Systematic review of risk factors, prognosis, and management of colorectal signet-ring cell carcinoma

Abstract

BACKGROUND

Colorectal signet-ring cell carcinoma (CSRCC) is a rare clinical entity which accounts for approximately 1% of all colorectal cancers. Although multiple studies concerning this specific topic have been published in the past decades, the pathogenesis, associated risk factors, and potential implications on treatment are still poorly understood. Besides the low incidence, historically confusing histological criteria have resulted in confusing data. Nevertheless, the rising incidence of CSRCC along with relatively young age at presentation and associated dismal prognosis, highlight the actual interest to synthesize the known literature regarding CSRCC.

AIM

To provide an updated overview of risk factors, prognosis, and management of CSRCC.

METHODS

A literature search in the MEDLINE/PubMed database was conducted with the following search terms used: ‘Signet ring cell carcinoma’ and ‘colorectal’. Studies in English language, published after January 1980, were included. Studies included in the qualitative synthesis were evaluated for content concerning epidemiology, risk factors, and clinical, diagnostic, histological, and molecular features, as well as metastatic pattern and therapeutic management. If possible, presented data was extracted in order to present a more detailed overview of the literature.

RESULTS

In total, 67 articles were included for qualitative analysis, of which 54 were eligible for detailed data extraction. CSRCC has a reported incidence between 0.1%-2.4% and frequently presents with advanced disease stage at the time of diagnosis. CSRCC is associated with an impaired overall survival (5-year OS: 0%-46%) and a worse stage-corrected outcome compared to mucinous and not otherwise specified adenocarcinoma. The systematic use of exploratory laparoscopy to determine the presence of peritoneal metastases has been advised. Surgery is the mainstay of treatment, although the rates of curative resection in CSRCC (21%-82%) are lower compared to those in other histological types. In case of peritoneal metastasis, cytoreductive surgery with hyperthermic intraperitoneal chemotherapy should only be proposed in selected patients.

CONCLUSION

CSRCC is a rare clinical entity most often characterized by young age and advanced disease at presentation. As such, diagnostic modalities and therapeutic approach should be tailored accordingly.

Core Tip: Although multiple studies concerning colorectal signet-ring cell carcinoma (CSRCC) have been published in the past decades, the pathogenesis, associated risk factors, and potential implications on treatment are still poorly understood. As such, the aim of this systematic review was to provide an updated overview of CSRCC by means of a synthesis of the current literature concerning this topic with emphasis on risk factors and clinical and histopathological features of the disease along with diagnostic and therapeutic features that should be tailored to this specific clinical entity.

INTRODUCTION

Colorectal signet-ring cell carcinoma (CSRCC) is a rare clinical entity that accounts for approximately 1% of all colorectal cancers (CRCs)[1]. According to the latest WHO classification of digestive system tumors, a tumor is defined as signet-ring cell carcinoma (SRCC) if > 50% of the tumor cells have prominent intracytoplasmic mucin, typically with displacement and molding of the nucleus[2].

Although multiple studies concerning this specific topic have been published since it was first described by Laufman and Saphir[3] in 1951, the pathogenesis, associated risk factors, and potential implications on treatment are still poorly understood. The low incidence of this disease, with historically confusing histological criteria, renders comparison between studies difficult to make, especially in the less recent literature[4]. Nevertheless, the rising incidence of CSRCC along with the relatively young age at presentation and associated dismal prognosis, highlight the actual interest in synthesizing the known literature regarding CRSRCC. In this context, we conducted a systematic review of the literature regarding CSRCC with emphasis on risk factors and the influence of the clinical and histopathological features on the management of the disease.

MATERIALS AND METHODS

A literature search in the MEDLINE/PubMed database was conducted according to the PRISMA guidelines[5]. The following search terms were used: ‘Signet ring cell carcinoma’ and ‘colorectal’. Only studies in the English language, published after January 1980, were eligible for inclusion. Studies were screened based on the abstract. Additional studies were retrieved by screening the references of each individual article. The exclusion criteria were the following: Case reports, studies in which the full text was not available, and studies in which SRCC was not clearly defined as a histological entity according to the WHO guidelines. Additionally, studies focusing on SRCC and mucinous adenocarcinoma (MA) as a single group, were excluded. The literature search was performed by Drubay V and Nuytens F. In case of doubt, additional advice was obtained from a third party (Piessen G). A last update of the literature search was performed in November 2020. Studies included in the qualitative synthesis were evaluated for content concerning epidemiology, risk factors, and clinical, diagnostic, histological, and molecular features along with the metastatic pattern and their influence on therapeutic management. Based on these findings, a narrative synthesis was performed. If possible, data on demographic, surgical, pathological, and molecular features as well as survival outcomes were extracted to present a more detailed overview of the literature. No funding was granted for this systematic review.

RESULTS

Study selection

Overall, 476 studies were identified through the database search. An additional 9 studies were identified through other sources. An overview of the study selection using a PRISMA flow diagram is shown in Figure 1. After screening and application of the exclusion criteria, 67 articles were included for qualitative analysis and integration in the narrative review. In total, 54 articles were eligible for detailed data extraction. An overview of the methodological details concerning these articles is summarized in Table 1.

Figure 1.

PRISMA flow diagram.

Table 1.

