Available online at www.sciencedirect.com

ScienceDirect EJSO 40 (2014) 12e26

www.ejso.com

Review

Intraperitoneal chemotherapy in advanced gastric cancer. Meta-analysis of randomized trials F. Coccolini a,b,*, E. Cotte b, O. Glehen b, M. Lotti a, E. Poiasina a, F. Catena c, Y. Yonemura d, L. Ansaloni a a

b

General Surgery Dept., Papa Giovanni XXIII Hospital, Bergamo, Italy General Surgery Dept., Centre Hospitalier Lyon Sud, Hospices Civils de Lyon and EMR 3738, Universite Lyon 1, France c General Surgery Dept., Ospedale Maggiore, Parma, Italy d General Surgery Dept., Kusatsu General Hospital, Yabase 1660, Japan Accepted 23 October 2013 Available online 5 November 2013

Abstract Introduction: An important component of treatment failure in gastric cancer (GC) is cancer dissemination within the peritoneal cavity and nodal metastasis. Intraperitoneal chemotherapy (IPC) is considered to give a fundamental contribute in treating advanced GC. The purpose of the study is to investigate the effects of IPC in patients with advanced GC. Material and methods: A systematic review with meta-analysis of randomized controlled trials (RCTs) of IPC þ surgery vs. control in patients with advanced GC was performed. Results: Twenty prospective RCTs have been included (2145 patients: 1152 into surgery þ IPC arm and 993 into control arm). Surgery þ IPC improves: 1, 2 and 3-year mortality (OR ¼ 0.31, 0.27, 0.29 respectively), 2 and 3-year mortality in patients with locoregional nodal metastasis (OR ¼ 0.28, 0.16 respectively), 1 and 2-year mortality rate in patients with serosal infiltration (OR ¼ 0.33, 0.27 respectively). Morbidity rate was increased by surgery þ IPC (OR ¼ 1.82). The overall recurrence and the peritoneal recurrence rates were improved by surgery þ IPC (OR ¼ 0.46 and 0.47 respectively). There was no statistically significant difference in lymph-nodal recurrence rate. The rate of haematogenous metastasis was improved by surgery þ IPC (OR ¼ 0.63). Conclusions: 1, 2 and 3-year overall survival is incremented by the IPC. No differences have been found at 5-year in overall survival rate. 2 and 3-year mortality rates in patients with nodal invasion and 1 and 2-year mortality rates in patients with serosal infiltration are improved by the use of IPC. IPC has positive effect on peritoneal recurrence and distant metastasis. Morbidity rate is incremented by IPC. Locoregional lymph-nodes invasion in patients affected by advanced gastric cancer is not a contraindication to IPC. Ó 2013 Elsevier Ltd. All rights reserved. Keywords: Gastric cancer; Intraperitoneal chemotherapy; HIPEC; Carcinosis; Nodal infiltration; Meta-analysis

Introduction Gastric cancer (GC) is the second leading cause of cancer death and the fourth most common cancer in the world.1,2 GC disseminates principally through the haematic torrent or through the peritoneal fluids. It has been * Corresponding author. General Surgery Dept., Papa Giovanni XXIII Hospital, Piazza OMS 1, 24127 Bergamo, Italy. Tel.: þ39 (0) 352673486; fax: þ39 (0)352674963. E-mail addresses: [email protected], [email protected] (F. Coccolini). 0748-7983/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejso.2013.10.019

demonstrated as peritoneal dissemination is more frequent than haematogenous metastases. The 40% of patients died for GC have hepatic metastases, while the 53e60% showed a disease progression and died with peritoneal carcinosis (PC). The two most important factors affecting prognosis in GC are the serosal invasion and the lymphatic spread.3e5 When gastric serosa is infiltrated, PC could be considered practically unavoidable.6 As a consequence, up to half of the patients with advanced GC will develop PC in spite even radical surgery.8e11 PC is already present in 5e20% of patients explored for potentially curative resection also

F. Coccolini et al. / EJSO 40 (2014) 12e26

in early gastric cancer.6,7 Surgical resection associated to systemic chemotherapy is the mainstay of treatment. Systemic chemotherapy improves median survival in advanced and/or metastatic GC to not more than 12 months.12e15 The same gain in term of survival has not been described with macroscopic PC16e19 due to the inadequate diffusion of systemic chemotherapy into the abdominal cavity.20 As many patients present with advanced-stage-disease, the research for more effective treatments is mandatory. An important component of treatment failure is cancer dissemination within the peritoneal cavity and nodal metastasis. In contrast to lymphatic and haematogenous dissemination, peritoneal spread should be regarded as a locoregional disease extension rather than systemic metastasis.21 Taking the natural history of GC into account, the use of intraperitoneal chemotherapy (IPC) as a targeted adjuvant treatment after surgery may be considered a rational prophylactic/therapeutic approach. Actually IPC has been progressively more used in advanced GC due to the appealing theoretical rationales. Although many different regimens of IPC exist, they all could be considered as the different applications of the same treatment method. IPC allows to reach an high intraperitoneal drug concentration and allows the drugs to directly act on the free-tumour-cells and peritoneal nodules. Drugs absorbed through the peritoneum enter the portal vein, and also have a chemotherapeutic effect on the liver.22 It remains controversial if the IPC has a positive effect on the lymph-node metastasis. The purpose of the present metaanalysis of randomized controlled trials (RCT) investigating the effects of IPC in patients with advanced GC.

