Abstract
Background/Aim: Signet-ring cell carcinoma (SRCC) is an uncommon histological variant of colorectal cancer (CRC). Knowledge is scarce due to its rarity. Our aim was to better evaluate the clinicopathologic and prognostic features of this little-known malignancy. Patients and Methods: Thirty-nine consecutive patients with non-metastatic colorectal SRCC undergoing curative resection at University Hospital of Parma between 2000 and 2018 were examined in this retrospective analysis. Results: Mean overall (OS) and disease-free survival (DFS) were 33.6 and 31.5 months, respectively. At univariate analysis, the lymph-related parameters (nodal status, Stage III, metastatic lymph node ratio and lymphovascular invasion) were significantly associated with shorter OS and poorer DFS. At multivariate analysis, Stage III and a metastatic lymph node ratio ≥25% were found to be the only independent prognostic factors significantly correlated with worse OS and DFS. Conclusion: Nodal and lymphatic status should be carefully pondered when planning the most appropriate management of patients with colorectal SRCC.
Colorectal cancer (CRC) is the second and third most common malignancy in women and men with an increasing incidence under the age of 40 and 50 years, respectively (1-3). Differently from mucinous CRC where mucin occupies more than 50% of the extracellular space, signet-ring cell carcinoma (SRCC) characteristically shows intracellular mucin accumulation and peripheral arrangement of the nuclei in more than 50% of tumor cells (4-8). SRCCs have been frequently described in the gastrointestinal tract (especially stomach) but also in the breast, lung, gallbladder, pancreas, ovary, cervix, urinary system and salivary glands (5, 9). The large bowel represents a further, rare site of occurrence. First described in the colon by Laufman and Saphir in 1951, SRCC has been reported to have an incidence of 0.3-4.6% among all subtypes of CRCs, with slight differences between Western countries, Asia and Africa (10-17). A recent review by Tajiri et al. highlighted the clinicopathological and genetic features of SRCC, confirming the general belief that this histological subtype is related to a worse prognosis (18); however, despite this, nowadays little is known about the factors impacting the survival of patients. In fact, the relative literature is scarce, heterogeneous and hampered by small numbers (6, 19-24). The aim of our study was to assess poor outcome’s predictors of this uncommon colorectal malignancy.
Patients and Methods
From a database of 2,945 CRC patients undergoing potentially curative resection between 2000 and 2018 at the University Hospital of Parma, Italy, we retrieved the clinical records of 39 consecutive SRCCs (1.32%). CRC clinical and pathological staging were assessed with contrast-enhanced full body computed tomography scan and according to the 8th edition of the American Joint Committee on Cancer (AJCC), respectively. Pelvic magnetic resonance imaging and/or endorectal ultrasound were also used in case of rectal cancer (RC). Metastatic lesions were excluded. The clinicopathologic characteristics mainly investigated were: sex, age, tumor location, lymph node (N) status, TNM Stage, lymph node ratio (LNR), budding grade, signet-ring cell percentage, lymphovascular (LVI) and perineural invasion (PnI). LNR and budding grade were evaluated as formerly described (<25% vs. ≥25% and low vs. high grade, respectively) (25-27). All patients were offered adjuvant chemotherapy, tailored on an individual basis; all RCs also received neoadjuvant chemoradiation. Follow-up was evaluated in terms of overall (OS) and disease-free survival (DFS). Concerning the cut-off value of Signet-Ring Cells (SRC) percentage, since this is a new, unpublished histopathologic parameter, based on a semi-quantitative evaluation we classified SRCs into two groups: group A (presenting from 51 to 70% of SRCs), and group B (≥71% of SRCs). Survival of subclasses was also studied.
Statistics. Categorical, ordinal and continuous variables were compared using the Chi-square, Kruskal–Wallis and Student’s t-test, respectively. Univariate analysis of survival was performed through Kaplan–Meier statistics (Tarone-Ware test for confrontations). Criteria independently related with poor prognosis having p-value <0.05 were eventually included in a multivariate Cox regression model with calculation of correspondent hazard ratio (HR) and 95% confidence interval (95%CI). More precisely, for multivariate analysis, in order to avoid the potential effect of collinearity of variables dealing with the status of lymph nodes and lymphatic vessels (N, Staging, LNR and LVI), three distinctive models were individually assessed: a) Stage III vs. I-II; b) LNR>0 vs. 0, LNR <25% vs. 0 and LNR ≥25% vs. 0; c) N2 vs. N1.
