Introduction

Pseudomyxoma peritonei (PMP) is a rare condition with a reported incidence of approximately one per million per year [1]. It is characterized by copious production of mucinous ascites that, over time, fills the peritoneal cavity. The majority of cases originate from a perforated appendiceal epithelial neoplasm [1]. Appendiceal neoplasms are uncommon, accounting for 0.4-1% of all gastro-intestinal malignancies ([2], [3]). The majority of appendix neoplasms are carcinoid tumours, with the next most common being epithelial neoplasms [4]. Frequently epithelial appendiceal neoplasms present insidiously, after an occult rupture, with the features of PMP. The earliest description of the condition was by Rokitansky in 1842 in a patient with a benign mucocele of the appendix [5]. Werth first introduced the term “Pseudomyxoma Peritonei” 42 years later in 1884 when he described a patient with a ruptured pseudo-mucinous ovarian cyst with implantation of the cystic contents on the peritoneal surfaces [6]. In 1901, Frankel was the first to describe PMP from a ruptured appendiceal cyst [7].

PMP has generally been considered a benign condition but its behaviour over time suggests that it should be considered, at best, a “borderline malignant” condition with inevitable disease persistence and progression.

Definitions

The present paper addresses the available scientific evidence concerning PMP management, in an attempt to build a consensus on controversial issues of a comprehensive approach of cytoreductive surgery (CRS) and perioperative loco-regional chemotherapy (PLC). PLC regimens include intra-peritoneal chemotherapy used in the operating room with hyperthermia and/or early postoperative intra-peritoneal chemotherapy (EPIC) within 7 days of surgery. Heated perioperative loco-regional chemotherapy has been referred to by many different nomenclatures: continuous hyperthermic peritoneal perfusion (CHPP), heated intra-operative intra-peritoneal chemotherapy (HIIC), hyperthermic intra-peritoneal chemotherapy (HIPEC) or intra-peritoneal hyperthermic chemotherapy (IPHC) . In this paper, HIPEC was the designated terminology.

Site of Origin of Pseudomyxoma Peritonei

Since the publications by Werth and Frankel there has been ongoing controversy as to the site of origin of PMP, particularly in women. Synchronous disease is found in the ovary and appendix in most females with PMP, and the disease is reported to be more prevalent in females. Recent morphological, immunohistochemical, molecular and genetic evidence support the theory that the majority of classical PMPs originate from a perforated mucinous tumour of the appendix ([1], [8], [9], [10], [11], [12], [13], [14]). In women, the ovarian mucinous tumours represent secondary spread from the appendix. Mukherjee [15] reported a similar male to female distribution in their series of patients with PMP. It therefore seemed unlikely that the ovary was the usual site of origin in females unless the male and female appendix behaved in a completely different manner.

Undoubtedly a small proportion of cases arise from other organs and it is possible that an ovarian primary mucinous tumour may be the commonest in this diverse group arising from the stomach, colon, pancreas, gallbladder, urachus and other intra-abdominal organs ([1], [16]).

Pathological Classification of Appendiceal Neoplasms

There has also been considerable confusion in the literature about the pathological classification of epithelial appendiceal neoplasms and their relationship to PMP ([1], [2], [9], [17], [18], [19]). High-grade colonic mucinous neoplasms, adenocarcinomas of the appendix and mucinous adenocarcinomas originating from any other intra-abdominal organ can simulate the clinical, radiological, and pathological features of PMP [8]. Additionally, there appears to be a spectrum of disease from low to high-grade mucinous appendiceal neoplasms, though the pathological appearances do not correlate with the clinical behaviour of the tumour in a significant number of cases [1].