13

Material and method

which patients with advanced GC (with or without PC) were randomly assigned to receive either surgery combined with IPC or surgery without IPC. All included patients must have histologically-proven gastric or gastro-oesophagealjunction adenocarcinoma and underwent potentially curative resection. Included studies consider both patients with locally advanced GC with macroscopic serosal invasion, and patients with peritoneal carcinomatosis but without distant metastasis. All forms of IPC in addition to surgery were included. No language restrictions have been applied. Eligibility for study inclusion into the metaanalysis and study quality assessment were performed independently by two authors (EC, FeCo). Study data were extracted onto standard forms independently by two authors (EC, FeCo). Discrepancies between the two investigators were resolved by discussion and evaluation of the question with a senior investigator. The final results were reviewed by three senior investigators (YY, LA, OG). The primary outcome measures for the meta-analysis were the impact of IPC on 1, 2, 3 and 5-year mortality and the effect of IPC on the mortality of patients with loco-regional nodal metastasis. For this last outcome were included data from studies in which at least the 80% of patients had loco-regional nodal metastasis at the time of intervention. The same criteria has been adopted to evaluate the impact of IPC on survival in patients with serosal invasion (secondary outcome) (including studies with at least the 80% of patients with serosal infiltration at the time of diagnosis). To evaluate the mortality rates studies were divided into two sub-groups: studies that included patients with and without PC. Moreover the impact of IPC on recurrence (overall, peritoneal, lymph-nodal) and haematogenous metastasis and morbidity, has been evaluated.

Literature search strategy

Assessment of risk of bias

Electronic searches were performed using Medline, Embase (1988eDecember 2012), PubMed (January 1980eDecember 2012), Cochrane Central Register of Controlled Trials (CCTR), Cochrane Database of Systematic Reviews (CDSR) and CINAHL from (1966e2012). The search terms were: ‘intraperitoneal chemotherapy’, ‘stomach’, ‘gastric cancer’, ‘carcinosis’, ‘randomized trial’, ‘meta-analysis’ combined with AND/OR. Research included also all the MeshTerms. No search restrictions were imposed. The reference lists of all retrieved articles were reviewed for further identification of potentially relevant studies. Review articles were also obtained to determine other possible studies. Duplicate published trials with accumulating numbers of patients or increased lengths of follow-up, were considered only in the last or at least in the more complete version.

There is a potential risk of overestimating the beneficial treatment effects of RCT with a resultant risk of bias. The risk of bias was assessed comprehensively according to guidelines of the Cochrane Collaboration23,24 and six items have been considered relevant (Table 1): 1) whether the method of allocation was truly random; 2) whether there was proper allocation concealment; 3) whether the groups were similar at baseline; 4) whether the eligibility criteria were documented; 5) whether loss to follow-up in each treatment arm was specified; 6) whether intention-to-treat analysis was conducted. Therefore the evaluation of the quality level of the study was conducted as follows: Positive answer to at least six questions was required for a trial to be rated as high quality. With a positive answer to five or four questions the study was considered of fair quality. With a positive answer to three or fewer questions the study was registered as low quality. When studies did not report adequate information to determine the above-mentioned assessment criteria, an attempt to obtain direct additional data from the investigators was made. Studies reported in

Selection criteria Studies which have been judged eligible for this systematic review and consequent meta-analysis are those in

14

F. Coccolini et al. / EJSO 40 (2014) 12e26

Table 1 Bias risk and quality assessment. Study (ref.) year

Randomization

Allocation concealment

Homogeneous baseline characteristic

Eligibility criteria

Loss to follow-up and drop-out described

Intention-to-treat analysis

Study quality

Koga11 1987 Hagiwara26 1992 Hamazoe9 1993 Fujimura28 1994 Sautner27 1994 Ikeguchi10 1995 Takahashi29 1995 Fujimoto30 1999 Rosen31 1998 Yu32 1998 Shimoyama33 1999 Tan41 2000 Yonemura34 2001 Wei35 2004 Zuo40 2004 Ding38 2007 Kuramoto7 2009 Deng39 2009 Yang36 2011 Miyashiro37 2011

YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES

Unclear YES Unclear Unclear Unclear Unclear YES Unclear YES Unclear Unclear Unclear Unclear Unclear Unclear Unclear YES Unclear Unclear YES

YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES

YES YES YES YES YES YES YES YES YES YES YES YES YES YES Unclear Unclear YES YES YES YES

YES YES YES YES YES YES YES YES YES YES YES YES YES YES Unclear YES YES Unclear YES YES

NO NO Unclear Unclear YES NO Unclear Unclear Unclear Unclear Unclear Unclear Unclear NO YES YES YES Unclear YES YES

Fair Fair Fair Fair Fair Fair Fair Fair Fair Fair Fair Fair Fair Fair Low Fair High Low High High

Chinese and Japanese languages were translated with the assistance of native speakers. Statistical analysis Data from the individual eligible studies were entered into a spread sheet for further analysis. Review Manager (RevMan) (Version 5.1. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2011) was used to perform the statistical analysis. Pooled odds ratios (OR) were calculated for discrete variables. The fixed-effects and random-effects models were used to calculate the outcomes.25,26 In case of significant statistical heterogeneity, only the results of the random-effects model were reported. Heterogeneity amongst the trials was determined by means of the Cochran Q value and quantified using the I2 inconsistency test.

on a central site and transmitted to treatment providers by telephone, fax or sealed opaque envelopes. In the remaining fifteen RCTs9e11,27,28,30,32e36,38e41 the information regarding approaches to allocation concealment could not be determined. The baseline features were similar between treatment groups in all 20 RCTs thus reducing the clinical heterogeneity. All RCTs but two38,40 specified the eligibility criteria for patients to be enrolled. All RCTs but two39,40 specified numbers lost to follow-up in each treatment group. Six RCTs7,27,40,36e38 analysed the data on an intention-to-treat (ITT) basis, whereby participants were analysed in the groups to which they were initially randomised. Four RCTs did not perform ITT analysis10,11,26,35 and in the remnant ten RCTs it was unclear.9,28e34,39,41 Blinding after allocation was impossible because of the nature of the trials. All twenty RCTs were considered to be at acceptable risk of bias in the important domains (Table 1).

