Efficacy of Modified Dose-dense Paclitaxel in Recurrent Cervical Cancer

Abstract

Objective

To examine survival outcomes of women with recurrent cervical cancer who received salvage chemotherapy with modified dose-dense paclitaxel (MDDP) monotherapy (paclitaxel 80 mg/m², administered on day 1, 8, and 15 without day 22).

Materials and Methods

A retrospective study was conducted to evaluate cause-specific survival after the first recurrence (SAR) of women with recurrent cervical cancer diagnosed between 2006 and 2014. Pooled analyses were performed to examine SAR in women who received MDDP monotherapy (n = 17) for any treatment line, compared with those who received salvage chemotherapy with paclitaxel-doublet (n = 18) and nonpaclitaxel regimens (n = 52).

Results

In the whole cohort, median SAR was 13.7 months including 63 (72.4%) events. MDDP monotherapy regimen was most commonly used in the second-line setting (35.3%) followed by the third/fourth lines (both, 23.5%). Among the women who received MDDP regimen, there were 6 (35.3%) women who received ≥6 cycles; there was 1 (5.9%) women who discontinued the regimen due to adverse effects (grade 3 transaminitis); regimen postponement was seen in 2 (1.4%) of 140 total cycles; and the response rate after the sixth cycle of this regimen was 29.4% (1 complete and 4 partial responses). On univariate analysis, MDDP usage had the highest 2-year SAR rate (MDDP 54.1%, paclitaxel-doublet 43.6%, and nonpaclitaxel regimens 28.1%; P_trend = 0.044). On multivariate analysis, MDDP monotherapy remained an independent prognostic factor for improved SAR compared with the nonpaclitaxel regimen (adjusted-hazard ratio, 0.50; 95% confidence interval, 0.26–0.95; P = 0.036).

Conclusion

Our results suggested that MDDP monotherapy is a tolerable and relatively effective regimen for recurrent cervical cancer.

Cervical cancer is the most common gynecologic malignancy in the world, estimated to have 527,600 new cases and 265,700 deaths in 2012.¹ Approximately 10% to 20% of patients with early-stage cervical cancer will experience recurrence,^2,3 while more than 30% to 40% of patients with locally advanced stages will relapse.⁴ Survival time after distant recurrence in cervical cancer appears to be dismal.⁵

Treatment paradigms in the management of recurrent cervical cancer include surgical excision, radiation, chemotherapy, clinical trials, and best supportive care.^5–7 Cisplatin is considered the most effective agent for cervical cancer, therefore, the recommended first-line regimens for recurrent disease are platinum-based combination regimens such as cisplatin + paclitaxel + bevacizumab, cisplatin + paclitaxel, and cisplatin+topotecan.^7–10 Nevertheless, given that most patients with recurrent disease have been previously treated with platinum agents during their primary therapy, sensitivity to cisplatin at the time of recurrence may be reduced.¹¹ In addition, prolonged use of platinum agents is challenging because of their cumulative toxicity; thus, there is a need for less toxic second-line or maintenance alternatives in the management of recurrent cervical cancer. For these reasons, noncisplatin agents are considered reasonable alternatives. Previous studies have shown that paclitaxel is not only efficacious and tolerable but may be used as the first-line single-agent chemotherapy for recurrent cervical cancer.^12–14 However, paclitaxel was administered every 3 weeks in these prior studies, and the use of a dose-dense regimen for recurrent cervical cancer has not been described.

The rationale for dose-dense chemotherapy is based upon the hypothesis that shortening the dose interval of cytotoxic agents with weekly administration will be more effective for cancer eradication¹⁵; however, dose-dense therapies were limited by poor patient tolerance with a high prevalence of adverse effects (discontinuation/postponement rate up to 53%, and adverse events up to 69%).^16–19 To decrease adverse effects, some clinicians have adopted a modified dose-dense regimen in which one day of paclitaxel is skipped each cycle. With no prior study clearly addressing the effectiveness of dose-dense paclitaxel monotherapy in recurrent cervical cancer, its role in this setting remains unknown. The aim of this study was to evaluate survival outcomes of women with recurrent cervical cancer who received salvage chemotherapy with modified dose-dense paclitaxel monotherapy.

MATERIALS AND METHODS

Study Design and Eligibility

Approval from the Institutional Review Board was obtained at the University of Southern California, and the institutional database for cervical cancer was utilized to identify the cases.²⁰ Eligibility criteria for this study included patients with a diagnosis of recurrent cervical cancer between January 1, 2006 and December 31, 2014 who received salvage chemotherapy at Los Angeles County Medical Center. These patients initially experienced complete remission after primary therapy. Patients were excluded from this study if they had no evidence of recurrent disease, had no clinical information at the time of cancer recurrence, or did not undergo salvage chemotherapy. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines were consulted for this retrospective cohort study.²¹

Clinical Information

The following information was abstracted from the medical records of eligible cases: (i) patient demographics including age at the initial recurrence, calendar year at recurrence, ethnicity, and comorbidities including hypertension, diabetes, and hypercholesterolemia; (ii) pathology results including histologic subtype, initial cancer stage, and anatomic site of the first cervical cancer recurrence; (iii) laboratory test results including complete blood counts obtained at the initial recurrence (white blood cell counts [×10⁹/L], hemoglobin levels [g/dL], and platelet counts [×10⁹/L]); (iv) chemotherapeutic details for recurrent disease (number of lines and cycles received, description of treatment regimens and responses, and reasons for discontinuation); and (v) survival outcomes including disease-free survival (DFS) and survival time after recurrence (SAR).

