Current status and future prospects of clinical trials on CRS + HIPEC for gastric cancer peritoneal metastases

Abstract

Purpose: There is no standard treatment for peritoneal metastases (PM) from gastric cancer (GC). The aim of this review is to evaluate the clinical trials on cytoreductive surgery (CRS) plus hyperthermic intraperitoneal chemotherapy (HIPEC) for GC PM.

Materials and methods: The published clinical trials on CRS + HIPEC for GC PM are critically evaluated, and survival and safety are the primary endpoints. In addition, the registered ongoing clinical trials are summarised.

Results: The natural course of GC PM is <5 months. CRS + HIPEC could improve the overall survival (OS). In prospective studies, the median OS was 11.0 months in the CRS + HIPEC group vs. 5.4 months in the CRS alone group. In case-control studies, the median OS was 13.3 months in the CRS + HIPEC group vs. 7.9 months in the CRS alone group. In cohort studies, the median OS after CRS + HIPEC was 13.3. The median 1-, 2- and 5-year survival rates after CRS + HIPEC were 50.0%, 35.8% and 13.0%, respectively. There is no statistically significant increase in serious adverse events that are directly attributed to CRS + HIPEC.

Conclusions: The combination of CRS and HIPEC is a promising integrated treatment strategy for GC PM that has encouraging initial results, calling for urgent further evaluation of this strategy in randomised control trials (RCTs).

Keywords:

• Clinical trials
• gastric cancer
• cytoreductive surgery
• hyperthermic intraperitoneal chemotherapy

Introduction

Gastric cancer (GC) accounts for 9% of all cancer-related deaths [1], and China accounts for over 40% of all new cases worldwide [1,2]. Approximately 30% of GC patients have regional spread at diagnosis, and the locoregional progression of GC generally results in peritoneal metastases (PM), which are characterised by the presence of tumour nodules of various sizes, numbers, and distributions on the peritoneal surface [3,4]. PM have a significant negative impact on the overall survival (OS) and quality of life (QoL) as a result of refractory ascites, progressive intestinal obstruction and uncontrollable abdominal pain. The traditional therapies for such patients are systemic chemotherapy, palliative surgery and the best supportive care. The prognosis for GC PM is poor because of its recurrent patterns, and it has a median survival of approximately half a year [5].

However, over the past three decades, significant improvements in the diagnosis and treatment of PM have been achieved as a result of the in-depth understanding of the fundamental mechanisms of PM and treatment technology improvements. It is now widely accepted in the oncology community that at least some PM are due to regional and not terminal disease, and proactive local-regional therapy that combines cytoreductive surgery (CRS) and hyperthermic intraperitoneal chemotherapy (HIPEC) has the advantages of surgical removal of a visible bulky tumour burden and regional hyperthermic chemotherapy to eradicate micrometastases and invisible free tumour cells [6].

Focussing on GC PM, this work aims to review the milestone developments from perspectives of published and ongoing clinical trials. The former depict the real-life natural course of GC PM, and the latter is critically evaluating the efficacy and safety issues of CRS + HIPEC. Finally, the registered clinical trials involving GC PM are also summarised to inspire future developments.

Materials and methods

Literature search strategy

A comprehensive search of electronic databases was performed using Medline, PubMed, Web of Science and Cochrane library with the search terms “gastric cancer”, “hyperthermic intraperitoneal chemotherapy” and “HIPEC”, from January 1, 1980 to August 1, 2016. The literature search also included all Mesh Terms, as spontaneously matched by PubMed. Review articles were also obtained to determine other relevant studies. Only the latest version was considered for duplicate published trials with accumulating numbers of patients or increased lengths of follow-up. The reference lists of all retrieved articles were examined for further identification of potentially relevant studies. A comprehensive search of registered clinical trials was performed in clinicaltrials.gov with the search terms “gastric” and “peritoneal”.

Selection criteria

All articles were assessed in two rounds by two independent experienced reviewers (Y. Li and Z.H. Ji). The first round of the literature search focussed on reading the titles, abstracts and keywords in published review articles, basic research articles, case reports and clinical trials, and any paper that did not include GC was excluded. Papers on prospective or retrospective clinical trials about CRS + HIPEC in treating GC PM were selected for the second round of assessment. In the second assessment, the full texts were reviewed, and the articles included in this review had over 10 cases reporting on the median and/or mean OS or 1-, 2-, 3- and 5-year survival rates, with or without morbidity and mortality information (Figure 1).

