Systemic chemotherapy and surgical cytoreduction for poorly differentiated and signet ring cell adenocarcinomas of the appendix

Abstract

Background:

Poorly differentiated and signet ring cell adenocarcinomas of the appendix represent a subset with aggressive tumor biology and poor outcomes with few studies evaluating the impact of systemic chemotherapy and cytoreductive surgery (CRS).

Patients and methods:

A retrospective chart review of patients with either poorly differentiated and signet ring cell appendiceal adenocarcinomas was completed from 1992 to 2010.

Results:

One hundred forty-two patients were identified. Seventy-eight patients with metastatic disease received chemotherapy. Radiographic response was 44%, median progression-free survival (PFS) was 6.9 months, and median overall survival (OS) was 1.7 years. In multivariate analysis, response to chemotherapy [hazard ratio (HR) 0.5; P = 0.02] predicted improved PFS, and complete CRS (HR 0.3; P = 0.004) predicted improved OS. Patients who underwent complete CRS (n = 26) had a median relapse-free survival (RFS) of 1.2 years and a median OS of 4.2 years. In multivariate analysis for this subset, complete cytoreduction score of 0 was significantly correlated with improved RFS (HR 0.07; P = 0.01) and OS (HR 0.02; P = 0.01).

Conclusions:

Systemic chemotherapy appears to be a viable treatment option for patients with metastatic poorly differentiated and signet ring cell appendiceal adenocarcinomas. Complete CRS is associated with improved RFS and OS, though part of this benefit likely reflects the selection of good tumor biology.

introduction

Appendiceal tumors encompass a rare and diverse group of neoplasms that have an age-adjusted incidence of ∼0.12 cases per 1 million individuals per year. Appendiceal tumors represent only 1% of all colorectal cancers diagnosed each year in the United States [1]. Little is known about the risk factors or etiology of epithelial tumors of the appendix, and historically, these tumors have been grouped together with colorectal cancers. However, appendiceal tumors, for which outcomes are strongly determined by histologic subtype, tend to have a biology and natural history that are markedly different from those of adenocarcinomas of the colon and rectum [2].

The primary treatment of patients with metastatic peritoneal dissemination of mucinous epithelial appendiceal neoplasms is complete cytoreductive surgery (CRS) followed by hyperthermic intraperitoneal chemotherapy (HIPEC) [3]. Retrospective single-institution studies of complete CRS and HIPEC in patients with disseminated peritoneal adenomucinosis (DPAM) and peritoneal mucinous carcinomatosis (PMCA) histological subtypes report median overall survival (OS) durations of 5–10 years [4–8]. Patients with PMCA have worse outcomes than patients with DPAM with 5-year OSs for patients who undergo complete CRS and HIPEC of 30%–35% [8, 9]. Studies also suggest worse outcomes for the subset of PMCA patients who have either signet ring cell or poorly differentiated tumor histology [9–11]. In one study of patients who were not candidates for CRS, poorly differentiated tumor histology [hazard ratio (HR) = 2.86, P = 0.03] and the presence of signet ring cells (HR = 3.99, P = 0.001) were both associated with shorter OS [11]. At present, the benefit of CRS with or without HIPEC for the subgroup of patients with either signet ring cell or poorly differentiated tumor histology is not known.

The majority of metastatic appendiceal adenocarcinomas demonstrates a relatively slow-growing biology and are primarily treated with CRS and HIPEC. Consequently, the role of systemic chemotherapy for these relatively slow-growing tumors has not been well investigated. An initial report from investigators at Memorial Sloan-Kettering showed no benefit in OS for patients treated with systemic chemotherapy [12]. More recent studies have begun to evaluate the role of systemic chemotherapy in this disease process, though the exact benefit still remains unclear [11, 13]. Because appendiceal adenocarcinomas with poor differentiation or signet ring cell histology appear to have a more aggressive biology, the potential effects of systemic chemotherapy in this subtype of appendiceal adenocarcinomas may be distinct from the effects in appendiceal adenocarcinomas with well or moderately differentiated histologies. To date, no study has evaluated the natural history, clinical behavior, or treatment response in this histologic subgroup of appendiceal adenocarcinomas. The goal of this retrospective study was to define the subset of poorly differentiated and/or signet ring cell histologies in appendiceal adenocarcinomas and to determine the effects of systemic chemotherapy and CRS on survival in this group of patients.