Definition

Cancer stage was evaluated based on the 2009 FIGO system.²² In this study, cancer stage was further grouped into the following: early-stage (stage IA1-IB1), locally advanced stage (stage IB2-IVA), and distant metastasis (stage IVB) based on a prior study.²⁰ Histologic subtypes were grouped as squamous cell, adenocarcinoma, adenosquamous, or other histology. Anemia (hemoglobin <10.0 g/dL), leukocytosis (white blood cell ≥10.0×10⁹/L), and thrombocytosis (platelet ≥400×10⁹/L) were defined based on prior studies.^23–25 DFS was defined as the time interval between the date of completion of the initial treatment and the date of the first recurrence of cervical cancer and divided in 2 groups (<12 vs. ≥12 mo). Anatomic recurrent site of cervical cancer was categorized as local (intrapelvic) or distant (extrapelvic with or without local metastasis).³

The modified dose-dense regimen was administered as follows: intravenous paclitaxel (80 mg/m² dose) administered on day 1, 8, and 15 without day 22, repeated every 28 days. A paclitaxel-doublet regimen was defined as combination therapy of paclitaxel (not dose-dense schedule) with a nonpaclitaxel agent. Nonpaclitaxel regimens were composed of chemotherapeutic agents other than paclitaxel. In our general practice, outcome of treatment response was determined after 6 cycles of chemotherapy administration by the Response Evaluation Criteria in Solid Tumors (RECIST) criteria.²⁶ SAR was defined as the time interval between the date of the first recurrence and either the date of death due to cervical cancer or the date of transfer to hospice due to disease progression in the setting of terminal illness.²⁷ Patients were censored if alive at the last follow-up visit or if they died of other disease. The coinvestigators entered the data into a de-identified data sheet, and the principal investigator of the study examined the accuracy, consistency, and quality of the data.

Systematic Literature Review

A comprehensive literature search, evaluating (i) the use of paclitaxel in recurrent cervical cancer and (ii) the use of dose-dense paclitaxel chemotherapy for gynecologic cancer, was conducted for a systematic review using public search engines PubMED/MEDLINE with entry keywords (i) “cervical cancer” AND “recurrence” AND “paclitaxel” and (ii) “dose-dense” AND “chemotherapy” AND “paclitaxel” searched on July 20, 2016. Eligible publications included prospective and retrospective observational studies, case-control studies, and randomized controlled trials reported in English language that had adequate descriptions of patient demographics, chemotherapy regimen, response, and follow-up. The references listed in each identified article were also reviewed and relevant articles that met the eligibility criteria were included in the study. Case reports, review articles, and publications lacking adequate information as above were excluded. Types of chemotherapy regimens, response rates, adverse events, and survival outcomes were abstracted for the descriptive analysis.

Statistical Analysis

The primary interest of the analysis was to examine SAR for the modified dose-dense paclitaxel regimen compared with paclitaxel-doublet regimens and nonpaclitaxel regimens. The secondary interest of the analysis was to examine the rates of treatment response, postponement, and adverse events with modified dose-dense paclitaxel treatment for recurrent cervical cancer. Continuous variables were examined for normality by the Kolmogorov-Smirnov test and expressed with mean (± SD) or median (range) as appropriate. Statistical significance of the continuous variables was assessed with the Kruskal-Wallis test or 1-way analysis of variance test as appropriate. Categorical or ordinal variables were expressed with number (%), and statistical significance was examined by the χ² test.

For evaluating SAR, the Kaplan-Meier method was used to construct the survival curves and the statistical difference between the curves was examined by a log-rank test for univariate analysis. A Cox proportional hazard regression model for multivariate analysis was performed to determine independent prognostic factors for SAR. Covariates affecting survival entered in the initial model were selected in an a priori manner. These included age at recurrence (<50 vs. ≥50 y), calendar year of recurrence (2006 to 2009 vs. 2010 to 2014), DFS (<12 vs. ≥12 mo), cancer stage at the initial diagnosis (early vs. locally advanced/distant metastasis), histology (squamous vs. nonsquamous), anatomic recurrence site (local vs. distant), and salvage chemotherapy regimen (modified dose-dense paclitaxel monotherapy, paclitaxel-doublet, or nonpaclitaxel).²⁸ Then, the least significant covariate was removed from the model until the final model retained only significant covariates (conditional backward method). The magnitude of statistical significance for survival analysis was expressed with a hazard ratio and 95% confidence interval. All statistical tests were 2-tailed, and a P-value <0.05 was considered to be statistically significant. Statistical Package for the Social Sciences (SPSS, version 24.0, Chicago, IL) was used for the analysis.

