Comparative efficacy evaluation of catheter intraperitoneal chemotherapy, normothermic and hyperthermic chemoperfusion in a rat model of ascitic ovarian cancer

Abstract

Objectives: The choice of an optimal administration route for intraperitoneal (IP) chemotherapy and a suitable chemotherapeutic regime in the treatment of ovarian cancer remains a controversy. We investigated survival outcomes according to catheter intraperitoneal chemotherapy (CIPC), normothermic and hyperthermic chemoperfusion (NIPEC and HIPEC) with cytostatic drugs dioxadet and cisplatin in rats with transplantable ascitic ovarian cancer.

Methods: Ascitic liquid containing 1 × 10⁷ tumour cells was inoculated to female Wistar rats and 48 hours after rats received dioxadet and cisplatin at the maximum tolerated doses. Dioxadet at doses 1.5, 30 and 15 mg/kg and cisplatin at doses 4, 40 and 20 mg/kg body weight were administered for CIPC, NIPEC and HIPEC, respectively. Rats in the control groups received physiological saline and CIPC with physiological saline was regarded as the untreated control. The antitumor activity of the drugs was evaluated as an increase in average life expectancy (ALE). Analysis of the data was based primarily on Bayesian statistics and included Kaplan–Meier method, log-rank test and hazard ratio (HR) estimation.

Results: Compared to the untreated control CIPC, NIPEC and HIPEC with dioxadet significantly increased ALE by ₁₀13₁₆, ₆15₂₄ and _1.717₃₅ days, whereas with cisplatin by ₆10₁₃, ₁₂24₃₇ and _–135₂₃ days, respectively.

Conclusions: Dioxadet and cisplatin show similar efficacy in the CIPC route. Compared with CIPC IP chemotherapy by chemoperfusions is more effective for both the drugs. Dioxadet in HIPEC showed highest survival benefit whereas largest effect during NIPEC is achieved with cisplatin.

Keywords:

• Ovarian cancer
• intraperitoneal chemotherapy
• NIPEC
• HIPEC
• cisplatin
• dioxadet

Introduction

Epithelial ovarian cancers (EOCs) rank the fifth place within overall oncological mortality in women and the first place among all gynaecological malignancies [1]. Current standards for the initial EOCs management include optimal debulking and adjuvant intravenous (IV) chemotherapy with platinum-based drugs and taxanes [2]. Despite the fact that large proportion of patients at early disease stages will respond to such treatment, most of the women, which are presented with advanced disease, are less likely to benefit from it with higher rates of recurrences and shorter progression free intervals [3]. Recurrences occurred in a six months period after last chemotherapy cycle are considered potentially resistant to platinum drugs and require second-line therapy, to which only 30% of the patients will respond [4]. As a result, roughly 80% of women are going to develop peritoneal carcinomatosis and ascites-associated complications [5].

The characteristic features of advanced and recurrent EOCs in most cases are the implantation metastases to the parietal and visceral peritoneum, the diaphragm dome, lesser and greater omentum, which prompt that if we are to improve the cure of EOCs, an adequate therapy should also meet the specifics of peritoneal malignancies [5–7]. Intraperitoneal (IP) chemotherapy has been discussed as the most promising strategy to challenge current EOCs treatment standards, though the techniques of its delivery vary [4]. Pharmacokinetic rationale behind IP chemotherapy enables higher doses of anticancer drugs to act locally on the diseased tissues, thus providing less systemic toxicity compared to IV chemotherapy [8]. Generally, three IP chemotherapy routes are described: catheter IP chemotherapy (CIPC) via IP catheter or port, normothermic (36.0–37.0 °C) IP chemoperfusion (NIPEC) with cytostatic solution being instilled and drained via system of pumps, and hyperthermic IP chemoperfusion (HIPEC) via same system under heated conditions (38.5–43.0 °C) [5,7].

Long-term observations from phase III clinical trials have established that 10-year survival rate for advanced stages of EOCs could be within reach for at least 50% of the patients by the means of IP chemotherapy [4]. However, lack of compelling evidence and conflicting results between major clinical trials have hindered its further clinical translation [9]. Moreover, no direct comparison on the efficacy between the three IP administration routes has been conducted. Neither has even been addressed the question of whether the outcome of a particular IP route could be compromised by the choice of antitumor drug. For instance, IV cisplatin is dose-limited by nephrotoxicity and exhibits neural, and ototoxicity [10]. Research has shown that IP delivery of cisplatin has similar toxicity profile due to its rapid clearance into blood plasma, and also a tendency to cause adhesions in the peritoneum [8,10]. While this safety profile does not necessarily change the outcome of HIPEC route, its administration via catheter or port could be rather hazardous.