Figure 1. Flow chart of the search process and results. A total of 335 relevant papers were selected from all the published literature; 109 duplicated papers were removed, 226 papers underwent the first round of assessment and 98 papers underwent the second round of assessment. Only 29 qualified papers were identified for inclusion in the final evaluation.

Data extraction and management

Two reviewers independently extracted data, including the publication time, type of design (randomised control, prospective, retrospective case-control or retrospective cohort), country, number of patients, treatment arms, HIPEC characteristics, survival outcomes (median OS; 1-, 2-, 3- and 5-year survival rates; and other survival-related parameters), mortality and morbidity and multivariate analysis results.

Statistical analysis

Descriptive statistics (simple counts, means and medians) were used to summarise the extracted data. The management of data and construction of tables was performed with Excel 2010 (Microsoft Corp., Redmond, WA, USA), and the drawing of figures was performed with Graph Pad Prime 6 (GraphPad Software, Inc. La Jolla, CA, USA).

Results

Importance of PM in GC patients

Several large-sample clinical studies across major regions of the world demonstrated that GC PM is a long-standing but ignored clinical problem (Table 1) [7–11]. These studies at different time periods and through different geographic regions have provided convincing evidence that PM is the most frequent and aggressive type of GC metastasis, and it remains the most important challenge in GC treatment.

Table 1. Information from large-sample clinical trials on postoperative recurrence of gastric cancer.

Ref.	Country	Study type	No. of pts	Patient features	Median follow-up (months)	Recurrence
						All forms	PC, N (%)
Roviello et al. [7]	Italy	Prospective	441	152 (34%) D1, and 289 (66%) D2–D3 lymphadenectomy	48 (2–140)	215	77 (35.8)
Sakuramoto et al. [8]	Japan	RCT	1059	998 (94.2%) R0 gastrectomy D2 lymphadenectomy	2.9	321	143 (44.5)
Lee et al. [11]	Korea	Retrospective case-control	1874	816 open gastrectomy, 1058 laparoscopic gastrectomy	50.3 (1.8–99.4)	49 laparoscopic group; 179 open group	23 (46.9) laparoscopic group; 113 (63.1) open group
Yu et al. [9]	Korea	RCT	458	Curative gastrectomy with at least D2 lymphadenectomy with adjuvant chemotherapy only or chemotherapy and radiotherapy	58.0 (4.5–100.6) or 59.3 (1.8–90.7)	133	74 (55.6)
Cao et al. [10]	US, China	Retrospective case-control	1058	414 from US, 644 from China, all with R0 resection	81 (2–226) in US; 63 (10–215) in China	76	15 (16)

PC: peritoneal carcinomatosis; RCT: randomise controlled trial.

The natural course of GC PM

Two large-sample studies, one institution-based [12] and one population-based [5], have provided convincing data on the natural history of GC PM, from two different time periods and geographic areas. In the institution-based study, performed in France in 2000 [12], the natural course of 125 patients with GC PM was first systematically described. The mean and median OS values were 6.5 months and 3.1 months. In the population-based study, conducted in the Netherlands in 2014 [5], the median OS values were 4.6 months for patients with PM as the only metastatic site and 3.3 months for those with PM combined with other metastases. Taken together, these studies firmly established that the natural course of GC PM is less than five months.

CRS + HIPEC trials for GC PM

In terms of clinical studies on CRS + HIPEC to treat GC PM, retrospective, cohort and prospective studies have been conducted throughout the world, demonstrating improved clinical efficacy and acceptable safety profiles. At the time of this review, a total of 1863 patients with GC PM were included in CRS + HIPEC studies, including 1659 patients in retrospective studies and 204 patients in prospective studies. The major features of these studies are summarised in Table 2 [13–41].

Table 2. Major characteristics of CRS + HIPEC studies for peritoneal carcinomatosis from gastric cancer.

