Outcomes of minimal access cytoreductive surgery (M-CRS) and HIPEC/EPIC vs. open cytoreductive surgery (O-CRS) and HIPEC/EPIC in patients with peritoneal surface malignancies: a meta-analysis

Introduction

In the last two to three decades the management of patients with peritoneal surface malignancies has transitioned from palliative care to therapeutic and often curative intent [1]. Cytoreductive surgery (CRS) for removal of macroscopic disease and intra-peritoneal chemotherapy to treat the microscopic remnant disease has become a standard for treatment of peritoneal surface malignancies [2], [3], [4], [5]. Although survival outcomes have improved, these complex surgical procedures accompany inevitable high morbidity and mortality in patients, which is comparable to other high risk open oncological procedures [6, 7]. The morbidity of these procedures can be reduced to an extent by better peri-operative optimisation of patients, still the complication rates are high and management becomes a challenge [8, 9].

The role of minimal access surgery like laparoscopic surgery, hand assisted laparoscopic surgery and robotic surgery has already been established in gastrointestinal and thoracic oncology. The oncological and survival outcomes are similar but the complication rates and morbidity is significantly lower in minimal access surgery [10], [11], [12], [13], [14], [15]. The role of minimal access surgery in peritoneal surface malignancies is evolving rapidly due to early referral of patients to tertiary centres with early stages and low peritoneal burden. Feasibility of laparoscopic CRS and HIPEC in low peritoneal carcinoma index (PCI) patients with low grade and borderline tumours have been described since 2011 [16, 17]. For high grade tumours like ovary and colorectal, literature is limited yet better peri-operative outcomes with minimal access surgery have been described [18].

The purpose of this meta-analysis is to assess the current literature and studies comparing open CRS and intraperitoneal chemotherapy procedure with minimal access CRS and to determine whether minimal access surgery is feasible and has better peri-operative outcomes compared to open procedure.

Materials and methods

Data assessment and inclusion and exclusion criteria

After elimination of duplicate abstracts, two investigators independently reviewed all abstracts and full text of articles which were regarded as potentially eligible for further consideration. Hand-searched reference lists of relevant articles were performed to identify further articles for analysis. Thereafter, eligible articles were selected for final analysis according to predefined inclusion and exclusion criteria. Disagreements were resolved by consensus.

Studies comparing outcomes of minimal access cytoreductive surgery like laparoscopic cytoreductive surgery, hand-assisted laparoscopic cytoreductive surgery or robotic cytoreductive surgery and HIPEC with open cytoreductive surgery and HIPEC were included. Studies which used EPIC instead of HIPEC were also included. Studies which did not have a comparative or control group with open CRS and HIPEC were not included. Studies with feasibility and peri-operative outcomes were focussed instead of survival outcomes due to limited follow up and data. Further inclusion and exclusion criteria is presented in the table in Figure 1. Authors of studies were not contacted for lack of data or missing data.

Figure 1:

The Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) flowchart.

Results

Search syntax described previously was used and total of 2,807 results were obtained (2,793 from PubMed and 14 from Cochrane library). Duplicate articles (n=6) were removed and 2,801 articles were reviewed. A total of 2,758 articles were excluded based on title and abstract review. Forty three articles were reviewed on the basis of their full text. According to exclusion criteria, 29 articles were rejected. Finally, 14 articles were included in this meta-analysis with total of 732 patients [15, 21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33] [Table 1]. Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) flowchart and selection criteria is shown in Figure 1. The assessment of the quality of studies in this meta-analysis was done by Newcastle-Ottawa Quality Assessment Scale and all studies showed a score of five stars and above indicating ‘good studies’.

There was a total 732 patients of which 319 (43.57 %) were in the study group i.e. minimal access CRS and HIPEC/EPIC and 413 (56.43 %) were in the control group i.e. open CRS and HIPEC/EPIC group. The median age was 55.7 (48.1–63.3) years in the M-CRS group and 53.3 (47.3–59.3) years in the O-CRS group. The female to male ratio was 1.31:1 and 1.67:1 whereas median BMI was 26.4 and 27.5 in the M-CRS and O-CRS groups, respectively. As only studies with matched proportion of population were considered, there was little heterogeneity in the demography. Other demographic and patient factors are given in the table (Table 2). Of all the 732 patients 48.9 % had appendix as primary, 35.1 % had colorectal primary, 5.9 % had ovarian primary and 5.5 % had mesothelioma as primary. Remaining patients had miscellaneous primaries. Majority of patients in the study had PCI <10. Disease characteristics of both groups are given in (Table 3).