In this regard, dioxadet, an alkylating agent from the ethylenimine class, has been proposed as a potential substitute for cisplatin-based chemotherapeutic regimes. Dioxadet has been approved in Russia for systemic and intracavitary (intrapleural and IP) administration in the treatment of various malignancies. This includes ascitic ovarian cancer, locally advanced and disseminated breast cancer and lung cancer. The drug is also used for endovascular and endolymphatic chemotherapy of tumoural lesions of retroperitoneal lymph nodes in patients with ovarian, cervical, uterine cancers and malignant lymphomas. Phase II clinical trial of dioxadet has demonstrated its pronounced antitumor efficacy during IV + IP administration with dose-limiting but reversible myelosuppression in women with advanced EOCs [11]. Its ability to penetrate tumour tissue on direct contact as well as cell-cycle independent activity leads to a renewed interest of dioxadet as a compound for IP regimes [11–14].

Our previous studies on rats with ascitic ovarian cancer have demonstrated that single IP administration of some cytostatics, including dioxadet, cisplatin, mitomycin C, melphalan and paclitaxel statistically significantly increased median survival compared to their IV administration, while the highest median survival was achieved with dioxadet and cisplatin [15]. In our other report on the safety and efficacy of cisplatin and dioxadet on the same rodent model it was confirmed that their administration in NIPEC and HIPEC mode also resulted in better survival compared to IV and IP chemotherapy. However, the toxicities associated with cisplatin were more severe [16]. The goal of the present study was to directly compare the efficacy of CIPC, NIPEC and HIPEC with dioxadet and cisplatin in a rat model of an ascitic ovarian cancer.

Methods and materials

Experimental design

Outbred female Wistar rats were purchased from “Rappolovo” animal nursery (Leningrad Region, Russia) and housed 5–7 in polypropylene cages under standard 24-h light–dark regimen (12L:12D) at 22 ± 2 °C with relative humidity 50–60%. Rats received standard laboratory crop from Laboratorkorm Ltd. (Moscow, Russia) and tap water ad libitum. All animal experiments were performed following the approval of the N.N. Petrov Research Institute of Oncology Ethic Committee within the frame of rules declared by the European Treaty ETS No. 123.

The rat ovarian cancer strain was obtained from N.N. Blokhin Russian Cancer Research Center (Moscow, Russia). Ascitic liquid containing 1 × 10⁷ tumour cells was diluted 1:4 with physiological saline and injected IP to healthy rats. Day of ovarian cancer inoculation was accepted for zero. Two hundred and eighty-eight rats were then randomised into total of nine groups for the three IP chemotherapy regimes with each antitumor drug as well as concomitant control to each regime (physiological saline). A single intervention proceeded 48 h after ovarian cancer inoculation. Dioxadet (Chemconsult, Saint Petersburg, Russia) and cisplatin (Pharmachemie, Haarlem, Netherlands) were administered at the maximum tolerated doses in CIPC via syringe, 1.5 and 4 mg/kg body weight, during NIPEC (36.5–37.5 °C) and HIPEC (40.5–41.5 °C), 30 and 40 as well as 15 and 20 mg/kg body weight, were infused for a continuous circulation using an experimental setting with pumps, respectively. Antitumor effects of the treatments were estimated as an increase in average life expectancy (ALE) compared to the control without treatment (IP saline injection) and to concomitant controls (CIPC, NIPEC and HIPEC with physiological saline). The diagnosis of ovarian cancer was confirmed by cytological examination of ascitic smear stained with haematoxylin and eosin. The methodology has been thoroughly described in our previous reports [12–14]. Nine rats died due to surgical complications and were excluded from the analysis.