Results Overall 1-year mortality Twenty RCTs, fulfilled the inclusion criteria and were included in the meta-analysis (publication dates 1987e2011). There were a total of 2145 patients (1152 randomized to receive radical resection þ IPC and 993 randomized to receive radical resection without IPC) (Table 2). Quality of trials There was good agreement between the reviewers (EC and FeCo) about the eligibility and quality of the studies. Table 1 demonstrates the quality of the 20 included RCTs.7,9e11,26e41 In five RCTs,7,26,29,31,37 the method of allocation concealment was adequate; randomisation was performed

Nine studies reported 1-year mortality,7,26e28,36,38e41 two randomized patients with PC.27,36 343 patients received the standard treatment and 283 surgery þ IPC (Fig. 1). There was no statistical heterogeneity between studies. In the random-effects model, the 1-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.31, 95%CI ¼ 0.19e0.48). Overall 2-year mortality Seven studies reported 2-year mortality,26,28e30,33,36,41 three included patients with PC.29,30,36 249 patients received the standard treatment and 277 surgery þ IPC

F. Coccolini et al. / EJSO 40 (2014) 12e26

15

Table 2 Summary of the 20 included studies. Study (Ref.) year

Number of patients

Study period

Stage I/II

III

IV

Intraperitoneal chemotherapy arm n ¼ 1152

Surgery arm n ¼ 993

n ¼ 32 HIPEC: MMC 64e100 mg in 8e12 lt. solution Temp: 40e45  C Time: 50e60 min. Technique: close Lost F-Up: 0/60 (both arms) n ¼ 25 Perfusion: MMC 500 mg þ 375 mg CH Technique: close Lost F-Up: 1/50 (both arms) n ¼ 42 HIPEC: MMC (10 mg/ml) Temp: 40e45  C Time: 50e60 min Technique: close Lost F-Up: 0/82(both arms) n ¼ 40 Perfusion: MMC 30 mg/kg þ CDDP 300 mg/kg in 10 lt. saline solution Time: 60 min. Hyperthermic: 41e42  C (n ¼ 22) Normothermic: 37e38  C (n ¼ 18) Technique: open Lost F-Up: 0/58 (both arms) n ¼ 33 EPIC (started between 10 and 28 POD) CDDP 90 mg/m2 in 2 lt. saline solution Time perfusion: 4 h. Treatment repeated monthly Technique: close Lost-F-Up: 3/67 (both arms) n ¼ 56 Perfusion: MMC 50 mg þ CH in 100 ml saline solution Temperature: room temperature Technique: close Lost-F-Up: 2/113 (both arms) n ¼ 78 HIPEC: MMC 80e100 mg/m2 in 8e10 lt. solution Temperature: 40e45  C Time: 50e60 min. Technique: open Lost-F-Up: 0/174 (both arms) n ¼ 46 Perfusion: MMC 50 mg þ 375 CH Technique: open Lost-F-Up: 0/91 (both arms) n ¼ 125 EPIC: MMC 10 mg/m2 in 1 lt. dextrose solution for 23 h (POD 1), 5-FU 700 mg/m2 in 1 lt. dextrose sol. (POD 2, 3, 4, 5) Temperature: 37  C Technique: close Lost-F-Up: 3/248 (both arms) (continued

n ¼ 27

1

Koga11 1987

60

1983e1985

ND

ND

ND

2

Hagiwara26 1992

50

1987e1991

ND

ND

0

3

Hamazoe9 1993

82

1983e1986

14

57

11

4

Fujimura28 1994

58

1988e1992

20

20

18

5

Sautner27 1994

67

1985e1989

0

42

25

6

Takahashi29 1995

113

1987e1992

0

76

37

7

Ikeguchi10 1995

174

1980e1989

ND

ND

ND

8

Rosen31 1998

91

1994e1997

0

79

12

9

Yu32 1998

248

1990e1995

93

84

71

n ¼ 25

n ¼ 40

n ¼ 18

n ¼ 34

n ¼ 57

n ¼ 96

n ¼ 45

n ¼ 123

on next page)

16

F. Coccolini et al. / EJSO 40 (2014) 12e26

Table 2 (continued) Study (Ref.) year

Number of patients

Study period

Stage I/II

10

Fujimoto30 1999

141

1987e1996

24

11

Shimoyama33 1999

46

1991e1993

12

Tan41 2000

51

13

Yonemura34 2001

14

IV

Intraperitoneal chemotherapy arm n ¼ 1152

58

59

21

24

1

1995e1998

ND

ND

ND

139

1988e1993

0

102

37

Zuo40 2004

82

1998e2000

18

54

10

15

Wei35 2004

156

1999e2001

41

93

22

16

Ding38 2007

78

2001e2003

20

33

25

17

Deng39 2009

85

2002e2004

25

37

24

18

Kuramoto7 2009

88

1995e2005

ND

ND

ND

n ¼ 71 HIPEC: MMC 10 mg/ml in 3e4 lt. solution Temperature: 43e45  C Time: 120 min. Technique: close Lost F-Up: 0/141 (both arms) n ¼ 30 Perfusion: MMC 10 mg in 0.5 lt. solution þ systemic chemother as in surgery arm Temperature: room temperature Technique: open Lost-F-Up: 0/46 (both arms) n ¼ 22 HIPEC: MMC 64 mg in 10 lt. solution Temperature: 45  C Time: 60 min. Lost-F-Up: 0/51 (both arms) n ¼ 48 HIPEC n ¼ 44 NIC Perfusion: MMC 30 mg þ CDDP 300 mg in 8e10 lt. saline solution. Temperature: 42  C and 37  C Time: 60 min Technique: open Lost-F-Up: 0/139 (both arms) n ¼ 46 EPIC (3 cycles between POD 21 and 28) CDDP-80e100 mg þ 5-FU 1 g in 2 lt solution Temperature: 41e43  C Time: 60 min. Technique: close n ¼ 101 HIPEC (n ¼ 52): 1 g/m2 5-FU in 1 lt. solutions Temperature: 43e45  C Time: 60 min. HIPEC þ EPIC (n ¼ 49): HIPEC þ 60 mg/m2 CDDP in 1 lt. solution for 24 h (POD 1) Lost-F-Up: 0/156 (both arms) n ¼ 41 EPIC: 60mg/m2 CDDP in 2.5e3.5 lt. solution Temperature: 42e44  C Time: 30 min (4 cycles) Technique: close Lost-F-Up: 0/78 (both arms) n ¼ 44 HIPEC: 1000e1500 mg 5-FU þ MMC20 mg in 3 lt. solution Temperature: 42e43  C Time: 60e90 min. EPIC: 5-FU 1000e1500 mg in 3 lt. solution Temperature: 42e43  C Time: 60 min. n ¼ 30 intraperitoneal lavage þ NIC n ¼ 29 NIC NIC: CDDP 100 mg in 1.5 lt. solution Time: 1 h. Technique: close Lost-F-Up: 0/88 (both arms)