These difficulties in pathological classification of the clinical entity of PMP have led to diverse reports in the literature and ongoing confusion as to the outcomes following intervention. Thus, many series include all cases of PMP, of whatever origin, and include patients with mucinous adenocarcinoma of the appendix whereas others have reported only on classical pseudomyxoma from appendiceal cystadenomas. Ronnett and colleagues, in a retrospective review of a series of patients who had undergone complete cytoreduction by Sugarbaker’s group, reported a pathological system commonly quoted in the literature [9]. They classified low-grade tumours as disseminated peritoneal adenomucinosis (DPAM) and high-grade tumours as peritoneal mucinous carcinomatosis (PMCA), with an intermediate group (IG) demonstrating a mixture of DPAM and PMCA [9]. Survival was significantly higher in the low-grade (DPAM) group as compared with the high-grade tumours (IG and PMCA) with actuarial 5-year survivals of 84%, 37.6% and 6.7% respectively. They were unable to show a statistically significant difference between the IG and PMCA groups and in subsequent articles have grouped these together ([8], [20]). Dichotomous categorisations of mucinous tumours of the appendix have been adopted by nearly all institutions treating PMP and they all demonstrated that PMCA had a less favourable prognosis, as shown in table 1.

The reported 5-year survival of DPAM varies from 74 to 100%. Loungnarath’s [22] excellent actuarial survival of 100% were in a group of only 8 patients with a median follow-up of 23 months. PMCA have a predictably lower 5-year survival ranging from 28-65%. HHMoran and Elias figures include only patients with complete cytoreduction with similar survival outcomes for low-grade tumours as in other studies but a marked difference in the high-grade group ([23], [24]).

In a recent study by Misdraji and colleagues, a different histopathologic classification of appendiceal neoplasm was proposed [18]. They classified the appendiceal mucinous tumours into low-grade appendiceal mucinous neoplasms (LAMN) and high-grade mucinous adenocarcinomas (MACA). The former diagnosis corresponds to mucinous cysto-adenoma and the latter to mucinous cysto-adenocarcinoma in the conventional classification system [26]. However, in this study, follow-up data were available only in 66 of 107 patients and patients underwent variety of treatments by different surgeons over a 50-year period. The procedures performed were mainly appendectomy with or without serial debulking procedures. Thirty patients with follow-up data had disease confined in the appendix and 7 patients had localized peritoneal disease involving the ovary or the fallopian tube. Although much progress has been made, greater knowledge regarding the histopathology of this disease and the surgical implications of these findings is needed.

Table 1. Studies describing pathological grade and 5 year survival for Pseudomyxoma Peritonei

Preoperative evaluation

The clinical presentation of PMP has been poorly defined due to few reports with large patient populations. The recent publication by Esquivel and Sugarbaker is the most enlightening [27]. In a series of 410 patients with appendiceal tumours, 217 had the diagnosis of and PMP with histological confirmation. Overall, 27% presented with suspected appendicitis, 23% presented with increasing abdominal distension and 14% presented with a new onset hernia. In women, a diagnosis of PMP was most commonly made while being evaluated for ovarian mass (39%). Nevertheless, the majority of patients are diagnosed at, or after, a laparotomy performed for either suspected appendicitis or peritonitis or gynaecological cancer.

More recently, the diagnosis of PMP is suspected in an increasing number of patients at imaging studies, particularly based on radiological features at CT-scan [28]. CT-scan or US-guided biopsy may be useful, although the relatively acellular material is often difficult to diagnose with certainty.

CT-scan findings are pathognomonic for PMP with appropriate radiologic techniques and a combination of oral, rectal and venous contrast. On CT-scan PMP appears of low attenuation but areas of high attenuation, due to solid elements within the mucinous material or compressed mesentery, can be seen [29]. Scalloping of visceral surfaces, particularly of the liver, distinguishes mucinous from fluid ascites on CT-scan [29].