Type of study and references	Country	Number of points	Treatment arms (Number of points)	HIPEC regimen
				Model	Drug and dose	Temperature (°C)	Duration (h)	Flow rate (ml/min)
Prospective studies
Rudloff et al. [15]	USA	17	CRS + HIPEC + FOLFOXRI (9) vs. FOLFOXRI alone (8)	Closed	L-OHP 460 mg/m^2	41	30	2000
Yang et al. [13]	China	68	CRS + HIPEC (34) vs. CRS alone (34)	Open	CDDP 120 mg + MMC 30 mg	43	60–90	500
Yang et al. [16]	China	28	CRS + HIPEC	Open	HCPT 20 mg or CDDP 120 mg + MMC 30 mg	43	90–120	200
Glehen et al. [17]	France	49	CRS + HIPEC	Closed	MMC 40–60 mg	40–43	90	500
Beaujard et al. [18]	France	42	CRS + HIPEC	Closed	MMC 40–60 mg	45 (41–49)	90	400–500
Retrospective studies
Wu et al. [19]	China	50	CRS + HIPEC	Open	LOB 50 mg/m^2 + DOC 60 mg/m^2	43	60	400
Tu et al. [20]	China	231	CRS + HIPEC	Closed	5-FU 1500 mg + CDDP 100 mg	43	60	600
Boerner et al. [21]	Germany	103	CRS + HIPEC (38) vs. CRS + SC (27) vs. Palliative treatment (31)	Closed	CDDP 75 mg/m^2 + DOX 15 mg/m^2	42–43	60	NR
Passot et al. [22]	France	127	CRS + HIPEC	Closed	CDDP, MMC, L-OHP, DOX, CPT-11	NR	NR	500–700
Desantis et al. [23]	France	14	CRS + HIPEC	Open	CDDP 50 mg/m^2/L, DOX 15 mg/m^2/L	43	60	800
Yarema et al. [24]	Ukraine	40	CRS + HIPEC + SC (20) vs. SC alone (20)	NR	CDDP 75 mg/m^2 + MMC 12.5 mg/m^2	42.3 ± 1.3	90	NR
Magge et al. [25]	USA	23	CRS + HIPEC	Closed	MMC 40 mg	42	100	800
Muller et al. [26]	Greece	26	NAC + CRS + HIPEC	Open	L-OHP 200 mg/m^2 + DOC 80 mg/m^2	NR	NR	NR
Konigsrainer et al. [27]	Germany	18	NAC + CRS + HIPEC	Open	CDDP 50 mg/m^2	42	90	NR
Canbay et al. [28]	Japan	194	BIPSC + CRS + HIPEC	Open	DOC 30 mg/m^2	NR	NR	NR
Schildberg et al. [29]	Germany	76	NAC + CRS + HIPEC (16) vs. NAC (60)	NR	NR	NR	NR	NR
Kang et al. [30]	Taiwan China	47	CRS + HIPEC (16) vs. CRS alone (31)	Open	CDDP 90–120 mg + MMC 30–40 mg + VP-16 60-80 mg	41–43	60	NR
Hultman et al. [31]	Sweden	18	NAC + CRS + HIPEC + EPIC	Open	CDDP 50 mg/m^2 + DOX 15 mg/m^2/L-OHP 460 mg/m^2	42–44	90 or 30	NR
Glehen et al. [14]	France	159	CRS + HIPEC and/or EPIC	Open/Closed	CDDP 50–100 mg/m^2 ± MMC 30–50 mg/m^2/L-OHP 360–460 mg/m^2 ± CPT-11 100–200 mg/m^2	40–43	60–120 or 30	NR
Shen et al. [32]	USA	62	CRS + HIPEC	Closed	MMC 40 mg	40–42.5	120	1000
Scaringi et al. [33]	France	26	CRS + HIPEC	Open	CDDP 200 mg/m2 + MMC 120 mg	41–43	90–120	NR
Zhu et al. [34]	China	22	CRS + HIPEC (10) vs. CRS alone (12)	NR	CDDP 250-300 mg + MMC 25-30 mg	43.0 ± 1.0	60	NR
Yonemura et al. [35]	Japan	107	Conventional CRS + HIPEC (65) vs. peritonectomy + HIPEC (42)	NR	CDDP 300 mg + MMC 30 mg + VP-16 150 mg	42–43	NR	10,000
Hall et al. [36]	USA	74	CRS + HIPEC (34) vs. CRS alone (40)	Closed	MMC 40 mg	40–41	120	800–900
Fujimura et al. [37]	Japan	15	CRS + HIPEC	Open	CDDP 150–250 mg + MMC 60–100 mg + VP-16 120–200 mg	42–43	60	NR
Hirose et al. [38]	Japan	37	CRS + HIPEC (17) vs. CRS alone (20)	Open	CDDP 100 mg + MMC 20 mg + VP-16 100 mg	41.0–44.5	NR	400–600
Fujimoto et al. [41]	Japan	66	CRS + HIPEC (48) vs. CRS alone (18)	Closed	MMC 30–40 mg	44.5–45	120	NR
Yonemura et al. [39]	Japan	83	CRS + HIPEC	Open	CDDP 300 mg + MMC 30 mg + VP-16–150 mg	42–43	60	30 000
Yonemura et al. [40]	Japan	41	CRS + HIPEC	Open	CDDP 300 mg + MMC 50 mg	41–43	40–60	100–300