patients and methods

patient and data collection

Review of The University of Texas MD Anderson Cancer Center (UTMDACC) tumor registry between September 1992 and January 2010 was conducted. A total of 584 patients with appendiceal adenocarcinomas were identified. For inclusion in this study, each patient’s pathology was reviewed at UTMDACC and found to have either signet ring cells or poor histologic differentiation. Histological groupings for analysis were poorly differentiated adenocarcinomas with or without focal signet ring cells, signet ring cell adenocarcinomas (defined by the World Health Organization as adenocarcinomas with >50% signet ring cells), and well or moderately differentiated adenocarcinomas with focal signet ring cells.

Patients’ medical records were reviewed for information regarding demographic data, tumor characteristics, treatment history, response to treatments, laboratory tumor markers [carcinoembryonic antigen, cancer antigen (CA) 125, and CA 19-9], disease progression, and survival. Complete cytoreduction was defined as patients with peritoneal metastases who had ≤2.5 mm of disease remaining following CRS. The complete cytoreduction cohort was further stratified into two groups: those with no visible remaining disease [complete cytoreduction score of 0 (CC-0)] and those with remaining visible disease ≤2.5 mm (CC-1). Available pathology specimens from patients who underwent complete CRS were rereviewed by a gastrointestinal pathologist (MT) for histological grade confirmation. Perioperative chemotherapy was defined as patients who received chemotherapy either neoadjuvantly or adjuvantly. Oligometastatic disease within the peritoneal cavity was defined as disease that involved no more than two localized sites of disease (e.g. patients with bilateral ovarian disease). Distant metastases were defined as nonperitoneal sites of disease that were discontiguous from nearby peritoneal sites.

For systemic chemotherapy evaluation, patients were required to have had visible disease on radiographic imaging before chemotherapy, to have received at least one cycle of chemotherapy, and to have had a follow-up radiologic imaging evaluation. Patients excluded from this study generally had been seen in consultation and recommended to initiate systemic chemotherapy but had not met the above criteria. Response to chemotherapy was categorized according to a review of radiologic reports and the treating physician’s clinical notes. Only patients with improvement in disease as shown on radiologic imaging were considered to have a response. Patients considered to have stable disease were required to have both radiographic and clinical stability, while patients with progressive disease could have either radiographic progression or clinical decline. Because tumor markers were not documented in all patients, tumor markers were not included in the determination of response. This retrospective analysis was approved by the MD Anderson institutional review board, and a waiver of consent was allowed.

statistical analysis

Progression-free survival (PFS), relapse-free survival (RFS), and OS were analyzed using the Kaplan–Meier method of product-limit estimation. For the overall study population, OS was calculated from the date of diagnosis to death. For the chemotherapy subgroup, PFS and OS were calculated from the date of initial chemotherapy for metastatic disease, and for the CRS subgroup, RFS and OS were calculated from the date of CRS. The log-rank test was used to calculate the survival difference for categorical variables. The chi-square test and Fisher’s exact test were used to assess the relationship between categorical variables. All P values were two-sided and P values <0.05 were considered statistically significant.

results

We identified 142 patients with appendiceal adenocarcinomas who had either poor tumor differentiation or signet ring cell histology. The median age was 52 years (range, 27–82 years). Patient characteristics are summarized in Table 1. The median follow-up was 4.5 years, and OS duration by disease stage at initial diagnosis was 9.7 years for stage II, 4.7 years for stage III, and 1.8 years for stage IV (Figure 1). All patients with metastatic disease had peritoneal carcinomatosis. Nonperitoneal distant metastases occurred in 12 (10%) patients.

Table 1.