RESULTS

Selection criteria are shown in Supplemental Figure S1 (Supplemental Digital Content 1, http://links.lww.com/AJCO/A181). There were 155 patients who experienced cervical cancer recurrence during the study period, 68 of whom were excluded due to nonchemotherapeutic management for recurrent disease. The remaining 87 cases received salvage chemotherapy for recurrent disease: modified dose-dense paclitaxel monotherapy (n = 17), paclitaxel-doublet (n = 18), and nonpaclitaxel regimens (n = 52).

Clinicopathologic demographics are shown in Table 1. For the whole cohort, median age at recurrence of cervical cancer was 52 years. The majority of patients were of Hispanic ethnicity (71.3%) and had received primary whole pelvic radiotherapy (89.7%). The most common medical comorbidity was hypertension (29.9%) followed by diabetes mellitus (13.8%). Anemia was observed in one fifth of patients (20.5%) at the time of recurrence. Leukocytosis and thrombocytosis were also detected in 15.5% and 16.7%, respectively. Tumor characteristics included predominantly squamous cell histology (74.7%) and advanced-stage (87.4%). The median DFS was 14.6 months. More than one third of disease recurrence occurred within 12 months of completing primary treatment (41.4%). The vast majority of patients had extrapelvic disease (80.5%) at the time of the first recurrence. Patient demographics were examined among the 3 groups and were statistically similar (all, P > 0.05). There was a higher proportion of dyslipidemic and hypertensive women in the modified dose-dense paclitaxel group among the 3 groups, but this did not reach statistical significance.

TABLE 1.

Patient and Tumor Demographics (N = 87)

Characteristics	All	Dose-dense PTX (17 [19.5%])	PTX Doublet (18 [20.7%])	Non-PTX (52 [59.8%])	P
Age at recurrence (y)	52 (25–76)	51 (25–76)	54 (43–60)	52.5 (31–75)	0.89
< 50	34 (39.1)	5 (29.4)	6 (33.3)	23 (44.2)
≥50	53 (60.9)	12 (70.6)	12 (66.7)	29 (55.8)
Ethnicity					0.45
White	3 (3.4)	0	0	3 (5.8)
African American	8 (9.2)	3 (17.6)	2 (11.1)	3 (5.8)
Hispanic	62 (71.3)	13 (76.5)	12 (66.7)	37 (71.2)
Asian	14 (16.1)	1 (5.9)	4 (22.2)	9 (17.3)
Hypertension					0.08
No	61 (70.1)	9 (52.9)	11 (61.1)	41 (78.8)
Yes	26 (29.9)	8 (47.1)	7 (38.9)	11 (21.2)
Diabetes					0.09
No	75 (86.2)	14 (82.4)	13 (72.2)	48 (92.3)
Yes	12 (13.8)	3 (17.6)	5 (27.8)	4 (7.7)
Hypercholesterolemia					0.07
No	79 (90.8)	13 (76.5)	17 (94.4)	49 (94.2)
Yes	8 (9.2)	4 (23.5)	1 (5.6)	3 (5.8)
Primary WPRT					0.55
No	9 (10.3)	2 (11.8)	3 (16.7)	4 (7.7)
Yes	78 (89.7)	15 (88.2)	15 (83.3)	48 (92.3)
WBC (×109/L)*	6.7 (2.8–38.0)	7.0 (3.0–22.0)	6.1 (3.5–13.2)	6.8 (2.8–38.0)	0.98
≥10.0	13 (15.5)	2 (13.3)	3 (17.6)	8 (15.4)
< 10.0	71 (84.5)	13 (86.7)	14 (82.4)	44 (84.6)
Platelet (×109/L)*	289 (124–913)	292 (175–688)	263 (202–913)	290 (124–707)	0.31
≥400	14 (16.7)	5 (33.3)	2 (11.8)	7 (13.5)
< 400	70 (83.3)	10 (66.7)	15 (88.2)	45 (86.5)
Hemoglobin (g/dL)*	11.0 (7.7–14.6)	11.4 (8.4–14.6)	10.8 (8.3–13.6)	11.0 (7.7–13.6)	0.15
≥10.0	66 (79.5)	10 (71.4)	12 (70.6)	44 (84.6)
< 10.0	17 (20.5)	4 (28.6)	5 (29.4)	8 (15.4)
Histology at diagnosis					0.53
Squamous cell	65 (74.7)	11 (64.7)	13 (72.2)	41 (78.8)
Adenocarcinoma	16 (18.4)	4 (23.5)	4 (22.2)	8 (15.4)
Adenosquamous	2 (2.3)	0	0	2 (3.8)
Other	4 (4.6)	2 (11.8)	1 (5.6)	1 (1.9)
Stage at initial diagnosis					0.27
Early-stage	11 (12.6)	2 (11.8)	4 (22.2)	5 (9.8)
Locally advanced stage	70 (80.5)	15 (88.2)	14 (77.8)	41 (80.4)
Distant metastasis	6 (6.9)	0	0	5 (9.8)
DFS (mo)	14.6 (1.9–113.1)	15.9 (4.3–78.4)	17.0 (5.5–107.3)	13.5 (1.9–113.1)	0.23
< 6	9 (10.3)	2 (11.8)	1 (5.6)	6 (11.5)
6–11.9	27 (31.0)	5 (29.4)	4 (22.2)	18 (34.6)
≥12	51 (58.7)	10 (58.8)	13 (72.2%)	28 (53.8%)
Recurrent anatomic site					0.53
Local metastasis	17 (19.5)	3 (17.6)	2 (11.1)	12 (23.1)
Distant metastasis†	70 (80.5)	14 (82.4)	16 (88.9)	40 (76.9)

Values are represented as median (range) or number (%).