III

Surgery arm n ¼ 993 n ¼ 70

n ¼ 16

n ¼ 29

n ¼ 47

n ¼ 36

n ¼ 55

n ¼ 37

n ¼ 41

n ¼ 30

F. Coccolini et al. / EJSO 40 (2014) 12e26

17

Table 2 (continued ) Study (Ref.) year 19

Yang36 2011

20

Miyashiro37 2011

Number of patients

Study period

Stage I/II

68

2007e2009

0

268

1993e1998

ND

III 0

ND

IV

Intraperitoneal chemotherapy arm n ¼ 1152

68

n ¼ 34 HIPEC: CDDP 120 mg þ MMC 30 mg in 6 lt. solution Temperature: 43  0.5  C Time: 60e90 min. Technique: open Lost-F-Up: 0/68 (both arms) n ¼ 135 Perfusion: CDDP 70mg/m2 for 2 h in 1 lt. solution. Chemotherapy: i.v. CDDP 70mg/m2 (POD 14), i.v. 5-FU 700 mg/m2 (POD 14, 16), Oral FU 267 mg/m2 daily, for 12 months starting in the 4th Post-Op week Lost-F-Up: 1/268 (both arms)

ND

Surgery arm n ¼ 993 n ¼ 34

n ¼ 133

Abbreviations: HIPEC ¼ hyperthermic intraperitoneal chemotherapy, EPIC ¼ early postoperative intraperitoneal chemotherapy, NIC ¼ normothermic intraperitoneal chemotherapy MMC ¼ Mytomicin C, CDDP ¼ cisplatin, CH ¼ activated carbon, POD ¼ post-operative day, 5-FU ¼ 5-fluoruracil, ND ¼ not declared.

(Fig. 1). There was acceptable statistical heterogeneity between studies. In the random-effects model, the 2-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.27, 95%CI ¼ 0.16e0.47). Overall 3-year mortality Nine studies reported 3-year mortality,26e29,36,38e41three enrolled patients with PC.27,29,36 311 patients received the standard treatment and 340 surgery þ IPC (Fig. 2). There was acceptable statistical heterogeneity between studies. In the fixed-effects model, the 3-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.29, 95%CI ¼ 0.20e0.41). Overall 5-year mortality Six studies reported 5-year mortality,7,9,10,27,34,37 two included patients with PC.27,37 379 patients received the standard treatment and 439 surgery þ IPC (Fig. 2). There was acceptable statistical heterogeneity between studies. In the random-effects model, there was no statistically significant difference in the 5-year mortality rate (OR ¼ 0.99, 95%CI ¼ 0.71e1.37). 1-Year mortality in patients with peritoneal carcinomatosis Two studies randomized patients with PC (68 patients: standard treatment and 67: surgery þ IPC)27,36 (Fig. 1). There was no statistical heterogeneity between studies. In the random-effects model, the 1-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.25, 95%CI ¼ 0.12e0.51).

2-Year mortality in patients with peritoneal carcinomatosis Three studies included patients with PC (161 patients: standard treatment and 161 surgery þ IPC)29,30,36 (Fig. 1). There was no statistical heterogeneity between studies. In the random-effects model, the 2-year mortality rate was favourable to the surgery þ IPC arm (OR ¼ 0.29, 95%CI ¼ 0.17e0.51). 3-Year mortality in patients with peritoneal carcinomatosis Three RCTs enrolled patients with PC (125 patients: standard treatment and 123 surgery þ IPC)27,29,36 (Fig. 2). There was statistical heterogeneity between studies. In the fixed-effects model, the 3-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.25, 95%CI ¼ 0.14e0.47). 5-Year mortality in patients with peritoneal carcinomatosis Two RCTs included patients with PC (167 patients: standard treatment and 168 surgery þ IPC)27,37 (Fig. 2). There was no statistical heterogeneity between studies. In the random-effects model, there was no statistically significant difference in the 5-year mortality rate (OR ¼ 1.05, 95% CI ¼ 0.67e1.55). 1-Year mortality in patients without peritoneal carcinomatosis Seven trials evaluated the 1-year mortality (276 patients: standard treatment and 215

18

F. Coccolini et al. / EJSO 40 (2014) 12e26

Figure 1. Overall mortality at 1 year (A) and at 2 years (B).

surgery þ IPC)7,26,28,38e41 (Fig. 1). There was no statistical heterogeneity between studies. In the randomeffects model, the 1-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.34, 95% CI ¼ 0.18e0.62). 2-Year mortality in patients without peritoneal carcinomatosis Four trials evaluated the 2-year mortality (88 patients: standard treatment and 116 surgery þ IPC)26,28,33,41 (Fig. 1). There was statistical heterogeneity between studies. In the random-effects model, the 2-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.24, 95%CI ¼ 0.08e0.76).

3-Year mortality in patients without peritoneal carcinomatosis Six trials evaluated the 3-year mortality (186 patients: standard treatment and 217 surgery þ IPC)26,28,38e41 (Fig. 2). There was no statistical heterogeneity between studies. In the fixed-effects model, the 3-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.31, 95%CI ¼ 0.20e0.47). 5-Year mortality in patients without peritoneal carcinomatosis Four studies evaluated the 5-year mortality (212 patients: standard treatment and 271 surgery þ IPC)7,9,10,34

F. Coccolini et al. / EJSO 40 (2014) 12e26

19

Figure 2. Overall mortality at 3 years (C) and at 5 years (D).

