TIMP-1 is a novel serum biomarker for the diagnosis of colorectal cancer: A meta-analysis

Abstract

Purpose: Tissue inhibitor of metalloproteinase-1 (TIMP-1) is a glycoprotein involved in cell survival and tumorigenesis. There have been some promising results regarding the diagnostic value of TIMP-1 for patients with colorectal cancer (CRC). The aim of the present study was to assess the diagnostic accuracy and clinical utility of serum TIMP-1 in CRC patients through meta-analysis.

Methods: A systematic search of online databases was performed to collect eligible studies. The pooled sensitivity, specificity, diagnostic odds ratio (DOR), and summary receiver operator characteristic (SROC) curve were generated from accuracy data using the random-effects model. Fagan's nomogram and the likelihood matrix were applied to estimate the clinical utility of TIMP-1.

Results: A total of 9 eligible studies with 1886 patients were included. Among the patients, 819 were pathologically diagnosed with CRC, whereas 1067 did not have adenomas or other cancers. The overall sensitivity, specificity, and DOR of TIMP-1 for the diagnosis of CRC were 0.65 (95% confidence interval (CI): 0.57-0.72), 0.87 (95% CI: 0.76-0.94), and 12.73 (95% CI 5.71-28.38), respectively. The area under the SROC was 0.77 (95% CI, 0.73-0.81), suggesting the potential diagnostic value of TIMP-1 in CRC patients. Among patients with a pretest CRC probability of 20%, posttest probabilities were 56% and 9% for positive and negative TIMP-1 results, respectively.

Conclusions: TIMP-1 expression exhibits an upper moderate diagnostic value in CRC, and TIMP-1 assessment may be useful as a noninvasive screening tool for CRC in clinical practice.

Fig 1. Flow diagram for the study selection process and quality assessment using QUADAS-2.

The quality of each eligible study was assessed by QUADAS-2. It summarized ‘‘risk of bias” and ‘‘applicability concerns” through judging each domain for each included study.

Fig 2. Results of Deeks’ funnel plot asymmetry test for the evaluation of potential publication bias.

Every point represents one study and the line is the regression line. The nonsignificant slope indicates that no significant bias was found. The p value is of 0.26 (Bias coefficient = -24.55).

Fig 3

Forest plots show the pooled sensitivity and specificity (a), the pooled positive likelihood ratio and negative likelihood ratio (b), the pooled diagnostic score and diagnostic odds ratio(c) for assessing the diagnostic value of TIMP-1in colorectal cancer. The forest plots show the pooled diagnosis index of TIMP-1 for the diagnosis of CRC. The individual study symbol is shown as point and the pooling symbol is shown as point. Inconsistency is used to quantify the heterogeneity caused by nonthreshold effect. For these studies, DerSimonian–Laird was used to pool these data.

Fig 4. Meta-regression and subgroup analyses for potential sources of heterogeneity.

The forest plots show the pooled diagnosis index of TIMP-1 for the diagnosis of CRC in specified covariates. The pooling symbol is shown as point. P-values of covariants in the meta-regression analysis are more than 0.05, which shows that none of the factor that is responsible for the heterogeneity.

Fig 5. Summary ROC curve of the TIMP-1 diagnostic value in colorectal cancer.

Every point represents an included study. The diamond shape represents the summary sensitivity and specificity. The AUC is 0.77, which implies an upper moderate diagnostic accuracy for diagnosing CRC.

Fig 6. Fagan’s Nomogram for the elucidation of posttest probabilities.

With a pretest probability of CRC of 20%, the posttest probabilities of CRC, given positive and negative TIMP-1 results, are 56% and 9%, respectively.

Fig 7. Likelihood matrix for the overall distribution of the studies.

Each point corresponds to a study. None of the included studies were found on the bottom left side of the matrix, which shows that they report reasonable sensitivity.