1 Introduction
Ovarian cancer (OC) is the most lethal gynecological malignancy worldwide. In 2020, it was estimated that there were 313 959 new cases and 207 252 fatalities globally [
1]. In China, there were approximately 52 100 new cases and 22 500 deaths due to this disease by 2022 [
2]. The significant challenge with OC is its late diagnosis, with approximately 75% of cases identified at advanced stages, leading to high relapse rates [
3]. An important aspect of patient management in recurrent ovarian cancer is the platinum-free interval, which is pivotal for patient categorization. Based on the platinum-free interval, recurrent cases are classified as platinum-sensitive (relapses occurring six months or later after the last platinum-based therapy) and platinum-resistant (recurrences within six months after the last platinum-based therapy) [
4].
The standard approach for treating platinum-resistant ovarian cancer (PROC) has traditionally been monotherapy with non-platinum cytotoxic agents. These include PEGylated liposomal doxorubicin (PLD), gemcitabine, paclitaxel, and topotecan. Recent monotherapy trials for PROC have shown response rates of 10%–15%, with a median progression-free survival (PFS) of approximately 3.5 months [
5–
9]. The modest response rates to existing treatments underscore the critical need for more effective therapies for PROC. Angiogenesis is essential for cancer growth and metastasis. Antiangiogenic therapies, including anti-VEGF antibodies and multireceptor tyrosine kinase inhibitors (TKIs), have been shown to be effective in treating OC [
10,
11]. Bevacizumab, a monoclonal antibody targeting VEGFA, has been shown to improve progression-free survival (PFS) and overall survival (OS) in OC when combined with frontline chemotherapy, as evidenced by the GOG218 and ICON7 trials [
12,
13]. The AURELIA study initiated the application of anti-angiogenic drugs in patients with recurrent ovarian cancer, significantly improving the PFS and objective response rate (ORR). Median PFS was 3.4 months with chemotherapy alone versus 6.7 months with bevacizumab-containing therapy [
14]. Anlotinib, a novel TKI targeting multiple kinases involved in tumor growth and angiogenesis, showed activity against OC [
15]. Clinical studies have revealed that anlotinib, combined with standard platinum-based chemotherapy, offers synergistic benefits in OC patients [
16–
18]. In addition, a recent phase II study indicated that anlotinib as a single agent for recurrent platinum-resistant or refractory ovarian carcinoma achieved a 31.2% ORR and 75% disease control rate (DCR) [
18]. Albumin-bound paclitaxel (ab-paclitaxel), an innovative formulation of paclitaxel that uses albumin as a solvent, exhibits enhanced anti-tumor effects and reduced toxicity compared to traditional paclitaxel. A phase II clinical trial of bevacizumab with albumin-bound paclitaxel in the United States patients with recurrent, platinum-resistant primary epithelial ovarian or primary peritoneal carcinoma demonstrated substantial antitumor activity, with a response rate of 50% and median PFS of 8.08 months. This indicates that the addition of antiangiogenic therapy to ab-paclitaxel can improve outcomes in platinum-resistant OC [
19]. Based on these insights, this study aimed to investigate the efficacy and safety of combining anlotinib with ab-paclitaxel in Chinese patients with platinum-resistant or refractory OC.
2 Materials and methods
2.1 Study design and participants
This phase II, single-arm, prospective study was conducted at Sun Yat-sen University Cancer Center, Guangzhou, China. Eligible participants were women aged 18–75 years with platinum-resistant ovarian cancer, defined as disease progression within six months of the last platinum treatment. Inclusion criteria included histological/pathological confirmation of the primary tumor, at least one line of platinum-based chemotherapy, with or without prior use of albumin-bound paclitaxel, measurable disease per RECIST (version 1.1), or non-measurable disease with elevated CA-125 and significant symptoms (per Rustin criteria), an ECOG performance status of 0–2, a minimum life expectancy of 3 months; and adequate bone marrow, liver, and renal function as defined by specific laboratory criteria. Full inclusion and exclusion details are provided in the supplementary material (Supplement 1).