The Kruskal-Wallis test or χ² test for P-values. Histology, stage were data from the initial diagnosis.

^*Values at recurrence and 3 cases are missing data for laboratory tests at recurrence.

^†Distant metastasis with or without local metastasis.

BUN indicates blood urea nitrogen; DFS, disease-free survival from the completion of the initial treatment; PTX, paclitaxel; WBC, white blood cell; WPRT, whole pelvic radiation therapy.

Details of modified dose-dense paclitaxel monotherapy treatment outcomes are shown in Table 2. No patient had received paclitaxel previously. All the patients except one had received prior platinum chemotherapy at the initial treatment and/or recurrent treatment. Most patients (83%) receiving this regimen received it as the second or greater line of therapy. There were 6 (35.3%) of 17 patients who received ≥6 cycles including 4 (23.5%) patients who received ≥12 cycles as maintenance chemotherapy. The response rate after 6 cycles of treatment was 29.4% (1 complete response and 4 partial responses). Among these 5 cases, 2 cases underwent surgical resection of their recurrent disease after MDDP monotherapy. In all cases except one, these treatments targeted pulmonary or hilar nodal metastases. Disease progression was the most common indication for treatment discontinuation (n = 15, 88.2%), and 1 patient (5.9%) discontinued chemotherapy due to an adverse event with grade 3 transaminitis. Postponement of chemotherapy was seen in 2 (1.4%) of 140 total cycles for the 17 patients; one for transaminitis and the other for a hypersensitivity reaction.

TABLE 2.

Treatment Patterns and Outcomes with Modified Dose-dense Paclitaxel Monotherapy (N = 17)

Characteristics	n (%)
Prior paclitaxel treatment	0
Prior cisplatin treatment
No	1 (6)
Yes	16 (94)
Total cycles of treatment
≤6	11 (64.7)
6–12	2 (11.8)
>12	4 (23.5)
Treatment line
1st line	3 (17.6)
2nd line	6 (35.3)
3rd line	4 (23.5)
4th line	4 (23.5)
No. cycles administered (median [range])
1st line	24 (2–36)
2nd line	4 (1–18)
3rd line	2 (1–3)
4th line	4.5 (3–21)
Response after 6 cycles treatment
Complete response	1 (5.9)
Partial response	4 (23.6)
Stable disease	0
Progressive disease	12 (70.6)
Reason for discontinuation
Disease progression	15 (88.2)
Liver toxicity	1 (5.9)
Complete response	1 (5.9)

DFS indicates disease-free survival.

Chemotherapeutic regimens for recurrent cervical cancer among patients who did not receive modified dose-dense paclitaxel monotherapy are shown in Supplemental Table S1 (Supplemental Digital Content 1, http://links.lww.com/AJCO/A181). Among these 70 cases, the majority received combination therapy (82.9%). Platinum was the most common agent used for salvage therapy (75.7%) followed by gemcitabine (38.6%) in this population. For individual chemotherapy regimens, gemcitabine combined with a platinum agent (cisplatin n = 25 and carboplatin n = 1, 35.1%) was the most common regimen followed by paclitaxel with a platinum agent (cisplatin n = 8 and carboplatin n = 5, 16.9%). Ten percent of regimens combined molecular targeted drugs with chemotherapeutic agents.

A survival analysis was performed with results shown in Table 3. Median SAR of all the patients was 13.7 months (range: 0.9 to 113.5 mo), and the median SAR of women who received the modified dose-dense paclitaxel regimen was 20.5 months (range: 4.3 to 113.5 mo). There were 48 (55.2%) patients who died from cervical cancer and 15 (17.2%) patients admitted to hospice for disease progression. Longer DFS was significantly associated with longer SAR (DFS <12 vs. ≥12 mo, 2-y SAR rates 25.5% vs. 44.0%; P = 0.005). Clinicopathologic factors associated with decreased SAR included older age at recurrence, nonsquamous cell histology, advanced-stage disease at initial cancer diagnosis, and distant metastasis at recurrence, although these did not reach statistical significance.

TABLE 3.