The pattern of disease accumulation is characteristics and should suggest the diagnosis. The circulation of peritoneal fluid in the abdominal cavity is dictated by gravity, pressure changes associated with respiration and physical boundaries of the peritoneal reflections. The mucin producing cells in PMP are poorly adherent and freely circulate with peritoneal fluid. This process has been termed as redistribution phenomenon. Accordingly, the predictable pattern of accumulation of mucinous acites instead of the individual deposits can suggest the diagnosis of early stage PMP. At first, the mucin secreting cells from the ruptured appendiceal tumor fall by gravity into the pelvis. Hence they are carried up the paracolic gutters to the sub-phrenic spaces. Therefore, the pouch of Douglas, the right and left sub-hepatic spaces and the surfaces of liver and spleen are the commonest sites involved at the earlier stage of the disease. Once PMP has involved the abdomino-pelvic regions listed above, it fills those sites where peristalsis is limited by peritoneal attachment (ileocecal region, Treitz ligament, sigmoid colon) and finally occupies the remaining abdominal cavity. When the peritoneal cavity is completely, or almost completely filled with PMP, CT-scan findings become less specific, as and the characteristic pattern of spread of PMP cannot be appreciated. In most cases the striking feature is the relative sparing of the small bowel and its mesentery and the small bowel compartimentalization in the central abdomen by a large omental cake and massive mucinous ascites. Paraumbilical, inguinal ed hiatus herniae may also be present [29].                 

The role of MRI imaging in this clinical setting is presently under investigation. According to a preliminary report, the use of MRI seems promising in disease staging and patient selection for cytoreductive surgery [30]. PET is of very limited value for low-grade mucinous lesions, such as pseudomyxoma [31].

When a patient presents with increasing abdominal girth as a result of presumed malignant ascites, diagnosis is usually established with a paracentesis or laparoscopy and biopsy. In all instances, paracentesis or laparoscopy with a biopsy should be done directly within the midline and through the linea alba. These sites can be excised as part of a midline abdominal incision. No lateral puncture sites or port sites should be used because they could cause the tumour to seed into the abdominal wall, greatly reducing the probability of disease eradication [32].

Most patients are referred from other surgical or gynaecological units and have usually had variable degrees of surgery prior to formal attempts at complete cytoreduction. Extensive previous attempts to reduce tumour load were shown by one large study to have a negative impact on survival [20]. It advisable, therefore, that the referring team perform the minimal amount of surgery required to establish a histological diagnosis prior to referral to a specialist centre.

Eligibility for surgical treatment

Surgical intervention should only be considered for two indications:

1. where complete tumor removal is likely to be achieved and can be combined with intra-peritoneal chemotherapy.
2. for the relief of symptoms (palliative surgery).

Careful selection for this expensive, time consuming and high-risk surgery is essential. However, the potential surgical morbidity, along with the spectrum of disease aggressiveness, makes preoperative selection of patients for cytoreductive surgery (CRS) and hyperthermic intra-peritoneal chemotherapy (HIPEC) critical. Current selection criteria predominantly revolves around clinical assessment and CT-scan imaging ([1], [32]). Moreover, the history of previous surgical procedures is relevant [20].

Candidates have to be medically fit in order to undergo safely to CRS with HIPEC. Patients affected by peritoneal carcinomatosis with performance score of 2 to 3 according to the Eastern Cooperative Oncology Group (ECOG) have shown significantly poorer overall survival after CRS and HIPEC, compared to those with ECOG score of 1 [31].

In an effort to correlate the role of previous surgery in the establishment of peritoneal implants, Sugarbaker introduced the Prior Surgical Score (PSS). In patients with a PSS=0, diagnosis of peritoneal carcinomatosis was obtained through biopsy or laparoscopy only. PSS=1 indicates only a previous exploratory laparotomy. PSS=2 indicates exploratory laparotomy with some resections. Usually this was a greater omentectomy and/or right colectomy. PSS= 3 indicates patients had an attempt at a complete cytoreduction. Patients with PSS scores of 0 through 2 had a significantly improved survival compared with those with a PSS of 3 ([20], [33]).