2.2 Procedures
Patients were administered 260 mg/m2 of albumin-bound paclitaxel over 30 min intravenously on day one every three weeks, up to a maximum of six cycles, along with 10 mg of oral anlotinib daily for the first 14 days of each 3-week cycle. After six cycles of treatment, if progression did not occur, oral anlotinib monotherapy was continued until disease progression, patient withdrawal, or emergence of unacceptable side effects.
To manage adverse events, dose modifications including interruptions and reductions were allowed. If necessary, the treatment could be paused for up to 14 days at a time, with the possibility of multiple interruptions. For ab-paclitaxel, two dose reductions were allowed (to 200 mg/m2 every three weeks, and then to 150 mg/m2 every three weeks); while for oral anlotinib, one dose reduction (to 8 mg once daily) was allowed in the case of severe toxicity.
2.3 Efficacy evaluation
The primary endpoint was objective responses of patients with measurable disease who had a complete or partial response according to RECIST version 1.1, and patients with non-measurable disease according to the Rustin criteria. Defining CA-125 response according to the Rustin criteria, a biological response based on CA-125 was defined as either a 50% or a 75% reduction in CA-125 levels. To reduce the chance of falsely predicting a response, the 50% CA-125 response definition required four CA-125 levels with a sequential decrease of at least 50%. The definition of a 75% CA-125 response requires only three CA-125 measurements, with a sequential decrease of at least 75%. For both the 50% and 75% response criteria, the last sample should be taken at a minimum of 28 days following the preceding sample. Doubling the CA-125 level from the upper limit of normal (ULN) or the lowest point (nadir) is adequate for diagnosing disease progression [
20–
25].
The secondary endpoints were progression-free survival, duration of response, proportion of disease control, and safety. Progression-free survival was defined as the interval from the start of treatment to disease progression or death for any cause (whichever occurred first) or last progress-free survival assessment for patients alive without progression. The duration of response was assessed in patients who achieved a response and was defined as the time from the date of the first documented response until the date of documented progression or death from any cause.
Disease control was defined as the proportion of patients who achieved complete response, partial response, or stable disease.
2.4 Statistical analysis
We employed Simon’s two-stage design, with a one-sided 5% α error and 90% power. Preliminary results from an ongoing phase 2 study of anlotinib monotherapy in platinum-resistant or-refractory ovarian cancer showed an ORR of 14.3%. Previous data have suggested a maximum ORR of 23% for ab-paclitaxel monotherapy in a similar setting. We hypothesized that combining anlotinib with ab-paclitaxel would yield a 48%. Accordingly, the first stage involved treating 16 evaluable patients, requiring at least five responses to proceed to the second stage, where 24 additional patients were enrolled, totaling 40. The regimen was deemed successful if 14 or more responses were observed overall.
The data analysis encompassed three groups: intention-to-treat, per-protocol, and safety. The intention-to-treat group included all enrolled patients; the per-protocol group comprised those who met all trial criteria and followed the protocol without major violations; and the safety population consisted of those who received at least one dose of the study drug, excluding patients without safety data. Efficacy was analyzed in the intention-to-treat and per-protocol groups, excluding patients lacking data for post-baseline efficacy assessments from the per-protocol analysis. Safety data were analyzed for the safety population. Response rates and 95% confidence intervals (CIs) were calculated using the Clopper-Pearson method, and the median duration of response and progression-free survival with their 95% CIs were calculated using the Kaplan–Meier method. The log-rank test was used to compare progression-free survival across different toxicity grades. Analyses were limited to data collected until December 31, 2023, and were conducted using SPSS (version 22.0).
