Incomplete Cytoreduction of Colorectal Cancer Peritoneal Metastases: Survival Outcomes by a Cytoreduction Score

Abstract

Background

The surgical management of peritoneal metastases from colorectal cancer has been a topic of controversial discussion for many decades. Peritonectomy and perioperative intraperitoneal chemotherapy added options for surgical treatment of this condition beyond palliative surgery. The most favorable outcomes are recorded when peritoneal metastases from colorectal cancer can be resected to no visible evidence of disease.

Methods

To determine if any benefit from surgical treatment of patients with colorectal peritoneal metastases can occur from incomplete resection of peritoneal metastases, we studied patients by the completeness of cytoreduction (CC) score. The CC-3 indicated a palliative resection, CC-2 gross residual disease, and CC-1 almost complete cytoreduction but visible residual disease. The impact of clinical-, pathologic-, and treatment-related variables on the survival of the three groups was compared.

Results

Eighty-five patients with long-term follow-up were available for study. The median age was 53 years (range 18–82). There were 60 males (70.6%). Symptomatic patients, those with bowel obstruction, and patients with positive retroperitoneal lymph nodes had significantly reduced survival. The median survival of the CC-3, CC-2, and CC-1 groups were significantly different (p = 0.0027). The 2-year or greater survivals of the three groups were 4.8%, 15.1%, and 38.7%, respectively.

Conclusions

If a near complete cytoreduction combined with hyperthermic intraperitoneal chemotherapy can be performed, short-term survival benefit could be observed.

Introduction

Attempts to accurately quantitate the extent of oligometastatic disease has always been a major challenge. There is general agreement that knowledgeable patient selection is essential to favorable outcomes. Colorectal cancer is a malignancy in which great interest in treatment of oligometastatic disease occurs [1, 2]. Selection of patients for a variety of interventions remains as a frequent and occasionally protracted discussion for the multidisciplinary team (MDT). For colorectal liver metastases, multiple treatment options exist [3, 4]. The size of metastatic nodules, their number, and location within the liver parenchyma are important for the formulation of treatment options [5]. The information gained from radiologic studies (CT, MRI, and PET-CT) usually provides the necessary information for the MDT. Also, for colorectal peritoneal metastases, the size of metastatic nodules, their number, and their location within the abdomen and pelvis are important for the formulation of treatment options. Attempts to quantitate peritoneal metastatic disease were originally described by Jacquet et al. [6]. The peritoneal cancer index (PCI) and completeness of cytoreduction (CC) score continue to be used in most of the publications regarding the management of peritoneal metastases. However, the precise radiologic quantitation of peritoneal metastatic disease as compared to hepatic metastases remains more problematic and much less accurate. Progress has occurred but much remains to be placed into practice [7].

The most reliable prognosticator for peritoneal metastases is the CC score [6]. It is determined at the completion of cytoreductive surgery (CRS) after the surgeon's best efforts have been expended. Its early use was in the management of appendiceal mucinous neoplasms. Because of the copious mucus that coated the visceral peritoneum, a complete CRS was defined both as CC-0 (visible complete tumor removal) and CC-1 (residual tumor nodules ≤2.5 mm). However, when Elias et al. [8] performed a statistical analysis of 523 patients with colorectal peritoneal metastases, there was a median survival of 33 months for the CC-0 group and 18 months for the CC-1 group. Clearly, for colorectal cancer, complete CRS to no visible evidence of disease was a strong favorable prognostic indicator [8].

In this article, we evaluate the overall survival of patients with colorectal peritoneal metastases who have an incomplete CRS. Variables that impact on the outcome of an incomplete CRS are determined. Also, benefits of CC-3 versus CC-2 versus CC-1 CRS are compared. Is an aggressive but incomplete CRS plus HIPEC associated with improved survival? Should the surgeon persist to achieve a CC-1 CRS plus HIPEC if this requires an extended time in the operating theater and an inevitable increase in morbidity and mortality?

Materials and Methods

Data Acquisition

All clinical data were prospectively recorded in a standardized database and/or secured files and then collected and statistically analyzed. All of the original clinical data from referring institutions regarding the primary cancer and its resection were available. Also, all records from the Washington Cancer Institute regarding the CRS and perioperative chemotherapy were available for analysis. Prior to commencing this study, permission was obtained from MedStar Health Research Institute Office of Research Integrity. All patients had a diagnosis of colon or rectal cancer with biopsy-confirmed peritoneal metastases at the time of CRS. At the time of primary cancer resection and at the time of CRS, there were no liver metastases. The primary outcome measure was overall survival defined as the time interval between CRS and death or current status as alive with disease or no evidence of disease. Patients with a CT prior to CRS that showed systemic disease or direct extension into pleural or pericardial space were excluded.

