. The authors speculated that large, and possibly necrotic, metastases would also take more time to heal. In light of our results, the owners of cats with larger npSCCs should be warned of the risk of more severe side effects [48].
Lastly, as already reported in other tumor types, size was closely associated with treatment outcome. In our case, the cut-off value that guaranteed a higher probability of CR was <1.7 cm [31,50]. Since ECT causes apoptosis and tissue necrosis, the association between tumor size and a higher risk of a toxicity score of >2 when the tumor size is >1 cm does not seem surprising. The larger the tumor treated with ECT, the larger the area that undergoes tissue necrosis, and the higher the chances of local toxicity reaching scores above 2. However as reported above, more studies should be performed to confirm this hypothesis.
The main limitation of the present study is the retrospective nature of the collected data. In addition, the use of different types of electroporators as well as different electroporation protocols due to the multicentric nature of the study prevents definitive conclusions from being drawn on the ideal ECT protocol for cats with npSCC. A study on a larger population using only one type of electroporator and a single protocol would be ideal. However, the development of a standardized ECT protocol for npSCCs in cats was not included among the aims of the present study and we believe that the variability of instrumentations and protocols reflects the reality of clinical practice. Nevertheless, it would be interesting to conduct a prospective study including a comparison of different electroporators and protocols used in order to establish an ideal ECT procedure that would guarantee the best outcome for the patient. Moreover, although the local toxicity scoring system has been applied effectively in previous studies, its subjective nature is another limitation of this study.

Conclusions
The results of our study confirm that ECT is an effective and safe treatment for feline npSCCs and it could thus be considered one of the treatments of choice, especially in low-stage tumors. In our study population, the statistical analysis identified that tumor size and local treatment response were the variables that most influenced survival time and tumor recurrence or progression. This thus confirms that early diagnosis is crucial for these locally invasive carcinomas located on the nasal planum, and that the results of the first treatment are crucial for a successful patient outcome. Institutional Review Board Statement: Ethical review and approval were waived for this study, due to the retrospective of the research. Animals were not treated as part of an experimental study but out of necessity and only the data were later selected and included in this study.
Informed Consent Statement: Not applicable. Owners signed informed consent for the procedure as standard procedure in the included facilities. No specific informed consent statement was signed for inclusion in this retrospective study.
Data Availability Statement: Data available on request due to restrictions, e.g., privacy or ethical. The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the multicentric nature of the study and the inclusion of various facilities from all over the world and to guarantee the privacy of the presented data.

Conflicts of Interest:
The authors declare no conflict of interest.