(Fig. 2). There was statistical heterogeneity between studies. In the random-effects model, there was no statistically significant difference in the 5-year mortality rate (OR ¼ 0.89, 95%CI ¼ 0.49e1.63). Overall recurrence Eight RCTs were included7,9,28,30e32,34,39 (Fig. 2), one included patients with PC (31). 413 patients received the standard treatment and 519 surgery þ IPC. There was no statistical heterogeneity between studies. In the randomeffects model, the overall recurrence rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.46, 95% CI ¼ 0.35e0.62).

Peritoneal recurrence Eleven RCTs reported the peritoneal recurrence rate,9,10,28,30e32,34,36,38,41 two included patients with PC (30, 36). 568 patients received the standard treatment and 650 surgery þ IPC (Fig. 3). There was statistical heterogeneity between studies. In the random-effects model, the peritoneal recurrence rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.47, 95%CI ¼ 0.35e0.63). Dividing RCTs into those who treated patients with and without PC: 104 patients received the standard treatment and 105 patients were treated with surgery þ IPC for PC. There was great statistical heterogeneity between studies. In the random-effects model, the peritoneal recurrence

20

F. Coccolini et al. / EJSO 40 (2014) 12e26

Figure 3. Recurrence rate: overall (A) and peritoneal (B).

rate was significantly favourable to the surgery plus IPC arm (OR ¼ 0.29, 95%CI ¼ 0.12e0.70). In the PCprophylaxis trials 464 patients received the standard treatment and 545 surgery þ IPC. There was acceptable statistical heterogeneity between studies. In the random-effects model, the peritoneal recurrence rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.50, 95% CI ¼ 0.37e0.68). Lymph-nodal recurrence Six RCts reported the lymph-nodal recurrence rate,7,9,28,30,32,34 two included 100% of patients with positive lymph-nodes at the intervention.30,7 327 patients received the standard treatment and 429 surgery þ IPC (Fig. 4). There was no statistical heterogeneity between studies. In the random-effects model, there were no statistically significant difference in lymph-nodal recurrence rate between the two groups (OR ¼ 0.47, 95% CI ¼ 0.35e0.63). Dividing studies into those who

included 100% of patients with lymph-nodal metastasis and the other four studies: 99 patients with 100% of lymph-nodal metastasis received the standard treatment and 130 surgery þ IPC. There was no statistical heterogeneity between studies. In the random-effects model, there was no statistically significant difference in lymph-nodal recurrence rate between the two groups (OR ¼ 0.95, 95%CI ¼ 0.40e2.27). In the remnant four studies, 228 patients received standard treatment and 299 surgery þ IPC. There was statistical heterogeneity between studies. In the random-effects model there was no statistically significant difference in lymph-nodal recurrence rate between the two groups (OR ¼ 0.94, 95%CI ¼ 0.38e2.30). Haematogenous metastasis Eight studies were included (387 patients: standard treatment and 514 surgery þ IPC)7,9,28,30,32,34,38,39 (Fig. 4). There was acceptable statistical heterogeneity between

F. Coccolini et al. / EJSO 40 (2014) 12e26

21

Figure 4. Recurrence rate: lymph-nodal (C) and haematogenous metastasis (D).

studies. In the random-effects model, the rate of haematogenous metastasis was significantly favourable to the surgery þ IPC arm (OR ¼ 0.63, 95%CI ¼ 0.40e0.98).

1-Year mortality in patients with nodal metastasis Three RCTs were included (83 patients standard treatment and 105 surgery þ IPC)7,26,41 (Fig. 5). There was no statistical heterogeneity between studies. In the random-effects model, there was no statistically significant difference (OR ¼ 0.51, 95% CI ¼ 0.22e1.17).

2-Year mortality in patients with nodal metastasis Four RCTs were included26,29,30,41 (Fig. 5). 181 patients received the standard treatment and 173 surgery þ IPC. There was no statistical heterogeneity between studies. In the random-effects model, the 2-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.28, 95%CI ¼ 0.17e0.45).

3-Year mortality in patients with nodal metastasis Three RCTs were included26,29,41 (Fig. 5). 111 patients received the standard treatment and 102 surgery þ IPC. There was no statistical heterogeneity between studies. In the random-effects model, the 2-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.16, 95%CI ¼ 0.08e0.29). 1-Year mortality in patients with serosal infiltration Three RCTs were included26,27,41 (Fig. 6). One study included patients with PC.27 88 patients received the standard treatment and 79 surgery þ IPC. There was no statistical heterogeneity between studies. In the random-effects model, the 1-year mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.33, 95%CI ¼ 0.15e0.72). 2-year mortality in patients with serosal infiltration Three RCTs were included26,29,41 (Fig. 6). One study included patients with PC.29 111 patients received the

22

F. Coccolini et al. / EJSO 40 (2014) 12e26

Figure 5. Mortality in patients with loco-regional lymph-nodal metastasis at 1 year (A), at 2 years (B) and at 3 years (C).

standard treatment and 102 surgery þ IPC. There was no statistical heterogeneity between studies. In the randomeffects model, the 2-years mortality rate was significantly favourable to the surgery þ IPC arm (OR ¼ 0.27, 95% CI ¼ 0.15e0.48). Morbidity Ten RCTs were included9e11,28e32,34,36 (Fig. 7). 558 patients received the standard treatment and 616 surgery þ IPC. There was statistical heterogeneity between studies. In the fixedeffects model, the morbidity rate was significantly favourable to the surgery arm (OR ¼ 1.82, 95%CI ¼ 1.29e2.57). Discussion PC is the most common form of evolution and recurrence in patients suffering from GC,Although PC has been for long considered as incurable and associated with extremely poor prognosis, in the last 20 years was established the concept of multimodal treatment which combines systemic chemotherapy, radical surgery and IPC. The principal aim of the multimodal treatment in GC is to prevent

and/or to treat efficiently PC. Some cohort studies demonstrated its efficacy in advanced GC with or without PC.42 Many RCT confirmed its efficacy in treating advanced GC with or without PC and in preventing the locoregional recurrence.7,9e12,26e41 The published RCT applied different forms of IPC. All demonstrated advantages compared to the standard treatment considered as radical surgery and systemic chemotherapy. In our opinion the different forms of IPC can be considered as a unique treatment. The different way to apply it, gives slightly different results, but the inclusion in the multimodal treatment approach gives to all of them the same attribute: the direct application of a chemotherapeutic agents on the macro/ microscopic tumour seeding or residual. This gives the advantage to treat directly, with limited systemic chemotherapy diffusion, the microscopic component of tumour which spilled during the surgical manipulation or which diffuses through the peritoneal fluid circulations. The possibility to apply IPC also in neoadjuvant setting has been recently investigated with promising results.43,44 The main differences between the different forms of IPC are principally due to the pharmacokinetic and pharmacodynamic of the different drugs in the different situations.