CRS: cytoreductive surgery; HIPEC: hyperthermic intraperitoneal chemotherapy; FOLFOXRI: oxaliplatin + irinotecan +5-fluorouracil; L-OHP: oxaliplatin; CDDP: cisplatin; MMC: mitomycin; HCPT: hydroxycamptothecine; LOB: lobaplatin; DOC: docetaxel; 5-FU: 5-Fluorouracil; SC: systemic chemotherapy; DOX: doxorubicin; CPT-11: irinotecan; VP-16: etoposide; NAC: neoadjuvant chemotherapy; EPIC: early post-operational intraperitoneal chemotherapy; BIPSC: bidirectional intraperitoneal and systemic induction chemotherapy; NR: not reported.

Technical analyses of these studies revealed several eminent features. In terms of the geographic distribution, 12 studies were from Europe (France 6, Germany 3, Sweden 1, Greece 1 and Ukraine 1), 13 from Asia (Japan 7, Mainland China 5 and Taiwan China 1), and 4 from America (US 4). In terms of the publication time, 15 studies were published during 2010–2016, 8 during 2000–2010, and 6 during 1990–2000. In terms of major CRS + HIPEC protocols, 14 studies adopted open HIPEC, 10 studies used closed HIPEC, 1 study used both open and closed techniques, and 4 studies did not specify the technical modality.

The most promising result is the survival outcome. Five prospective studies and 24 retrospective studies were available for survival analysis, including the median OS; 1-, 2-, 3- and 5-year survival rates, and other survival-related parameters (Table 3) [5,12–41]. There are slight differences in the median OS among different types of studies. For the five prospective studies, the median OS was 11.0 months (range 10.0–11.3 months) in the CRS + HIPEC group vs. 5.4 months (range 4.3–6.5 months) in the CRS alone group. For the eight retrospective case-control studies, the median OS was 13.3 months (range 8.0–24.5 months) in the CRS + HIPEC group vs. 7.9 months (range 5.0–11.0 months) in the CRS alone group. For the 16 retrospective cohort studies, the median OS after CRS + HIPEC was 13.3 (range 6.6–37.0 months). The median 1-, 2- and 5-year survival rates after CRS + HIPEC were 50.0% (range 41.2%–83.4%), 35.8% (range 19.9%–69.0%) and 13.0% (range 6.4%–31.0%), respectively.

Table 3. Survival outcomes of CRS + HIPEC treating peritoneal carcinomatosis from gastric cancer.