The current best method to assess operability is contrast-enhanced CT-scan. Radiological features predicting the likelihood to adequate remove all the peritoneal tumor deposits have been described by Jacquet and validated by Sulkin ([28], [29]). In the study by Jacquet two radiologic findings predicted complete vs. suboptimal cytoreduction: the segmental obstruction of the small bowel and tumor masses >5cm on small bowel and its mesentery exclusive of distal ileum. A statistical approach using a tree-structured diagram showed that patients with both these features on preoperative CT scan, had an 88% probability of incomplete resection. Patients without these two CT findings had a 92% probability of complete resection [28].

The role of laparoscopy for PMP has never been specifically addressed. The experiences with patients affected by carcinomatosis from colorectal cancer has led to the conclusion that laparoscopy is accurate to score peritoneal involvement and to assess the complete resectability of carcinomatosis in patients for which there is inadequate or contradictory information concerning disease extent [34].

Finally, in recent years the prognostic value of markers in patients undergoing CRS and HIPEC has been extensively addressed. The role of markers in predicting the completeness of the cytoreduction was addressed in the paper by Baratti: normal preoperative CA125 correlated to the likelihood to achieve adequate CRS, but only at univariate analysis; in the same study, increased baseline CA19.9 was an independent predictor of worse progression-free survival at multivariate analysis [35]. CEA and CA19.9 were investigated by van Ruth in 63 patients. Only weak statistical association between baseline CA19.9 levels and disease-free interval was found, but elevated CA19.9 after the procedure or rising during follow-up was related to disease recurrence [36]. Survival was related to preoperative CEA and CA19.9 among 532 patients studied by Carmignani. Both markers, measured at the time of disease recurrence, correlated to survival after a second cytoreduction with HIPEC [37]. In the recent paper by Alexander-Sefre recurrence-free interval was statistically reduced for patients with increased CEA, but also for patients with at least one positive marker (among CEA, CA125 and CA19.9) [38].

Evidence supporting intensive loco-regional multimodality treatment as the standard of care for pseudomyxoma peritonei

Traditionally patients with PMP have been treated with repeated interval debulking procedures for relief of symptoms, but with limited expectation of long-term survival and no prospect of cure. However, accurate historical controls of uniformly treated patients are scarce, partly due to the rarity of the disease. In 1994, Gough reported a 10-year survival of 32% in 56 PMP patients who underwent serial debulking procedures and selectively treated with intra-peritoneal radiotherapy or chemotherapy between 1957 and 1983 [39]. In 2005, Miner reported a 10-year survival of 21% in 97 PMP patients treated with serial debulking, systemic chemotherapy and/or delayed intermittent intra-peritoneal 5-fluorouracil over a 22-year period [21]. Misdraji reported on 107 patients with a median survival of about 7.5 years, and a 20-year survival of 25% after serial debulking and perioperative intra-peritoneal chemotherapy. The number of patients in this group who received aggressive locoregional treatment is not known [18].

Although a subset of patients may remain asymptomatic for many years, the disease almost always recurs and patients often re-present with gastrointestinal obstructive symptoms. Over time each repeated debulking procedure becomes more ineffective and sometimes more dangerous due to the risk of bowel injury and subsequent fistula formation [1]. In addition, in some patients the disease may not remain indolent throughout its clinical course. Yan and colleagues showed that some patients underwent transitions from a less aggressive to a more aggressive histopathologic type over time and with repeated surgical interventions [41].

The structured approach described herein has been shown in multiple studies to be an independent variable for survival. Sugarbaker published a large series of 385 patients in 1999 [20]. Of these, 205 received HIPEC. He showed survival advantages in those who had complete vs incomplete cytoreduction (80% vs 20%) and in those with low-grade vs high-grade tumours (80% vs 28%) but did not comment on whether the introduction of HIPEC made any difference to survival. Glehen analysed the data from the same institution over a 30-year period and found a survival benefit for those receiving HIPEC in addition to early postoperative intra-peritoneal chemotherapy (27.2% vs 7.3% 5-year survival) but he was specifically looking at patients with incomplete cytoreduction [40]. Most recent updates by Sugarbaker demonstrated a median survival of 156 months and 5- and 10-year survival of 72% and 55% in 501 PMP patients. In this study, the uniform treatment approach has shown improved 10-year survival, as compared with historical controls [44]. Taken together these data suggest that treatment of PMP by means of CRS and HIPEC is supported by a "(Type 3)", as evidence is available from non-randomised studies, with external controls allowing comparisons.

