Efficacy and toxicity

For the treatment of FMC, adjuvant chemotherapy is usually recommended in cases with tumor showing evidence of lymphatic or blood vessels invasion, while some recommend chemotherapy for all cases [6]. Doxorubicin is the mostly used chemotherapy agent in canine and feline mammary carcinoma. To the author’s knowledge, this is the first study to compare the efficacy and the toxicities between vinorelbine and doxorubicin in FMC. In the present study, 7 cats received vinorelbine for first-line adjuvant chemotherapy and 19 cats were administered with doxorubicin as their first chemotherapy agent. Although the response rate of VRL group was higher than DOX group (75% versus 18%), there is no significant difference in median TTP for two groups (115 days for VRL group and 102 days for DOX group, P =0.949). However, the nature of this study design may cause biased results. This study was not a randomized controlled trial. Cats in VRL group were treated with vinorelbine prospectively, while cats in DOX group were enrolled to this study retrospectively. In addition, most cats in DOX group were administered doxorubicin every 3 to 4 weeks, but cats in VRL group were treated with vinorelbine every 1 to 2 weeks. Hence, cats in VRL group had much more frequent follow-up, and therefore any progression events such as local recurrence, distant metastasis or even pleural effusion would be detected earlier. On the other hand, because of the nature of retrospective design of DOX group, there was lack of standard and regular follow-up schedule, which would result in overestimating TTP for DOX group. This might be the reason why there was much higher response rate in VRL group but similar TTP in two groups. Because TTP is not affected by rescue or subsequent therapy, it is an important endpoint to compare the efficacy between different therapies. As the result, the efficacy

of vinorelbine treated in FMC was similar to doxorubicin. In the aspect of survival time, median OST was numerically longer for VRL group (352 days) than DOX group (284 days), although there was no statistical significance (P =0.948). This could be due to the small sample size and relative high percentage of patients in VRL received rescue chemotherapy.

According to a research by Novosad et al [31], 67 cats with FMC underwent adjunctive doxorubicin treatment and the disease-free interval was 183 days and the overall survival time was 331 days for cats with pulmonary metastasis. Mauldin et al used doxorubicin-based chemotherapy to treat 14 cats with advanced stage of mammary adenocarcinoma; median survival time was 90 days in that study [25]. These results indicated that the outcome of our study was comparable to previous studies. In a study by McNeill et al [26], disease-free interval (DFI) was 676 days and OST was 848 days for 36 cats with mammary carcinoma received adjuvant doxorubicin therapy after surgery;

both DFI and OST in that study were much longer than VRL group or DOX group in our study or even longer than previous studies. Nevertheless, in that study, only 23% cats had evidence of lymph node metastasis and all of which were removed at the time of surgery.

In contrast, there was almost 60% cats had pulmonary metastasis in both groups in our study and this would result in a relative unfavorable outcome.

In the present study, gastrointestinal toxicities were the most common adverse events in DOX group and the incidence of vomiting (48%) and anorexia (48%) were both significant higher than VRL group. Besides, GI toxicity was also the main reason of dose reduction for DOX group. Most of cats (70%) in DOX group received 25 mg/m² as their starting dosage and three cats experienced dose reduction. Some oncologists may use 1 mg/kg (approximately 20 mg/m²) as starting dosage in cats to reduce the incidence of toxicity. However, a decreased of dosage may also decrease the efficacy of chemotherapy.

26

In a previous study [36], different dosing schemes of doxorubicin were used to treat tumor-bearing cats and compare the toxicities. Vomiting and anorexia were the most common reported complaints in that study. There were 24 episodes of vomiting in 129 doses (18.6%) for cats received 20 mg/m² doxorubicin, and 12 episodes of vomiting in 58 doses (20.6%) for cats administered 25 mg/m² doxorubicin. Most (50% to 67%) of vomiting episodes were classified as grade 2 in both groups. The incidence of anorexia was 10.9% and 10.3% for 20 mg/m² and 25 mg/m² doxorubicin, respectively. Most of anorexic episodes (about 60%) were classified as grade 1 in both groups. The results of that study suggested that a lower dosage of doxorubicin may not be associated with reducing risk of GI toxicity. Despite the fact that there was similar dosage in our study and the previous study, the incidence of vomiting and anorexia were both higher. The reason for this difference might be that most of patients in our study were advanced stage.