Type of study and references	Median OS (months)	1-year OS (%)	2-year OS (%)	3-year OS (%)	5-year OS (%)	Median OS by CC			Independent prognostic factors
						CC0 (months)	CC1 (months)	CC2–3 (months)
Population-based studies
Thomassen et al. [5]	4.6	NR	NR	NR	NR	NR	NR	NR	NR
Sadeghi et al. [12]	3.1	NR	NR	NR	NR	NR	NR	NR	NR
Prospective studies
Rudloff et al. [15]	11.3 vs. 4.3	44.4	33.3	NR	NR	NR	NR	NR	NR
Yang et al. [13]	11.0 vs. 6.5	41.2 vs. 29.4	14.7 vs. 5.9	5.9 vs. 0	NR	12.0 vs. 11.0^a	NR	8.2 vs. 4.0	CRS + HIPEC, syn-PC, CCR0-1, SC >6 cycles, no SAE
Yang et al. [16]	NR	50.0	42.8	NR	NR	43.4	9.4	8.3	PCI <20, CCR
Glehen et al. [17]	10.3	48.1	19.9	NR	16.0%	21.3^a	NR	6.6	Ascites, CCR
Beaujard et al. [18]	NR	48.0	33.0	NR	NR	NR	NR	NR	NR
Retrospective studies
Wu et al. [19]	14.3	58.0	40.0	32.0	NR	23.5^a	NR	8.0	NR
Tu et al. [20]	37	83.4	69.0	38.7	NR	NR	NR	NR	CCR0-1, TM (N), SC ≥6 cycles
Boerner et al. [21]	17.2 vs. 11.0 vs. 7.7	71.1 vs. 33.3 vs. 29.0	35.8 vs. 16.9 vs. 3.2	24.1 vs. 0 vs. 0	6.4 vs. 0 vs. 0	NR	NR	NR	TNM stage, ascites CCR
Passot et al. [22]	13.0	NR	NR	NR	14.0	NR	NR	NR	Age <56 year, HIPEC
Desantis et al. [23]	13.3	NR	NR	21.6	21.6	NR	NR	NR	NR
Yarema et al. [24]	12.0 vs. 8.0	68.8 vs. 25.0	NR	NR	NR	NR	NR	NR	Histological type, performance status operation time, >3 peritonectomies, previous SC, CCR
Magge et al. [25]	9.5	49.6	NR	17.9	NR	NR	NR	NR	PCI, CC
Muller et al. [26]	19.0	NR	38.0	NR	NR	NR	NR	NR	NR
Konigsrainer et al. [27]	8.9	NR	NR	NR	NR	NR	NR	NR	PCI >12^c
Canbay et al. [28]	15.8	66.0	32.0	NR	10.7	NR	NR	NR	NR
Schildberg et al. [29]	24.5 vs. 10.1	NR	NR	NR	NR	NR	NR	NR	PCI, CCR, pathological response to NIPS
Hultman et al. 2013[31]	10.3	NR	NR	NR	NR	11.9	NR	NR	NR
Kang et al. 2013[30]	14.6 vs. 8.0^d	NR	NR	NR	NR	NR	NR	NR	Male, number of anastomoses, number of positive lymph nodes, CCR
Glehen et al. [14]	9.2	43.0	NR	18.0	13.0	15.0	6.0	4.0	HIPEC, T stage
Shen et al. [32]	6.1	NR	NR	NR	NR	NR	NR	NR	Female, CC
Scaringi et al. [33]	6.6	NR	NR	NR	NR	15.0	NR	3.9	Performance status, CCR, complications, tumour histology.
Zhu et al. [34]	10.0 vs. 5.0	NR	NR	NR	NR	NR	NR	NR	NR
Yonemura et al. [35]	11.5	NR	NR	NR	6.7	19.2	NR	7.8^b	No independent factors
Hall et al. [36]	8.0 vs.7.8	27.0 vs. 41.0	23.0 vs. 28.0	NR	6.0 vs. 17.0	36.3	11.2	3.3	CCR, peritonectomy
Fujimura et al. [37]	15.4	57.0	21	NR	NR	20	NR	9.6	NR
Hirose et al. [38]	11.0 vs. 6.0	44.4 vs. 15.8	NR	NR	NR	NR	NR	NR	NR
Fujimoto et al. [41]	14.1 vs. 8.2	54.0 vs. 11.0	NR	41.5 vs. 0	31.0 vs. 0	NR	NR	NR	NR
Yonemura et al. [39]	NR	43.0	NR	NR	11.0	NR	NR	NR	NR
Yonemura et al. [40]	14.5	NR	NR	28.5	NR	NR	NR	NR	NR

^a CC0–1.

^b CC2–3.

^c Negative factor.

^d Mean OS.

CRS: cytoreductive surgery; HIPEC: hyperthermic intraperitoneal chemotherapy; Syn-PC: synchronous PC; CCR: completeness of cytoreduction; SC: systemic chemotherapy; SAE: serious adverse event; PCI: peritoneal carcinomatosis index; NIPS: neoadjuvant intraperitoneal and systemic chemotherapy; NR: not reported.

Prognostic factors have also been explored in these studies. Multivariate analyses were conducted in 16 studies to identify independent prognostic factors (Table 3). In terms of the frequency, the following parameters were identified as independent prognostic factors: the completeness of cytoreduction (CCR) in 12 studies; peritoneal cancer index (PCI) in four studies; HIPEC, systemic chemotherapy and TNM stage each in three studies; and ascites, tumour histopathology, performance status, peritonectomy and postoperative complications each in two studies.