Overview of articles eligible for data extraction with methodological details

Ref.	Type of article	Time of inclusion	% (SRC/total study population)	Database	Compared to
Tawadros et al[13], 2015	Retrospective cohort study	2004-2010	< 40 yr: 3%, > 40 yr: 0.87% (NA)	SEER	MA and OA of the rectum
Song et al[25], 2017	Retrospective cohort study (pT1 cancer)	1988-2011	62/21463 (0.3%)	SEER	MA and OA
Wei et al[41], 2016	Retrospective case series (SRC component)	11/2008-01/2015	39 group A, 22 group B	NA	Group A (> 50% SRC) and group B (< 50% SRC, but with SRC component)
Wu et al[9], 2018	Retrospective cohort study	1988-2012	4402 colon + 853 rectum/24171 overall SRC	SEER	Other SRCC locations (bladder, breast, esophagus, gallbladder, lung, pancreas, small intestine)
Hartmann et al[16], 2013	Retrospective case series	2002-2012	53/4760 (1.1%)	NA	40 (75%) mucin-rich SRCC vs 13 (25%) mucin-poor SRCC and 23 (43%) MSI-H SRCC vs 30 (57%) MSS SRCC
Pande et al[54], 2008	Retrospective cohort study (SRC component < 50%)	01/2002-05/2006	39/753 (5.1%)	NA	MA, CMA, > 50% SRRC, < 50% SRC component, OA in a random sample size of 100 patients
van Oudheusden et al[59], 2015	Retrospective cohort study	04/2005-12/2013	20/351 (5.7%): CRS + HIPEC in 16/20 (80%)	NA	CRS + HIPEC in SRCC vs CRS + HIPEC in other differentiation (252/331)
Wu et al[10], 2019	Retrospective cohort study	2004-2015	4140/399791 (1.0%)	SEER	MA and NMA
Hyngstrom et al[19], 2012	Retrospective cohort study	1998-2002	2260/244794 (1%)	NA	MA and OA
Hugen et al[17], 2015	Retrospective cohort study	1989-2010	1972/196757 (1.0%)	NA	MA and OA
Shi et al[61], 2019	Retrospective cohort study	2010-2014	1932/173460	SEER	Non-SRCC
Kang et al[12], 2005	Retrospective cohort study	1991-2000	1522/164628 (0.9%)	NA	MA and OA
Ciarrocchi[4], 2014	Retrospective cohort study	2004-2009	1484	NA	Rectal SRCC vs non-rectal colonic SRCC
Huang et al[26], 2016	Retrospective cohort study (<35/>35 y with SRCC)	01/1988-12/2011	NA	SEER	NA
Ling et al[50], 2017	Retrospective cohort study	1988-2011	622/69543 (0.9%)	NA	MA and OA of the rectum
Simkens et al[38], 2016	Retrospective cohort study	01/2005-12/2014	385/5516 (7%)	NA	MA and OA with peritoneal metastases
Razenberg et al[62], 2015	Retrospective cohort study (peritoneal metastases +)	01/2005-12/2012	311/4277 (7%)	NA	MA and OA with peritoneal metastases
Wu et al[11], 2017	Retrospective case series	1988-2012	292 (stage II and III rectal SRCC)	NA	NA
Thota et al[30], 2014	Retrospective cohort study	01/1995-12/2008	206/36260 (0.6%)	VACCR	MA and NMA
Tamhankar et al[51], 2016	Retrospective case series	01/2011-12/2013	170/1487 (11.4%)	NA	NA
Nitsche et al[33], 2013	Retrospective cohort study	1998-2012	160/28056 (0.6%)	SEER	MA and OA
Fu et al[58], 2016	Retrospective cohort study	1998-2010	94/3568 (2.6%)	NA	Resectable metastatic SRRC vs resectable metastatic non-SRCC
Kakar and Smyrk[44], 2005	Retrospective case series	1985-2000	72	NA	NA
Yun et al[63], 2017	Retrospective cohort study (+ propensity score matching)	09/1994-12/2013	71/12631 (0.56%)	NA	OA
Sung et al[21], 2008	Retrospective cohort study	01/1995-12/2006	65	NA	MA
Chen et al[6], 2004	Retrospective cohort study	01/1979-07/2001	61/2619 (2.3%)	NA	MA and NMA of the rectum
Wang et al[43], 2016	Retrospective case series	09/2008-07/2014	59/6625 (0.89%)	NA	NA
Nissan et al[18], 1999	Retrospective case matched study/cohort study	1986-1997	40/5350 (0.7%)	NA	CG/non-signet-ring CRC
Liang et al[31], 2018	Retrospective case series	01/1990-12/2010	37/2643 (1.4%)	NA	NA
Korphaisarn et al[46], 2019	Retrospective cohort study (metastatic CRC)	03/1994-11/2015	35/665 (5.3%), 28(4.2%) SRC component	NA	OA with SRC component and OA
Lee et al[52], 2007	Retrospective cohort study	10/1996-12/2005	35/5022 (0.7%)	NA	MA and poorly differentiated OA
Bittorf et al[8], 2004	Retrospective cohort study	01/1978-12/1999	34/4458 (0.8%)	NA	Non-SRCC
Messerini et al[32], 1995	Retrospective case series	01/1985-12/1993	34/2995 (1.1%)	NA	NA
Bademci et al[28], 2019	Retrospective case series	2009-2018	34/3180 (1.06%)	NA	NA
Chua et al[53], 2009	Retrospective case series (peritoneal carcinomatosis)	01/1997-12/2008	33	NA	NA
Kakar et al[45], 2012	Retrospective cohort study	NA	33	NA	MA and OA
Barresi et al[27], 2016	Retrospective case series	2003-2014	32/3248	NA	NA
Chew et al[22], 2010	Retrospective cohort study	01/1999-12-2005	30/2764 (1.1%)	NA	MA and OA
Nitsche et al[1], 2016	Retrospective cohort study	01/1982-10/2012	30/3479 (0.9%)	NA	MA and OA
Anthony et al[7], 1996	Retrospective case series	09/1971-12/1993	29/3690 (0.8%)	NA	NA
Song et al[64], 2009	Retrospective cohort study	08/1994-03/2007	25/2079 (1.2%)	NA	MA and NMA
Belli et al[20], 2014	Retrospective case series	03/2007-06/2013	22/842 (2.6%)	NA	NA
Zhang et al[65], 2020	Retrospective cohort study	2006-2013	21 vs 90	NA	MA
Wang et al[39], 2019	Retrospective cohort study	2004-2013	21/90 (SRC/SRC + MAC)	NA	MA
Song et al[66], 2019	Retrospective cohort study	2000-2010	20/299	NA	Predominant SRCC (> 50% SRCC) vs mixed SRC vs cluster predominant vs strip predominant
Mizushima et al[36], 2010	Retrospective cohort study	1993-2007	19/5884 (0.32%)	NA	Well or moderately differentiated OA and poorly differentiated MA
Foda et al[42], 2018	Retrospective cohort study	01/2007-12/2011	19/150 (12.6%)	NA	MA, OA, OA with mucinous component
Lee et al[29], 2015	Retrospective case matched study	01/2003-12/2011	15 vs 75 NMA	NA	NMA
Imai et al[67], 2013	Retrospective cohort study	04/1998-03/2000	15/250 (6%)	NA	Well or moderately differentiated OA, poorly differentiated OA and MA
Psathakis et al[15], 1999	Retrospective case matched study/cohort study	01/1979-12/1997	14/1600 (0.88%)	NA	Non-SRCC
Pozos-Ochoa et al[37], 2018	Retrospective cohort study	1995-2015	12	NA	MA with SRC component
Sasaki et al[14], 1998	Retrospective cohort study	01/1963-12/1996	11/1595 (0.7%)	NA	Mucinous adenocarcinoma (KRAS mutation screening in 30 control advanced colorectal carcinomas)
Ooi et al[23], 2001	Retrospective case series	1989-1999	9/3000 (0.3%)	NA	NA
Secco et al[68], 1994	Retrospective cohort study	01/1979-12/1986	4/352 (1.1%)	NA	MA and OA

SRC: Signet-ring cell; SRCC: Signet-ring cell cancer; MA: Mucinous adenocarcinoma; OA: Ordinary/conventional adenocarcinoma; NMA: Non-mucinous adenocarcinoma; CMA: Component of mucinous adenocarcinoma; CRS: Cytoreductive surgery; HIPEC: Hyperthermic intraperitoneal chemotherapy; MSI-H: Microsatellite instability-high; MSS: Microsatellite stable; NA: Not applicable; CG: Control group; SEER: Surveillance, Epidemiology and End Results program database; VACCR: Veterans Affairs Central Cancer Registry; CRC: Colorectal cancer; MAC: Mucinous adenocarcinoma.

Epidemiology and risk factors

CSRCC has a reported incidence between 0.1%-2.4%[6,7]. Significantly higher incidence rates of up to 5% in Egypt and 18.5% in Jordan and Lebanon have been reported. Potential overlap in histologic diagnosis of MA with CSRCC, tertiary center referral bias, and nutritional (dietary habits) as well as hereditary factors have been suggested as potential reasons for these isolated differences in incidence rate[8]. After the stomach (63.4%), the colon is the second most affected organ by SRCC histology (18.2%), followed by the esophagus (5%) and the rectum (3.5%)[9-11]. Interestingly, in a report based on the Surveillance, Epidemiology and End Results program database (SEER), Kang et al[12] showed that the incidence of CSRCC has increased by 0.2% from 1990 till 2000. In a multivariate analysis of rectal SRCC, Tawadros et al found that especially in patients under the age of 40, the incidence has risen by 3%, entailing a 3.6-fold increased risk of CSRCC compared to the general population[13]. Higher detection rates and more widespread implementation of the histological criteria for CSRCC, along with an increased diagnostic rate of SRCC-related diseases such as inflammatory bowel disease (IBD), are likely to be responsible for this observed increase in the incidence rate[12]. Several authors have indeed identified IBD and more specifically ulcerative colitis as a potential risk factor for CSRCC[13,14]. One study mentioned that 14.3% of patients with CSRCC had a history of IBD[15]. Hartman et al[16] found that IBD-associated intestinal cancers displayed significantly more signet-ring cell (SRC) characteristics than other subtypes of adenocarcinoma (OA; 28% vs 4%, P < 0.001).