The parameter of duration of surgery from incision to closure was reported by all 14 studies. The statistical analysis for duration of surgery in minutes between the M-CRS and O-CRS groups showed significant heterogeneity by the Higgins and Thompsons method with I² 95.99 % (95 % CI for I² 94.54 to 97.05, p<0.0001) hence random effects model of DerSimonian and Laird was used to calculate relative risk. Funnel plot and Egger’s regression test were used to test publication bias which was 0.7372 (p=0.8672, 95 % CI −8.668 to 10.143). The Standardized Mean Difference for duration of surgery was −0.0643 (95 % CI −0.993 to 0.865, p=0.892). No significant difference was seen between two groups (Figure 2A).

Figure 2:

Forrest plots of relative risk and standardized mean difference of different peri-operative outcomes between M-CRS and O-CRS group. (A) Duration of surgery, (B) estimated blood loss, (C) blood transfusions; (D) bowel recovery.

Median estimated blood loss was reported by nine studies. The Standardized Mean Difference (SDM) for estimated blood loss in mL between M-CRS and O-CRS was SMD=−2.379, p<0.001 (95 % CI −2.952 to −1.805) with random effects model (I² 79.30 %, 95 % CI 61.22 to 88.95, p<0.0001). Publication bias was −2.6251 (95 % CI −7.4531 to 2.2029 with p=0.2109). There was significantly less blood loss in the M-CRS group (Figure 2B).

Number of patients who required blood transfusions were reported in seven studies out of 14. The relative risk for blood transfusion in M-CRS and O-CRS was 0.598 with significant p-value of p=0.011 (95 % CI 0.402 to 0.889) fixed effects model with I² 0.00 % (95 % CI 0.00 to 32.90, p=0.8835). Publication bias 0.07921 (95 % CI −1.5120 to 1.6704, p=0.8842). There were significantly less number of blood transfusions in the M-CRS group as compared to O-CRS group (Figure 2C).

Parameter of bowel recovery in terms of number of days to pass flatus or number of days to start oral feeding were reported by six studies. The standardized mean difference for bowel recovery was SMD=−0.843, p=0.01 (95 % CI −1.487 to −0.2) with I² 85.24 % (95 % CI 69.76 to 92.79, p<0.0001) and publication bias −4.3775 (95 % CI −12.5461 to 3.7910, p=0.2110). There was significantly better bowel recovery in the M-CRS group (Figure 2D).

Number of days in hospital post-surgery was reported in all studies. The Standardized Mean Difference was SMD=−2.348, p<0.001 (95 % CI −3.178 to −1.519) by random effects model. I² 93.78 % (95 % CI 91.15 to 95.63, p<0.0001) and publication bias −5.6364 (95 % CI −11.6817 to 0.4089 p=0.0650). Significantly less number of days in hospital was spent in the M-CRS group (Figure 3A).

Figure 3:

(A) Forrest plot of SMD of hospital stay in days in M-CRS and O-CRS groups, (B) Forrest plot of relative risk of completeness of cytoreduction score 0 in M-CRS and O-CRS groups.

Morbidity of procedures was reported in 13 studies. Grade I and Grade II Clavien–Dindo morbidities were considered minor morbidity and Grade III and Grade IV Clavien–Dindo morbidities were considered major morbidity. The relative risk of minor morbidity was 0.592 (95 % CI 0.401 to 0.872 with p=0.008) with fixed effects model with I² 0.00 % (95 % CI 0.00 to 27.34, p=0.8589) and publication bias 0.1479 (95 % CI −1.1630 to 1.4589, p=0.8043). The relative risk for major morbidity was 0.474 (95 % CI 0.273 to 0.821 with p=0.08) with fixed effects model with I² 0.00 % (95 % CI 0.00 to 17.17 with p=0.9052) and publication bias −0.1432 (95 % CI −1.3937 to 1.1073, p=0.8014). The relative risk of total morbidity (Grade I to IV) was 0.538 (95 % CI 0.395 to 0.731 with p<0.001) with I² 0.00 % (95 % CI 0.00 to 34.48, p=0.7905) and publication bias 0.1626 (95 % CI −0.9860 to 1.3113, p=0.7611). This shows significantly less minor morbidity, major morbidity and total morbidity in M-CRS group as compared to O-CRS group (Figure 4A–C).