Multivariate Analysis for Survival After Recurrence

	SAR			Univariate		Multivariate (Conditional Backward)
Characteristics	N	Median (mo)	2-y (%)	HR (95% CI)	P	HR (95% CI)	P
Age at recurrence (y)					0.81
< 50	34	13.4 (2.1–113.5)	39.0	1		—	—
≥50	53	13.7 (0.9–78.3)	34.3	1.07 (0.64–1.76)		—	—
Year of recurrence					0.19
2006–2009	28	11.3 (0.9–113.5)	25.5	1		—	—
2010–2014	59	14.3 (2.1–78.3)	41.9	0.71 (0.43–1.19)		—	—
DFS (mo)					0.005
≥12	51	16.0 (0.9–78.3)	44.0	1		1	—
< 12	36	7.0 (2.1–113.5)	25.6	2.01 (1.22–3.32)		1.96 (1.18–3.26)	0.009
Stage at diagnosis					0.22
Early stage	11	17.6 (9.7–33.9)	52.1	1		—	—
Locally advanced/distant	76	11.4 (0.9–113.5)	33.8	1.63 (0.74–3.60)		—	—
Histology					0.25
Squamous cell	65	12.5 (2.1–78.3)	35.2	1		—	—
Nonsquamous cell	22	15.7 (0.9–113.5)	40.0	0.71 (0.40–1.28)		—	—
Recurrence site					0.10
Local metastasis	17	13.9 (0.9–113.5)	42.4	1		—	—
Distant metastasis*	70	12.5 (3.5–46.2)	13.2	1.82 (0.92–3.61)		—	—
Salvage chemotherapy regimen					0.044
Nonpaclitaxel	52	10.1 (0.9–78.3)	28.1	1		1	—
Modified dose-dense paclitaxel	17	20.5 (4.3–113.5)	54.1	0.50 (0.26–0.96)		0.50 (0.26–0.95)	0.036
Paclitaxel-doublet	18	20.8 (2.7–47.7)	43.6	0.55 (0.28–1.09)		0.62 (0.31–1.22)	0.16

Log-rank test for univariate analysis. Cox proportional hazard regression test for multivariate analysis. Clinically meaningful variables as listed in Table were initially entered in the model, and conditional backward method was used to retain only significant variables in the final model.

Significant P-values are emboldened.

^*Distant metastasis with or without local metastasis.

CI indicates confidence interval; DFS, disease free survival from diagnosis; HR, hazard ratio; SAR, survival after recurrence.

On univariate analysis, the modified dose-dense paclitaxel regimen had the highest 2-year SAR rate among the 3 treatment groups: modified dose-dense paclitaxel 54.1%, paclitaxel-doublet 43.6%, and nonpaclitaxel regimens 28.1% (P_trend = 0.044; Fig. 1). On multivariate analysis (Table 3), women who received the modified dose-dense paclitaxel monotherapy had a statistically significantly improved SAR compared with women who did not receive this regimen for recurrent cervical cancer (adjusted-hazard ratio, 0.50; 95% confidence interval, 0.26–0.95; P = 0.036). In addition, DFS <12 months remained an independent prognostic factor for decreased SAR (adjusted-hazard ratio, 1.96; 95% confidence interval, 1.18–3.26; P = 0.009).

FIGURE 1.

Survival time after recurrence based on chemotherapy types. The Kaplan-Meier method for survival curves for survival after recurrence of cervical cancer. Log-rank test for P-value. mdd-PTX indicates modified dose-dense paclitaxel monotherapy; non-PTX, nonpaclitaxel regimen; PTX, paclitaxel.

Systematic Review

Results of the systematic review for the effectiveness of paclitaxel regimens in recurrent cervical cancer are summarized in Tables 4 and 5.^{7,8,10,12–14,29–48} There were 26 studies examining the efficacy of paclitaxel in recurrent cervical cancer (selection schema, Supplemental Fig. S2, Supplemental Digital Content 1, http://links.lww.com/AJCO/A181). No study in the search examined the efficacy of dose-dense paclitaxel monotherapy for recurrent cervical cancer. The response rate ranged from 17% to 31% (median, 25%) for paclitaxel monotherapy regimens (Table 4). The response rate was up to 68% in combination paclitaxel regimens (median: 46%, range: 20% to 68%). Table 5 outlines the toxicity profile for paclitaxel regimens in recurrent cervical cancer. Hematologic and neurological toxicities were rare for paclitaxel monotherapy compared with combination therapy.

TABLE 4.

Efficacy of Paclitaxel in Recurrent Cervical Cancer: Systematic Review of Literature