Three Groups of Patients with Incomplete Cytoreduction

In this article, incomplete cytoreduction of peritoneal metastases was defined as visible cancer nodules or plaques that remained after the best efforts at CRS were completed. In order to evaluate possible benefits of CRS plus perioperative chemotherapy, we used the CC score to define three groups of patients to quantitate the extent of residual colorectal cancer [6]. The palliative CC-3 surgery group had a specific goal for treatment. Long-term survival was not expected. These patients were symptomatic. The surgery was often an attempt at relief of intestinal obstruction or resection of a dominant and painful tumor mass. Peritonectomies were rarely performed, and dissection was kept to a minimum. HIPEC was not used unless debilitating ascites was considered likely in the future [9]. Often, these patients having palliative surgery were unfit for major cytoreduction. Residual tumor nodules were greater than 2.5 cm in diameter, or there was a layering of residual tumor often at multiple sites.

The CC-2 group of patients also had a major abdominal incision. However, the goal of surgery was to remove as much abdominal and pelvic colorectal cancer as possible. Peritonectomies and visceral resections were used. Vigorous irrigation and HIPEC were used in an attempt to prevent tumor cell entrapment (TCE). However, despite the surgeon's best efforts, a mass or layering of cancer remained at one or sometimes more anatomic sites. The extent of disease was limited as much as possible, but gross disease 0.25–2.5 cm remained.

In the CC-1 group of patients, the CRS started with a long xiphoid to pubis incision. Peritonectomy procedures and visceral resections were performed so that tumor nodules or plaques were removed to ≤0.25 cm. Efforts were made to remove residual disease, but despite the best efforts of CRS, visible disease remained. In contrast to the CC-2 groups where isolated anatomic sites with a moderate extent of disease remained behind, all residual disease was of minimal extent. In many patients, residual disease was on the small bowel or small-bowel mesentery. A common site for residual small tumor nodules was the junction of the small bowel and small-bowel mesentery. This anatomic site of disease requires a small-bowel resection for complete removal [10].

Clinical Features Determined

From the hospital admission records for CRS plus HIPEC, patients were determined as asymptomatic versus asymptomatic. A planned second-look following neoadjuvant chemotherapy, a rising carcinoembryonic antigen (CEA), or progress of peritoneal disease by CT were the indications for CRS in asymptomatic patients. Bowel obstruction was a common indication for referral of a symptomatic patient for CRS plus HIPEC. Patients with chronic obstruction on total parenteral nutrition, acutely obstructed patients with recent vomiting, and patients with impending obstruction accompanied by abdominal pain were included in the bowel-obstructed group. The interval of time in months was recorded between the primary colorectal cancer resection or ileostomy construction and the CRS. The time to complete CRS plus HIPEC procedure was recorded. The PCI was determined at the time of CRS or estimated from the operative notes and pathology reports [6]. In the palliative surgery group, a determination of the PCI was not possible in some patients. The PCI was an assessment of the distribution and extent of peritoneal dissemination in 13 abdominopelvic regions recorded by the surgeon at the time of abdominal exploration with the assessment continued throughout the CRS. For data analysis, the patients were grouped by PCI as 9 through 20 or greater than 20.

Pathologic- and Treatment-Related Variables

The presence or absence of disease extension to para-aortic or common iliac lymph nodes was determined histologically at the time of surgery. The presence or absence of signet ring morphology cells was determined from the pathology report of the CRS. If an abdominal colectomy with Hartmann closure of the rectum or a proctocolectomy was performed, this was recorded as total colectomy.

Perioperative Chemotherapy

The cytoreductive surgical procedure was followed by HIPEC and EPIC in most patients except in the palliative CC-3 group [11]. For HIPEC, a curled peritoneal dialysis catheter (Covidien, Mansfield, MA, USA) was used to infuse the chemotherapy solution, and four outflow catheters (Closed Wound Suction Evacuation Kit, Bard, Covington, GA, USA) used to drain the chemotherapy solution to complete the recirculation through the heat pump. Two drugs were administered intraperitoneally in 1.5 L/m² of 1.5% dextrose peritoneal dialysis solution. The two drugs were doxorubicin at 15 mg/m² and mitomycin C at 15 mg/m² [12, 13, 14, 15, 16]. The chemotherapy solution was administered as rapidly as possible over approximately 5 min with the infused solution maintained between 41.5 and 43.5°C within the whole abdomen by a heat pump (Belmont Surgical Instruments, Billerica, MA, USA). A standardized open abdomen technique with manual distribution of the chemotherapy solution was used [17]. All HIPEC treatments were for 90 min. At the initiation of the HIPEC, 5-fluorouracil at 400 mg/m² and leucovorin at 20 mg/m² were infused through separate veins over 8 min. Following completion of the HIPEC, repair of seromuscular tears, bowel anastomoses, and abdominal closure were performed.