Statistical analysis

In accordance with modern tendency we insist on applying Bayesian approach to statistical data analysis together with orthodox frequentist approach, and even instead of it. There is a lot of corresponding publications by foremost biostatisticians including American Statistical Association (ASA) and U.S. Food and Drug Administration (FDA) [17]. Our approach could be referred to as “harmonizing statistical evidences and predictions” as it combines not only confidence interval (CI) estimates for parameters, for the standardised effect size d_C, for the achieved power (1 − β), and planned sample sizes, but also prediction intervals (PIs) for them and for the p values, as well as estimates of posterior probabilities for the null hypothesis P(H_o|D) and Bayes factors BF₀₁. To present point estimates of parameters and CIs we prefer to use the format with subscripts invented by John Tukey and recommended by Louis and Zeger [18]. Its advantage consists in compactness and visibility, for instance, _0.71.2_1.8 against 1.2 (with 95% CI from 0.7 to 1.8). Information about the used software and their appointments is presented in Table 1.

Table 1. Software applied for the statistical analysis.

Program	Version	Intention	URL	References
PAST	3.1	Point and interval estimations, normality tests, significance testing using bootstrap and Monte Carlo	http://folk.uio.no/ohammer/past/	[19]
JASP	0.7.5.5	Bayesian t-test and Cohen’s dC estimation	https://jasp-stats.org/	[20]
BF Calculators		Bayes factor for grouped or two-sample t-tests	http://pcl.missouri.edu/bayesfactor	[21]
G*Power	3.1.9.2	Power analysis	http://www.gpower.hhu.de/	[22]
LePrep	2.1.0	Probabilities of replication and prediction intervals	http://lmrs.univ-rouen.fr/Persopage/Lecoutre/PAC.htm	[23]
R (packages “ggplot2”, “survival”)	3.4.0	Kaplan–Meier plots, log-rank test and hazard ratios using Cox proportional hazards estimation	https://www.R-project.org	[24]

Monte Carlo and bootstrap resampling procedures were applied preferably. Levene’s test was employed to check the assumption of equal variances followed by Tukey’s post hoc multiple comparison. The data were then further tested for statistical homogeneity using one-way ANOVA. Bonferroni’s correction was used to address the multiple comparison problem. Pairwise comparison was carried out using randomised Student’s t-test. The observed effects were verified with Bayes factor (BF) and tested for practical (respectively clinical) significance using Cohen’s d_C. Probability of replication (P_srep) along with PIs were calculated for both non-standardised (D) and standardised (d_C) metrics. A priori power analysis was carried out to estimate an adequate sample size to achieve the desired power (1 − β = 0.95) for the chosen level of probability of type I error (α* = 0.05). Survival analysis included Kaplan–Meier’s method and two-sided log-rank test. Hazard ratios (HR) were estimated using Cox’s proportional hazards model. Cumulative hazard plots were analysed for violation of proportional hazards assumption. All statistical analyses (Table 1) were performed using PAST 3.1 [19], JASP 0.7.5.5 [20], BF Calculators [21], G*Power 3.1.9.2 [22], LePrep 2.1.0 [23] and R 3.4.0 [24].

Results

In all rats implantation-take achieved 100% and lead to ascites and peritoneal carcinomatosis. Death of the animals in control without treatment occurred in 7–14 days after ovarian cancer inoculation. The ALE in the control without treatment accounted for ₁₂15₁₉ days. Contrarily, CIPC, NIPEC and HIPEC with dioxadet increased this rate by 13 (p = 1.7 × 10⁻¹³), 15 (p = 3.1 × 10⁻⁷) and 31 days (p = 1.6 × 10⁻¹⁴) or 85%, 97% and 202%, respectively (Table 2). Kaplan–Meier’s plots are shown in Figure 1. In cisplatin-treated rats this increase accounted for 10 (p = 4.4 × 10⁻⁹), 24 (p = 2.6 × 10⁻¹¹) and 19 days (p = 1.8 × 10⁻⁷) or 66%, 160% and 126%, respectively (Table 2).

Figure 1. Survival outcomes in control groups and according to treatment with cisplatin vs. dioxadet by the means of catheter intraperitoneal chemotherapy (CIPC), normothermic and hyperthermic chemoperfusion (NIPEC and HIPEC) in rats with ascitic ovarian cancer.

Table 2. The results of harmonised statistical analysis of survival in rats with ascitic ovarian cancer in the course of the various IP chemotherapy routes.