Characteristics of study patients (N = 142)

Variable	Frequency (%)
Median age, years	52
Sex
Male	62 (44)
Female	80 (56)
Disease stage
I	2 (1)
II	16 (11)
III	18 (13)
IV	106 (75)
Number of patients with metastatic diseasea	123 (87)
Sites of metastatic disease at presentation
Peritoneal disease	123 (100)
Ovaries	44 (36)
Uterus	22 (18)
Prostate	1 (1)
Testicle	2 (2)
Extraperitoneal disease	12 (10)
Liver	5 (4)
Distant lymph node	3 (2)
Lung	3 (2)
Bone	1 (1)
Histology
Poorly differentiated	114 (80)
Signet ring cell adenocarcinoma	19 (13)
Well of moderately differentiated adenocarcinoma with focal signet ring morphology	9 (6)
Presence of extracellular mucin	74 (52)
Presence of goblet cell carcinoid component	21 (15)
Complete cytoreductive surgery	26 (21)
No residual disease	15 (58)
Residual disease ≤2.5 mm	11 (42)
HIPEC	9 (35)
Perioperative chemotherapy	17 (81)
Oligometastatic disease (<3 sites)	8 (31)
Elevated tumor markers
CEA	22/72 (31)
CA 19-9	3/29 (11)
CA 125	29/51 (56)

^aMetastatic disease at any point during disease course.

HIPEC, hyperthermic intraperitoneal chemotherapy; CEA, carcinoembryonic antigen; CA, cancer antigen.

Figure 1.

Overall survival by disease stage.

Eighty percent of patients had poorly differentiated histology, while 13% had signet ring cell histology. Nine (6%) patients had well or moderately differentiated adenocarcinoma with a focus of signet ring cell morphology. These nine patients had a higher rate of localized disease at presentation (44% versus 16%, P = 0.4) and a longer median OS (12.0 versus 2.6 years, P = 0.03) than did patients who had signet ring cell adenocarcinoma. When patients with poorly differentiated histology were compared with patients with signet ring cell adenocarcinoma, the median OS was similar (2.0 versus 2.6 years, P = 0.8). Neither the presence of extracellular mucin (2.0 versus 2.5 years, P = 0.5) nor the presence of a goblet cell carcinoid component (4.1 versus 2.0 years, P = 0.1) had an impact on median OS.

chemotherapy group

Seventy-eight patients with metastatic disease received first-line systemic chemotherapy and had a follow-up radiologic evaluation (Table 2). Sixty-three patients had poorly differentiated adenocarcinomas, 10 patients had signet ring adenocarcinomas, and 5 patients had well or moderately differentiated adenocarcinoma with focal signet ring morphology. The overall response rate to chemotherapy was 44%. Stable disease and progressive disease occurred in 42% and 14% of patients, respectively. The median PFS for this cohort of patients was 6.9 months (95% CI 5.3–9.2) with a median OS of 1.7 years (95% CI 1.4–2.3) (Figure 2).

Table 2.

First-line systemic chemotherapy regimens for metastatic disease^a

Regimen	N (%)
Platinum based	48 (62)
5-FU + oxaliplatin + bevacizumab	22
5-FU or capecitabine + oxaliplatin	17
5-FU or capecitabine + cisplatinum	5
5-FU + oxaliplatin + cetuximab	2
Capecitabine + carboplatin	2
Irinotecan based	13 (17)
5-FU + irinotecan + bevacizumab	4
Cisplatin + irinotecan	4
5-FU + irinotecan	3
5-FU + irinotecan + cetuximab	2
Other	17 (22)
5-FU or capecitabine	14
Carboplatin + paclitaxel	2
5-FU + doxorubicin	1

^aSeventy-eight patients received first-line systemic chemotherapy for metastatic disease.

5-FU, fluorouracil.

Figure 2.

Kaplan–Meier estimate of progression-free survival and overall survival probabilities for metastatic patients treated with chemotherapy.

There was no significant difference among chemotherapy regimens in response rate, PFS, or OS based upon chemotherapy regimen. There was a nonsignificant trend towards increased PFS in patients who received irinotecan-based chemotherapy as the first-line treatment (1.0 versus 0.5 years, P = 0.07), but this trend was not associated with increased OS (1.8 versus 1.7 years, P = 0.6). The addition of a targeted therapy to first-line chemotherapy occurred in 48% of patients. Twenty-eight patients received therapy targeting the vascular endothelial growth factor receptor, and 6 patients received therapy targeting the epidermal growth factor receptor. No benefit from these biological agents was seen for the various efficacy end points analyzed.

Second-line chemotherapy was administered to 45 patients, and a response was seen in 6 patients (13%). The median PFS was 2.8 months. Twenty two of these patients had previously demonstrated a response to first-line chemotherapy.