To administer EPIC, the Tenckhoff catheter and closed-suction drains were maintained after the HIPEC [11]. EPIC administration was initiated on the first postoperative day. A 1-L chemotherapy solution containing 5-fluorouracil at 600 mg/m² was administered intraperitoneally. Fifty meq of sodium bicarbonate was added to the carrier solution which was 1.5% dextrose peritoneal dialysis solution. This was administered as rapidly as possible by gravity flow into the peritoneal space. At 23 h, the drains and Tenckhoff catheter were unclamped, and fluid drained as completely as possible from the peritoneal space. This procedure was repeated for 5 consecutive days.

Systemic Chemotherapy and Perioperative Chemotherapy

All patients having surgery for peritoneal metastases in this study had prior systemic chemotherapy. At the time for this intervention, the MDT determined that there were no additional systemic chemotherapy options. Preoperative systemic chemotherapy was used in an attempt to downstage the extent of disease. Systemic chemotherapy was not used as a contraindication to CRS [18].

The perioperative chemotherapy in this group of patients was standardized. The palliative CC-3 resections did not receive HIPEC or EPIC. The CC-2 cytoreductions received HIPEC or EPIC but not both. The perioperative chemotherapy had a limited goal of reducing TCE into resection sites [19]. The patients with CC-2 and CC-1 cytoreduction all received HIPEC. If systemic chemotherapy had not been extensive, EPIC with 5-fluorouracil was used if thought to be safe. Again, the rationale for using perioperative chemotherapy in these patients, most of whom had extensive peritonectomy and visceral resections, was to limit TCE.

In CC groups 1, 2, and 3, an indication for HIPEC was ascites present in large volume when opening the abdominal incision. If this occurred, the indication was to use HIPEC even with a palliative CC-3 resection [9].

Morbidity and Mortality

The morbidity and mortality over the long time span of this study was determined by a prospective evaluation of adverse events [20, 21]. Class 1 and 2 adverse events are not tabulated. Class 3 adverse events required an invasive intervention. Class 4 required return to the operating theater, and class 5 indicated postoperative death.

Follow-Up

The follow-up was by clinical visits to an experienced oncologist every 3 months for 3 years. A CT scan of the chest, abdomen, and pelvis was performed every 6 months for 3 years and then yearly for the next 7 years for a total of 10 years of radiologic follow-up. The records regarding subsequent palliative systemic chemotherapy for recurrence were not available for analysis.

We elected not to attempt to report disease-free survival. After CRS, patients were rarely followed up at our institution, and the data regarding disease-free data were inaccurate. A definitive date for recurrent disease was difficult to establish and was often never confirmed by biopsy.

Statistical Analysis

Univariate descriptive statistics analysis was conducted using a nonparametric procedure. The Kaplan-Meier method estimated overall survival. p value was set at <0.05. Data were analyzed using SAS version 9.4. For Table 1, Fisher's exact test was used to detect differences between groups. A hazard ratio showed differences of each feature separately within a CC group. The χ² test was used to detect differences in incidences of right colectomy in males versus females.

Table 1.

Clinical-, pathologic-, and treatment-related variables and their effect on survival in 85 colorectal cancer patients who had an incomplete cytoreduction of peritoneal metastases

Effect of variables on survival
variable	N (%)	median survival, months	2-year survival %	hazard ratio (95% CI)	p value
Age, years
18–50	33 (38.8)	8.0	18.2	Reference	0.9836
>50	52 (61.2)	10.5	13.5	1.0 (0.65, 1.56)
Gender
Male	60 (70.6)	11.0	16.7	0.7 (0.47, 1.20)	0.2260
Female	25 (29.4)	8.0	12.0	Reference
Symptomatic
Yes	52 (61.2)	6.5	5.8	2.2 (1.39, 3.44)	0.0007
No	33 (38.8)	19.0	30.3	Reference
Bowel obstruction
Yes	38 (44.7)	6.0	2.6	2.1 (1.36, 3.35)	0.0010
No	47 (55.3)	16.0	23.4	Reference
Interval primary surgery to CRS, months
0–12	51 (60.0)	12.0	17.7	Reference	0.3202
>12	34 (40.0)	8.0	11.8	1.2 (0.81, 1.94)
Hours for CRS (N = 75)
2–8	42 (56.0)	8	14.3	Reference	0.6938
>8	33 (44.0)	13.0	18.2	0.9 (0.57, 1.45)
PCI (N = 74)
9–20	33 (44.6)	8.0	24.2	Reference
>20	41 (55.4)	10.0	12.2	1.1 (0.68, 1.73)	0.7503
Retroperitoneal lymph nodes
Positive	5 (5.9)	5.0	0	3.3 (1.28, 8.72)	0.0139
Negative Signet ring cells	80 (94.1)	11.0	21.3	Reference
Signet ring cells
Present	5 (5.9)	5.0	0	2.4 (0.94, 5.96)	0.0677
Absent	80 (94.1)	10.5	21.3	Reference
Total colectomy
Yes	14 (16.5)	6.5	0	1.7 (0.97, 3.14)	0.0625
No	71 (83.5)	12.0	22.5	Reference
Subgroup by CC
CC-3	21 (24.7)	6.0	0	2.6 (1.47, 4.69)	0.0011
CC-2	33 (38.8)	8.0	12.1	1.5 (0.93, 2.52)	0.0937
CC-1	31 (36.5)	17.0	35.5	Reference