Historically, SRCC has been associated with young age at presentation, which is confirmed by more recent reports[6,8,12,17]. Several other authors, however, were unable to demonstrate any association between CSRCC and younger age[4,15,18]. Median age at presentation varies from 40 to 70 years old. Small study samples are probably accountable for this major variability. Ciarrocchi[4] showed that especially in the case of rectal SRCC, younger patients were more frequently affected. Psathakis et al[15] stated that although the proportion of CSRCC in young patients is high, the overall median age and range are similar to those of OA.

Concerning gender distribution, the literature is also characterized by major heterogeneity. Most authors mentioned a male predominance or no significant difference at all[6,17,19,20]. One report showed that in patients with rectal SRCC, the male-to-female ratio was 2:1[4]. In another study, a multivariate analysis identified female gender as an independent prognostic factor for worse OS in CSRCC compared to patients with MA[21].

Concerning racial differences, only Tawadros et al[13] found that patients with rectal SRCC were more frequently from another race than black or white. In a review comparing Asian and Western data of patients with CSRCC, however, no significant differences in demographic data could be withheld[22].

With regard to genetic predisposition, most authors found no correlation between CSRCC and a positive family cancer history[15,23,24]. Whether this is the result of small study samples or because of current clinicopathological features, remains to be determined[24]. In a study sample of 61 patients, only Song et al[25] found a positive family cancer history in 49.2% of the patients. In half of these cases, it concerned gastric cancer or CRC.

Out of all of the above-mentioned risk factors, only young age and male gender were described as negative prognostic factors[26,27]. Patients with CRSCC under the age of 35 display significantly worse cause-specific survival (CSS) as compared to older patients. More aggressive tumor biology in this specific subset of young patients, as well as a delay in diagnosis, caused by a low index of suspicion, both for patients and physicians, have been hypothesized as potential reasons for these findings[26].

Clinical and diagnostic features

Most studies including a large number of patients observed that CSRCC is predominantly located in the proximal (right) colon and rectum[4,8,17,18]. Only a few authors report a more frequent location in the left hemi-colon or rectosigmoid, mostly in studies containing smaller numbers of patients[20,28,29]. In rectal SRCC, tumors are more frequently located at the level of the lower rectum, which results in a higher rate of abdominoperineal resection[6].

Table 2 shows an overview of the literature reporting on tumor characteristics at the time of diagnosis. CSRCC is frequently characterized by the presence of advanced disease stage at the time of diagnosis, including large tumor size and high rates of lymph node- and peritoneal metastasis[6,7,15,17]. At the time of diagnosis, the mean tumor size varies from 4.6 to 8 cm, with the vast majority of patients (61%-100%) already being affected by stage III or IV disease[7,13,14,23,30]. As a consequence, patients with CSRCC more often require emergency procedures[8,15]. A change in bowel habits has been described as the most frequent presenting symptom in two small studies[31,32]. Ooi et al[23] on the other hand found that rectal bleeding and bowel obstruction were the two most common presenting symptoms, both in 33% of the cases.

Table 2.