Uni- and multivariate analyses for this cohort are presented in Table 3. The only variable significantly correlated with PFS in both uni- and multivariate analyses was achieving a response to chemotherapy. For OS, the presence of nonperitoneal distant metastases (P = 0.045) and achieving a complete CRS (P = 0.009) were significant on univariate analysis. Although achieving a response to chemotherapy was of borderline significance in univariate analysis (P = 0.07), it was not significant in multivariate analysis (P = 0.21). In multivariate analysis for OS, only undergoing a complete CRS (HR = 0.26, P ≤ 0.001) was statistically significant.

Table 3.

Uni- and multivariate analysis of PFS and OS data for the chemotherapy cohort

	Univariate			Multivariate
	Months	95% CI	P value	HR	95% CI	P value
PFS
Age (25–45)	5.33	3.16–11.54	0.421	0.98	0.95–1.02	0.30
Poorly differentiated	6.97	5.33–9.20	0.892	0.74	0.21–2.63	0.64
Signet ring cell adenocarcinoma	6.88	>3.68	0.927	0.62	0.15–2.56	0.51
Extraperitoneal disease	4.24	>2.46	0.222	0.98	0.43–2.23	0.97
Complete CRS	11.54	>6.97	0.229	0.62	0.28–1.35	0.23
Response	9.56	7.52–17.06	0.002	0.49	0.27–0.89	0.02
OS
Age (25–45)	1.5	0.89–4.05	0.725	0.99	0.97–1.02	0.73
Poorly differentiated	1.73	1.49–2.22	0.991	2.81	0.75–10.56	0.13
Signet ring cell adenocarcinoma	2.55	>0.515	0.951	1.53	0.36–6.45	0.57
Extraperitoneal disease	0.75	>0.6	0.045	1.10	0.48–2.53	0.84
Complete CRS	2.75	>1.73	0.009	0.26	0.10–0.66	0.004
Response	2.22	1.73–3.86	0.069	0.68	0.36–1.29	0.24

PFS, progression-free survival; OS, overall survival; CI, confidence interval; HR, hazard ratio; NR, not reached; CRS, cytoreductive surgery.

complete cytoreduction

Within the cohort of patients who had metastatic disease, 26 patients had a complete CRS, resulting in CC-0 in 15 patients and CC-1 in 11 patients. Twenty-two patients had poorly differentiated adenocarcinomas and 4 patients had signet ring cell adenocarcinomas. Pathology was available for independent review of cases from 17 patients, and this histological review confirmed either poorly differentiated adenocarcinoma or signet ring cell adenocarcinoma in all patients.

The effect of complete CRS on OS is shown in Figure 3. The median RFS after resection was 1.2 years (95% CI 1.05–7.08), and the median OS was 4.2 years (95% CI > 2.55). In univariate analysis, a CC-0 resection, compared with a CC-1 resection, was significantly associated with improved RFS (2.2 versus 0.9 years, P = 0.01) and OS (10.4 versus 2.3 years, P ≤ 0.001) (Table 4). In addition, oligometastatic disease demonstrated a trend favoring improved median OS (10.40 versus 3.67 years, P = 0.09). The chance of achieving a CC-0 resection appeared greater in patients with oligometastatic disease, as 8 of 10 patients with oligometastatic disease had CC-0 resections. Neither HIPEC (carried out in 35% of cases) nor perioperative chemotherapy (given in 81% of cases) had a statistically significant effect on RFS or OS. Of the 17 patients receiving perioperative chemotherapy, 5 received neoadjuvant chemotherapy and 12 patients received adjuvant chemotherapy.

Figure 3.

Effect of complete cytoreduction on overall survival in patients with metastatic disease. *Incomplete cytoreduction defined by patients not receiving surgery or who did not obtain a complete cytoreduction score of 0 or CC-1 resection.

Table 4.

