Sir,

We are most grateful to Professor Park, McMillan and Roxburgh for their interest and valuable comments on our manuscript titled ‘Tumour-infiltrating inflammation and prognosis in colorectal cancer: systematic review and meta-analysis’ (Mei et al, 2014). They cited several recently published studies with consistent results, pointing out some other important relationships among inflammatory cell infiltrate, the tumour microenvironment and immune response in colorectal cancer (CRC), which were not discussed in detail in our original publication because of the length of the publication. Therefore, some major concerns, such as the following, must be addressed.

First, the analytical methods used in our publication for the generalised tumour inflammatory infiltrate were relatively standardised ones and included the Jass classification, the Klintrup—Makinen (K–M) criteria and Crohn’s-like reaction criteria. The pooled hazard ratios and 95% CIs for overall survival, cancer-specific survival and disease/recurrence-free survival in a subset of highly generalised tumour inflammatory infiltrate were <1, indicating a robust survival marker for CRC. However, conflicting results (with HRs and 95% CI across 1) were noted among individual studies as heterogeneity originated from local inflammatory reaction grading systems, patient characteristics, follow-up schemes and some other factors, which was especially evident among individual T-cell subtypes. More detailed stratifications by tumour location, stage, grade and other microenvironmental components should be proposed.

Second, there have been conflicting reports regarding the relationship between inflammatory cell infiltrate and local inflammatory response for CRC prognosis, justifying the need for further analyses. In our meta-analysis, we did not include some of the mentioned studies because of the following reasons.

  1. 1)

    absence of time-to-event (survival) data for high-grade over low-grade immune cell inflammation (Klintrup et al, 2005);

  2. 2)

    sharing of the same cohort (Richards et al, 2012a, 2012b);

  3. 3)

    investigating the outcome of tumour inflammatory cell infiltrate in primary operable invasive ductal breast cancer (Mohammed et al, 2012);

  4. 4)

    study publication after the deadline of August 2013 (Vayrynen et al, 2013; Richards et al, 2014).

To minimise variation between studies, currently, standardised and robust methods for assessment of the generalised inflammatory cell infiltrate used in clinical practice are urgently needed. Forrest et al (2014) developed an automated, computer-aided scoring method that proved to be more facilitated, objective, accurate, reproducible and cost-effective than the manual method. We assumed that some larger prospective studies could be proposed to validate the robustness of association between not only the generalised inflammatory cell infiltrate but also the subsets of T lymphocytes as well and CRC survival.