Results
Table I shows the clinicopathological data of the examined population. Among the 39 analyzed patients, there were 14 men (35.9%) and 25 women (64.1%) with a mean age of 69.9 years (SD: ±14.4). In 28 patients the tumor was in the right colon (71.8%), in 7 in the left colon (17.9%) and in 4 in the rectum (10.3%). All SRCCs were free of metastasis (M0). Six patients presented with Stage I (15.4%), 9 with Stage II (23.1%) and 24 with Stage III cancers (61.5%). At pathology, 15 patients did not have lymph node involvement (N0, 38.5%), while 24 subjects presented nodal metastasis (N+, 61.5%). Of these, 7 were N1 (17.9%), while 17 were N2 (43.6%). Excluding N0 cases, mean LNR was 27% (SD: ±32.7); applying a 25% cut-off, 10 patients had LNR<25% (25.6%) whereas 14 (35.9%) had LNR≥25%. All SRCCs showed poor differentiation (G3). Tumor budding was high-grade in 3 (7.7%) and low-grade in 36 (96.3%) patients. Considering the SRCs percentage, 5 patients (12.8%) were classified in group A and 34 (87.2%) in B. LVI and PnI were observed in 18 (46.2%) and 2 patients (5.1%), respectively. At a mean 3.8-year follow-up, there were 21 alive (53.8%) and 18 dead (46.2%) subjects. Among the alive ones, 17 (43.6%) developed secondary metastases. Most metastases occurred in the peritoneum (9, 23.1%) followed by liver (4, 10.3%), lung (2, 5.1%), brain (1, 2.6%) and bone (1, 2.6%) (Table II). OS and DFS were 33.6 and 31.5 months, respectively. At univariate analysis, involvement of all the lymph-related parameters (N, TNM Stage, LNR and LVI) appeared related to a poor prognosis (Table III). In fact, compared to TNM Stage I and II, Stage III was significantly associated with shorter OS (40.2 vs. 84.6 months; p<0.001) and DFS (30.6 vs. 85.8 months; p<0.001) (Figure 1 and Figure 2, respectively; Table III). The presence of LVI was also associated with a worse OS (27.8 vs. 79.9 months; p=0.007) and DFS (26.9 vs. 68.9 months; p=0.022) (Table III). Multivariate analysis is shown in Figure 3. Concerning OS, an increased hazard ratio (HR) was found for tumor Stage III vs. I-II (HR=7.227, 95%CI=1.618-32.278), and LNR ≥25 vs. 0 (HR=13.227, 95%CI=2.493-70.182), indicating a significantly independent higher risk of death in such subjects. Regarding DFS, an increased HR was found in subjects with Stage III vs. I+II (HR=8.764, 95%CI=1.990-38.601), LNR <25% vs. 0 (HR=5.095, 95%CI=1.006-25.794) and LNR ≥25% vs. 0 (HR=17.549, 95%CI=3.496-88.091) (Figure 3). No significant HR for OS and DFS was identified in N2-when compared to N1-cases.
Kaplan–Meier curves showing differences in overall survival between Stage III and Stage I/II patients with colorectal signet ring cell carcinoma.
The Kaplan–Meier method showed statistically significant differences in disease-free survival between Stage III and Stage I/II patients with colorectal signet ring cell carcinoma.
Multivariate analysis showing independency among prognostic factors of overall and disease-free survival. Statistically significant values (p<0.05) are written in bold type and are located at the right side of Hazard Ratio column.
Clinicopathological of the 39 patients with signet-ring cell colorectal cancer.
Sites of metastases from signet-ring cell carcinomas (SRCCs).
Univariate analysis of clinicopathological parameters on overall and disease free-survival.