3 Results
3.1 Participants
Between June 1, 2021, and April 1, 2023, a total of 58 patients with ovarian cancer were screened, and 44 were enrolled and included in the intention-to-treat analysis. However, one patient was excluded from the per-protocol analysis because of the absence of a post-baseline efficacy assessment (Fig. 1). The baseline demographic and clinical characteristics are summarized in Table 1. The predominant tumor pathology was high-grade serous adenocarcinoma, and most patients had stage III or IV disease at the time of diagnosis. Eighteen patients (40.9%) received 3–6 lines of chemotherapy. Twenty patients (45.5%) had received prior antiangiogenic treatment and PARP inhibitors. A total of 29 patients had measurable lesions, while 15 had non-measurable diseases. At the time of data cut-off on December 31, 2023, the median follow-up period was 20.5 months (interquartile range (IQR) 14.9–28.3 months).
By the cut-off date, 39 of 44 patients (88.6%) discontinued treatment, with five patients (5/44, 11.4%) still receiving treatment (Fig. 2). The main reason for discontinuation was disease progression, accounting for 27 patients (27/39, 69.3%), while adverse events were the cause of treatment discontinuation in three patients (3/39, 7.7%). Other reasons for discontinuation included transferring to other treatments (four patients, 4/39, 10.3%) and the COVID-19 pandemic precluding the administration of medication on time (five patients, 5/39, 12.8%).
3.2 Efficacy
Of the first 16 assessable patients, 13 patients had measurable lesions and three had non-measurable lesions. Among those with measurable lesions, objective responses were observed in eight patients (61.5%), consisting of one complete response (CR) and seven partial responses (PR), while the remaining five patients achieved stable disease (SD) and none experienced progressive disease (PD). For patients with non-measurable lesions, two out of three demonstrated responses, resulting in an overall response rate of 62.5% (10/16). These findings met the objective response threshold for the first stage of Simon’s two-stage design, prompting continuation of the trial to full accrual.
For the efficacy assessment, 43 of the 44 enrolled patients were evaluable; one patient with non-measurable disease was excluded because of patient withdrawal after one cycle, and no post-treatment data were obtained. Among the 29 patients with measurable lesions, 16 (55.2%; 95% CI 21.9–51.2) achieved objective responses, including CR in two patients and PR in 14 patients. Thirteen patients exhibited SD and none had PD. In the cohort of 14 patients with non-measurable lesions, 9 (64.3%; 95% CI 16.6–58.4) showed responses based on the CA-125 criteria, four stable disease, and one disease progression. Overall, responses were observed in 25 of the 44 patients (56.8%; 95% CI 41.1–71.3) in the intention-to-treat population, and the per-protocol (PP) analysis reflected a response rate of 58.1% (25 out of 43 patients; 95% CI 42.2–72.6), as detailed in Table 2.
Disease control was achieved in 100% of patients with measurable lesions (29 of 29; 95% CI 36.7–63.3) and in 92.9% of patients without measurable lesions (13 of 14; 95% CI 66.1–99.7), as shown in Table 2. The median duration of response was 7.8 months (IQR 1.5–3.8; 95% CI 1.4–2.6). Tumor shrinkage was observed in all 29 patients with measurable lesions who underwent at least one post-baseline efficacy assessment (Fig. 3). The average maximal reduction in target lesion size from baseline was −0.38% (SD = 0.27). Detailed response information according to the RECIST 1.1 criteria is available in the supplementary material (Supplement 2).
In addition, among the 29 patients with measurable lesions, we assessed efficacy using the Rustin CA-125 criteria. All 29 patients had elevated pretreatment CA-125 levels. Responses were observed in 20 patients (68.9%; 95% CI 0.28.7–0.519). The CA-125 levels were recorded for all 44 patients, including baseline values and readings before each treatment course. The responses evaluated using the Rustin criteria can be found in the supplementary material (Supplement 3).
As of the data cut-off date, 27 (61.3%) out of 44 patients experienced disease progression, while 12 (27.3%) showed no progression, with 5 of these patients still undergoing treatment. The median PFS was 9.8 months (95% CI 7.0–12.6; Fig. 4). The 12-month PFS rate was 36.3% and 12-month OS rate was 72.0%, and the 24-month PFS rate was 13.2% and 24-month OS rate was 42.3%, respectively.