Overview of articles reporting on tumor characteristics at the time of diagnosis

Ref.	Mean tumor size (cm)	LNM	Stage III or IV	% M+
Tawadros et al[13], 2015	5.2 vs 4.8 MA vs 4 OA (P < 0.001)	60.8% vs 51.4% MA vs 35% OA (P < 0.001)	NA	24.3% vs 15.2% MA vs 12.3% OA (P < 0.001)
Song et al[25], 2017	NA	35.5% vs 11.6% MA vs 18.4% AC (P < 0.001)	NA	NA
Wei et al[41], 2016	NA	NA	54/61 (88.5%)	NA
Wu et al[9], 2018	NA	NA	NA	NA
Hartmann et al[16], 2013	NA	NA	57% mucin-rich vs 100% mucin-poor (P = 0.005); 44% MSI-H vs 87% MSS (P = 0.0012)	NA
Pande et al[54], 2008	NA	NA	74.2% vs 65% MA vs 70.5 CMA vs 63% OA (P = 0.75)	17/39 (43.6%) vs 30% MA vs 32.4% CMA vs 35% OA
van Oudheusden et al[59], 2015	NA	N2: 62.5% vs 36.1% (P = 0.04)	NA	NA
Wu et al[10], 2019	NA	63.3% vs 45.1% MA vs 36.1% OA (P < 0.001)	3231 (78%) vs 52.3% MA vs 44.7% OA (P < 0.001)	1595 (38.5%) vs 20.3% vs 19% (P < 0.001)
Hyngstrom et al[19], 2012	NA	NA	80% vs 52% MA vs 44% OA (P < 0.01)	NA
Hugen et al[17], 2015	NA	NA	75.2% vs 47.7% (MA) vs 43.6% OA P < 0.0001	NA
Shi et al[61], 2019	NA	60.8% vs 35.1% (P < 0.001)	76.04 vs 44.05% (P < 0.001)	39.13% vs 19.08% (P < 0.001)
Kang et al[12], 2005	NA	NA	80.9 % vs 52.8% (MA; P < 0.0001) vs 49.5% OA (P < 0.0001)	NA
Ciarrocchi[4], 2014	5.6 ± 3.1 (5.4 ± 4.9 rectum vs 5.7 ± 2.7 colon; P = 0.235)	NA	1123 (76%) (144 (72%) vs 979 (76%; P = 0.193)	NA
Huang et al[26], 2016	> 4 cm: 67.9%	74.90%	III: 74.9%	NA
Ling et al[50], 2017	> 5 cm: 45.3% vs 43.2% MA vs 31.1% OA (P = 0.372 vs MA. P < 0.001 vs OA)	NA	Stage III: 79.4% vs 57.1 MA (P < 0.001) vs 52.8 OA (P < 0.001)	NA
Simkens et al[38], 2016	NA	NA	NA	NA
Razenberg et al[62], 2015	NA	62% vs 60% MA vs 59% OA. Nx: 30% vs 25% MA vs 28% OA (P < 0.0001)	100%	NA
Wu et al[11], 2017	53.5 ± 25.1	NA	Stage III: 241/292 (82.5%)	NA
Thota et al[30], 2014	NA	46.5% vs 38% MA vs 27% NMA. Unknown: 32.5% vs 14% MA vs 29% NMA	61.2% vs 44.6% MA vs 44.5% NMA. Unknown: 22.3% vs 16% MA vs 32% NMA	NA
Tamhankar et al[51], 2016	NA	NA	91.20%	39.40%
Nitsche et al[33], 2013	NA	71% vs 44% (P < 0.001)	86% vs 48% (P < 0.001)	35% vs 18% (P < 0.001)
Kakar and Smyrk[44], 2005	> 5 cm: 52/72 (72%)	NA	55/72 (76.4%)	NA
Yun et al[63], 2017	NA	NA	69% vs 43% (P < 0.001)	NA
Sung et al[21], 2008	6.64 ± 2.4 vs 10.35 ± 55.2 MA (P = 0.608)	57/65 (88%) vs 130/266 (49%) (P < 0.0001)	58/65 (89%) vs 138/266 (52%) MA (P < 0.0001)	NA
Chen et al[6], 2004	5.7 ± 3.8 vs 4.3 ± 1.8 MA vs 3.8 ± 1.7 (NMA; P < 0.001)	NA	90% vs 69% (MA; P = 0.013) vs 49% (NMA; P < 0.001)	25/61 (41%) vs 44/144 (30%; MA) vs 311/2414 (13%) NMA
Wang et al[43], 2016	NA	52/59 (88.1%)	NA	14/59 (23.7%)
Nissan et al[18], 1999	NA	NA	60% vs 60% CG vs 45% CRC	30%
Liang et al[31], 2018	NA	26 (70.3%)	33/37 (89.2%)	16/37 (43%)
Korphaisarn et al[46], 2019	NA	NA	NA	100%
Lee et al[52], 2007	NA	NA	79.9% vs 55% MA vs 64.7% poorly differentiated OA (P = 0.003)	37.1% vs 15% MA vs 25% poorly differentiated OA
Bittorf et al[8], 2004	NA	NA	20/24 (83%) resected specimens vs 2089/4231 (49%) resected specimens (P = 0.002)	15/34 (44%) vs 912/4458 (21%; P = 0.002)
Messerini et al[32], 1995	> 5 cm: 22/34 (64.7%)	NA	Dukes stage C: 21/34 (61.8%). D: 3/34 (8.8%). Combined: 24/34 (70.6%)	NA
Bademci et al[28], 2019	4.6 ± 1.8	17 (68%)	22 (64.7%)	NA
Chua et al[53], 2009	NA	Colorectal: 13/15 (87%). Appendiceal: 11/18 (61%)	NA	NA
Kakar et al[45], 2012	NA	NA	79% vs 62% MA vs 34% OA (P < 0.001)	NA
Barresi et al[27], 2016	NA	NA	28/32	NA
Chew et al[22], 2010	NA	25 (89%) vs 101 (61%) MA vs 1320 (52%) OA (P = 0.002)	38 (94%) vs 112 (67%) MA vs 1426 (56%) (P < 0.001)	13 (43%) vs 45 (27%) MA vs 520 (20%) (P = 0.009)
Nitsche et al[1], 2016	NA	NA	26/30 (87%) vs 48% MA vs 55% OA (P < 0.001)	12/30 (40%) vs 25% MA vs 22% OA (P = 0.015)
Anthony et al[7], 1996	6 (0.8-15)	13/29 (45%)	21/29 (72%)	8/29 (28%)
Song et al[64], 2009	5.52 ± 3.36 vs 5.52 ± 1.823 MA vs 4.62 ± 268 NMA (P < 0.001)	18 (72%) vs 59% MA vs 51% NMA	20 (75%) vs 59% MA vs 52% NMA (P < 0.01)	NA
Belli et al[20], 2014	7 ± 3.6	NA	20/22 (91%)	10/22 (45%)
Zhang et al[65], 2020	SRCC + MA: > 5 cm: 50%	SRCC + MAC: 74.4%	SRCC + MAC: 83.3%	SRCC + MAC: 23.3%
Wang et al[39], 2019	NA	NA	NA	38% vs 20%
Song et al[66], 2019	NA	80% vs 65.1 mixed vs 47.1 cluster vs 32.1% strip (P < 0.001)	85% vs 65.1% mixed vs 48.4% cluster vs 35.8% strip (P = 0.006)	15% vs 7.0% mixed vs 9.0% cluster vs 6.2% strip (P = 0.603)
Mizushima et al[36], 2010	7.2 ± 3.8 vs 4.4 ± 2.2 (well/moderately differentiated) vs 5.6 ± 2.7 (poorly differentiated/MA)	14/19 (73.7%)	16/19 (84.21%) (SRC differentiation)	7/19 (36.8%) vs 14.5% well/moderately differentiated vs 29.9% (poorly differentiated/mucinous) (SRC differentiation)
Foda et al[42], 2018	NA	84.2% vs 57.2 MA vs 60.7% OA with mucinous component vs 42.6% OA	84.2% vs 57.1% MA vs 60.7% AWNC vs 44.7% OA	NA
Lee et al[29], 2015	7.59 ± 2.9 vs 5.37 ± 2.19 (P = 0.003)	12/15 (80%) vs 60/75 (80%; P = 0.994)	80% vs 80%	NA
Imai et al[67], 2013	NA	9/15 (64.3%) vs 46.8% MA vs 78.4% poorly differentiated OA vs 54% well or moderately differentiated OA	11/15 (73.3%) vs 50% vs 81.1% vs 54%	NA
Psathakis et al[15], 1999	NA	NA	13/14 (92.8%) vs 30/56 (50%; P < 0.05)	NA
Pozos-Ochoa et al[37], 2018	NA	75% vs 50% (P = 0.040)	91.7% vs 75% (P = 0.248)	NA
Sasaki et al[14], 1998	> 8 cm: 11/11 (100%) vs 6/29 (21.4%; P < 0.00001)	NA	9/22 (81.8%) vs 17/29) (58.6%) (stages I-III/stage IV P < 0.01)	NA
Ooi et al[23], 2001	NA	NA	Dukes stage C or D: 9/9 (100%)	2/9 (22%)
Secco et al[68], 1994	NA	NA	Duke C/D: 100% vs 56.4% MA vs 42.4% OA	NA

LNM: Lymph node metastases; SRCC: Signet-ring cell cancer; MA: Mucinous adenocarcinoma; OA: Ordinary/conventional adenocarcinoma; NA: Not applicable; NMA: Non-mucinous adenocarcinoma; CMA: Component of mucinous adenocarcinoma; MAC: Mucinous adenocarcinoma; CRC: Colorectal cancer; AWNC: Adenocarcinoma with mucinous component.

Locally advanced and metastatic disease at the time of diagnosis has been attributed to the specific clinical course and a more frequent proximal location of CSRCC, resulting in a later onset of symptoms as compared to left-sided colon cancer[8,15]. Nitsche et al[33] stated that other histological types of tumors of the right hemi-colon are not associated with such a delay in diagnosis. Because of a primarily submucosal infiltration pattern in SRCC, a fecal occult blood test is often negative[34]. Moreover, the submucosal infiltration associated with the histopathological characteristics of SRCC is translated into an endoscopic and radiographic morphology that frequently mimics that of IBD[34]. In this context, endoscopic deep-tissue biopsies or even endoscopic-ultrasound-guided biopsies have been recommended to obtain adequate histological proof[35]. Because of no adenoma-carcinoma sequence in early CRSCC, meticulous endoscopic inspection for subtle superficial lesions has been proposed[34]. Despite the importance of early diagnosis for CSRCC, Sasaki et al[14] found that the implementation of this recommendation into clinical practice appears to be difficult since the average duration between symptoms and diagnosis was only 2.8 mo.

Because of the increased risk of peritoneal carcinomatosis (PC) in the case of CSRCC, other authors have suggested the systematic use of a staging laparoscopy with peritoneal lavage and cytology to determine the presence of peritoneal (micro) metastases, especially since the sensitivity of CT scan for PC in this context tends to be low[35,36]. Few studies have evaluated the role of carcinoembryonic antigen (CEA) as a potential marker for CSRCC. Except in one small study containing 12 patients, no significant difference in CEA levels could be identified between CSRCC and OA[22,29,37].

Histologic and molecular features

According to the latest WHO classification of digestive system tumors, adenocarcinoma of the colon is defined as CSRCC if > 50% of the tumor cells have prominent intracytoplasmic mucin, typically with displacement and molding of the nucleus[2]. Carcinomas with SRCs in < 50% of the tumor are categorized as having a SRC component. SRCCs are now classified as poorly cohesive carcinomas.