References	Type*	Design	N	Regimen	Cycle	RR (%)	PFS	OS
Single regimen
Curtin et al12	R/P Non-SCC	Phase II	42	PTX (170/135 mg/m2 q21)	5	31	4.8	NA
McGuire et al13	R/P SCC	Phase II	55	PTX (170/135 mg/m2 q21)	2	17	3.4	NA
Kudelka et al14	R/P SCC	Phase II	32	PTX (250/m2 q21)	4	25	3.5	7.3
Combination regimen
Symonds et al29 (CIRCCa)	R/P	Phase II	35	PTX (175 mg/m2 q21) + CBDCA (AUC5 q21)	6	45	6.7	NA
Kitagawa et al30 (JCOG0505)	R/P/M	Phase III	126	PTX (175 mg/m2 q21) + CBDCA (AUC5 q21)	6	63	6.2	17.5
			127	PTX (135 mg/m2 q21) + CDDP (50 mg/m2 q21)		59	6.9	18.3
Vergote et al31	R/M	Retro	34	PTX (60 mg/m2 weekly) + CBDCA (AUC 2.7 weekly) + G-CSF	18	51	7	13
Penson et al32 (GOG 240)	R/P/M	Phase III	114	PTX (135/175 mg/m2 q21) + CDDP (50 mg/m2 q21)	6	45	NA	14.3
			111	PTX (175 mg/m2 q21) + TOPO (0.75 mg/m2 q21)		27		12.7
Tewari et al10	R/P/M	Phase III	229	PTX (135/175 mg/m2 q21) + CDDP (50 mg/m2 q21) + w/, w/o Bev 15 mg/kg	6	45–50	7.6	15
			223	PTX (175 mg/m2 q21) + TOPO (0.75 mg/m2 q21) + w/, w/o Bev 15 mg/kg	6	27–47	5.7	12.5
Garces et al33	R/M	Retro	154	PTX (175 mg/m2 q21) + CBDCA (AUC5 q21)	6	35	5.2	10.6
Torfs et al34	R/M	Retro	43	PTX (90 mg/m2 day1.8 q 21) + CBDCA (AUC4 day 1.8 q21)	6	58	5	11
			22	PTX (60 mg/m2 weekly) + CBDCA (AUC 2.7 weekly)		36	NA	10
Downs et al35	R M	Phase II	21	PTX (175 mg/m2 q28) + CBDCA (AUC5 q28) + IFO (2 g q28)	4	33	5	10
Monk et al8 (GOG 204)	R/P	Phase III	103	PTX (135 mg/m2 q21) + CDDP (50 mg/m2 q21)	6	29	5.8	12.9
Mabuchi et al36	R	Retro	28	PTX (175 q28) + CBDCA (AUC5–6 q28) or PTX (80 day1.8.15 q28) + CBDCA (AUC2 day1.8.15 q28)	5	68	7	NA
Kosmas et al37	R/M	Phase I/II	42	PTX (175 mg/m2 q21) + CDDP (40 mg/m2 q21) + IFO (2.5 g day 1–2 q21)	6	62	7	16.5
Pectasides et al38	R/M	Phase II	51	PTX (175 mg/m2 q21) + CBDCA (AUC5 q21)	6	53	6	13
Mountzios et al39	R/M	Phase II	76	PTX (175 mg/m2 q21) + CDDP (70 mg/m2 q21) + IFO (1.5 g day 1–3 q21)	6	27	7.9	15.4
Secord et al40	R/M	Phase II	15	PTX (80 day 1.8.15 q28) + CBDCA (AUC2 day 1.8.15 q28)	4	20	3.4	7.6
Choi et al41	R	Phase II	53	PTX (135 mg/m2 q21) + CDDP (40 mg/m2 q21) + IFO (1.5 g day 1–3 q21)	4.8	47	8	19
Tinker et al42	R/M	Retro	25	PTX (155–175 mg/m2 q21) + CBDCA (AUC5–6 q21)	4	40	3	21
Moore et al7	R/P	Phase III	130	PTX (175 mg/m2 q21) + CDDP (50 mg/m2 q21)	5	47	4.9	9.9
Tiersten et al43	R/P/M	Phase II	15	PTX (175 mg/m2 q21) + TOPO (1 mg/m2 day 1–5 q21)	5	54	3.8	8.6
Dimopoulos et al44	R/M	Phase II	57	PTX (175 mg/m2 q28) + CDDP (75 mg/m2 q28) + IFO (1.5 g day 1–3 q28)	6	46	8.3	18.6
Papadimitriou et al45	R/M	Phase II	34	PTX (175 mg/m2 q21) + CDDP (75 mg/m2 q21)	6	47	5	9
Rose et al46	R/P	Phase II	41	PTX (135 mg/m2 q21) + CDDP (75 mg/m2 q21)	6	46	5.4	10
Piver et al47	R/M	Phase II	20	PTX (135 mg/m2 q21) + CDDP (75 mg/m2 q21)	6	45	10.5	13
Zanetta et al48	R/P	Phase II	45	PTX (175 mg/m2 q21) + CDDP (75 mg/m2 q21) + IFO (5g day 1 q21)	4	67	NA	NA

^*Type of tumors.

AUC indicates area under the curve; Bev, bevacizmab; CBDCA, carboplatin; CDDP, cisplatin; IFO, ifosfamide; M, metastasis cervical cancer; NA, not available; Non-SCC, non–squamous cell carcinoma histology; OS, overall survival; P, persistent cervical cancer; PFS, progression-free survival; PTX, paclitaxel; R, recurrent cervical cancer; retro, retrospective; RR, response rate; SCC, squamous cell carcinoma histology; TOPO, topotecan; w/, with; w/o; without.

TABLE 5.