3.3 Safety
All the 44 patients were included in the toxicity evaluation. The main toxicities included neutropenia (41/44 patients (93.2%)), thrombocytopenia (26/44 (59.1%)), gum pain (30/44 (68.2%)), hand-foot syndrome (28/44 (63.6%)), hypertension (26/44 (59.1%)), and numbness in the hands and feet (22 /44 (50%)). The most common grade 3 or 4 adverse events were gum pain (10/44 (22.7%)), hypertension (10/44 (22.7%)), hand-foot syndrome (4/44 (9.1%)), and thrombocytopenia (4/44 (9.1%)) (Table 3). There have been no reports of febrile neutropenia, fistulation, or fatal adverse events. One patient reported treatment-related deep vein thrombosis in the lower extremities that required hospitalization for anticoagulation. Only three patients (6.8%) stopped treatment because of adverse effects.
A dose reduction was observed in some patients. Nine out of 44 (20.5%) required dose reductions related to albumin-bound paclitaxel, with eight of these (88.9%) requiring only one reduction and one (11.1%) requiring two reductions. Regarding the schedule, two patients (22.2%) had their first dose reduction in the second cycle and seven (77.8%) after the second cycle. Meanwhile, anlotinib dose reductions were necessary for 12 patients (27.3% of 44), all of whom required only a single reduction. The first dose reduction of anlotinib occurred during the second cycle in three patients (3 of 12, 25%), in the third cycle in five patients (41.7%), and after the seventh cycle in four patients (33.3%). The main reasons for reducing the anlotinib dose included hand-foot syndrome (4, (33.3%)), poorly controlled hypertension (3, 25%), fatigue (2, 16.7%), rash (1, 8.3%), hemorrhage (1, 8.3%), and weight loss (1, 8.3%). Detailed reasons for dose adjustments of albumin-bound paclitaxel and anlotinib are provided in the supplementary material (Supplement 4).
3.4 Exploratory analysis
Among the enrolled patients, four patients (4/44, 9.1%) had previously used albumin-bound paclitaxel, with a median albumin-bound paclitaxel free interval of 20.5 months (range 4.3–39.9 months) since the last time albumin-bound paclitaxel. In the four patients, three achieved SD with a PFS of 1.8, 22.2, and 6.9 months, respectively; one patient had PR with a 1.6-month PFS. Twenty patients had previous use of antiangiogenic reagents, 17 patients (17/20, 85%) had bevacizumab, and three patients (3/20, 15%) had apatinib (a TKI). The responses were observed in 11 of 20 (55.0%) patients with previous use of antiangiogenic reagents and in 14 of 23 (60.8%) patients without antiangiogenic reagent use. The median PFS was 9.8 months (95% CI 5.9–13.7) and 10.4 months (95% CI 7.1–13.7), respectively. Among the 20 patients (20/44, 45%) who had previously used PARP inhibitors and 23 patients who had not, responses were observed in 11 of 20 (55%) patients with prior PARPi use and in 14 of 23 patients (60.9%) without prior PARPi use, and the median PFS was 8.8 months (95% CI 5.9–13.7) and 10.9 months (95% CI 8.9–12.9), respectively; anlotinib dose reductions occurred in four patients (4/20, 20%) with previous antiangiogenic reagents and in eight patients (8/24, 33.3%) without.
4 Discussion
This study demonstrated that combining the VEGFR TKI anlotinib and ab-paclitaxel is an effective regimen, achieving an overall response rate of 58.1% in patients with platinum-resistant or platinum-refractory ovarian cancer. The median PFS was 9.8 months (95% CI 7.0–12.6), with a median response of 7.4 months (95% CI 2.3–12.0). To the best of our knowledge, this is the first study on the combination treatment of ab-paclitaxel and a VEGFR TKI.