The term ‘linitis plastica’ finds its origin as a macroscopic description for a sclerotic and rigid stomach wall and has little to no relevance in the context of CSRCC. However, it has often been confounded with MA and SRCC[7]. In reality, only 16%-20% of all ‘linitis plastica’ cases involve SRCC[32]. In general, CSRCC is associated with poor differentiation: 51%-90% of CSRCCs are classified as poorly differentiated or undifferentiated[12,26,38,39]. Although the cut-off for CSRCC has been set at > 50% SRCs, Ogino et al[40] have stated that tumors with a SRC component less than 50% display molecular features that are similar to those that contain more than 50% of SRCs. This similarity seems to be of clinical importance since another study found that the dismal prognosis in patients with tumors containing > 50% of SRCs did not differ from that of patients with tumors that had less than 50% of SRCs[41]. As a consequence, it has been recommended that the pathology report should mention any SRC component, even minor, as this could be relevant for the subsequent treatment and prognosis. Additionally, Hartman et al[16] suggested taking into account whether the SRCs are either mucin-rich or -poor: Although most studies have demonstrated that CSRCC, in general, is associated with a high rate of lymphovascular invasion (LVI), Hartman et al[16] found that even after adjustment for stage, especially mucin-poor CSRCC is associated with a high degree of LVI, more advanced stages, and worse OS[8,16,21,42]. An overview of studies describing the rate of LVI invasion or molecular features is listed in Table 3.

Table 3.

Overview of studies describing rate of lymphovascular invasion or molecular features

Ref.	LVI +	MSI-H	KRAS	E-cadherin expression
Wei et al[41], 2016	NA	NA	9/54 (16.7%) KRAS. 3/54 (5.6%) NRAS. 2/54 (3.7%) BRAF
Hartmann et al[16], 2013	L: 55% mucin-rich vs 100% mucin-poor (P = 0.002); 56% MSI-H vs 73% MSS (P = 0.25); V: 8% mucin-rich vs 46% mucin-poor (P = 0.004); 4% MSI-H vs 27% MSS (P = 0.06)	23/53 (43%)	16/53 (30%)	NA
Nitsche et al[33], 2013	L: 79% vs 36% (P < 0.001). A: 38% vs 18.37% (P < 0.001)	20% vs 30% (P = 0.644), 155/160 and 15368/25172 missing data	NA	NA
Kakar and Smyrk[44], 2005	NA	22/72 (31%)	NA	NA
Yun et al[63], 2017	L: 60.6% vs 24.8% (P < 0.001). V: 36.6% vs 13.4% (P < 0.001)	NA	NA	NA
Sung et al[21], 2008	Vascular invasion: 58/65 (89%) vs 137/266 (52%; P < 0.0001)	12/63 (19%) vs 21/95 (22%)	NA
Wang et al[43], 2016	NA	NA	NA	23/59 (39%) E-cadherin+ vs 36 (61%) E-cadherin- (P = 0.010)
Nissan et al[18], 1999	Blood vessel invasion: 7.5% vs 22.5% vs 11.9%; LVI: 42.5% vs 2.5% vs 3.5% (P < 0.001); neural invasion: 17.5% vs 7.5% control group vs 4.7% CRC (P < 0.001)	NA
Korphaisarn et al[46], 2019	NA	Deficient MMR: 12.1% vs 9.5% OA with SRC component vs 4.9% (P = 0.10)	Wild type: OR = 7.7, 95%CI: 2.7-22 (P < 0.001)	NA
Lee et al[52], 2007	56% vs 17.4% MA vs 43.3% poorly differentiated OA (P < 0.001)	NA	NA	NA
Bittorf et al[8], 2004	V: 75% vs 34 % and L: 92% vs 54% (P < 0.001)	NA
Bademci et al[28], 2019	22 (88%)	NA	NA	NA
Kakar et al[45], 2012	NA	24% vs 27% MA vs 12% OA (P = 0.52)	52% vs 27% MA vs 40% OA (P = 0.04)	NA
Barresi et al[27], 2016	17/32	7/32	NA	NA
Chew et al[22], 2010	Vascular emboli: 18 (60%) vs 33 (20%) MA vs 587 (23%; P < 0.001)	NA
Nitsche et al[1], 2016	L: 20/30 (67%) vs 23% vs 25% (P < 0.001). Angio-invasion: 5/30 (17%) vs 5% vs 9% (P = 0.18)	1/2 (50%) (P = 0.006) (data only on 251 patients available)
Belli et al[20], 2014	19/22 (86%)	NA
Zhang et al[65], 2020	SRCC + MA: 45.5%	NA	NA	NA
Song et al[66], 2019	75% vs 30.2% mixed vs 21.3% cluster vs 8.6% strip (P < 0.001)	11.1% vs 34.2% mixed vs 23.7% cluster vs 34.2% strip (P = 0.110)	NA	NA
Foda et al[42], 2018	89.5% vs 62.5% MA vs 60.7% OA with mucinous component vs 59.6% OA	NA	NA	26.3% E-cadherin + vs 78.6% MA vs 85.7% OA with mucinous component vs 89.4% OA
Lee et al[29], 2015	7/15 (46.7%) vs 25/75 (33.3%) (P = 0.336)	3/11 (27.3%) vs 10/6 (16.7%; P = 0.486)
Pozos-Ochoa et al[37], 2018	58.3% vs 65% (P = 0.706)	NA	NA	NA
Sasaki et al[14], 1998	NA	NA	4/11 (36.4%) vs 11/29 (37.9%) vs 18/30 (60%) OA

LVI: Lymphovascular invasion; L: Lymphatic invasion; V: Vascular invasion; SRC: Signet-ring cell; SRCC: Signet-ring cell cancer; MA: Mucinous adenocarcinoma; OA: Ordinary/conventional adenocarcinoma; MSI-H: Microsatellite instability-high; MSS: Microsatellite stable; NA: Not applicable; CRC: Colorectal cancer.

Several molecular features have been identified to be responsible for the high infiltrative potential of CSRCC, resulting in high rates of LVI and lymph node- and peritoneal metastasis: First of all, the mucopolysaccharide produced by SRCC tends to inhibit discrimination by host immunocytes, resists encapsulation, and thus favors LVI and peri-intestinal infiltration[22,31]. Second, only 26%-39% of CSRCCs retain E-cadherin expression[42,43]. As E-cadherin plays an important role in cell-to-cell adhesion, an increased potential for invasion and metastasis is observed in its absence[22,33]. A retrospective study among 59 patients with CSRCC demonstrated that loss of E-cadherin expression was associated with significantly worse survival[41]. Finally, a higher activity of acid phosphatase and β-glucuronidase in CSRCC is associated with an increased disintegration of the fibrous tissues surrounding the tumor[31].