F. Coccolini et al. / EJSO 40 (2014) 12e26

23

Figure 6. Mortality in patients with serosal invasion at 1 year (A) and at 2 years (B).

The survival and recurrence rate of patients affected by advanced GC depends strongly on the presence of PC and on loco-regional nodal infiltration. Nodal invasion could be theoretically considered as extra-peritoneal disease demonstrating the cancer systemic diffusion beyond the peritoneal cavity. With this assumption, theoretically IPC would have limited or no efficacy due to its presumed only loco-regional effect. The effect of IPC on nodal metastasis has never been clearly demonstrated. Two previous RCT analysed the effect of IPC on survival relatively to the positivity of lymph-nodes at pathological examination.10,34 Yonemura et al. found on one hand no statistically significant differences in negative-nodes patient group. On the other hand they found an increase in survival with the application of IPC in positive-node patients group. The survival was moreover increased by the application of hyperthermic IPC.34 Ikeguchi et al. demonstrated the limited positive effect on survival of IPC in patients with positive lymph-nodes. They divided patients into three subgroups: no lymph-nodes metastases, 1e9 positive lymph-nodes and more than 10 positive lymph-nodes.10 Authors demonstrated the appreciable but not statistically significant (66% in surgery þ IPC group vs. 44% in control group) difference in 5-year survival in 1e9 positive lymph-nodes patients subgroup.10 The present meta-analysis demonstrated positive effect on mortality in patients with loco-regional lymph-nodal invasion: 2 and 3-year mortality rates are positively affected by the use of IPC. As a counterpart, the rate of nodal recurrence is not so positively affected by the IPC either in patients with 100% of positive loco-regional lymph-nodes either in remnant patients group. The positive effect on nodal metastasis is absolutely in agreement with the haematogenous metastasis rate reduction. It remains

hard to explain the absence of significant effect of IPC on 1-year mortality. Serosal infiltration could be considered as a strong indicator of tumour cells diffusion to the peritoneal cavity. This cell diffusion can be classified as “on-going peritoneal carcinomatosis”. Many studies included patient case-mix that comprehends many patients with no evident PC but with serosal infiltration. Also in ours daily practice we face doubt cases with no evident PC but with evident and “high-suspicion” serosal infiltration. These patients group represent a challenge because of the impossibility to definitively include them into either the PC-prophylaxis or into the PC-treatment group. They could be considered as an intermediate undefined group. The use of IPC in these cases could be controversial as no definitive evaluation of the effect of IPC in preventing and treating PC have been published. The previously published meta-analysis45e48 suggested the efficacy in treating/preventing PC of IPC but they did not analyse the different groups separately. For this reason the analysis of IPC effect on prevention/ treatment of PC has been conducted either separating studies that consider the prevention and the treatment of PC either considering them together. In 1, 2 and 3-year mortality analysis, IPC demonstrated to offer good chance to patients affected by advanced GC. The gain in terms of survival is confirmed considering the two groups either separately either together. The 5-year mortality analysis demonstrated to be not statistically significant. This data differs from the last published meta-analysis which found 5 and 9-year survival benefit.48 The last published metaanalysis and the previous published ones analysed different primary outcomes. However all confirmed the efficacy of IPC in increasing the survival rates at different time points.

24

F. Coccolini et al. / EJSO 40 (2014) 12e26

Figure 7. Morbidity.

The present study included a higher number of studies than the previous ones. Contemporarily it didn’t differentiate between the different form of IPC administration; as stated before we considered all the IPC administrations as declinations of the same treatment included in a multimodal diagnostic-therapeutic pathway. Reduced 1 and 2-year mortality have been also demonstrated by analysing only studies including patients with serosal invasion. Overall and peritoneal recurrence rate and haematogenous metastasis rate are positively influenced by IPC. These data agree with those about the efficacy on nodal metastasis. None of the previous meta-analysis analysed the different pattern of recurrence. These results suggest the possibility to enlarge the selection criteria for surgery þ IPC in advanced GC either with either without PC. In fact the high rate of loco-regional and systemic recurrence contributes to maintain the indications to surgery þ IPC as a kind of “desperate” alternative in many centers. As demonstrated by the present and previous metaanalysis as well as randomized trials, IPC plays a positive and fundamental role in treating GC and in preventing PC. Survival rates are increased by IPC even if morbidity has been demonstrated to be at risk of increasing with IPC after surgery.21,46 Four published meta-analysis found different results.21,46e48 In the present study the morbidity analysis showed as surgery þ IPC increased the rate of post-operative complications. Data about morbidity however are not definitive because of the impossibility to completely analyse the real impact of IPC on postoperative complications. As many randomized trials and meta-analysis about the effect of IPC as in GC have never been published for other diseases. The majority of them give positive results in terms of mortality and recurrence rate in patients who generally experiment a poor prognosis with relatively scarce chance of cure. IPC offers evidence-based good possibilities to improve the limited weapons we have to face with a such aggressive disease. However, no definitive indications on its use have been still given by expert panels. The present paper will contribute to solve the controversies about the

efficacy of IPC on nodal metastasis. In our opinion, in agreement with the results of the present study, IPC should start to be considered as part of common clinical practice in the treatment of advanced GC with or without PC. A big effort should be done to investigate more the physiopathology of this form of treatment. Conclusion 1, 2 and 3-year overall survival is incremented by the IPC. No differences have been found at five years in overall survival. 2 and 3-year mortality in patients with nodal invasion and 1 and 2-year mortality in patients with serosal invasion are improved by IPC. A positive effect of IPC has been found on overall and peritoneal recurrence and on distant metastasis. Morbidity rate is incremented by IPC. Loco-regional lymph-nodes invasion in patients affected by advanced gastric cancer is not a contraindication to IPC.