Groups and metrics	Notation	Diox-CIPC	Cpt-CIPC	Diox-NIPEC	Cpt-NIPEC	Diox-HIPEC	Cpt-HIPEC
Intervention	n	79	52	14	12	13	14
	M1	252831*,**	222527*,**	213039*,**	273952*,**	374656*	203548*
Concomitant control	n	79		11		14
	M2	131517		4,71525		213039
Effect (difference)	D = M1 − M2	101316	61013	61524	122437	–1,71735	–13523
	95% PI	(8.5; 17.5)	(5.4; 14.0)	(1.7; 29.0)	(6.5; 42.5)	(–9.3; 42.2)	(–20.8; 30.8)
p value	Pval	1.7 × 10^−13	4.4 × 10^−9	0.0033	0.00062	0.74	0.58
	95% PI	(0.0; 5 × 10^−7)	(10^−16; 5 × 10^−4)	(3 × 10^−7; 0.37)	(5 × 10^−8; 0.17)	(2 × 10^−5; 0.84)	(8 × 10^−4; 0.99)
Bayes factor	BF10	3·10^9	2.17·10^6	11.9	44.17	1.25	0.39
Probability of replication	Psrep	1.00	0.99	0.79	0.88	0.45	0.15
Standardized effect size	dC	0.91.31.6	0.71.11.5	0.41.32.2	0.71.62.6	–0.070.721.5	–0.530.20.95
	95% PI	(0.81; 1.8)	(0.6; 1.6)	(0.13; 2.7)	(0.4; 3.2)	(–0.38; 1.9)	(–0.87; 1.32)
Achieved power at α = 0.05	1 − β	0.991.001.00	0.970.991.00	0.180.880.99	0.410.950.99	0.050.440.96	0.270.080.67
Required n at α = 0.05 and (1 − β) = 0.95	n1 = n2	121731	112445	717136	111345	13525305	30651∞

Diox: dioxadet; Cpt: cisplatin; CIPC: catheter intraperitoneal chemotherapy; NIPEC: normothermic intraperitoneal chemoperfusion; HIPEC: hyperthermic intraperitoneal chemoperfusion; Concomitant control: administration of physiological saline – for CIPC: IP injection; for NIPEC: normothermic perfusion; for HIPEC: hyperthermic perfusion; n: sample size; M₁ and M₂: the mean values of Life Expectancy in days with and without treatment; D: mean difference; PI: prediction interval; P_val: the observed P-value; BF10: the Bayes Factor of H1 against H0; dC: Cohen’s effect size; (1 – β): achieved statistical power; α: significance level; n1 and n2: required sample size under desired α and (1 – β). Left and right subscripts are the lower and upper 95% confidence or prediction limits. *Statistical significance compared to the untreated control; **Statistical significance compared to the concomitant control.

Hyperthermic perfusion with physiological solution alone increases ALE by 14 days (p = 3.6 × 10⁻⁵) compared to the untreated control (HR _0.182.84_0.67). Yet HIPEC was the only mode where compared to its concomitant control no statistically significant increase in ALE for neither cisplatin (p = 0.58), nor dioxadet (p = 0.07) has been confirmed. In both cases 95% CIs for mean differences cover the indifferent value D₀ = 0 thus indicating the shift towards null hypothesis of no difference in means (Table 2). These data are supported by weak BF₁₀ = 0.08 and BF₁₀ = 0.44, respectively. The effect size could also be considered as negligible with d_C = 0.2 and d_C = 0.72 (Table 2).

Figure 2. Peritoneal carcinomatosis after intraperitoneal transplantation of ovarian cancer in rat. The autopsy of the rats reveals merging peritoneal tumour implants of greater and lesser omentum, mesentery, diaphragm dome, and miliary involvement of visceral and parietal peritoneum.

Figure 3. Smear of the ascitic fluid in rat with transplanted ovarian cancer. (A) The image depicts cyst of ovarian adenocarcinoma, cell mitosis and cell necrosis. (B) Haemorrhagic ascetic fluid contains large malignant cells with basophilic cytoplasm, nucleus mitosis, extensive cytoplasm vacuolisation and pseudopods on the external cytoplasmic membrane.

CIPC was associated with favourable survival for either of the applied cytostatics (Figure 1). Nevertheless, the results of dioxadet-CIPC were slightly more promising (HR = _0.50.72_1.02): 13 vs. 10 days (p = 1.7 × 10⁻¹³ vs. p = 4.4 × 10⁻⁹) with d_C = 1.3 and d_C = 1.1. It has been indicated that the statistical results for CIPC with the both drugs were also the most sensitive with the achieved power 1 − β = 1.00 and 1 − β = 0.99.