Several studies have found a significantly higher rate of microsatellite instability (MSI; 30-43%) in CSRCC, compared to OA[29,31]. An association between MSI and the more frequent right-side location of CSRCC has previously been suggested[33]. Parallel to the characteristics of MSI in OA, an association between MSI CSRCC and a more frequent right-sided location, larger tumor size, older age, and female gender has been demonstrated[21,44]. In contrast, however, to MSI characteristics in OA, the majority of recent studies were unable to identify MSI as a favorable prognostic factor for survival in CSRCC, even when adjusted for stage[21,27,44]. A BRAF mutation is identified in 22%-30% of CSRCCs and was associated with an impaired 5-year survival in the subgroup of microsatellite stable tumors[40,45]. In a study by Kakar et al[45], a BRAF mutation was found in 43% of MSI CSRCC tumors. In the same study, the authors hypothesized that the high rate of loss of heterozygosity observed in MSI CSRCC (80%) could counterbalance the potential beneficiary effect of MSI on survival. The presence of KRAS mutations in 52% of patients was not associated with any specific clinical feature or change in oncological outcome[45]. Two other studies, on the other hand, found a lower rate of KRAS-mutations in CSRCC (11.4%-36.4%) compared to OA (60%)[14,46]. Since a lower rate of KRAS mutations is presumably associated with a non-polypoid carcinoma etiology, as frequently observed in IBD-associated tumors, Sasaki et al[14] and Kim et al[47] suggested that the carcinogenesis in CRSCC is different from that of OA. The identification of IBD as a risk factor for CSRCC, as stated above, seems to support this hypothesis. Moreover, CSRCC has more molecular features (positivity for trefoil factor and MUC2, E-cadherin negativity, and Bcl-2 amplification) in common with SRCC of, for instance, the stomach than it has with OA[33]. In an attempt to synthesize the different molecular characteristics, one study found two distinct molecular profiles of CSRCC: An MSI/CIMP (CpG Island Methylator Phenotype)/BRAF-mutated/PD-L1+ hypermethylated genotype, which accounts for most of the tumors found in the proximal colon and a hypomethylated genotype which is more often found in the distal colon[48].

Finding the primary tumor in the case of an SRCC with diffuse PC can sometimes pose a challenge, especially if endoscopy is not conclusive. Immunohistochemical markers such as CK20 or CDX2 are not reliable to differentiate whether the primary SRCC is located in the lower or upper gastrointestinal (GI) tract in addition to the fact that these markers are frequently negative for SRCC. In this context, Ma et al[49] demonstrated that the special AT-rich sequence binding protein (SATB2) alone and in combination with CDX2 is more frequently expressed in SRCC of lower GI tract origin.

Metastatic pattern and prognosis

As already mentioned, CSRCC is more frequently diagnosed at an advanced stage (61%-100% stage III or IV) compared to MA or OA[10,13,15,30]. Ciarrocchi[4] found that local invasiveness was more frequent in patients with rectal tumors, while the rate of metastatic disease was higher in colon tumors. The extended local invasion is reflected by the higher rate of circumferential resection margin positivity (19%) in rectal SRCC compared to rectal tumors involving other histological types (4%)[50]. Given its histopathological features, CSRCC predominantly metastasizes to the peritoneum rather than the liver or lung, as is the case in OA. In stage IV CSRCC, 43%-86% of patients are affected by PC while liver metastases occur only in 5%-14% of cases[22,29,51-53]. Ovarian metastases are found in 4.8%-7.2% of cases[8,15]. Table 4 shows an overview of the literature reporting on survival outcomes for CSRCC. Although more favorable in comparison to gastric or pancreatic SRCC, the overall prognosis of CSRCC is still poor, with most studies reporting a 5-year OS rate between 0%-46% and a median survival between 8 and 48 mo[4,8,10,11,54]. CSS at 5 years varies from 11% to 52%[1,22,26]. Once again, Hartman et al[16] demonstrated the importance of differentiating between mucin-rich vs -poor CSRCC since mucin-rich tumors were associated with an OS of 66% whereas this was 19% in mucin-poor tumors. Most studies found that SRCC histology has a significantly increased risk of recurrence after curative treatment, with rates varying between 33% and 79%[7,22,29,33]. Chew et al[22] were unable to demonstrate any difference in locoregional or systemic recurrence rates between CSRCC and MA or OA. In the study by Ciarrocchi, after adjustment for confounding factors, rectal SRCC was found to have a worse prognosis compared to non-rectal SRCC (hazard ratio [HR] = 0.79)[4]. These findings, however, have to be interpreted with caution, since the SEER database on which the latter study was based did not include a standardization concerning adjuvant treatment for rectal cancer.

Table 4.

Overview of articles reporting on survival outcomes for colorectal signet-ring cell carcinoma

Report	5-yr OS (SRC vs other)	5-yr OS in stage III tumors	5-yr DFS % (SRC vs other)
Song et al[25], 2017	5-yr CSS: 85.6% (colon: 89.2%, rectum: 73.3%) vs 90.7% MA (P = 0.17) (colon: 90.5%; rectum: 91.6%) vs 93.3% AC (P = 0.013) (colon: 93.8%; rectum: 92.2%)	NA	NA
Wu et al[9], 2018	5-yr CSS: Colon: 33.2%; rectum: 28.1%	NA	NA
Hartmann et al[16], 2013	66% mucin-rich vs 19% mucin-poor (P = 0.0035); 75% MSI-H vs 48% MSS	NA	66% mucin-rich vs 0% mucin-poor (P = 0.0001); 73% MSI-H vs 31% MSS
van Oudheusden et al[59], 2015	3-yr OS: 0% vs 44%	NA	NA
Wu et al[10], 2019	26.8% vs 50.6% MA vs 60.2% OA (HR: 1.592, 95%CI: 1.558-1.627, P < 0.001)	NA	NA
Hyngstrom et al[19], 2012	NA	Relative survival: Colon IIIa: 0.81 vs 0.87 MA vs 0.86 OA; IIIb: 0.49 vs 0.66 MA vs 0.65 OA; IIIc: 0.21 vs 0.42 MA vs 0.45 OA; rectum IIIa: 0.91 vs 0.82 MA vs 0.86 OA; IIIb: 0.55 vs 0.56 MA vs 0.64 OA; IIIc: 0.3 vs 0.44 MA vs 0.48 OA	NA
Hugen et al[17], 2015	Relative survival: Colon: 31% vs 58% MA vs 57% OA; rectum: 20% vs 54% MA vs 57% OA	Relative survival: Colon: 36% vs 54% MA vs 57% OA; rectum: 25% vs 45% MA vs 54% OA	NA
Kang et al[12], 2005	26.8% vs 58.1% MA (P < 0.0001) vs 62.9% OA (P < 0.0001)	IIIa: 68.3% vs 87.3% MC vs 83.8% OA; IIIb: 46% vs 64.3% MC (P = 0.0002) vs 63.6% OA (P <0.0001); IIIc: 19% vs 29.2% MC (P <0.0001) vs 30% OA (P <0.0001)	NA
Ciarrocchi[4], 2014	32% (32% rectum vs 33% colon)	NA	NA
Huang et al[26], 2016	5-yr CSS: 52.2% (< 35 yr poorer CSS than > 35 yr, P < 0.001 at cutoff of 30 y, P < 0.001 at cutoff of 35 y, P = 0.015 at cutoff 40 yr)	NA	NA
Ling et al[50], 2017	NA	34.% vs 53.9% (MA) vs 63.1% (OA; P = 0.001)	NA
Wu et al[11], 2017	39%	NA	NA
Thota et al[30], 2014	NA	19% vs 41% MA vs 47% (P = 0.0002)	NA
Nitsche et al[33], 2013	40.3% (95%CI: 32.1-48.5) vs 58.7% (95%CI: 58.0-59.3)	NA	NA
Fu et al[58], 2016	9.66% vs 24.04% (non-SRCC)	NA	NA
Kakar and Smyrk[44], 2005	41% MSI-H vs 34% MSS (P = 0.3)	NA	NA
Sung et al[21], 2008	3-yr CSS: 33% vs 74% MA (P < 0.0001)	NA	NA
Chen et al[6], 2004	23.3 % vs 40.3% (MA) vs 54.5 % (NMA) (P < 0.001)	32.5% vs 38.3% (MA) vs 41.5% (NMA)	NA
Wang et al[43], 2016	3-yr OS: 62.7%	NA	NA
Liang et al[31], 2018	10.80%	Mean survival time in stage III tumors: 37.1 ± 3.9 (22-80) mo	NA
Lee et al[52], 2007	25.3% vs 68.6% MA vs 53.7% poorly differentiated OA (P < 0.001)	14.5% vs 61.5% MA vs 46.3% (P < 0.001)	12.3% vs 59.9% MA vs 46.8% poorly differentiated OA (P < 0.001)
Bittorf et al[8], 2004	14.7% vs 53.6% (P < 0.001)	20.8% vs 62.6%	NA
Messerini et al (1995)[32]	9.10%	NA	NA
Bademci et al[28], 2019	NA	NA	16 (47%) disease-free during follow-up
Kakar et al[45], 2012	33% vs 50% MA vs 63% (P = 0.09)	III/IV: 27% vs 43% MA vs 57% OA (P = 0.03)	NA
Chew et al[22], 2010	5-yr CSS: 11.1% (0-22.9%) vs overall 57.4% (55.4%-59.4%; P <0.001)	NA	No statistical difference (P = 0.356)
Nitsche et al[1], 2016	5-yr CSS: 21% ± 8% vs 61% ± 3% MA vs 67% ± 1% OA (P < 0.001)	5-yr CSS for stage III: 15% ± 10% vs 60% ± 5% MA vs 65% ± 2% OA (P < 0.001)	NA
Anthony et al[7], 1996	22%	NA	NA
Song et al[64], 2009	27% vs 51% MA vs 69% (P < 0.01)	NA	NA
Zhang et al[65], 2020	31.3% ± 12.9% vs 58.1% ± 7.7% (P = 0.018)	NA	NA
Wang et al[39], 2019	37.3% (95%CI: 14.4%-61.2%) vs 62.9% (95%CI: 48.5%-74.3%; P = 0.021)	NA	NA
Song et al[66], 2019	SRCC predominant MAC with worse survival (P < 0.001)	NA	NA
Mizushima et al[36], 2010	24.1% vs 77.5% well/moderately differentiated vs 57.7% poorly differentiated/MA	NA	NA
Foda et al[42], 2018	10.5% vs 23.2% MA vs 42.9% AWMC vs 55.3% OA	NA	3-yr DFS: 11.1% vs 28.3% MA vs 60.7% OA with SRC component vs 63% OA
Lee et al[29], 2015	46% vs 88.7% (P > 0.001)	NA	NA
Pozos-Ochoa et al[37], 2018	3-yr CSS: 28.6% vs 80.7% (P = 0.017) (II, III, and IV: 100%, 33%, and 0% vs 100%, 78%, and 73%, P = 0.017)	NA	NA
Sasaki et al[14], 1998	0% vs 76.5 % (3 mo-85 mo) (P < 0.0001)	NA	NA
Ooi et al[23], 2001	12%	NA	NA
Secco et al[68], 1994	0% vs 28% MA vs 45% OA	NA	NA