Patients with advanced stage may relatively have poor nutrition or hydration status and caused higher incidence of chemotherapy-related toxicities.

Neutropenia was the most common adverse event for VRL group in contrast. The incidence of neutropenia was 44% for VRL group and significant higher than DOX group (0%). All cats received vinorelbine with a starting dosage of 11.5 mg/m², and all cats experienced neutropenia during their therapies. Neutropenia occurred after the first dose of vinorelbine in 4 cats and after the second dose in 2 cats. Dose escalation (12.6 mg/m²) of second dose of vinorelbine was performed in only one cat, but grade 4 neutropenia then happened to this cat. The dosage of third dose was therefore reduced to 11.4 mg/m² for this cat; however, neutropenia still occurred at this dosage. Based on previous phase I study [34], the maximum tolerated dose for tumor-bearing cats was 11.5 mg/m² and recommended as starting dosage. Neutropenia was also a common type of adverse events in that study, and 2 cats experienced worsening grades of neutropenia at de-escalated

dosages. One possible explanation mentioned in that study was that there was cumulative dosing and detrimental effects on the bone marrow. Another possible explanation was that there might be a double neutrophil nadir, which was similar to lomustine, an alkylating nitrosourea compound commonly used in treating lymphoma and histiocytic sarcoma in small animals. Although the incidence of neutropenia caused by vinorelbine was high in our study, none of these cats experienced sepsis and all of the neutrophil counts normalized within 7 days. As for DOX group, the incidence of neutropenia was 0.

Incidence of neutropenia was 12% for cats treated with 25 mg/m² doxorubicin in a previous study by Reiman et al., and 71% of neutropenic episodes happened in 0-14 days after the treatment. The nadir (time of the lowest neutrophil count) of doxorubicin usually occurs at 5 to 10 days after the treatment in small animals [16], [19]. All DOX group cats in our study were administered doxorubicin every 3-4 weeks and CBC were evaluated at the day of each treatment. Therefore, clinicians might miss the nadir and resulted in an underestimated incidence of neutropenia.

Renal injury was an important and major concern when giving and dosing chemotherapy, especially using doxorubicin in cats. Doxorubicin can cause renal tubular necrosis, renal infarction or interstitial nephritis in cats and lead to decreasing creatinine clearance [32]. Nephrotoxicity can be seen with cumulative doxorubicin dosages of 130–

320 mg/m² [32]. Although there was significant elevation of creatinine in DOX group after treatments in our study, the changes of creatinine level were not obvious. The possible explanation of rare azotemia episodes in DOX group was that the cumulative dosage in our study was much lower than previous studies (median cumulative dosage:

45 mg/m², range 20-100 mg/m²).

28

5.2 Prognostic factors

Although the tumor behavior of FMC is generally aggressive, time to progression and overall survival time can be quite various. Higher T stage, presented with vascular or lymphatic tumor invasion, status of metastasis and histopathological subtype of solid or cribriform are negative prognostic factors to time to progression in previous research [6], [31]. Numerous studies have identified factors that will influence survival times, including tumor size, WHO stage, histopathological grade, proliferation profile and surgical approach [6], [24], [27], [37].

In this study, neuter status and biological response to chemotherapy were the only two factors that significantly influence TTP in univariate analysis, and these two factors remained significance in multivariate analysis. It was not surprising that responders of chemotherapy had longer TTP than non-responders. But it was interesting that intact female cats in our study had longer TTP compared to neutered female cats (119 days versus 63 days; HR, 5.377; 95% CI, 1.406-20.562; P =0.021 and 0.014 for univariate and multivariate analysis, respectively). All but one intact female cats had undergone tumor removal with concomitant OHE. Sex hormones are associated with mammary tumor development in small animals and humans [17], [38], [40]. A study by Overley et al showed that cats spayed before 1 year of age have a significantly decreased risk of developing mammary carcinoma [33]. Estrogen receptor (ER) and progesterone receptor (P4R) are two main hormonal receptors implicated in mammary tumor development.