Uni- and multivariate analysis of RFS and OS data for the cytoreductive surgery cohort

	Univariate			Multivariate
	Years	95% CI	P value	HR	95% CI	P value
RFS
Age (<45 years)	2.3	0.5–NR	0.23	0.95	0.86–1.05	0.31
Poor differentiation	1.2	0.9–NR	0.79	0.39	0.10–1.46	0.16
Oligometastatic disease	2.0	1.1–NR	0.21	0.71	0.12–4.31	0.71
Perichemotherapy	1.2	0.9–NR	0.90	0.22	0.04–1.25	0.09
HIPEC	1.2	1.2–NR	0.63	0.43	0.05–3.38	0.43
CC-0	2.2	1.2–NR	0.01	0.07	0.01–0.53	0.01
OS
Age (<45 years)	10.4	2.6–NR	0.41	1.01	0.90–1.14	0.84
Poor differentiation	13.6	1.8–NR	0.56	0.44	0.12–1.59	0.21
Oligometastatic disease	10.4	4.2–NR	0.09	1.17	0.18–7.60	0.87
Perichemotherapy	3.7	2.3–NR	0.88	0.12	0.01–1.59	0.11
HIPEC	3.0	2.3–NR	0.58	1.76	0.31–10.07	0.52
CC-0	10.4	4.2–NR	<0.001	0.02	<0.01–0.37	0.01

RFS, relapse-free survival; OS, overall survival; CI, confidence interval; HR, hazard ratio; NR, not reported; Perichemotherapy, perioperative chemotherapy; HIPEC, hyperthermic intraperitoneal chemotherapy; CC-0 complete cytoreduction score of 0.

The postrelapse survival was longer in patients who had a CC-0 resection, 29 months, as compared with patients who had a CC-1 resection, 14 months, P = 0.08 (Figure 4). The rate of repeat complete CRS does not appear to explain this difference, as repeat complete CRS was carried out in only five patients (33%) who had an initial CC-0 resection and two patients (18%) who had an initial CC-1 resection.

Figure 4.

Effect of residual disease on the relapse-free survival and overall survival of patients receiving complete cytoreductive surgery.

In the multivariate analysis, a CC-0 resection was significantly correlated with improved RFS (HR = 0.07, 95% CI 0.01–0.53, P = 0.01) and OS (HR = 0.02, 95% CI 0–0.37, P = 0.01) (Table 4). Use of perioperative chemotherapy was of borderline significance for both RFS (HR = 0.22, 95% CI 0.04–1.25, P = 0.09) and OS (HR = 0.12, 95% CI 0.01–1.59, P = 0.11).

discussion

This study reports on the largest collection of poorly differentiated and signet ring cell appendiceal adenocarcinomas to date. It is the first to specifically evaluate the role of systemic chemotherapy and CRS in this more aggressive histological subgroup. Our results have two important treatment implications. First, the use of systemic chemotherapy appears to be a valid approach for this appendiceal adenocarcinoma subset. Second, CRS offers a survival benefit, especially in patients in whom a CC-0 resection is achieved.

Our retrospective analysis confirms earlier reports that poorly differentiated adenocarcinomas and signet ring cell adenocarcinomas represent a subset of appendiceal epithelial neoplasms with poor outcomes. The median OS for patients with this histology and metastatic disease was 1.7 years, which is markedly less than previously reported OS for patients with PMCA [14]. Both systemic chemotherapy and complete CRS appear to be beneficial for patients with appendiceal adenocarcinoma with poorly differentiated and/or signet ring cell histologies. However, complete CRS was achieved in a minority of patients, and those who achieved a CC-0 resection appeared to derive a large benefit. Interestingly, the long interval between disease relapse and death for the CC-0 resection subgroup suggests that successful surgery appears, in part, to reflect the selection of good tumor biology.

Patients with either signet ring cell adenocarcinoma or poorly differentiated adenocarcinoma had similarly poor OS. Patients whose appendiceal adenocarcinoma had only a subset of signet ring cells had an improved OS, but this subgroup was small, representing only 6% of our patients. Prior studies have noted a worse outcome for patients with either poor differentiation or signet ring cells; however, no previous study has directly compared outcomes associated with these two histologic features [9–11]. Despite the poor outcome for this subset of appendiceal adenocarcinomas, the development of extraperitoneal disease still remains rare, occurring in only 10% of patients. However, those patients who did develop distant extraperitoneal metastases appeared to have a worse OS. Our findings are similar to those of a prior small retrospective study of 30 patients that compared well-differentiated appendiceal adenocarcinomas with moderate or poorly differentiated appendiceal adenocarcinomas [15]. In that study, a higher rate of extraperitoneal metastases (0% versus 43%) and worse mean OS (26 versus 13 months, P < 0.01) was seen in patients with a moderate or poorly differentiated histology. Interestingly, no molecular differences in microsatellite instability, p53 overexpression, or K-ras mutations were seen between these two histological groups [15].