Ab-paclitaxel, a novel solvent-free formulation of paclitaxel, has shown promising safety and efficacy profiles in the treatment of recurrent ovarian cancer. Teneriello
et al. reported that ab-paclitaxel is highly active as a single agent (260 mg/m
2 administered intravenously for 30 min on day 1 of a 21-day cycle for six cycles or until disease progression) in patients with platinum-sensitive recurrent ovarian cancer [
26]. The ORR was 64% (15 CR and 13 partial PR among the 44 assessable patients), and the median PFS was 8.5 months. Coleman
et al. have shown that for the weekly treatment with ab-paclitaxel at a dose of 100 mg/m
2 on days 1, 8, and 15 in 51 patients with platinum- and taxane-resistant ovarian cancer, the ORR was only 23% (1 CR and 10 PR), with a median PFS of 4.5 months, overall survival of 17.4 months. Seventeen patients (36%) had a PFS > six months [
27]. This response rate is not better than that of other single agents for platinum-resistant recurrent ovarian cancer. Similarly, single-agent antiangiogenic TKIs have yielded ORRs of 0–30% in platinum-resistant diseases [
18,
28–
32].
Recently, investigations have shown that antiangiogenic therapy with ab-paclitaxel further improves outcomes in platinum-resistant ovarian cancer. A phase II study by Tillmanns
et al. reported that ab-paclitaxel with bevacizumab showed an ORR of 50% (4 CR and 20 PR), and patients received a median of six treatment cycles (range 1–31 cycles). The median PFS was 8.08 months, the 6-month progression-free rate was 62.5%, and the median overall survival was 17.15 months. Grade 3–4 adverse events included gastrointestinal disorders, neutropenia, and hypertension [
19]
.The efficacy of our study, which combined ab-paclitaxel and anlotinib, seemed to be better than or consistent with that of similar previous investigations. The current ORR is higher than that (50%) observed in Tillmanns
et al.’ study, which combined ab-paclitaxel and bevacicumab [
18]. The PFS was also longer than that reported in other studies, including anlotinib combined with other drugs [
16–
18,
32–
36]. For instance, Chen
et al. reported that in platinum-resistant ovarian cancer with anlotinib combined with pemetrexed, the ORR was 32% and the median PFS was 8.9 months [
18]. Another phase II study, anlotinib combined with TQB2450 (target PD-L1) showed that the ORR was 47.8%, and the median PFS was 7.8 months [
16]. The ORR of the current study is also consistent with ANNIE study with niraparib plus anlotinib, in which the ORR was 50.0%, median PFS and overall survival were 9.2 months and 15.3 months [
16]. The additional details of several recent studies are presented in Table 4.
In the present study, the combination of ab-paclitaxel and anlotinib showed a manageable toxicity profile. No new safety signals were observed when compared with the respective single-agent safety profiles. Neutropenia and thrombocytopenia were the most frequently reported grade 3 or 4 hematological toxicities that might be attributed to ab-paclitaxel. For more convenient administration, our protocol used a schedule of every three weeks for ab-paclitaxel. The safety results are similar to those of previous studies of ab-paclitaxel [
19,
26,
27], in which ab-paclitaxel was usually administered weekly. In general, grade 3–4 hematological toxicities are manageable with growth factors or recombinant human thrombopoietin and dose reductions. Once every three weeks administration of ab-paclitaxel is more convenient than weekly administration for patients. Adverse events such as hypertension and hand-foot syndrome, typically associated with antiangiogenic agents, were mainly grade 1 to 2 in this trial, which is consistent with the incidences reported in previous studies [
16–
18,
28–
37]. Previous clinical trials have indicated a dosage of 12 mg of oral anlotinib daily for the first 14 days of each 21-day cycle in the treatment of various solid tumors [
18,
33]. Considering the protocol amendment of the ANNIE study on anlotinib initial dose reduction (from 12 mg to 10 mg) [
16], we opted for a starting dose of 10 mg for anlotinib. Compared with previous safety profiles, fewer patients required anlotinib dose reduction and discontinued treatment in this study. None of the patients experienced intestinal fistula or visceral bleeding. No treatment-related deaths were recorded.