Figure 4:

Forrest plots of relative risk of morbidity between M-CRS and O-CRS groups. (A) Minor morbidity, (B) major morbidity, (C) total morbidity.

The parameter of completeness of cytoreduction score [CC score] were reported directly or indirectly in all 14 studies. Both groups had low peritoneal burden with mean PCI of <10. The relative risk for CC0 score for M-CRS vs. O-CRS groups was 1.064 (95 % CI 0.992 to 1.140 with p=0.083) with I² 0.00 % (95 % CI 0.00 to 40.92, p=0.8570) and publication bias 0.4821 (95 % CI −1.3639 to 2.3281, p=0.4668). No significant difference is seen between M-CRS and O-CRS groups hence strengthening the feasibility and oncological equivalence of both methods (Figure 3B).

The oncological and survival outcomes are shown in the table (Table 4). All studies didn’t have long follow up hence the interpretation of the outcomes like disease free survival and overall survival is limited. Sheng-Chi Chang et al. in their study report 4 (11.8 %) early peritoneal recurrence (<12 months) in the M-CRS group [median follow-up 24.8 (1.1–56.3) months] and 10 (29.4 %) in the O-CRS group [median follow up 47 (4.9–61.3)]. He reports 3 years overall survival of 78.5 % and 63.8 % in the M-CRS and O-CRS groups respectively [21]. Frederic Mercier with median follow up of 33 (19.3–41.6) months in M-CRS and 19.9 (1.05–35) months in O-CRS reports 1 year and 5 year DFS of 100 % and 91.04 % in M-CRS and 100 % and 62.5 % in O-CRS groups respectively. He also reports 5-year survival of 100 % in both M-CRS and O-CRS groups [27]. L. Rodriguez-Ortiz in their study report DFS and OS similar in both groups but Park SY et al. report better DFS in M-CRS group with similar OS in both groups [28, 32]. Chong Wang et al. mentions that there was no difference in DFS in both groups, OS was not reached due to limited follow up. Sang Hun Ha reports 3-year peritoneal recurrence free survival, DFS and OS of 51 %, 39 %, 67  and 53 %, 33 % and 66 % in M-CRS and O-CRS groups [30, 33]. 30 day mortality in the M-CRS group was 0 % in 13 of 14 studies, only in one study by Ha et al. there was 90-day mortality of one patient (2.4 %) [33]. Thirty-day mortality in the O-CRS group was 1(7.1 %) by Park et al., 4(16.6 %) by Koti et al. and 90-day mortality was 1(4.8 %) in the study by Ha et al. [25, 32, 33] Overall, the survival outcomes were not evaluated statistically but limited number of studies that evaluated survival outcomes showed that they were similar.

Discussion

Peritoneal surface malignancies comprise of a heterogenous group of malignancies which can be primary or secondarily spread to peritoneum from other organs with unique proclivity of peritoneal dissemination [34]. Cytoreductive surgery and intra-peritoneal chemotherapy in the form of HIPEC/EPIC/PIPAC has been proven to improve overall survival in these patients [35]. Even in rare primary malignancies like sarcomas, GIST, etc. with peritoneal only dissemination, CRS and HIPEC has shown benefit [36]. At present, open CRS and HIPEC has major morbidity of 22–34 % and mortality rates of 0.8–4.1 % [37]. With the standardisation of this procedure, it becomes necessary to reduce the surgery related morbidity with the use of modern technologies without affecting the oncological outcomes. Minimal access surgery has revolutionised this concept in various oncological fields by significantly decreasing morbidity and mortality. Earlier there were concerns regarding increased peritoneal dissemination of cancer cells by laparoscopy but prospective randomised trials have shown no difference in port site or peritoneal recurrences [38].

The minimal access approach has a theoretical advantage of decreased adhesion formation post-surgery which may show benefit according to ‘Fibrin entrapment hypothesis’ [39]. Also decreased hospital stay and complications in the post-operative period leads to decreased over all costs for the patient and hospital [25, 40, 41]. Another important advantage of minimal access surgery is decreased time to adjuvant chemotherapy due to overall early recovery of patients and has shown to increase disease free survival [21]. This has a lesser role in low grade malignancies like appendiceal mucinous neoplasms and multi cystic mesothelioma but has a major impact on patients with high grade primaries like ovary and colorectal malignancies [42, 43].