Frequency of Grade 3 to 4 Toxicity in Paclitaxel Regimens for Recurrent Cervical Cancer: Review of Literature

References	Hematologic (%)*	Neutropenia (%)	Thrombocytopenia (%)	Neurotoxicity (%)	Hypersensitivity (%)	Liver Toxicity (%)
Single regimen
Curtin et al12	8	74	2	2	NA	NA
McGuire et al13	7	36	0	0	NA	0
Kudelka et al14	2	28	0	6	NA	NA
Combination regimen
Symonds et al29 (CIRCCa)	9	12	9	0	NA	NA
	6	31	0	6	NA	NA
Kitagawa et al30 (JCOG0505)	44	67	25	5	NA	NA
	31	84	3	0	NA	NA
Vergote et al31	81	38	19	NA	NA	NA
Penson et al32 (GOG 240)	NA	56	NA	6	3	NA
Tewari et al10 (w/o Bev)	NA	26	NA	NA	NA	NA
Garces et al33	43	18	9	0.7	2	NA
Torfs et al34	42	63	42	0	14	NA
	55	50	46	0	10	NA
Downs et al35	19	95	30	0	NA	NA
Monk et al8 (GOG 204)	17	78	7	2	5	0
Mabuchi et al36	NA	78	NA	NA	NA	NA
Kosmas et al37	7	83	9	0	0	NA
Pectasides et al38	22	36	14	6	2	0
Mountzios et al39	10	26	10	4	0	0
Secord et al40	NA	NA	NA	NA	NA	NA
Choi et al41	8	13	6	5	NA	0
Tinker et al42	32	32	12	0	NA	NA
Moore et al7	28	67	4	3	NA	NA
Tiersten et al43	47	27	13	13	NA	NA
Dimopoulos et al44	13	26	7	3	2	0
Papadimitriou et al45	18	15	3	9	0	0
Rose et al46	27	83	20	7	0	0
Piver et al47	35	55	15	15	NA	NA
Zanetta et al48	73	91	49	0	0	2

^*Anemia.

Bev indicates bevacizmab; NA, not available; w/o, without.

Results of the systematic review for dose-dense paclitaxel for gynecologic cancer are shown in Supplemental Tables S2–S3 (Supplemental Digital Content 1, http://links.lww.com/AJCO/A181). There were 14 relevant studies identified in the literature search (Supplemental Fig. S3, Supplemental Digital Content 1, http://links.lww.com/AJCO/A181), the majority of which were on ovarian cancer (64.3%). There were 4 studies for cervical cancer but these all examined combination therapy with carboplatin. In a pooled analysis for all gynecologic cancer types, the response rate for dose-dense paclitaxel regimens was 53% to 86% (median, 67%). Among the 4 cervical cancer studies examined, response rate was similar to other cancer types (median: 56%, range: 53% to 68%).

DISCUSSION

The key finding of this study is that women who received modified dose-dense paclitaxel monotherapy for recurrent cervical cancer had better survival outcomes compared with women who did not receive this regimen. In addition, few patients on this regimen developed adverse events, suggesting that modified dose-dense paclitaxel monotherapy may be an effective and safe chemotherapeutic option for women with recurrent cervical cancer.

The idea that dose-dense metronomic administration of paclitaxel might produce superior clinical outcomes compared with the standard administration was first reported in the breast cancer literature.^49,50 Since then, multiple studies have explored the effectiveness of dose-dense paclitaxel chemotherapy on survival outcomes in various types of malignancies.^19,51,52 Currently, dose-dense paclitaxel has been implemented as a treatment for ovarian cancer (Supplemental Table S2, Supplemental Digital Content 1, http://links.lww.com/AJCO/A181). JGOG-3016, a randomized phase III clinical trial for ovarian cancer, reported that dose-dense paclitaxel combined with carboplatin improved overall survival (100.5 vs. 62.2 mo) and progression-free survival (28.0 vs. 17.2 mo) compared with a conventional paclitaxel and carboplatin regimen, given every 3 weeks.¹⁹ The dose-dense regimen resulted in increased hematologic toxicity, especially anemia (69% vs. 44%), compared with the every 3 week regimen. GOG-262, in which patients received standard carboplatin every 3 weeks combined with either dose-dense or conventional paclitaxel regimens, also showed the effectiveness of dose-dense chemotherapy in advanced ovarian cancer.¹⁸ That is, among patients not taking bevacizumab in this study, progression-free survival was superior in the dose-dense paclitaxel group compared with the conventional group (14.2 vs. 10.3 mo).¹⁸ Given these encouraging results in other malignancies, evaluating the use of a dose-dense paclitaxel regimen in recurrent cervical cancer is warranted.

Various chemotherapeutic agents have been previously evaluated, and cisplatin, carboplatin, paclitaxel, and topotecan are considered active agents for recurrent cervical cancer (Table 4).^13,19,31,32 The most effective regimens for treating recurrent or metastatic cervical cancer include cisplatin and/or paclitaxel.^7–10 The median response rate for paclitaxel monotherapy was 25% and progression-free survival for paclitaxel-containing regimens was 3.5 months. Recent randomized trials from GOG-240 demonstrated that a topotecan + paclitaxel + bevacizumab regimen was not inferior to a cisplatin + paclitaxel + bevacizumab regimen. No significant difference in overall survival was seen; however, the trend for response rate (45% vs. 27%) and overall survival (14.3 vs. 12.7 mo) suggest that the cisplatin + paclitaxel + bevacizumab regimen is superior.³² In addition, a phase III randomized trial (JCOG-505) showed that the response rate (63% vs. 59%) and overall survival (17.5 vs. 18.3 mo) were similar between carboplatin + paclitaxel and cisplatin + paclitaxel regimens.³⁰ These studies, therefore, endorse the importance of paclitaxel in the management of recurrent cervical cancer.