Results

From a prospectively maintained database on 249 patients with colorectal cancer having surgery for peritoneal metastases, we identified 85 patients (34.1%) who had an incomplete cytoreduction or palliative surgical procedure in the absence of hematogenous metastases. The overall survival of these 85 patients is shown in Figure 1. HIPEC and/or EPIC was used in these patients as described in Materials and Methods. The data were accumulated over a 30-year time period (1989–2019). The median age was 53 years (18–82). Sixty-one percent of the patients were older than 50 years (Table 1). There was an unusual preponderance of male patients (70.6%). Fifty-two patients (61.2%) were symptomatic. The indications for surgery in the 33 (38.8%) asymptomatic patients were planned second-look (24), abnormal follow-up CT (6), colostomy closure (2), and rising CEA (1). The variable prognostic implication within this group of asymptomatic patients was not determined. There were 52 (61.2%) symptomatic patients. A common cause of symptoms was overt or impending bowel obstruction with abdominal pain present in 34 patients (44.7%). The interval from primary colorectal surgery to CRS was more than 12 months in 34 patients (40.0%). The median number of hours for CRS alone, with EPIC or with HIPEC, was 8 h (range 2–16). A PCI of 9–20 occurred in 33 patients (44.6%). A PCI greater than 20 was recorded in 41 patients (55.4%) with data regarding PCI available in 74 patients. In some patients with palliative surgery, it was not possible to estimate the PCI. The mean PCI was 20. Colorectal cancer extension into retroperitoneal lymph nodes was documented in 5 patients (5.9%). This finding had a significant impact on survival (HR 3.3, p = 0.0139). Signet ring morphology was reported in 5 patients (5.9%) with a borderline statistical significance (p = 0.0677). A total colectomy was performed in 14 patients (16.5%). Patients requiring total colectomy had a borderline reduced survival (p = 0.0625). There were 21 patients (24.7%) in the palliative surgery (CC-3) group, 33 patients (38.8%) in the CC-2 group, and 31 patients (36.5%) in the CC-1 group. The overall survival of the CC-3 versus CC-2 versus CC-1 group is shown in Figure 2. The differences in survival were significant (p = 0.0027).

Fig. 1.

Overall survival of 85 patients treated for colorectal peritoneal metastases who had incomplete cytoreduction.

Fig. 2.

Overall survival of 85 patients treated for colorectal peritoneal metastases comparing CC-1 versus 2 versus 3. CC-1 indicates residual tumor nodules less than 0.25 cm, CC-2 tumor nodules 0.25–2.5 cm, and CC-3 tumor nodules greater than 2.5 cm.

Comparison of Characteristics of CC-3, CC-2, and CC-1 Groups

A survey of the clinical features of each of these three groups of patients was made in an attempt to determine what caused the CC-3, CC-2, or CC-1 group to occur (Table 2). By Fisher's exact test, the incidence of age, gender, symptoms, and PCI did not differ between groups. Bowel obstruction occurred in 25.8% of CC-1 patients but was significantly more common in CC-2 (57.6%) and CC-3 (52.4%) of patients (p = 0.0273). The interval between primary resection and CRS was less in the CC-1 group as compared to CC-2 and CC-3 (p = 0.0462). Also, as would be expected a greater surgical effort was apparent in the CC-1 group (p = 0.0045). The percentage of patients having a 9-h or greater CRS was 65.5% for the CC-1, 37.9% for the CC-2, and 17.6% for the CC-3 group.

Table 2.