There is adequate evidence to judge the safety profile of CRS + HIPEC for treating GC PM (Table 4) [13,15–20,25,27–29,31,35–39]. In terms of adverse events, the most frequent adverse events include pleural effusion, ileus, sepsis, wound infection, occlusion, hypoalbuminemia and anastomosis leakage.

Table 4. Mortality, morbidity, and serious adverse events (SAEs) of CRS + HIPEC.

Type of study and references	Number of points	Mortality (%)	Morbidity (%)	SAEs (Grade III/IV)
Yonemura et al. [39]	83	1.2	7.2	Bowel perforation 3 points, bone marrow suppression 2 points, renal dysfunction 1 point
Hirose et al. [38]	37	3.1	35.2	Renal dysfunction 5 points, anastomotic leakage 1 point, pancreatic fistula 1 point, adhesive ileus 1 point
Beaujard et al. [18]	42	3.6	9.6	Ileus 2 points, peritonitis 2 points
Fujimura et al. [37]	15	0	50.0	Intra-abdominal bleeding 5 points, intestinal perforation 2 points, intra-abdominal abscesses 2 points
Glehen et al. [17]	49	4	27.0	Occlusion 2 points, ileocolic fistula 2 points, pancreatic fistula, biliary peritonitis, peritonitis and intraperitoneal abscess 1 point
Hall et al. [36]	74	0	35.0	Sepsis 5 points, anastomotic leak 4 points, wound infection 3 points, leukopoenia 3 points, myocardial infarction 1 point, pulmonary embolism 1 point, atrial fibrillation 1 point
Yonemura et al. [35]	107	2.8	21.5	Anastomotic leakage 6 points, small bowel perforations 2 points, intra-abdominal abscesses 5 points, pancreatic fistula 1 point, chylous ascites 1 point, wound abscess 1 point, bleeding 1 point
Yang et al. [16]	28	0	14.3	Ileus 2 points, respiratory failure 1 point, bile leakage with sepsis, generalised peritonitis and abscess 1 point
Glehen et al. 2010 [14]	159	6.5	27.8	Digestive fistula 25 points, multiorgan failure 2 points, septic shock 2 points, respiratory complications 2 points, cardiopulmonary embolism 1 point, cardiac arrhythmia 1 point, haematologic toxicity 1 point
Yang et al. 2011 [13]	68	0	14.7	Wound infection and sepsis 2 points, respiratory failure 1 point, gastrointestinal bleeding 1 point, intestinal obstruction 1 point
Hultman et al. [31]	18	5.6	27.8	Anastomotic leakage 3 points, thromboembolism 3 points, intracranial haemorrhage 2 points, bowel perforation 2 points, intra-abdominal abscess 2 points
Konigsrainer et al. [27]	18	0	46.0	Pleural effusion 1 point, wound infection 1 point, sepsis 1 pt
Canbay et al. [28]	194	3.9	23.6	Intestinal obstruction 10 points, wound infection sepsis 9 points, renal failure 6 points, anastomotic leakages 6 points, intra-abdominal bleeding 1 point
Schildberg et al. [29]	76	6.3	31.0	Pneumonia 2 points, anastomosis stenosis 1 point
Magge et al. [25]	23	4.3	52.2	Sepsis 5 points, ileus 4 points, anastomotic leak 3 points, pancreatic leak 2 points, enterocutaneous fistula 1 point, bleeding 1 point
Rudloff et al. [15]	17	0	11.0	Pleural effusion 3 points, wound infection 2 points, anastomotic leak 2 points, endoscopic oesophageal perforation 1 point
Wu et al. [19]	50	0	23.1	Hypoalbuminemia 4 points, intestinal obstruction 3 points, septicaemia 2 points, intestinal leakage 1 point, diarrhoea 1 point, vomit 1 points
Tu et al. [20]	231	0.9	6.9	Multiple organ failure 2 points, hypoalbuminemia 1 points, electrolyte disturbance 1 point

Although the published survival data are promising, the results should be interpreted with caution because of the low quality of these studies. Among five prospective studies included in this review, only one randomised control trial (RCT) with sample size of 68 patients compared CRS + HIPEC vs. CRS alone, one RCT with sample size of 17 patients compared CRS + HIPEC + postoperative systemic chemotherapy vs. systemic chemotherapy alone, and the other three studies were prospective cohort studies with a total sample size of 119 patients. Of 24 retrospective studies, 16 studies were retrospective cohort studies or cases reports with a sample size ranging from 14 to 231 patients. Small sample size, lack of a control group and retrospective design are the major causes of a low quality of evidence. Therefore, the results from these studies are promising and should be the basis for further high quality research.