As prior cisplatin chemotherapy during initial treatment may decrease cisplatin sensitivity in the recurrent setting, other nonplatinum chemotherapy agents may be potential options. Our study explored the utility of modified dose-dense paclitaxel monotherapy in recurrent cervical cancer and found that this regimen was associated with better survival compared with nonpaclitaxel regimens. Patients who received prior cisplatin-based treatment may be good candidates for single-agent paclitaxel. Furthermore, our dose-dense regimen, which skips day 22 administration, has proven to be well tolerated despite the fact that half of patients (47%) were undergoing their third or fourth lines of chemotherapy. Because we did not examine the classic dose-dense paclitaxel regimen (day 1, 8, 15, and 22), it remains unknown if similar findings will be seen in the classic regimen and merits further investigation. However, the response rate of modified dose-dense paclitaxel monotherapy (29.4%) was higher than the rate identified for conventional every 3 week paclitaxel monotherapy in our systematic review (median 25%).

The cytotoxic activity of paclitaxel relies on the duration of exposure, so theoretically, weekly administration provides enhancement of antitumor activity over regimens with longer dosing intervals.¹⁹ According to the Norton-Simon hypothesis, the rate of regrowth during treatments will be proportional to the rate of tumor growth. Per the Gompertzian model, smaller tumors grow more quickly than larger ones, so if tumors are given less time to regrow between treatments, they remain small with rapidly dividing cells, facilitating chemotherapy-induced cell death.¹⁵ A shorter interval of paclitaxel administration, therefore, should more effectively reduce cancer cellular proliferation and will result in progressively smaller tumor burden. In addition, paclitaxel chemotherapy may inhibit angiogenesis, resulting in increased tumor drug penetration. Multiple prior studies have shown that paclitaxel reduced the extent of intratumoral angiogenesis in a dose-dependent manner and also suppressed expression of vascular endothelial growth factor in tumor cells.^53,54 Oncogenic HPV products E6 and E7 up-regulate tumor angiogenesis via the vascular endothelial growth factor pathway, therefore, combination use of bevacizumab and modified dose-dense paclitaxel may enhance the activity of antiangiogenesis. Moreover, others theorize that dose-dense chemotherapy is associated with a less toxic immune environment, reducing immunosuppression in tumor microenvironments and triggering the recruitment of macrophages and the antitumor T-cell response determined by interleukin 2 and interferon γ secretion.⁵⁵ Tumor-associated macrophages and antitumor T cells play important roles in antitumor immunity, inducing apoptosis via programmed death 1 and cytotoxic T lymphocyte–associated antigen 4 ligand, thus suppressing tumor proliferation and inhibiting tumor growth and metastasis.^56,57

Our study highlights the potential efficacy of salvage chemotherapy with dose-dense paclitaxel monotherapy in recurrent cervical cancer. The National Comprehensive Cancer Network guidelines have given category 1 evidence for bevacizumab, when used in combination with cisplatin and paclitaxel, and this has become the standard of care as first-line therapy for recurrent cervical cancer (www.nccn.org/professionals/physician_gls/pdf/cervical.pdf). Therefore, potential patient candidates for modified dose-dense paclitaxel monotherapy in this setting include women who are not able to tolerate this combination therapy, women who developed adverse events to cisplatin, and women who have contraindications to bevacizumab use. Of note, in our study, all patients treated with modified dose-dense paclitaxel monotherapy recurred before bevacizumab was approved by the United States Food and Drug Administration in August 2014, (www.cancer.gov/about-cancer/treatment/drugs/fda-bevacizumab) and thus, our study patients were rarely exposed to bevacizumab.

The main limitation of our study is its retrospective nature and possible confounding variables. Regardless, the majority of patients treated with modified dose-dense paclitaxel in this study were receiving it as second-line or third-line therapy, and nonetheless, were noted to have a relatively longer survival after recurrence compared with patients in the GOG 240 or JCOG-505 trials.^30,32 In our study, using a pooled analysis including any line of chemotherapy (salvage therapy and maintenance therapy) may have influenced the results. As evidenced by the phase II GOG trials, response rates for second-line chemotherapy range from 0% to 20% at best.^13,58–60 A relatively high response rate for modified dose-dense paclitaxel monotherapy in this study warrants consideration of patient selection bias. Although the appropriate study design is to perform a comparison of each treatment line, ideally prospectively, our limited sample size did not permit this approach. Because the majority of our study population was Hispanic, generalizability to different population remains unknown. Lastly, we used the date of hospice transfer due to terminal cancer stage as the survival endpoint because we not able to follow-up the final survival status after hospice transfer. Although this allocation can result in time-lead bias, we believe that hospice transfer can be a surrogate marker for survival endpoint due to minimal life-expectancy after hospice transfer.²⁷

The importance of the modified dose-dense paclitaxel regimen is its feasibility, tolerability, and potential survival benefit. This allows us to incorporate its use as a second or greater line of therapy for recurrent cervical cancer, a situation for which options have been very limited. These findings also should be validated in a different population via a prospective study to examine the safety and efficacy of modified dose-dense paclitaxel monotherapy.