Many previous studies have documented that most feline mammary carcinomas are ER and P4R negative, although slightly more than one-third are P4R positive. Dogs with ER-positive or P4R-ER-positive mammary carcinoma were significantly associated with low proliferation index and histopathological grade 1 (low grade) [17]. One study indicated

that expression of hormonal receptors of FMC is associated with lower malignancy and better prognosis [22], which is same as human medicine. In addition, a significant correlation of hormonal receptor (ER and P4R) positivity with absence of ovariectomy has been reported in both dogs and cats [17], [22]. Intact female dogs with ER-positive tumors or with increased peri-surgical serum 17b-estradiol (E2) represent a subset of dogs with mammary carcinomas likely to benefit from OHE [17], but these results are not established in feline medicine yet. Based on those studies, ER and P4R expression may be higher for intact female cats in our study and most of which underwent concomitant OHE and lead to longer TTP. Further screen test for ER and P4R expression and randomized controlled studies will be needed to confirm the hypothesis.

Tumor size (T stage), clinical stage, lymphatic or vascular invasion, histopathological subtype, grade of tumor and surgical approach were all not associated with TTP in our study. This result was opposed to previous studies. One possible explanation was that there were small and uneven distribution samples in the study. A non-randomized and non-prospective study usually has selection bias of cases. Most cases were advanced TNM stage and presented with lymphatic or vascular invasion in this study.

Another possible reason was that not all cases had complete information of histopathology result; approximately half of cases lacked information for surgical margin, tumor grade and subtype.

In the univariate analysis of prognostic factors for survival time, only biological response to chemotherapy was significant and this factor remained significant in multivariate analysis. Biological responders had longer survival times than non-responders (446 days versus 232 days; HR, 4.027; 95% CI, 1.004-16.152; P =0.012 and 0.049 for univariate and multivariate analysis, respectively). Based on previous studies, the objective response rate was about 40% to 50% for cats with mammary carcinoma

30

treated with doxorubicin-based chemotherapy, but this relatively high response rate did not reflect on survival benefits in those retrospective studies [38]. However, no further analysis was done to compare the survival time between responders versus non-responders and between cats with macroscopic disease versus microscopic disease in those studies. In our study, cats with biological response to chemotherapy got significant longer survival time. Since approximate half of biological responders (46%) were Non-CR/non-PD, which were cats with microscopic, non-measurable disease initially and maintained with non-measurable disease at the end of treatment, one possibility was that cats with microscopic disease initially may have longer survival time, and lead to longer survival time for whole biological response group. Nevertheless, no significant difference of survival time between cats with macroscopic disease and with microscopic disease was observed in our analysis. Thus, our results demonstrated that cats with biological response had survival benefits.

Cats with unilateral tumors was trending toward significance when compared for cats with bilateral tumors in univariate analysis (446 days versus 254 days, P =0.069).

Complete and unilateral mastectomy in cats with mammary carcinoma showed a significantly better tumor control in a previous study [20]. As a result, bilateral mastectomy would be recommended for cats with bilateral mammary gland tumors.

However, bilateral mastectomy performed as a single procedure can result in wound dehiscence because of high skin tension. Most surgeons therefore tend to undergo staged unilateral mastectomy, but metastasis may occur between twice surgeries. In addition, the cost and duration of recovery of bilateral mastectomy would be higher and longer, and those factors may decrease owners’ willingness of aggressive surgery. It was also suspected that there might be more aggressive tumor behavior for bilateral mammary tumors than unilateral tumors. Rare studies in veterinary medicine compared the survival

time between cats with bilateral and with unilateral mammary carcinoma. Contralateral mammary carcinoma can be either a metastatic lesion or the second primary cancer, and occurs either synchronously or metachronously. According to results of human medicine, several studies indicated women with bilateral breast cancer had worse prognosis than women with unilateral breast cancer [7]. In a large cohort study by Hartman et al [11], women with synchronous bilateral breast cancer had a higher mortality from breast cancer than women with unilateral disease. This evidence in human may be another reason why cats with bilateral mammary tumor were trending toward shorter survival time in our study.

It is interesting that ulceration of tumors was not significant to survival in univariate analysis (P =0.196) but this factor became significant in multivariate analysis (HR, 5.192;

95% CI, 1.137-23.706; P =0.034). Ulceration of tumors were associated to aggressive tumor behavior and high histopathological grade [27], which is an important prognostic factor of survival time. In addition, ulceration of tumors would cause pain, poor skin hygiene or infection and lead to poor quality of life as well as early euthanasia.

Tumor size, status of metastasis and even grading were not significant prognostic factors for survival time in our study, while those factors were strongly correlated to prognosis in other reports. As mentioned previously, lack of histopathological diagnosis in some cases, most of patients were advanced stage and high grade, small sample size and the nature defects of retrospective were the possible reasons.