Critical appraisal of the ongoing clinical trials in GC PM

We have identified 101 registered clinical trials in clinicaltrials.gov; 28 are ongoing clinical trials focussing on GC PM. Of the 28 ongoing clinical trials, 5 are on the diagnosis or evaluation of GC PM, 15 are on the treatment for GC PM, and 8 are on prophylaxis for GC PM (Supplemental Table 1).

Among five diagnostic trials, three used peritoneal lavage as the study object, and the major objective was to try to identify tumour cells or effective tumour makers to diagnose and evaluate PM.

The 15 therapeutic trials focussed on intraperitoneal (IP) administration of cytotoxic or immunologic drugs, including nine using HIPEC with/without IP and intravenous (IV) chemotherapy, five using IP with/without IV chemotherapy, and one using pressurised IP air-flow chemotherapy. Hyperthermic or normothermic IP administration of anti-tumour drugs combined with conventional systemic chemotherapy is the mainstream model of adjuvant chemotherapy for GC PM. However, only six trials combined CRS and HIPEC with/without IP or IV adjuvant chemotherapy for a curative intent, which means most oncologists still regard GC PM as a terminal incurable condition rather than a locoregional curable metastasis such as liver metastasis from colorectal cancer.

In eight prophylactic trials, three different strategies were used to prevent PM in GC patients, including extensive peritoneal lavage in three trials, HIPEC in three trials and postoperative IP chemotherapy in two trials.

In terms of the study design, 9 of 27 trials had a randomised parallel assignment, and the remaining 18 had non-randomised single group assignment. There were only two multi-centre RCTs, the GASTRIPEC and GASTRICHIP studies.

The GASTRIPEC study is a multi-centre German trial on GC PM patients with the objective of comparing neoadjuvant chemotherapy + CRS + HIPEC + postoperative chemotherapy vs. neoadjuvant chemotherapy + CRS + postoperative chemotherapy. This trial started on May 4, 2014, with established enrolment of 180, and it is still recruiting participants.

The GASTRICHIP study is a multi-centre French trial on advanced GC patients with a high risk for PM, and the objectives are to compare the five-year OS rates and recurrence-free or locoregional-free survival rates in GC patients treated with either curative gastrectomy + adjuvant HIPEC or with curative gastrectomy alone. This study started at May 2013 with an established enrolment of 322, and it is still recruiting participants and opening new centres at present. In 2016, Oliver Glehen reported their partial results at the PSOGI meeting in Washington DC. The first safety evaluation in November 2015 of 50 inclusions demonstrated that there was no difference in the mortality and morbidity between the Gastrectomy + HIPEC and Gastrectomy alone groups. The second safety evaluation in October 2016 of 120 inclusions found that one patient died within 60 days after surgery in the Gastrectomy + HIPEC group and no patients died in the Gastrectomy alone group. The severe morbidity was 28.4% in the Gastrectomy + HIPEC group and 26.2% in the Gastrectomy alone group.

Discussion

In the field of understanding and treating GC PM, what successes have we gained and what questions remain unanswered? From the comprehensive review of the current literature, several facts are evident.

First, there has been considerably increasing awareness of the importance of GC PM. The fundamental data on natural course of the disease obtained from France [12] and the Netherlands [5] demonstrated that the prognosis of GC PM is even worse than conventionally expected; it has an OS of less than five months. Moreover, PM are the most frequent and aggressive sites of metastasis in GC. It has been repeatedly demonstrated that the currently adopted conventional treatments could not improve the OS for such patients. These studies repeatedly imply that, at least for GC treatment, there are fundamental flaws in the concept, design and techniques in the currently universally adopted standardised curative gastrectomy with D2 lymphadenectomy. Under the current treatment paradigm, due attention has been paid to local control by superb resection techniques, and adequate attention has also been paid to systematic controls by improving systemic chemotherapeutic regimens, but far less attention has been paid to regional controls.

Therefore, there should be concept changes in the theory of GC metastasis. After curative resection, GC metastasis seems to follow a pattern of stepwise progression in that peritoneal metastasis is the first, key dissemination site and that other forms of GC spread occur later. Therefore, GC should not be regarded as a disease with early systemic metastasis. Instead, GC should be considered a disease with early intra-abdominal spread that is mostly in the form of PM, which is associated over time with a steadily increasing incidence of distant spreading [42]. Based on these in-depth understandings of GC metastasis, more attention should be paid to GC PM in terms of both basic research and clinical practice.