Results of CRS with HIPEC versus serial debulking should be interpreted with knowledge that these treatment strategies have not been compared directly ([53], [54], [55], [56]). A phase III trial would be ideal. However, it will be difficult to do in the current setting because it would compare a potentially curative treatment option with a palliative procedure, thus patients are not likely to accept randomization. Also it may not be practical, as a sufficient number of patients and a long-term follow-up of at least 10 years are required to demonstrate statistical differences. However, further data from a randomized trial in a multi-institutional setting to clarify the efficacy of HIPEC would be desirable. The optimal study would be randomizing patients to CRS with HIPEC versus CRS without HIPEC after complete cytoreduction.

State of the art of the methodology

The aim of surgery in PMP is complete cytoreduction as described by Sugarbaker [42]. This involves up to six different peritonectomy procedures in combination with visceral resections as required, to remove all visible tumour, or if this was not possible, to leave tumour deposits less than 2.5mm (2.5mm being the maximum direct penetration of locally applied chemotherapy). In brief peritonectomy includes a greater omentectomy and splenectomy, left upper quadrant peritonectomy, right upper quadrant peritonectomy, lesser omentectomy and cholecystectomy, appendicectomy or right hemicolectomy, total colectomy, partial or total gastrectomy, and pelvic peritonectomy with anterior resection of the recto-sigmoid colon. Females require hysterectomy and bilateral salpingo-oophorectomy.

To date, 37 English language articles have been published regarding CRS combined with locoregional chemotherapy regimens including HIPEC and/or early postoperative intra-peritoneal chemotherapy (EPIC) within 7 days of surgery [45]. Most are serial publications reporting accumulating numbers of patients or increased length of follow-up. Ten most recent complete updates from eight institutions are to be considered ([22], [25], [44], [46], [52]). There are no randomized controlled trials or comparative studies. All ten articles are observational studies without control groups. All reports originated from specialized tertiary referral centers. Overall 863 PMP patients were included for assessment. Five studies included ≥100 patients ([44], [46], [47], [48], [49]) and the remaining 5 series had <100 patients ([22], [25], [50], [51], [52]). Four studies reported PMP originating from appendiceal epithelial neoplasms exclusively ([44], [46], [49], [52]). Another 4 studies reported PMP originating mainly from appendiceal epithelial neoplasms, representing 86% to 97% of their sample population ([22], [47], [50]). In the remaining 2 studies the origin(s) of PMP could not be ascertained ([25], [51]). The rate of complete cytoreduction varies from 40-91% with the 2 largest studies, one European and the other Northern American, both having complete macroscopic tumour removal in 65% of cases ([20], [23]). This supports the benefits of centralising this aggressive surgical treatment at institutions with an interest in PMP.Effectiveness of CRS and PIC on survival and recurrence is demonstrated in table 2.

Table 2. Effectiveness of cytoreductive surgery combined with perioperative intra-peritoneal chemotherapy for pseudomyxoma peritonei.

No study has reported on the effect of hyperthermic compared to normothermic intra-peritoneal chemotherapy in patients in whom a complete cytoreduction was performed. The types of chemotherapy agents utilized, the dosage, the temperature or duration of intra-peritoneal chemotherapy have not been subject to randomised trials but have been chosen on knowledge of the agents’ intra-peritoneal pharmacokinetics.  The commonly used intra-operative agents are mitomycin alone, cisplatin alone, 5-fluorouracil alone or a combination of these and are usually administered for 30-120 minutes. For EPIC, 5-fluorouracil, cyclophosphamide and mitomycin C are most frequently used for up to 6 days (see table 3).

Table 3. Literature review of different modalities of perioperative intra-peritoneal chemotherapy for pseudomyxoma peritonei combined with cytoreductive surgery (CRS).

