Immune cell infiltration is a common feature of many human solid tumors. Recently, increased infiltration of tumors with cytotoxic CD8-positive T cells has been correlated with prolonged survival in multivariate analyses in a variety of epithelial cell cancers, including small-cell lung carcinoma; carcinomas of the endomentrium, bile duct, esophagus, and urothelium; pancreatic adenocarcinoma; and hepatocellular carcinomas. In this issue of Journal of Clinical Oncology, Mlecnik et al suggest that immune cell infiltration of colorectal cancer by cytotoxic CD8-positive and memory CD45ROpositive T cells has prognostic discriminatory power that is superior to standard staging systems (ie, American Joint Committee on Cancer– International Union Against Cancer–TNM). The authors examined more than 400 patients with colorectal cancer, quantitated the density of tumor-infiltrating T cells, devised an immune scoring system (higher scores reflecting greater CD8-positive and CD45RO-positive T-cell density), and correlated the immune scores with clinical outcome. The results demonstrate two key findings: patients with high immune scores have increased disease-free and overall survival as compared with patients whose tumors demonstrate low immune scores, and the immune score was superior in predicting disease outcome as compared with a host of important prognostic clinical parameters, including TNM staging. The authors further observed that there was an inverse correlation between immune cell density and tumor stage. In patient samples with the greatest immune cell density, the majority (60%) were tumor in situ or T1 stage tumors. In contrast, only 18% of T4 tumors demonstrated high-density T-cell infiltrates. Moreover, there were no tumor in situ or T1 stage disease specimens showing weak immune cell density scores, whereas 45% of T4 tumors demonstrated low immune scores. The data suggest that even with minimal tumor invasion, patients with a low immune score will be likely to experience a disease relapse. For these patients, surgery may not be curative. Remarkably, the prognostic significance of the immune score was retained regardless of whether the tumor tissue was derived from a patient with stage I or stage IV colorectal cancer. The study by Mlecnik et al is the culmination of several observations this investigative group has made over the last few years. Their initial observation, that a high density of memory CD8-positive T cells was associated with decreased evidence of early metastasis and increased survival in colorectal cancer, was followed by the development of an immune gene signature, associated with type I adaptive immunity, that correlated with decreased cancer recurrence. Moreover, in the same study, the investigators demonstrated that the type, density, and location of the T-cell infiltrate in the tumor were prognostically important. Further evaluations defined the importance of the memory T-cell phenotype and demonstrated the loss of coordinated functional immunity with the evolution of the tumor to a metastatic phenotype. A focused investigation of patients with earlystage colorectal cancer suggested that a multimarker panel of CD45RO-positive and CD8-positive T cells and cytotoxicity-related genes could predict prognosis even in these patients with minimal disease. Taken together, these studies laid the foundation for the immune score presented by Mlecnik et al as a clinical prognostic marker at any stage of colorectal cancer. Is the immune score presented in the Mlecnik et al study definitive? There are a few issues that remain to be addressed. The first issue is a better definition of the contribution of T regulatory cells to an immune score in colorectal cancer. CD4-positive T regulatory cells, as assessed by FOXP3 expression, have been shown to potentially modulate the antitumor response by suppressing the activity of CD8positive cytotoxic T cells and have been demonstrated to be a prognostic marker for disease outcome in several types of cancers. A high density of intratumoral FOXP3-positive T regulatory cells has been associated with poor outcomes in multiple solid tumors, including ovarian, pancreatic, and hepatocellular carcinoma. However, in colorectal cancer, studies of infiltrating FOXP3-expressing cells suggest the opposite: that regulatory T cells in the tumor are associated with an improved prognosis. Investigators evaluating nearly 1,000 tumor specimens from patients with stages II and III colorectal cancer demonstrated in multivariate analysis that highdensity intratumoral FOXP3-positive T cells were associated with improved survival (P .001). In a much smaller evaluation of 40 colorectal cancers, increased intratumoral T regulatory cells were correlated with limited versus metastatic disease. An additional study assessed the tumors of approximately 50 patients with relapsed colorectal cancer undergoing chemotherapy, and found that high intratumoral FOXP3-positive T-cell infiltrate was associated with benefits in overall survival (P .0005) and progression-free survival (P .0009). Colorectal cancer may be uniquely suited to a beneficial prognostic response to infiltrating T regulatory cells. In states of chronic gastrointestinal inflammation, such as inflammatory bowel disease, absolute mucosal T regulatory cell numbers are increased as a JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L S VOLUME 29 NUMBER 6 FEBRUARY 2