Second, CRS + HIPEC offer promising benefits for treating GC PM. As an integrated treatment strategy using the combination of surgical resection, cytotoxic chemotherapy, hyperthermic tumour ablation and hydrodynamic washing [43,44], CRS + HIPEC is the only possible treatment strategy that offers concrete clinical benefits for patients with GC PM, and it has attracted increasing and expanding attention from the oncology community across the world, particularly over the past decade. Although the published trials have not yielded unequivocal conclusions, evidence has been obtained to support the idea that CRS + HIPEC is promising for GC PM. After 30 years of exploration and development, adequate evidence is available showing that CRS + HIPEC could improve the median OS up to approximately 13 months, double the median OS associated with conventional therapy. Moreover, CRS + HIPEC could offer a survival benefit for selected GC PM patients with acceptable safety profiles.

Third, to maximise the efficacy and minimise the adverse events of CRS + HIPEC, careful patient selection is critical. Well-planned, structured pre-operative examinations are essential, including a performance status evaluation, serum levels of tumour makers, and medical imaging assessments, as reported in a recent expert consensus statement [6]. PCI and CCR, which were first described by Sugarbaker et al. [45,46], are two key prognostic factors for a better OS, and they are considered as inclusion criteria written into expert consensus of different countries [6] and large specialised PM institutions [22,47,48]. Therefore, a whole set of screening approaches is important, including routine physical examination, performance status evaluation, blood tests of the nutrition status, major organ function parameters, tumour markers, and specialised medical imaging studies [6]. Although efforts have been made to better define resectability before implementing surgical exploration, surgical exploration, preferably laparoscopic surgery, provides accurate evaluation of the resectability [22].

Fourth, the optimal technical approaches to HIPEC remains an unanswered question. In terms of the technical aspect of CRS + HIPEC, there is no evidence supporting the choice of open vs. closed HIPEC. Both techniques are acceptable in the current clinical practice. Similarly, in terms of drugs used in HIPEC, both the traditional mitomycin C- and cisplatin-based regimens as well as platinum- or docetaxel-based drugs are acceptable [49]. The selection of cytotoxic drugs used for HIPEC is mainly based on their depth of infiltration and their cytotoxic effects, but it is not strictly related to the pathology of the primary tumour [7,48,50]. There is still no standard protocol for the HIPEC duration, flow rate, temperature, drug selection or technique. Additionally, there is no control study evaluating these HIPEC-related parameters. In the present review, we found that an open HIPEC using mitomycin C- and cisplatin-based drug regimens for 60–90 min at 40–43 °C with a flow rate of 500 ml/min was the most frequently used approach.

Fifth, important issues remain unsettled in the field of CRS + HIPEC for GC PM in terms of clinical trials. The most important issue is in additional multi-centre trials and broader international collaborations. Multi-centre collaborations for RCTs are required to generate higher levels of evidence. Currently, most trials are retrospective or prospective single-centre trials. Furthermore, at present, there is only a single-centre RCT conducted by Yang et al. [13] that was conducted in China, and two registered multi-centre RCTs (GASTRIPEC and GASTRICHIP) are ongoing in Europe. Many countries have developed specialised peritoneal cancer centres and have accumulated sufficient surgical experience. It is now time for international collaborations in multi-centre RCTs, and the field is well prepared for developing guidelines to use CRS + HIPEC as part of GC PM treatment. In addition, standardised protocols are required for widespread application of the technique.

In conclusion, the combination of CRS and HIPEC is a promising integrated treatment strategy for GC PM that has encouraging initial results and calls for urgent further evaluation of this strategy in RCTs.

Acknowledgements

This work was supported by the Key Discipline Development Fund of Beijing Shijitan Hospital, Capital Medical University (2016fmzlwk), and the Special Fund for the Capital Characteristic Clinical Medicine Development Project (Z161100000516077) (both to Yan Li).

Disclosure statement

The authors declare that there are no conflicts of interest to report for this work.

Additional information

Funding

The Key Discipline Development Fund of Beijing Shijitan Hospital, Capital Medical University (2016fmzlwk), and the Special Fund for the Capital Characteristic Clinical Medicine Development Project (Z161100000516077) (both to Yan Li).