With the emergence of poly ADP-ribose polymerase inhibitors (PARPi) in ovarian cancer treatment, maintenance therapy with PARPi is widely used in platinum-sensitive recurrent ovarian cancer [
38,
39]. Is it possible for patients with platinum-resistant ovarian cancer to benefit from maintenance therapy? No data were available for this question. To explore this, we designed that after six cycles of combined agents, the patient continued to take oral anotinib as maintenance therapy if the disease did not progress. In our study, 35 (71.4%) patients received oral anotinib as maintenance therapy, which last for a median duration of 4.6 months, which is similar to that in BRCAm cohort of OReO study, among which, PARPi re-challenge produced a median PFS of 4.3 months [
40]. The 12-month PFS rate in our study was 36.3%, which could be associated with anotinib maintenance therapy. These data show that patients with platinum-resistant recurrent ovarian cancer may benefit from maintenance therapy.
In the literature, there are few data on the efficacy of anti-angiogenic agents after previous use of anti-angiogenic agents or PARPi in platinum-resistant ovarian cancer. Our results showed that in patients with previously used anti-angiogenic drugs, the ORR and median PFS were close to those in patients without prior use (ORR 55.0% vs. 58.4%; mPFS 9.8 vs. 10.4 months). Similarly, in patients with prior use of PARPi, the ORR and mPFS were also close to that in patients without prior use (ORR 55.0% vs. 58.4%; mPFS 8.8 vs. 10.9 months). In the entire cohort of our study, the 12-month OS rate reached 72.0% and the 24-month OS rate reached 42.3%. Notably, the related toxicity did not accumulate. Thus, our post-hoc exploratory analysis suggests that patients with prior anti-angiogenic treatment or PARP inhibitors still benefited from this ab-paclitaxel plus anlotinib protocol.
In addition, assessment of disease response presents a unique challenge in patients with ovarian cancer. As is commonly applied, the RECIST criteria will not accurately describe responses in patients with small volumes of subclinical peritoneal disease below the RECIST criteria threshold. The use of CA-125, which can predict disease recurrence, is limited by varying methodologies, laboratory variability, error, and the fact that not all patients with recurrent ovarian cancer have elevated CA-125 levels. However, in clinical practice, there is still a need for efficacy assessment in patients with non-measurable diseases, and elevated CA-125 levels are usually helpful in this setting according to the Rustin criteria. Rustin criteria differ slightly from the simpler GCIG CA-125 criteria which are now more commonly used. One may also use the simpler GCIG CA-125 criteria to assess responser in the recurrent patients with non-measurable disease [
41]. In our study, a response rate of 64.3% was observed in patients with elevated CA-125 levels, reflecting a response to treatment for small-volume disease not measurable by RECIST. A definitive statement regarding the utility of CA-125 versus RECIST, or the combination of both, to accurately assess response cannot be made because this study was not designed or powered to compare the methodologies as a measure of response.
However, this study had some limitations. One limitation is that it was a single-arm study with no control group for comparison; thus, selection bias could not be ruled out because of the non-randomized design. Additionally, we included 15 patients without measurable lesions and assessed their efficacy based on CA-125 levels in accordance with the Rustin criteria. Although research has shown a significant correlation between CA-125 tumor responses and imaging-based tumor responses in certain cases, the reliability of CA-125 for disease assessment is still debated. Despite its potential benefits, the findings suggest that CA-125 may not be reliably used as the sole criterion for determining treatment response and guiding overall patient management, highlighting the complexity of assessing treatment efficacy.
In conclusion, our study showed that the combination therapy of ab-paclitaxel with anlotinib offers promising efficacy and manageable toxicity in patients with platinum-resistant ovarian cancer. These findings also support the need to further validate the effectiveness and safety of this treatment regimen. Moreover, a randomized trial will be conducted in future research to determine whether albumin-bound paclitaxel and anlotinib are superior.
PDF
(1175KB)
