Differential significance of tumour infiltrating lymphocytes in sporadic mismatch repair deficient versus proficient colorectal cancers: A potential role for dysregulation of the transforming growth factor-β pathway

Multimarker phenotype predicts adverse survival in patients with lymph node-negative colorectal cancer

BACKGROUND

The heterogeneity of stage II colon cancer underlines the need for identifying high-risk, lymph node-negative patients. The objective of this study was to define a multimarker prognostic model of 5-year survival in patients with lymph node-negative, mismatch repair (MMR)-proficient colorectal cancer (CRC).

METHODS

Immunohistochemistry for 13 tumor markers was performed on 587 lymph node-negative, MMR-proficient CRC samples by using a tissue microarray. Immunoreactivity was evaluated semiquantitatively. A receiver-operating characteristic-based approach was used to detect clinically relevant tumor markers and to determine cutoff scores for tumor positivity. Univariate and multivariate analyses stratified by pathologic T3 (pT3) or pT4 tumor classification were performed.

RESULTS

In univariate analysis, the absence of CD8+ tumor infiltrating lymphocytes (TILs) (P < .001), loss of p27 (P = .006), positive urokinase-type plasminogen activator (uPA) expression (P = .002), and positive uPA receptor (uPAR) expression (P = .037) were associated with an adverse prognosis. In multivariate analysis, CD8 (P = .001), p27 (P = .031), and uPA (P = .014) were independent prognostic factors. The multimarker phenotype of negative CD8, loss of p27, and positive uPA expression led to significantly worse survival compared with all other combinations of these features. Stratified by pT3 or pT4 stage, CD8 (P = .006) and uPA (P = .011) had independent prognostic value. Combined CD8 negativity and uPA positivity led to a more adverse prognosis in both patients with pT3 tumors and patients with pT4 tumors (P < .001). No difference was observed in the length of survival between patients with pT3 tumors who had CD8 negativity and uPA positivity and patients with pT4 tumors (P = .267).

CONCLUSIONS

The multimarker phenotype of the absence of CD8+ TILs, loss of p27, and positive uPA expression was predictive of an adverse prognosis in patients with lymph node-negative, MMR-proficient CRC. The current findings suggested that a subgroup of patients with high-risk, lymph node-negative pT3 tumors should be considered for adjuvant therapy.

Value of staining intensity in the interpretation of immunohistochemistry for tumor markers in colorectal cancer

The purpose of this study was to determine whether staining intensity in conjunction with the percentage of positive tumor cells should be used as an indicator of protein expression detected by immunohistochemistry. A tissue microarray of 1,197 colorectal cancers was immunostained for p53, Her2/neu, epidermal growth factor receptor (EGFR), adenomatosis polyposis coli (APC), and beta-catenin. Immunoreactivity was described by the percentage of positive tumor cells (percent positivity) and by the staining intensity (weak, moderate, strong). The interobserver reproducibility of both was evaluated by two pathologists. The association of T stage, N stage, tumor grade, vascular invasion, and survival with percent positivity, staining intensity, and the combination of both was assessed. In univariate analysis, protein expression assessed by percent positivity resulted in 11 significant associations between the proteins and clinico-pathological features. Eight of these 11 were also demonstrated using only the degree of staining intensity. However, more than half of the associations identified by percent positivity alone were lost when staining intensity was also analyzed in combination with the percentage of positive tumor cells. A scoring method based on percent positivity, rather than on staining intensity, for p53, Her2/neu, EGFR, APC, and beta-catenin is reproducible and appears to be sufficient for establishing associations of the selected tumor markers with most clinico-pathological features.