Comparison of clinical- and treatment-related features of three groups of patients with colorectal peritoneal metastases who had an incomplete CRS

Feature	Survival within CC groups			Incidence between CC groups, p value (Fisher's exact test)
	CC-3, N (%)	CC-2, N (%)	CC-1, N (%)
Age >50 years Hazard ratio (95% CI) p value	12 (57.1) 1.3 (0.54, 3.28) 0.5323	20 (60.6) 1.1 (0.53, 2.22) 0.8230	20 (64.5) 0.8 (0.39, 1.77) 0.6410	0.8855
Gender − male Hazard ratio (95% CI) p value	17 (81.0) 0.6 (0.21, 1.96) 0.4268	22 (66.7) 0.4 (0.16, 0.81) 0.0132	21 (67.8) 1.1 (0.49, 2.25) 0.9050	0.5180
Symptoms present Hazard ratio (95% CI) p value	14 (66.7) 2.9 (0.91, 9.2) 0.0724	23 (69.7) 1.6 (0.74, 3.41) 0.2331	15 (48.4) 2.7 (1.26, 5.57) 0.0105	0.1982
Bowel obstruction Hazard ratio (95% CI) p value	11 (52.4) 2.8 (0.97, 8.23) 0.0559	19 (57.6) 1.3 (0.64, 2.67) 0.4688	8 (25.8) 7.2 (2.5, 20.7) 0.0003	0.0273
CRS interval >12 months Hazard ratio (95% CI) p value	10 (47.6) 1.3 (0.54, 3.20) 0.5498	17 (51.5) 0.9 (0.45, 1.84) 0.8020	7 (22.6) 1.3 (0.54, 3.02) 0.5778	0.0462
Hours ≥9 Hazard ratio (95% CI) p value	3/17 (17.6) 2.7 (0.69, 10.7) 0.1545	11/29 (37.9) 0.9 (0.42, 1.93) 0.7848	19/29 (65.5) 1.5 (0.66, 3.37) 0.3309	0.0045
Peritoneal Cancer Index >20 Hazard ratio (95% CI) p value	9/15 (60.0) 1.0 (0.31, 2.89) 0.9297	19/30 (63.3) 0.6 (0.29, 1.35) 0.2329	13/29 (44.8) 1.3 (0.61, 2.78) 0.5038	0.3536
≥2-year survival* Median survival, months	1 (4.8) 6 5 (15.1) 8	12 (38.7) 17	0.0085 0.0027**

^*Different from product-limit survival estimates due to different estimate function.

^**p value based on the log-rank test of survival curves.

We searched within the CC groups for clinical- or treatment-related features that indicated a reduced or more favorable outcome. Surprisingly, women in the CC-2 group had a reduced survival (HR 0.4, p = 0.0132). Within the CC-1 group, an absence of symptoms indicates an improved outcome (HR 2.7, p = 0.0105). Also, within the CC-1 group, an absence of bowel obstruction was a robust clinical feature, indicating an improved outcome (HR 7.2, p = 0.0003). Symptoms and bowel obstruction were more commonly present in CC-2 and CC-3 groups, and when these clinical features were present, they indicated a reduced survival within the CC-2 and CC-3 groups. As shown in Table 2, the 2-year or greater survival was 4.8% in the CC-3, 15.1% in the CC-2, and 38.7% in the CC-1 group (p = 0.0085). The median survival was 6 months, 8 months, and 17 months in CC-3, CC-2, and CC-1, respectively (p = 0.0027).

The poorer outcome of women within the CC-2 cytoreduction group was not readily apparent. Data on CC-2 group patients were again reviewed. A search was made for a prognosis feature associated with the women of the CC-2 group as compared to men. None of the women in this group had a preliminary diagnosis of ovarian cancer or a debulking procedure as their initial surgery. Eleven of 11 women had a right colon primary cancer. Twelve of 22 males had a primary right colon cancer (p = 0.0074).

Morbidity and Mortality

There were three deaths in this group of 85 patients. The first was a 56-year-old male with small-bowel obstruction from recurrent rectal cancer. On March 17, 2011, he had a low anterior resection for a T3N2M0 cancer. He was treated by postoperative radiation therapy with capecitabine. He entered an outside hospital on May 17, 2012, with intestinal obstruction. He had extensive venous thrombosis, requiring a vena cava filtration device. He had an emergent transfer to the Washington Cancer Institute. Upon exploring the abdomen, 1½ L of purulent material was found to fill the abdomen and pelvis. Biopsies were taken to confirm the presence of peritoneal metastases. A loop ileostomy was constructed, and the abdomen irrigated until clear. The abdomen was left open using a vacuum pack apparatus. His sepsis progressed, and he died 1 week following exploration. The patient was in the palliative CC-3 surgery group.