There have been a number of recent reports in the literature addressing the morbidity and mortality in patients with PMP treated by CRS and HIPEC (Table 4). One confounding feature, particularly applicable to survival, of the published reports is that some report all cases, from all causes, treated with combination therapy whilst others only include favourable cases of appendiceal origin that had complete cytoreduction combined with intra-peritoneal chemotherapy. The overall morbidity rate varied from 33-56%. Hematological toxicity rate varied from 4-9%. Blood loss ranged from 2100 to 8000 cc. Mean operation duration ranged from 6.0 to 12.6 hours. Re-operation rates for postoperative adverse events were 11% and 21% as reported in two studies. The overall mortality rates ranged from 0-18%.The median and mean hospital stay ranged from 16 to 21 days and 26 to 29 days, respectively (see table 4).

Table 4. Morbidity and mortality of cytoreductive surgery combined with hyperthermic intra-peritoneal chemotherapy (HIPEC) for pseudomyxoma peritonei. (mean*)

One aspect that has not been fully addressed is a strategy for the many patients whose tumours were preoperatively considered unlikely to be completely removable, either due to tumour extent and distribution, or as a result of serious co-morbidity or age. There is increasing evidence that many of these, in addition to patients where complete tumour removal is impossible at laparotomy, benefit from a major palliative resection with reasonable intermediate-term survival of 43% at 2 years and 15% at 5 years and improved quality of life ([43], [49]).  In these situations an approach involving extended right hemicolectomy, greater omentectomy and splenectomy with an ileocolic anastomosis or on occasions a total colectomy and end ileostomy may be advisable [56]. Glehen recommended combination of comprehensive surgical debulking with HIPEC except for patients with signet ring histology or lymph node involvement in their experience of 174 patients with incomplete cytoreduction [43].

Assessment of Quality of Life

No studies were identified focusing on the quality of life after CRS and IPHC for PMP alone. Two studies by McQuellon reported the short-term and long-term quality of life outcomes in patients who underwent CRS and HIPEC for peritoneal carcinomatosis, including PMP as a subset of the study population ([57], [58]). They found that 64 patients in the short-term study decreased overall quality of life after surgery compared with baseline, but then returned to baseline or better within 3 to 6 months of surgery . A follow-up study of seventeen 3-year-survivors demonstrated that more than 90% of patients had minimal to no limitations of activity and had functional assessments that compared favorably to national reference values for their respective age groups.  

Conclusion

Recent evidence suggests that optimal surgical resection (complete cytoreduction if possible) combined with HIPEC as popularised by Sugarbaker is the most fundamentally based strategy for PMP. A combined approach makes both common and scientific sense, in that surgery attempts to remove all macroscopic disease whilst HIPEC addresses residual microscopic disease. This treatment strategy is a complex procedure, associated with significant morbidity and mortality with a substantial institutional and individual “learning curve” phenomenon [59].

The clinical results for cytoreduction and HIPEC in PMP show good survival for those patients with low-grade histology amenable to complete cytoreduction. Poorly defined and often confusing terms in the literature have resulted in most series being a heterogeneous population. This, combined with the relatively low incidence has limited the quality of the evidence with a lack of randomised controlled trials [4]. This deficiency is not unique to PMP surgery as few, if any, major surgical techniques are amenable to randomisation. Recent publications in the last 15 years with increasing numbers of medium to large case series reports in the literature reflect an improved awareness and understanding of the disease. This trend should continue with the development of centralised treatment centres throughout the world such that the quality of care and information available for patients with PMP should ultimately improve.

An emerging network of specialized centres may facilitate multicentre studies on aspects of chemotherapy type, duration and temperature to help allay the criticisms of many surgical, and in particular, medical oncologists on the lack of hard scientific evidence in PMP management.

Meanwhile good evidence is rapidly accumulating and surgical nihilism is no longer acceptable in this inexorably progressive, universally fatal, but eminently treatable disease.