Discussion

The autopsy of the rats from the control group revealed carcinomatosis of visceral and parietal peritoneum, greater and lesser omentum, mesentery, metastatic affection of lymph nodes in the abdominal cavity (Figures 2 and 3). The biological pattern of the transplanted ovarian cancer in rats was similar to stages III and IV of ovarian cancer in humans, including redistribution phenomenon of the tumour growth described in progression of peritoneal carcinomatosis. Thus, we believe that our ovarian cancer model in rats can be regarded as a biological equivalent of optimally debulked ovarian cancer in women [14]. In our research we used antitumor agent dioxadet as a promising alternative for various IP routes. Phase II clinical trials of dioxadet have indicated its high antitumor efficacy in malignant tumours of various localisations, however, the most significant effect has been achieved in patients with metastatic ovarian cancer by combining IV and IP treatment modes [11]. The pronounced therapeutic efficacy of the drug is due to intramolecular crosslinking of the tumour cell DNA regardless of the cell cycle; low molecular weight, amphiphilic properties and slow clearance from the peritoneum [12]. According to our data, the largest efficacy of dioxadet could be achieved in HIPEC route. While temperatures above 40 °C only minimally impact the intact cells, in tumour cells it contribute to the DNA replication disorders, changes in the cell membrane structure and denaturation of oncoproteins. On tissue level, changes in blood flow, oxygenation, as well as in metabolic and energetic profile are observed. Chemoperfusion with heated cytostatic solution increases the drug uptake in tumours and hence, cytotoxicity [25]. However, the use of HIPEC is only adequate for patients undergoing cytoreductive surgery and carries all surgery-associated risks, not to mention technical difficulties and toxicological side effects (weight loss, haemorrhages, diarrhoea, gastroptosis) [5,13]. On the other hand, CIPC could be implemented at any stage of the treatment. Though less effective than NIPEC and HIPEC, the magnitude of survival benefit from CIPC is still much favourable compared to IV chemotherapeutic regimes as shown in our previous report [12].

In our study, rats treated by CIPEC with both cisplatin and dioxadet had a 66–85% increase in ALE (10 and 13 days). HIPEC with dioxadet had the highest survival compared to the control without treatment (31 days; p = 1.6 × 10⁻¹⁴), whereas no statistical significance was confirmed compared to the concomitant control (HIPEC with physiological saline). This could be explained by the high efficacy of hyperthermia itself with physiological solution (14 days; p = 3.6 × 10⁻⁵). Statistically, NIPEC with cisplatin was the most effective drug-to-route combination (HR = _0.20.2_1.4). To date, cisplatin is the most commonly used drug both in IV and IP chemotherapy [26]. When administered IP, its concentration in the blood corresponds to that of IV chemotherapy. As a result, most of the activated platinum localised on tumour implants enters the tissue via bloodstream, regardless of IV or IP mode of administration, while preserving the same or even higher toxicity [8]. This could mean, that despite theoretically effective combination of cisplatin with hyperthermia, its toxicity hazards the outcome of survival in rats with ascitic ovarian cancer [27]. Perhaps, toxicity of cisplatin during NIPEC might be much tolerable as indicated by our results. We believe that dioxadet could be a better substitute for the current cisplatin-paradigm in the treatment of advanced EOCs.

Conclusions

Our data correspond with the review of the literature and suggest that treatment outcome of peritoneal carcinomatosis in patients with advanced EOCs could be significantly improved by the means of IP chemotherapy. In our opinion, an integrated approach to the treatment of these patients is necessary. Cytoreductive surgery should be supplemented with intraoperative HIPEC and followed by adjuvant IP chemotherapy instead of adjuvant IV treatment courses. Dioxadet has some advantages in IP chemotherapy against the most commonly used today cisplatin, which indicates the prospects of dioxadet for future introduction into clinical practice as a primary drug for HIPEC and CIPC in patients with advanced and recurrent ovarian cancer. We also encourage further research to evaluate the role of cisplatin in NIPEC route.

Acknowledgements

The authors are grateful to Mrs. A.P. Sapozhnikova (N.N. Petrov Research Institute of Oncology) for technical assistance in preparing the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was partially financially supported by Government of Russian Federation [Grant 074-U01].