The second patient was a 66-year-old male who presented in April of 1996 to his local physician with intestinal obstruction. A diagnosis of right colon cancer led to a right colon resection on April 10, 1996. In December of 1996, ascites developed. CT-directed biopsy showed peritoneal metastases. CRS with HIPEC was performed on February 10, 1997. The cytoreduction was scored as CC-3. A large amount of cancer at the ileocolic resection site required a repeat right colon resection with anastomosis. On postoperative day 12, he complained of shortness of breath, and chest X-ray showed an infiltrate at the base of his left lung. Antibiotics were started. The pulmonary infiltrate progressed, requiring transfer to the surgical intensive care unit. A CT scan was performed of the abdomen and pelvis on postoperative day 19 because abdominal pain was unrevealing. On postoperative day 28, the midline incision developed purulent drainage. A limited abdominal exploration showed a purulent collection adjacent to the ileocolic anastomosis but no suture line leak. On postoperative day 30, a dressing change following oral blue dye revealed an anastomotic leak. A diverting loop ileostomy was performed. On the 34th postoperative day, he underwent a tracheostomy. His pulmonary status continued to deteriorate, and he expired on his 42nd postoperative day with a diagnosis of adult respiratory distress syndrome and anastomotic leak.

The third postoperative death was in a 78-year-old woman who underwent a sigmoid colon resection for perforated diverticulitis on February 14, 1997. A villous adenoma was present within the specimen. Mucinous ascites developed and then intestinal obstruction. On December 18, 1997, she underwent a CC-3 cytoreduction which required a total abdominal colectomy. After an uneventful first seven postoperative days, she developed a massive left frontal cerebral infarct. MRI of the brain showed no metastases. She rapidly deteriorated and died on her 10th postoperative day.

Grade IV adverse events with return to the operating room were required in 3 patients, two of whom went on to die. A grade III adverse event was recorded in 27% of patients. Median length of the stay in the hospital was 18 days with a range of 10–42 days.

Discussion

Proper Level of Surgical Persistence to Achieve a Best Outcome

One of the dilemmas that repeatedly faces the cytoreductive surgeon regards the reasonable level of surgical persistence in the operating theater. Most would agree that a CC-0 CRS is worth many hours in the operating theater and an increased risk of morbidity and mortality. However, what is the reasonable level of surgical persistence if it becomes clear as the dissection proceeds that a CC-1 CRS will be the final CC score and cure is not possible? In some situations, only a CC-2 CRS is possible with a reasonable postoperative outcome. In other situations, a CC-1 CRS is possible but only with a high level of surgical persistence and many hours in the operating theater. Our data show the prolonged survival (38.7% 2-year survival) is expected if the colorectal cancer peritoneal metastases are reduced to occasional cancer nodules visible but less than 0.25 cm in diameter and no layering of cancer. The survival with the CC-1 CRS is limited but in marked contrast to the results expected with a CC-2 or CC-3 CRS (p = 0.0027). This decision must be left with the operating surgeon. Hopefully, the data from this article will help the cytoreductive surgeon in real time formulate an optimal outcome for the individual patient. If a more reliable HIPEC that would control residual tumor nodules up to 0.25 cm were available, this decision would be much easier.

Completeness Cytoreduction Defined Differently in Appendiceal and Colorectal Peritoneal Metastases Patients

When the CC score was designed around the year 1995, we were attempting to quantitate in a meaningful way the CRS for appendiceal mucinous neoplasms [6]. In this disease, a complete CRS that would result in long-term survival included both CC-0 and CC-1 scores. This designation included residual mucinous tumor nodules up to 0.25 cm in diameter. With these less invasive mucinous tumors, visible accumulations up to the CC-1 size limit were not eradicated in a majority of patients [22]. It was postulated that HIPEC with mitomycin C penetrated into the small but visible residual mucinous tumor nodules that remained on the surfaces of the small bowel and small-bowel mesentery. The mucus acted as a chemotherapy reservoir that prolonged exposure [23]. The rapid penetration of chemotherapy solution combined with its retention resulted in eradication of nearly all mucinous nodules of 0.25 cm or less. The end result was that both CC-0 and CC-1 CRS with parietal peritonectomy plus HIPEC became the goal for treatment of peritoneal metastases of appendiceal mucinous neoplasms [24].

It was inevitable that the technology that was and continues to be so successful for peritoneal metastases from mucinous appendiceal neoplasms would be applied to other diseases, especially colorectal cancer. However, the target for CRS plus HIPEC was now minimal residual disease from an invasive malignancy. In this disease, tumor nodules are vascularized or sequestered in lymphatic organelles. Free cancer cells are not a prominent feature of peritoneal dissemination from colorectal cancer [25].