SRC: Signet-ring cell; SRCC: Signet-ring cell cancer; MA: Mucinous adenocarcinoma; OA: Ordinary/conventional adenocarcinoma; NMA: Non-mucinous adenocarcinoma; MSI-H: Microsatellite instability-high; MSS: Microsatellite stable; NA: Not applicable; CSS: Cause-specific survival; AWMC: Adenocarcinoma with mucinous component.

Historically, the question has been asked whether the worse prognosis of CSRCC is related to the advanced stage at the time of diagnosis or whether the tumor biology is inherently more aggressive[7]. Several studies evaluated this topic through stage-per-stage and multivariate analysis as well as using a propensity score matching technique. Although some authors found that SRCC histology was not an independent prognostic factor for survival, most studies including a larger number of patients found that CSRCC was associated with a worse stage per stage survival compared to MA and OA, especially in stage III disease[1,8,15,17,19,36]. In a study containing 78 patients with peritoneal metastasis from colorectal origin, SRCC was found to be independently associated with worse survival[55]. Ishihara et al[56] identified a proximal colonic CSRCC location as an independent prognostic factor for increased survival. This study, however, was not included in the qualitative synthesis of this review, since in their analysis, SRCC was grouped with poorly differentiated adenocarcinoma and MA.

Implications for therapeutic management

There are no specific guidelines for the treatment of CSRCC, other than those applicable for CRC in general. Surgery is considered the mainstay of treatment, but as expected in the context of advanced disease at the time of diagnosis, the rates of curative resection in CRSCC (21%-82%) are lower compared to those of other histological types of CRC[8,15,33]. Concerning early CRC, Song et al[25] found that SRC histology in the case of a pT1 tumor was associated with a significantly higher rate of lymph node metastasis compared to OA (35.5% vs 11.6%). In light of these results, the authors do not recommend local excision as a definitive surgical treatment in the case of SRCC histology[25]. The role of (neo-)adjuvant chemo- or radiotherapy for CSRCC has not been well studied. The majority of the less recent literature including databases going several decades back in time, as well as studies based on large national databases such as the SEER database, frequently suffer from incomplete data concerning (neo)-adjuvant treatment. In the context of higher rates of stage III disease, patients with CSRCC more frequently undergo (neo-)adjuvant therapy[18,19,22]. In a small study with 22 patients diagnosed with CSRCC, (neo-)adjuvant radio- or chemotherapy was not found to result in any survival benefit[20]. Hugen et al[17] and Belli et al[20] were among the first to evaluate the role of adjuvant chemotherapy in a larger cohort of 1972 patients with CSRCC. In this study, 51.6% of stage III patients were treated with adjuvant chemotherapy, which significantly improved survival (5-year-relative survival: 52%) compared to the group of patients who did not receive adjuvant treatment (5-year-relative survival: 30%). Moreover, in a multivariate analysis, the added value of adjuvant chemotherapy in CSRCC was comparable to that of OA[17]. For rectal SRCC specifically, Jayanand et al[57] found that SRCC histology was, surprisingly, a positive predictive factor for a pathological complete response after neoadjuvant chemoradiation. Although this study contained only 6 patients with SRC histology out of 248 rectal cancer patients, their results seem to be confirmed by Tamhankar et al[51] who reported a relatively high pathological complete response rate of 21.6% after neo-adjuvant chemoradiation in a group of 37 rectal SRCC patients. Two studies evaluating the role of (preoperative) radiotherapy in rectal SRCC, found an improved CSS compared to surgery alone in stage III, but not stage II, disease[11,50].

The value of surgery in metastatic CSRCC has been investigated by Fu et al[58] who compared the surgical and oncological outcomes of 94 patients with synchronous resectable metastatic CRSCC to those of 3474 patients without SRC histology. The authors found that SRCC histology was associated with a significantly higher rate of invalid surgery (defined as disease recurrence within 6 mo after surgery) compared to non-SRC histology (24.5% vs 13%, P = 0.001). Because of a frequent peritoneal metastatic pattern associated with CSRCC, a high level of difficulty associated with curative surgery in these patients has been hypothesized as one of the reasons for the increased rate of invalid surgery. As expected, patients with CSRCC in this study were affected by a significantly poorer OS (HR = 1.445, 95% confidence interval: 1.156-1.808, P = 0.001)[58]. The authors suggested that in light of these findings, the selection of patients with resectable metastatic CSRCC for surgical therapy should be highly individualized since the added value of surgery in this context seems limited.