Acknowledgement We would like to thank Dr. He Ping, from the Chengdu Hospital (Sichuan province, China) for the precious collaboration. Conflict of interest All authors declare to have no conflict of interest.

References 1. Kelley JR, Duggan JM. Gastric cancer epidemiology and risk factors. J Clin Epidemiol 2003;56(1):1–9. 2. Parkin DM, Pisani P, Ferlay J. Global cancer statistics. CA Cancer J Clin 1999;49(1):33–64. 3. Nakamura K, Ueyama T, Yao T, et al. Pathology and prognosis of gastric carcinoma. Findings in 10,000 patients who underwent primary gastrectomy. Cancer 1992;70:1030–7.

F. Coccolini et al. / EJSO 40 (2014) 12e26 4. Takahashi T, Hagiwara A, Sawai K, Kondo S, Yamaguchi T. Intensive intraoperative local chemotherapy for lymph node and peritoneal metastases in gastric cancer. Onkologie 1991;14:152–7. 5. Yu CC, Levison DA, Dunn JA, et al. Pathological prognostic factors in the second British Stomach Cancer Group trial of adjuvant therapy in resectable gastric cancer. Br J Cancer 1995;71:1106–10. 6. Ikeguchi M, Oka A, Tsujitani S, Maeta M, Kaibara N. Relationship between area of serosal invasion and intraperitoneal free cancer cells in patients with gastric cancer. Anticancer Res 1994;14:2131–4. 7. Kuramoto M, Shimada S, Ikeshima S, et al. Extensive intraoperative peritoneal lavage as a standard prophylactic strategy for peritoneal recurrence in patients with gastric carcinoma. Ann Surg 2009;250: 242–6. 8. Gunderson LL, Sosin H. Adenocarcinoma of the stomach: areas of failure in a re-operation series (second or symptomatic look) clinicopathologic correlation and implications for adjuvant therapy. Int J Radiat Oncol Biol Phys 1982;8:1–11. 9. Hamazoe R, Maeta M, Kaibara N. Intraperitoneal thermochemotherapy for prevention of peritoneal recurrence of gastric cancer. Final results of a randomized controlled study. Cancer 1994;73:2048–52. 10. Ikeguchi M, Kondou A, Oka A, Tsujitani S, Maeta M, Kaibara N. Effects of continuous hyperthermic peritoneal perfusion on prognosis of gastric cancer with serosal invasion. Eur J Surg 1995;161:581–6. 11. Koga S, Hamazoe R, Maeta M, Shimizu N, Murakami A, Wakatsuki T. Prophylactic therapy for peritoneal recurrence of gastric cancer by continuous hyperthermic peritoneal perfusion with mitomycin C. Cancer 1988;61:232–7. 12. Wagner AD, Unverzagt S, Grothe W, et al. Chemotherapy for advanced gastric cancer. Cochrane Database Syst Rev 2010 Mar 17; (3):CD004064. 13. Van Cutsem E, Moiseyenko VM, Tjulandin S, et al. V325 study group. Phase III study of docetaxel and cisplatin plus fluorouracil compared with cisplatin and fluorouracil as first-line therapy for advanced gastric cancer: a report of the V325 study group. J Clin Oncol 2006;24:4991–7. 14. Cunningham D, Starling N, Rao S, et al. Upper gastrointestinal clinical studies group of the national cancer research institute of the United Kingdom. Capecitabine and oxaliplatin for advanced esophagogastric cancer. N Engl J Med 2008;358:36–46. 15. Roth AD, Fazio N, Stupp R, et al. Swiss group for clinical cancer research. Docetaxel, cisplatin, and fluorouracil; docetaxel and cisplatin; and epirubicin, cisplatin, and fluorouracil as systemic treatment for advanced gastric carcinoma: a randomized phase II trial of the Swiss group for clinical cancer research. J Clin Oncol 2007;25:3217–23. 16. Ross P, Nicolson M, Cunningham D, et al. Prospective randomized trial comparing mitomycin, cisplatin, and protracted venous-infusion fluorouracil (PVI 5-FU) with epirubicin, cisplatin, and PVI 5-FU in advanced esophagogastric cancer. J Clin Oncol 2002;20:1996–2004. 17. Baba H, Yamamoto M, Endo K, et al. Clinical efficacy of S-1 combined with cisplatin for advanced gastric cancer. Gastric Cancer 2003;6(Suppl. 1):45–9. 18. Yabusaki H, Nashimoto A, Tanaka O. Evaluation of TS-1 combined with cisplatin for neoadjuvant chemotherapy in patients with advanced gastric cancer. Gan To Kagaku Ryoho 2003;30:1933–40. 19. Preusser P, Wilke H, Achterrath W, et al. Phase II study with the combination etoposide, doxorubicin, and cisplatin in advanced measurable gastric cancer. J Clin Oncol 1989;7:1310–7. 20. Al-Shammaa HA, Li Y, Yonemura Y. Current status and future strategies of cytoreductive surgery plus intraperitoneal hyperthermic chemotherapy for peritoneal carcinomatosis. World J Gastroenterol 2008;14:11. 21. Yan TD, Black D, Sugarbaker PH, et al. A systematic review and meta-analysis of the randomized controlled trials on adjuvant intraperitoneal chemotherapy for resectable gastric cancer. Ann Surg Oncol 2007;14:2702–13. 22. Speyer JL, Sugarbaker PH, Collins JM, et al. Portal levels and hepatic clearance of 5-fluorouracil after intraperitoneal administration in humans. Cancer Res 1981;41:1916–22.