The clinical data showed that residual colorectal cancer nodules up to 0.25 cm (CC-1) were poorly controlled by CRS and HIPEC with mitomycin C [8]. Other HIPEC regimens were used but were not of more benefit [11]. For colorectal peritoneal metastases, it was established that long-term favorable results with CRS and HIPEC needed a new threshold for complete cytoreduction. For colorectal peritoneal metastases, optimal CRS required resection to no visible evidence of disease. Small but visible cancer nodules that remained after the best effort at CRS would progress to a terminal condition in almost all patients. In order to accommodate this revised requirement for optimal CRS, the selection of patients for surgical treatment of colorectal versus mucinous appendiceal tumors changed. The threshold for colorectal PCI was defined at 17, while favorable results with appendiceal mucinous neoplasms were achieved with PCI at 30 and above [22, 24, 26]. In summary, the CC score required for a curative approach to colorectal peritoneal metastases was no visible evidence of disease versus 0.25 cm residual disease on the small bowel and its mesentery for mucinous appendiceal tumors.

For some surgeons, the implications of this requirement for a CC-0 CRS for colorectal peritoneal metastases was that small cancer nodules on the small bowel or its mesentery became a contraindication for elective treatment. Many groups required preoperative laparoscopy to rule out disease on the small bowel or small-bowel mesentery prior to CRS plus HIPEC [27]. CT or MRI will not provide accurate information regarding the presence versus absence of small peritoneal tumor nodules [7].

We reviewed our data on patients with incomplete cytoreductions for colorectal peritoneal metastases in order to test this new hypothesis regarding selection of patients for elective surgery. Our data establish that incomplete cytoreduction (including CC-1) is far inferior to complete removal of all disease visible to the naked eye. However, is it of no benefit especially in combination with HIPEC and systemic chemotherapy? The data in this manuscript can be interpreted to show that incomplete CC-1 CRS can result in limited survival benefit. In our group of CC-1 patients, all had visible cancer nodules present after the best efforts at CRS with HIPEC. By laparoscopy, it is unlikely that these patients would have been selected for CRS plus HIPEC. The CRS on this group of patients was time consuming with a mean of 10 h per patient. Yet, there was no mortality in this group of 31 patients. Their median survival was 17 months despite a mean PCI of 20. The 2-year survival was 35.5%.

When to Stop a CRS for Colorectal Cancer Peritoneal Metastases

Inevitably, despite careful patient selection in some patients with colorectal cancer peritoneal metastases, the surgeon must accept that complete CRS will not be possible. Many factors may come together to make this possibility a reality. Crucial anatomic sites that may be associated with visible evidence of persistent disease are the small bowel and its mesentery, the porta hepatis, the omental bursa, the pelvic sidewall, and the trigone of the bladder. Sometimes, the cancer at these technically difficult/impossible anatomic sites can be reduced to a CC-1 extent. An important surgical decision must be made when incomplete CRS becomes an unfortunate reality. Should the surgeon stop all further cytoreduction as soon as this observation is made? Or is there benefit to the resection of as much abdominal and pelvic disease as is possible to achieve a CC-1 CRS. From the perspective of an aggressive tumor biology expressing itself, invasion into retroperitoneal lymph nodes, the signet ring histology, or the requirement for total colectomy may prompt an early cessation of further cytoreduction. In contrast, if the surgeon foresees that a CC-1 cytoreduction is possible, the decision to move ahead with an incomplete resection of colorectal cancer may be a reasonable goal. In the 36.5% of our patients who had this near complete cytoreduction, the median survival was 17 months, and the 2-year survival 38.7%.

If one persists in cytoreduction and there are sites of CC-2 disease, the possible benefit to survival is substantially reduced. In 38.8% of our patients, nodules between 0.25 and 2.5 cm remained despite the best efforts of CRS. In these patients, despite surgery greater than or equal to 9 h in 37.9% of patients, median survival was only 8 months, and 2-year survival 15.1%. Perhaps this limited benefit may be reasonable in the young and fit patients with no other treatment options but not as standard of care.

The CC-1 Cytoreduction Is Different from a Debulking Procedure

A debulking procedure seeks to reduce the extent the disease with a minimum of surgical dissection and therefore with a minimum of risk. It is often added to a palliative procedure to relieve intestinal obstruction. There may be a large extent of disease that remains behind at some sites and little or no disease at other places. A CC-1 CRS is an incomplete cytoreduction but carries precise requirements. There is a uniformity of the residual disease. Although CC-1 disease may remain after the best efforts of CRS, there is no confluence of cancer (layering) that remains at any site. The visible nodules that are left are less than 0.25 cm in diameter. If there is a large extent of disease at any site, the CC score for colorectal cancer must be CC-2 or CC-3.