Because PC is frequently diagnosed early on in the process of CSRCC and is less prone to the effect of systemic chemotherapy, several studies have evaluated the value of intraperitoneal chemotherapy for PC in CSRCC[38,53,59]. Parallel to the conclusion of Fu et al[58], van Oudheusden et al[59] recommended that cytoreductive surgery (CRS) with hyperthermic intraperitoneal chemotherapy (HIPEC) should only be proposed to patients with CSRCC and PC if no other relative contra-indications such as extensive peritoneal disease, liver metastases, or poor general condition, are present, since it is questionable whether the limited survival benefit of CRS with HIPEC outweighs the associated reduction in quality of life. Although a peritoneal cancer index of < 20 was respected and both groups were comparable regarding the rate of neoadjuvant chemotherapy, the authors found that in patients with PC from CSRCC, the median survival after CRS and HIPEC was only 14.1 mo vs 35.1 mo in non-CSRCC patients[59]. It should be noted, however, that in the CSRCC group, a macroscopically complete cytoreduction was obtained in 87.5% of patients vs 97.2% in the non-CSRCC group (P = 0.04). Moreover, groups were not comparable regarding the rate of N2 stage disease (62.5% CSRCC vs 36.1% non-CSRCC, P = 0.04). Among 385 patients with CSRCC and peritoneal metastases, Simkens et al[38] demonstrated that CRS with HIPEC resulted in an absolute median survival benefit of 18 mo compared to supportive care only, with an additional survival benefit of supplementary systemic chemotherapy. Interestingly, the same study showed that in patients with CSRCC and PC with concomitant liver or lung metastases, systemic chemotherapy did not result in any survival benefit compared to best supportive care. Finally, in a study of 33 patients with SRCC and PC, Chua et al[53] found that CRS with HIPEC in patients with colonic SRCC resulted in a median survival of 13 mo compared to 18 mo (P = 0.75) in patients who received systemic chemotherapy only. In 18 patients with appendiceal SRCC on the other hand, CRS and HIPEC were associated with a median survival of 27 mo vs 15 mo in patients treated with systemic therapy only (P = 0.12). These findings led to the conclusion that a clear distinction between appendiceal SRCC and colorectal SRCC should be made, as this could have a significant influence on the potential benefits of CRS and HIPEC and thus on the optimal treatment strategy that should be implemented[53].

Few to no clinical studies so far have specifically focused on potential immunological, molecular, or genetic therapeutic agents for CSRCC. Based on its distinct molecular characteristics, some potential therapeutic pathways have been hypothesized[60]. Because a significant amount of CSRCC is MSI-H, the benefit of immune checkpoint inhibitors should be investigated[29,31,60]. However, since MSI in CSRCC does not appear to be associated with a similar prognostic benefit as it is in OA, the question remains how this could affect the mechanism of action and potential benefit of these checkpoint inhibitors. Since one study demonstrated a relatively high incidence of BRAF mutations in CSRCC, the use of a BRAF inhibitor could be evaluated as a potential therapeutic agent[60]. A major phase II study evaluating the interest of several combinations of BRAF, EGFR, and MEK inhibitors in the treatment of CRC (NCT01750918) reached study completion recently[48,60]. A subanalysis in function of the different histological subtypes could answer some questions concerning the role of these inhibitors in the treatment of CSRCC. Interestingly, Alvi et al[48] also found that 25% of CSRCCs are characterized by a KIT actionable mutation, which could as such potentially benefit from a KIT inhibitor (imatinib).

Regarding follow-up of patients with CSRCC, most studies agree that in light of the demographic characteristics and the aggressive biological behavior of CSRCC, a more frequent follow-up is warranted compared to OA, especially to detect recurrent peritoneal disease at an early stage[33].

DISCUSSION

The incidence of CSRCC has risen in the past decades, although it remains a rare histological subtype of CRC. Patients with CSRCC more frequently present at a younger age, with tumors predominantly located in the right colon and rectum. CSRCC is often characterized by locally advanced or even metastatic disease at the time of diagnosis. In contrast to OA, and facilitated by its biological and molecular features, CSRCC predominantly metastasizes into the peritoneal cavity. As such, oncological outcomes for CSRCC are very poor with a 5-year OS between 0% and 46%. According to the latest WHO definition, the amount of SRCs in the histological specimen should be mentioned in the pathology report since even a minor SRC component appears to entail prognostic and therapeutic implications similar to those of well-defined CSRCC. Furthermore, the pathological analysis should specify the presence of mucin-rich or -poor SRCC, as the latter is characterized by significantly worse biological behavior and survival. CSRCCs are often MSI-H, although this is not translated into improved survival as is the case in OA. As such, the role of immunological therapeutic agents in the case of MSI-H CSRCC remains unclear but will be a topic in future studies. Adjuvant chemotherapy and neo-adjuvant chemoradiation are found to offer improved long-term outcomes for CSRCC. For patients with PC, most authors suggest reserving CRS and HIPEC for well-selected patients in whom the reduction in quality of life, associated with this therapeutic option, does not outweigh the survival benefit, which was proven to be limited. In light of its aggressive biological behavior and metastatic pattern, a more frequent follow-up after surgical therapy is warranted, with the focus on early detection of peritoneal recurrent disease.

CONCLUSION

CSRCC is a rare clinical entity most often characterized by young age and advanced disease at presentation. As such, diagnostic modalities and therapeutic approach should be tailored accordingly.

ARTICLE HIGHLIGHTS

Research background

Colorectal signet-ring cell carcinoma (CSRCC) is a rare clinical entity which accounts for approximately 1% of all colorectal cancers with, however, a rising incidence with relatively young age at presentation and associated dismal prognosis.

Research motivation

Although multiple studies concerning this specific topic have been published in the past decades, the pathogenesis, associated risk factors, and potential implications on treatment are still poorly understood. Besides the low incidence, historically confusing histological criteria have resulted in confusing data. This highlights the actual interest to synthesize the known literature regarding CSRCC.

Research objectives

The aim of this systematic review was to provide an updated overview of risk factors, prognosis, and management of CSRCC.

Research methods

A literature search in the MEDLINE/PubMed database was conducted with the following search terms used: ‘Signet-ring cell carcinoma’ and ‘colorectal’. Studies in English language, published after January 1980, were included. Studies included in the qualitative synthesis were evaluated for content concerning epidemiology, risk factors, and clinical, diagnostic, histological, and molecular features, as well as metastatic pattern and therapeutic management.

Research results

In total, 67 articles were included for qualitative analysis, of which 54 were eligible for detailed data extraction. CSRCC has a reported incidence between 0.1%-2.4% and frequently presents with advanced disease stage at the time of diagnosis. CSRCC is associated with an impaired overall survival (5-year OS: 0%-46%) and a worse stage-corrected outcome compared to mucinous and not otherwise specified adenocarcinoma. Surgery is the mainstay of treatment, although the rates of curative resection in CSRCC (21%-82%) are lower compared to those of other histological types. In case of peritoneal metastasis, cytoreductive surgery with hyperthermic intraperitoneal chemotherapy should only be proposed in selected patients.

Research conclusions

The incidence of CSRCC has risen in the past decades, although it remains a rare histological subtype of CRC. Patients with CSRCC more frequently present at a younger age, with tumors predominantly located in the right colon and rectum. CSRCC is often characterized by locally advanced or even metastatic disease at the time of diagnosis. CSRCC predominantly metastasizes into the peritoneal cavity. As such, oncological outcomes for CSRCC are very poor with a 5-year OS between 0% and 46%.

Research perspectives

In light of its aggressive biological behavior and metastatic pattern, a more frequent follow-up after surgical therapy is warranted, with the focus on early detection of peritoneal recurrent disease. The role of immunological therapeutic agents in the case of MSI-H CSRCC remains unclear but will be a topic in future studies.