The role of systemic chemotherapy has not been delineated in appendiceal epithelial neoplasms and has generally been utilized in patients who are not candidates for surgical cytoreduction. A phase II study evaluating the use of concurrent mitomycin C and capecitabine in 40 patients with advanced unresectable DPAM or PMCA suggested a role for systemic chemotherapy [13]. In that study, 6 patients (15%) demonstrated radiographic reductions in disease, and 18 patients (46%) demonstrated disease stabilization. However, only 32% of patients in that study had PMCA histology, and efficacy end points were not stratified by histologic subtype. A recent retrospective review of patients with PMCA who were not candidates for CRS reported a radiographic reduction or symptom improvement in 13 (24%) of 52 patients and a median PFS of 7.6 months [11]. In that study, poorly differentiated histology was present in only 15 patients (28%). In our study, semiquantitative radiographic response was seen in 44% of patients. The median PFS was 6.9 months, and chemotherapy response correlated with an improved PFS in the multivariate analysis. Interestingly, patients who received second-line chemotherapy appeared to have less benefit from systemic chemotherapy, with a response rate of 13% and a median PFS of 2.8 months. These results appear comparable to the results demonstrated with oxaliplatin- or irinotecan-based regimens used in colorectal cancer [16]. The differential impact of systemic therapy between the first- and second-line therapy suggests that, for this subset of appendiceal adenocarcinomas, systemic chemotherapy is affecting the natural behavior of this cancer.

In our study, complete CRS was associated with a significant increase in OS (4.2 versus 1.7 years, P < 0.001). In a prior study from Glehen et al. [10], patients with appendiceal epithelial neoplasms who underwent a CC-0 or CC-1 resection had a 5-year OS of 80%. However, in the histologic subset of patients with poorly differentiated or signet ring cell histology in that study, the 5-year OS was only 20%. In our study, only 21% of our patients achieved a CC-0 or CC-1 resection. Complete CRS was significantly associated with an improvement in OS, and this benefit was most pronounced in patients who had a CC-0 resection. The achievement of a CC-0 resection appears to denote a subset of poor differentiation and signet ring appendiceal adenocarcinomas with markedly improved outcomes (median RFS, 2 years; median OS, 10.4 years). The dramatic difference in median time from disease relapse to death (8.4 years) in this subgroup suggests that achieving a CC-0 resection may reflect not only a therapeutic benefit from surgery but also a selection process that favors patients with good disease biology.

We acknowledge that our study has several limitations and that our findings are hypothesis generating. In particular, this is a single-institution retrospective analysis that spans a long time period. Although this study represents the largest collection of appendiceal patients with poor differentiation and signet ring cell histologic subtype, the interpretation of several variables is limited by the small subgroups analyzed. Owing to the long duration of this study, radiologic images were not available for our review, and radiographic response assessment was thus based on the treating physician’s documentation and radiographic reports. Also, the high percentage of patients with metastatic disease in this study likely reflects a referral bias favoring patients with more severe disease at our institution.

The rarity of appendiceal adenocarcinomas makes this disease difficult to study, and it is extremely unlikely that prospective randomized clinical trials studying either CRS or systemic chemotherapy in this histologic subtype will be conducted. It is also one of the first analyses to specifically evaluate the role of systemic chemotherapy and CRS in this more aggressive histologic subgroup. Despite the limitations of this study, we believe that the results have two important treatment implications. First, the use of systemic chemotherapy for these patients appears to be a valid approach for this subset of appendiceal adenocarcinomas. At present, the optimal systemic chemotherapy regimen remains unclear, although over 90% of our patients received a fluorouracil-based chemotherapy combination. Second, CRS should be considered in patients in whom a complete cytoreduction resection is considered feasible. Validation of our results in additional datasets is needed.

funding

Ruth L. Kirschstein National Research Service Award (NRSA) Institutional Research Training Grants (T32 CA009666 to CL).

disclosure

The authors declare no conflict of interest.