Role of HIPEC Cannot Be Established

The use or the absence of use of HIPEC and/or EPIC was not used as a variable in this data analysis. A standardized perioperative chemotherapy administration plan was followed. The goal of the HIPEC was to minimize TCE [19] and subsequent debilitating ascites [9]. The survival or local control advantage of HIPEC and/or EPIC would require a randomized trial.

Placement of Patients into CC-3, CC-2, or CC-1 Groups

An important issue regarding the utility of these data regards the timing of placement of patients into the 3 CC groups. If this knowledge regarding the CC score were available to the MDT prior to surgery, it would be of great value as a selection factor for proceeding with CRS and HIPEC. If the CC score became available at the finish of the CRS procedure, the information was available “after the fact” and of no value in patient selection for surgery. We have no prospective data regarding the timing of placement of patients into the CC groups. The authors accept the fact that the goals of surgery and the fitness of patients for a major CRS was not uniform throughout the cohort of 85 patients. This may be a limitation of the study but does not, in our opinion, change the interpretation of the data regarding the CC score. Most but not all of the CC-3 patients were identified as requiring palliative intervention. However, in a majority of patients despite knowledgeable preoperative radiology, the placement of patients into CC-2, CC-1, or CC-0 groups only become apparent as the CRS proceeded. Definitely, most of the CC-2 and CC-1 patients were taken to the operating theater with CC-0 as a goal for CRS. It is possible that laparoscopy prior to CRS would have changed the timing for a designation of the CC score [27]. This may be a limitation of the strategy used in these 85 patients. However, the determination of possible CC-1 status rather than CC-2 is probably not possible by preoperative laparoscopic examination. If a CC-1 CRS with median survival of 17 months and a 2-year survival of 38.7% is a reasonable goal in colorectal cancer peritoneal metastases patients, preoperative laparoscopy would not be considered a practice changing event.

Reduced Survival of Females in the CC-2 Group

Unexpectedly, a significantly poorer survival was observed in females in the CC-2 group. In an attempt to find an explanation for this statistics, we again studied our data. Women who have a diagnosis of ovarian cancer and undergo a debulking procedure show a reduced survival [28]. None of these 11 female patients had a preliminary diagnosis of ovarian cancer. Patients with right colon cancer have a reduced prognosis as compared to left colon cancer [29]. This has been attributed to the higher incidence of BRAF mutations associated with primary right colon cancer [30]. In our CC-2 female patients, 11 of 11 had right colon cancer as compared to 12 of 22 male CC-2 patients. This may help explain the reduced survival in the female CC-2 group.

Summary

For colorectal peritoneal metastases, the benefits of a complete cytoreduction with HIPEC are obvious, and in that, approximately 35% of patients are long-term survivors [1, 2, 8]. With incomplete removal of peritoneal metastases from colorectal cancer, there are no long-term survivors. Does this indicate that all resection should cease as soon as the surgeon determines that his/her best efforts cannot achieve a complete cytoreduction? To answer this important question, the survival of patients with CC-1, CC-2, and CC-3 cytoreductions were compared. Median and ≥2-year survival of the 3 groups were significantly different with p value of 0.0027 and 0.0085, respectively. If these patients with incomplete CRS had symptoms, especially bowel obstruction or retroperitoneal lymph nodes positive for cancer, the outcome was significantly reduced. When comparisons were made between groups, an interval between primary cancer surgery ≥12 months was more frequent in the CC-2 and CC-3 groups (p = 0.0462).

Although precise information will not be available preoperatively, the impact on survival of these clinical and histologic features regarding survival after incomplete CRS and HIPEC are crucial to judgments regarding the likelihood of success with this treatment plan. They must be considered in the risk/benefit assessment made by the MDT. Also, they can be a part of the informed decision-making process and assist in the construction of a collaborative surgeon-patient relationship [31]. With bowel obstruction, positive retroperitoneal lymph nodes, and a prolonged interval between primary surgery and CRS, alternative treatment options may need to be considered.

Statement of Ethics

This clinical research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. Prior to initiating this study, permission was obtained from the MedStar Georgetown Office of Research Integrity (Study 00001900). MedStar Georgetown IRB does not require permission to obtain patient informed consent for retrospective studies.

Conflict of Interest Statement

The authors have no conflicts of interest.

Funding Sources

Administrative and secretarial support was provided by the Foundation for Applied Research in Gastrointestinal Oncology.

Author Contributions

Paul H. Sugarbaker was responsible for conception and design; acquisition of data; analysis and interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; statistical analysis; obtaining funding; administrative, technical, or material support; and supervision. David Chang was responsible for analysis and interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content, and statistical analysis.

Data Availability Statement

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.