25

23. Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions: the cochrane collaboration. Chirchester (UK): John Wiley & Sons; 2008. 24. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7(3):177–88. 25. Demets DL. Methods for combining randomized clinical trials: strengths and limitations. Stat Med 1987;6(3):341–50. 26. Hagiwara A, Takahashi T, Kojima O, et al. Prophylaxis with carbonadsorbed mitomycin against peritoneal recurrence of gastric cancer. Lancet 1992;339:629–31. 27. Sautner T, Hofbauer F, Depisch D, Schiessel R, Jakesz R. Adjuvant intraperitoneal cisplatin chemotherapy does not improve long-term survival after surgery for advanced gastric cancer. J Clin Oncol 1994;12(5):970–4. 28. Fujimura T, Yonemura Y, Muraoka K, et al. Continuous hyperthermic peritoneal perfusion for the prevention of peritoneal recurrence of gastric cancer: randomized controlled study. World J Surg 1994; 18(1):150–5. 29. Takahashi T, Hagiwara A, Shimotsuma M, Sawai K, Yamaguchi T. Prophylaxis and treatment of peritoneal carcinomatosis: intraperitoneal chemotherapy with mitomycin C bound to activated carbon particles. World J Surg 1995;19(4):565–9. 30. Fujimoto S, Takahashi M, Mutou T, Kobayashi K, Toyosawa T. Successful intraperitoneal hyperthermic chemoperfusion for the prevention of postoperative peritoneal recurrence in patients with advanced gastric carcinoma. Cancer 1999;85(3):529–34. 31. Rosen HR, Jatzko G, Repse S, et al. Adjuvant intraperitoneal chemotherapy with carbon-adsorbed mitomycin in patients with gastric cancer: results of a randomized multicenter trial of the Austrian Working Group for Surgical Oncology. J Clin Oncol 1998;16(8): 2733–8. 32. Yu W, Whang I, Chung HY, Averbach A, Sugarbaker PH. Indications for early postoperative intraperitoneal chemotherapy of advanced gastric cancer: results of a prospective randomized trial. World J Surg 2001;25(8):985–90. 33. Shimoyama S, Shimizu N, Kaminishi M. Type-oriented intraoperative and adjuvant chemotherapy and survival after curative resection of advanced gastric cancer. World J Surg 1999;23(3):284–91. 34. Yonemura Y, de Aretxabala X, Fujimura T, et al. Intraoperative chemohyperthermic peritoneal perfusion as an adjuvant to gastric cancer: final results of a randomized controlled study. Hepatogastroenterology 2001;48(42):1776–82. 35. Wei G, Fang GE, Bi JW, et al. Efficacy of intraoperative hypothonic peritoneal chemo-hyperthermia combined with early postoperative intraperitoneal chemotherapy on gastric cancer. Chinese J Cancer 2005;24(4):478–82. 36. Yang XJ, Huang CQ, Suo T, et al. Cytoreductive surgery and hyperthermic intraperitoneal chemotherapy improves survival of patients with peritoneal carcinomatosis from gastric cancer: final results of a phase III randomized clinical trial. Ann Surg Oncol 2011;18(6): 1575–81. 37. Miyashiro I, Furukawa H, Sasako M, et al. Gastric Cancer Surgical Study Group in the Japan Clinical Oncology Group. Randomized clinical trial of adjuvant chemotherapy with intraperitoneal and intravenous cisplatin followed by oral fluorouracil (UFT) in serosa-positive gastric cancer versus curative resection alone: final results of the Japan Clinical Oncology Group trial JCOG9206-2. Gastric Cancer 2011; 14(3):212–8. 38. Ding WT, Duan DM, Sun GS, Gu C. Clinical observation on early intraperitoneal hypothermic perfusion chemotherapy in the treatment of stomach carcinoma after operation. Tumor 2007;27(7):585–7. 39. Deng HJ, Wei ZG, Zhen L, Li GX, Huang XC, Quing SH. Clinical application of perioperative continuous hyperthermic peritoneal perfusion chemotherapy for gastric cancer. J South Med Univ 2009;29(2): 295–7. 40. Zuo Y, Xu M, Shen D, Lu WD, Lu JF. Postoperative intraperitoneal hyperthermic chemoperfusion combined with intravenous

26

41.

42.

43.

44.

F. Coccolini et al. / EJSO 40 (2014) 12e26 chemotherapy for 82 advanced gastric cancer patients. Chin J Oncol 2004;26(4):247–9. Tan CQ, Wang XJ, Xu YK. Clinical studies on prevention of postoperational intraperitoneal neoplasm seeding of advanced gastric cancer by surgery combined with intraperitoneal perfusion of hyperthermic mitomycin C. J Pract Oncol 2000;15(3):165–7. Glehen O, Gilly FN, Arvieux C, et al. Association Franc¸aise de Chirurgie. Peritoneal carcinomatosis from gastric cancer: a multiinstitutional study of 159 patients treated by cytoreductive surgery combined with perioperative intraperitoneal chemotherapy. Ann Surg Oncol 2010;17(9):2370–7. Yonemura Y, Elnemr A, Endou Y, et al. Multidisciplinary therapy for treatment of patients with peritoneal carcinomatosis from gastric cancer. World J Gastrointest Oncol 2010 Feb 15;2(2):85–97. European Union Network of Excellence (EUNE) for Gastric Cancer Steering Group. Gastric cancer in Europe. Br J Surg 2008;95:406–8.

45. Xu DZ, Zhan YQ, Sun XW, Cao SM, Geng QR. Meta-analysis of intraperitoneal chemotherapy for gastric cancer. World J Gastroenterol 2004;10(18):2727–30. 46. Huang JY, Xu YY, Sun Z, et al. Comparison different methods of intraoperative and intraperitoneal chemotherapy for patients with gastric cancer: a meta-analysis. Asian Pac J Cancer Prev 2012;13(9):4379– 85. 47. Sun J, Song Y, Wang Z, et al. Benefits of hyperthermic intraperitoneal chemotherapy for patients with serosal invasion in gastric cancer: a meta-analysis of the randomized controlled trials. BMC Cancer 2012;12:526. 48. Mi DH, Li Z, Yang KH, et al. Surgery combined with intraoperative hyperthermic intraperitoneal chemotherapy (IHIC) for gastric cancer: a systematic review and meta-analysis of randomised controlled trials. Int J Hyperthermia 2013;29(2):156–67.