Tumor-infiltrating memory T-lymphocytes for prognostic prediction in cancer patients: a meta-analysis

Abstract

Purpose: Pathological evidence has continually supported the prognostic value of tumour-infiltrating T lymphocytes for several solid tumours across diverse patient cohorts. To investigate the clinical relevance of tumour-infiltrating memory T lymphocytes, we investigated relevant publications to identify the significance of memory tumour-infiltrating T lymphocytes (TILs) in predicting survival in cancer patients and analysed the influence of variations in the tumour stage and the infiltrated location thereon.

Methods: Relevant publications that assessed the clinical relevance of memory TILs to the patient's overall survival (OS) were investigated. Disease-free survival (DFS) was also evaluated where possible. Hazard ratios (HR) and 95% confidence intervals (CI) were assessed using a random-effects model. Heterogeneity was investigated with stratified analysis and I(2) statistics.

Results: In total, 16 relevant publications, including 4248 cancer patients, were analysed. In the pooled analysis, intra-tumour accumulation of memory TILs correlated positively with favourable clinical outcomes for both OS (pooled HR, 2.06; 95% CI, 1.76-2.40) and DFS (pooled HR, 2.31; 95% CI, 1.67-3.19). Various effects of the tumour stage and the anatomical region where the cells infiltrated were identified in survival prediction, using memory TILs.

Conclusions: Overall, infiltration of memory TILs can serve as a biomarker for survival prediction in cancer. Additional heterogeneous effects of the associated factors should be considered when categorizing high-risk patients.

Keywords: Cancer; memory T cells; prognosis; survival; tumour-infiltrating lymphocyte.

Figure 1

Flowchart of study identification and inclusion.

Figure 2

Forest plot of hazard ratio of memory tumour-infiltrating T lymphocyte (TIL) density for overall survival. Hazard ratios (HR) and 95% confidence intervals (CI) from each paper are shown as horizontal black bars with red dots. The pooled estimation is illustrated with a black diamond; the centre represents the pooled HR and the horizontal range indicates the 95% CI. The HRs are defined as the ratio of lower CD45RO density to higher CD45RO density, except for the 2006 Galon evaluation, which defined CD3⁺CD45RO⁺ cells as memory TILs. Therefore, a HR greater than unity represents a prolonged overall survival associated with a higher density of CD45RO⁺ TIL infiltration. CT, centre of tumour site; IM, invasive margin; CT/IM, both sites measured. OM, evaluated with tissue from omental metastasis; OV, tissue from primary lesion of ovarian cancer.

Figure 3

Forest plot of hazard ratio of memory tumour-infiltrating T lymphocyte (TIL) density for disease-free survival. Hazard ratios and 95% confidence intervals (CI) for disease-free survival associated with low density versus high density of CD45RO memory TIL infiltration. CT, centre of tumour site; IM, invasive margin; CT/IM, both sites measured.

Figure 4

Forest plot of hazard ratios for studies on the location of memory TILs. Pooled hazard ratios and 95% confidence intervals (CI) for the association of memory T cells and overall survival. The density of memory T cells was evaluated for primary tumour sites or other sites as indicated. CT, centre of tumour site; IM, invasive margin; CT/IM, both sites measured; OM, evaluated with tissue from omental metastasis; OV, tissue from primary lesion of ovarian cancer; CRC, colorectal cancer.

Figure 5

Forest plot of hazard ratios for studies on the staging of tumours. Pooled hazard ratios and 95% confidence intervals (CI) for cancer patients in stage I+II and those in stage III+IV. GC, gastric cancer, IUAC stage; EA, oesophageal adenocarcinoma, UICC stage; CRC, colorectal cancer, IUAC stage.

Figure 6

Funnel plot of hazard ratios and standard errors (SE) for the individual studies. No significant publication bias was observed.