Discussion
The incidence of SRCC in our CRC population (1.32%) was similar to the one reported in the world literature (0.3%-4.6%) (10-17) with only one series showing an unusual higher rate (18%) (28). The reported sex prevalence is heterogeneous in the literature (9, 12, 12, 21, 29-31); in our work, the female to male ratio was 25 to 14 (1.8:1). In line with some previous studies, in our cohort, SRCC showed higher incidence among elderly patients (mean age: 69.9 years); on the contrary, other authors demonstrated younger patients (mean age: 39 to 42 years) to be the most affected target-population advocating for a major local aggressiveness and predisposition to peritoneal carcinomatosis (9, 30-36). In our study, the right colon was more frequently involved (28 cases, 71%) in comparison with left colon and rectum (7 and 4 cases, respectively): this is in line with some literature, while other articles found a more frequent left-sided localization (12, 29, 31, 37, 38). In our case, the choice to consider lesions of the entire transverse colon as right-sided tumors may have been the determinant factor. Compared to non-SRCCs, SRCCs were found to have worse behavior and poorer prognosis with a 33.6-month OS and 31.5-month DFS. These data are consistent with the literature where SRCC is considered a predictor of poor outcome (19, 20). Actually, analysis of the clinicopathological and survival data of our patient population demonstrated an important local aggressiveness since all SRCCs showed poor differentiation (G3), frequent nodal metastases (Stage III: 24 of 39 patients, 61.5%), LVI (50%) and predisposition to develop peritoneal metastases (9 of 39 subjects) (Tables I and II). Moreover, the 18 SRCC patients belonging to Group B (those with a percentage of SRCs exceeding 70%) also showed a high-grade tumor budding. The impact of nodal and lymphatic metastasis on prognosis of our patients appears worth noting. At univariate analysis, in fact, lymph-related parameters (N, Stage, LNR and LVI) were significantly associated with poorer OS (p=0.001, <0.001, <0.001 and 0.007, respectively) and DFS (p<0.001, <0.001, 0.001 and 0.022, respectively) showing a difference of approximately 40 months of life expectancy (Table III, Figures 1 and 2). Furthermore, at multivariate analysis, such lymph-related parameters were found to be the only independent prognostic factors (Figure 3). More precisely, Stage III and LNR≥25% were statistically significant independent prognostic factors predictive of poor OS whereas Stage III and the entire LNR classification (that is LNR>0, LNR <25% and LNR≥25%) appeared independently associated with worse DFS (Figure 3). The comparison between N1 and N2 group did not show any significant difference in prognosis probably due to the small number of patients. Our results on SRCCs prognosis are in agreement with the previously reported studies (12, 19, 35, 37-39). Some authors have hypothesized the loss of E-cadherin expression and activation of Wingless-related integration site cascade as the leading molecular culprits for the described local tumor severity with development accompanied with nodal metastasis and peritoneal recurrences (27, 40). Further analyses (including multi-center trials) on larger populations are needed for corroborating our interesting findings on this rare colorectal malignancy.
Conclusion
SRCC is a rare form of CRC. Prognosis of affected patients appears dismal with a mean OS of less than 3 years after surgery. In our study, the lymph-related parameters (N, TNM Stage and LNR) were found to be the only independently related to worse survival (both OS and DFS). As for SRCC, the impact of nodal and lymphatic status on prognosis should be carefully pondered when planning the most appropriate treatment and follow up.
Footnotes
Authors’ Contributions
All the Authors agreed with the content of the article. Dr. Annicchiarico and Dr. Costi conceived the research. Dr. Annicchiarico and Dr. Virgilio wrote the manuscript. Dr. Morini and Dr. Romboli reviewed the literature. Dr. Crafa performed histology. Dr. Leonardo provided the survival data. Dr. Riccò and Dr. Virgilio conducted statistics. Dr. Dell’Abate and Dr. Costi supervised the project.
Conflicts of Interest
The Authors declare no conflicts of interest in relation to this study.
- Received October 13, 2020.
- Revision received October 25, 2020.
- Accepted October 26, 2020.
- Copyright © 2020 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