Initially the use of laparoscopy was limited to calculating PCI, to evaluate candidates for complete cytoreduction and therapeutic drainage of mucinous ascites for palliation and symptom control [44, 45]. In 2005, Ferron et al. first described feasibility of laparoscopic CRS and HIPEC in animals [46]. In 2011, Esquivel et al. were first to show feasibility of laparoscopic CRS and HIPEC in patients with peritoneal surface malignancies [16]. Since then many case series and small retrospective studies have shown feasibility of minimal access cytoreductive surgery and HIPEC. Further in this regard, Koti et al. and Fagotti et al. have shown feasibility of robotic CRS and HIPEC in their studies [25, 31].

To our knowledge this meta-analysis is first of its kind which directly compares outcomes of minimal access vs. open CRS and HIPEC/EPIC across different peritoneal surface malignancies. This meta-analysis includes studies with laparoscopic, hand assisted laparoscopic and robotic CRS and HIPEC/EPIC procedures. The PSOGI international collaborative registry and ASPSM multi-institution analysis describe outcomes of patients undergoing laparoscopic CRS and HIPEC in various peritoneal surface malignancies but lack direct comparison with open CRS procedures [17, 18].

In this meta-analysis peri-operative outcomes were compared in terms of duration of surgery, estimated blood loss, number blood transfusions, days for bowel recovery (time to pass flatus/time to start orals), hospital stay and morbidity [(minor=Grade I and Grade II) (major=Grade III and Grade IV) and (total=Grade I–IV)] according to Clavien–Dindo classification [47]. Technical feasibility outcomes were compared in terms of completeness of cytoreduction score 0. Survival outcomes were not evaluated statistically.

A Arjona-Sanchez et al. in the PSOGI international collaborative registry have described outcomes of 143 patients with laparoscopic CRS and HIPEC and have shown similar peri-operative outcomes to our meta-analysis of 319 patients of M-CRS group [17]. Our meta-analysis shows that peri-operative outcomes in terms of blood loss, blood transfusion, bowel recovery, hospital stay and morbidity are significantly better in the M-CRS group while duration of surgery have no significant difference. These findings are similar to the peri-operative outcomes of COLOR II trial and COREAN trial for colorectal cancers and MISSION trial for ovarian cancer comparing minimal access surgery to open surgery [48], [49], [50]. The oncological safety of procedure in terms of CC0 had no significant difference in both groups when the median PCI was low i.e. <10. This infers that minimal access approach can be used in patients with low PCI.

As many of the studies in this meta-analysis were retrospective cohort studies, the follow up period was less and survival outcomes like disease free survival, peritoneal recurrence free survival and overall survival could not be compared across both groups with accuracy. But, few studies like Chang et al., Cho CY et al., Mercer et al., Rodriguez-Ortiz et al., Wang et al. and Ha et al. have shown that disease free survival and overall survival in both groups were similar [21, 22, 27, 28, 30, 33]. Interestingly, Chang et al. have shown less early peritoneal recurrence (defined as peritoneal recurrence in less than 12 months) in the M-CRS group than O-CRS group (11.8 vs. 29.4 %) [21]. Similar results are shown by Salt et al. and Rodríguez-Ortiz et al. [26, 28]. Also, Park et al. have shown better 3-year DFS in the M-CRS group than the O-CRS group (53.8 vs. 27.3 %) in patients in whom complete cytoreduction was achieved (CC0) [32]. These finding may be attributed to the higher pressures reached during the minimal invasive HIPEC procedure which has shown to increase penetration and cytotoxicity of chemotherapeutic agents [51], [52], [53].

Finally, we believe that the strength of our meta-analysis is that a fair number of patients (732 patients, M-CRS 319 and O-CRS 413) were included in the study and comparative analysis was made. Also, the demographic characteristics of the patients in both groups were comparable since only studies with matched population were included. The major limitation of this meta-analysis is that, majority of studies barring a few are retrospective cohort studies. This is due to paucity of comparative work, relatively novel approach of minimal access in cytoreductive surgeries and lack of randomised controlled trials. Another limitation is that survival outcomes were not calculated statistically as only few studies mentioned survival outcomes and different studies had different primaries and histology.

Conclusions

This meta-analysis shows that minimal access cytoreductive surgery and HIPEC/EPIC is feasible and has better peri-operative outcomes compared to open procedure in patients having limited PCI peritoneal surface malignancies. However, further collaborative, prospective studies and randomised trials are warranted to confirm the results and for better assessment of survival outcomes.