Search

DNA mismatch repair protein deficient colon cancer frequently displays reduced CDX2 expression, and recent literature has suggested that negative CDX2 expression is a poor prognostic biomarker in colon cancer. We have recently demonstrated that SATB2 is an immunohistochemical marker that is complementary to CDX2. Using a tissue microarray approach, we evaluated SATB2 and CDX2 immunohistochemical expression in 514 patients with colonic adenocarcinoma including 146 with mismatch repair protein deficient tumors and correlated expression with histopathologic variables, molecular alterations, and survival. Overall, SATB2-negative and/or CDX2-negative expression was identified in 33% of mismatch repair protein deficient tumors compared with only 15% of mismatch repair protein proficient tumors (p < 0.001) and in 36% of BRAF V600E mutated compared with only 13% of BRAF wild-type tumors (p < 0.001). Both SATB2-negative and CDX2-negative colonic adenocarcinomas more often displayed lymphatic invasion, venous invasion, and perineural invasion (all with p < 0.05). SATB2-negative expression was also more frequently identified in tumors with mucinous or signet ring cell differentiation (p < 0.01 for both). In a multivariable analysis of survival in patients with mismatch repair protein deficient tumors (n = 131), only tumor stage (p = 0.01) and SATB2-negative and/or CDX2-negative expression (p = 0.009) independently predicted disease-specific survival. Of the 99 patients with stage II or III mismatch repair protein deficient tumors, death from disease only occurred in patients with either SATB2-negative or CDX2-negative tumors, and no patients with SATB2-positive/CDX2-positive tumors developed recurrence or died of disease. SATB2 and CDX2 expression had no effect on patient survival in mismatch repair protein proficient, BRAF-mutated, or KRAS-mutated tumors. In summary, our results suggest that SATB2 and CDX2 are prognostic biomarkers in patients with mismatch repair protein deficient colon cancer and that inclusion of SATB2 and CDX2 immunohistochemistry may be helpful as part of a comprehensive pathologic risk assessment in mismatch repair protein deficient colon cancer.

We evaluated a consecutive series of signet ring cell colorectal carcinomas in an attempt to correlate the histopathologic pattern of infiltration with molecular alterations and prognosis. Of the 4760 primary colorectal carcinomas surgically resected between the years 2002 and 2012, 53 (1%) were composed of >50% signet ring cells. Of the 53 signet ring cell carcinomas, 40 (75%) were composed of >50% extracellular mucin with signet ring cells floating within pools of mucin and were subclassified as mucin-rich signet ring cell carcinomas. Thirteen (25%) carcinomas were characterized by diffusely infiltrating carcinomas with minimal to no extracellular mucin and were subclassified as mucin-poor signet ring cell carcinomas. All 13 mucin-poor signet ring cell carcinomas were either stage III or IV, whereas many cases of mucin-rich signet ring cell carcinoma were stage I or II (17 cases) (P=0.005). Compared with mucin-rich tumors, mucin-poor signet ring cell carcinomas more frequently demonstrated adverse histologic features such as lymphatic invasion (13/13, 100% vs. 22/40, 55%; P=0.002), venous invasion (6/13, 46% vs. 3/40, 8%; P=0.004), and perineural invasion (11/13, 85% vs. 9/40, 23%; P=0.0001). Twenty-three of 53 (43%) signet ring cell carcinomas demonstrated high levels of microsatellite instability (MSI-H). Twenty-two of 23 (96%) MSI-H signet ring cell carcinomas were mucin rich; only 1 MSI-H signet ring carcinoma was mucin poor (P=0.0033). Mucin-poor signet ring cell carcinoma had significantly reduced overall and recurrence-free survival compared with mucin-rich signet ring cell carcinomas (P=0.0035 and 0.0001, respectively), even when adjusting for tumor stage. Mucin-poor signet ring cell carcinoma had a higher propensity for peritoneal dissemination (5/13, 38%) compared with mucin-rich signet ring cell carcinoma (5/40, 12.5%), although this was not statistically significant (P=0.052). Finally, MSI-H and microsatellite-stable signet ring cell carcinomas had similar overall and recurrence-free survival (P=0.2266 and 0.1055, respectively), even when adjusting for tumor stage. In conclusion, we identified a unique subset of signet ring cell colorectal carcinoma with diffuse infiltration and minimal to no extracellular mucin (mucin-poor signet ring cell carcinoma), which lacks MSI-H and has a dismal prognosis with an aggressive clinical course often with peritoneal dissemination. Further, our results confirm that MSI does not affect survival in colorectal signet ring cell carcinomas.