Predictive Value of Carcinoembryonic Antigen in Symptomatic Patients without Colorectal Cancer: A Post-Hoc Analysis within the COLONPREDICT Cohort

We aimed to assess the risk of cancer in patients with abdominal symptoms after a complete colonoscopy without colorectal cancer (CRC), according to the carcinoembryonic antigen (CEA) concentration, as well as its diagnostic accuracy. For this purpose, we performed a post-hoc analysis within a cohort of 1431 patients from the COLONPREDICT study, prospectively designed to assess the fecal immunochemical test accuracy in detecting CRC. Over 36.5 ± 8.4 months, cancer was detected in 115 (8%) patients. Patients with CEA values higher than 3 ng/mL revealed an increased risk of cancer (HR 2.0, 95% CI 1.3-3.1), CRC (HR 4.4, 95% CI 1.1-17.7) and non-gastrointestinal cancer (HR 1.7, 95% CI 1.0-2.8). A new malignancy was detected in 51 (3.6%) patients during the first year and three variables were independently associated: anemia (OR 2.8, 95% CI 1.3-5.8), rectal bleeding (OR 0.3, 95% CI 0.1-0.7) and CEA level >3 ng/mL (OR 3.4, 95% CI 1.7-7.1). However, CEA was increased only in 31.8% (95% CI, 16.4-52.7%) and 50% (95% CI, 25.4-74.6%) of patients with and without anemia, respectively, who would be diagnosed with cancer during the first year of follow-up. On the basis of this information, CEA should not be used to assist in the triage of patients presenting with lower bowel symptoms who have recently been ruled out a CRC.

Risk of gastrointestinal cancer in a symptomatic cohort after a complete colonoscopy: Role of faecal immunochemical test

BACKGROUND

Faecal immunochemical test (FIT) has been recommended to assess symptomatic patients for colorectal cancer (CRC) detection. Nevertheless, some conditions could theoretically favour blood originating in proximal areas of the gastrointestinal tract passing through the colon unmetabolized. A positive FIT result could be related to other gastrointestinal cancers (GIC).

AIM

To assess the risk of GIC detection and related death in FIT-positive symptomatic patients (threshold 10 μg Hb/g faeces) without CRC.

METHODS

Post hoc cohort analysis performed within two prospective diagnostic test studies evaluating the diagnostic accuracy of different FIT analytical systems for CRC and significant colonic lesion detection. Ambulatory patients with gastrointestinal symptoms referred consecutively for colonoscopy from primary and secondary healthcare, underwent a quantitative FIT before undergoing a complete colonoscopy. Patients without CRC were divided into two groups (positive and negative FIT) using the threshold of 10 μg Hb/g of faeces and data from follow-up were retrieved from electronic medical records of the public hospitals involved in the research. We determined the cumulative risk of GIC, CRC and upper GIC. Hazard rate (HR) was calculated adjusted by age, sex and presence of significant colonic lesion.

RESULTS

We included 2709 patients without CRC and a complete baseline colonoscopy, 730 (26.9%) with FIT ≥ 10 µgr Hb/gr. During a mean time of 45.5 ± 20.0 mo, a GIC was detected in 57 (2.1%) patients: An upper GIC in 35 (1.3%) and a CRC in 14 (0.5%). Thirty-six patients (1.3%) died due to GIC: 22 (0.8%) due to an upper GIC and 9 (0.3%) due to CRC. FIT-positive subjects showed a higher CRC risk (HR 3.8, 95%CI: 1.2-11.9) with no differences in GIC (HR 1.5, 95%CI: 0.8-2.7) or upper GIC risk (HR 1.0, 95%CI: 0.5-2.2). Patients with a positive FIT had only an increased risk of CRC-related death (HR 10.8, 95%CI: 2.1-57.1) and GIC-related death (HR 2.2, 95%CI: 1.1-4.3), with no differences in upper GIC-related death (HR 1.4, 95%CI: 0.6-3.3). An upper GIC was detected in 22 (0.8%) patients during the first year. Two variables were independently associated: anaemia (OR 5.6, 95%CI: 2.2-13.9) and age ≥ 70 years (OR 2.7, 95%CI: 1.1-7.0).

CONCLUSION

Symptomatic patients without CRC have a moderate risk increase in upper GIC, regardless of the FIT result. Patients with a positive FIT have an increased risk of post-colonoscopy CRC.

The fecal hemoglobin concentration, age and sex test score: Development and external validation of a simple prediction tool for colorectal cancer detection in symptomatic patients

Prediction models for colorectal cancer (CRC) detection in symptomatic patients, based on easily obtainable variables such as fecal haemoglobin concentration (f-Hb), age and sex, may simplify CRC diagnosis. We developed, and then externally validated, a multivariable prediction model, the FAST Score, with data from five diagnostic test accuracy studies that evaluated quantitative fecal immunochemical tests in symptomatic patients referred for colonoscopy. The diagnostic accuracy of the Score in derivation and validation cohorts was compared statistically with the area under the curve (AUC) and the Chi-square test. 1,572 and 3,976 patients were examined in these cohorts, respectively. For CRC, the odds ratio (OR) of the variables included in the Score were: age (years): 1.03 (95% confidence intervals (CI): 1.02-1.05), male sex: 1.6 (95% CI: 1.1-2.3) and f-Hb (0-<20 µg Hb/g feces): 2.0 (95% CI: 0.7-5.5), (20-<200 µg Hb/g): 16.8 (95% CI: 6.6-42.0), ≥200 µg Hb/g: 65.7 (95% CI: 26.3-164.1). The AUC for CRC detection was 0.88 (95% CI: 0.85-0.90) in the derivation and 0.91 (95% CI: 0.90-093; p = 0.005) in the validation cohort. At the two Score thresholds with 90% (4.50) and 99% (2.12) sensitivity for CRC, the Score had equivalent sensitivity, although the specificity was higher in the validation cohort (p < 0.001). Accordingly, the validation cohort was divided into three groups: high (21.4% of the cohort, positive predictive value-PPV: 21.7%), intermediate (59.8%, PPV: 0.9%) and low (18.8%, PPV: 0.0%) risk for CRC. The FAST Score is an easy to calculate prediction tool, highly accurate for CRC detection in symptomatic patients.

Development and external validation of a faecal immunochemical test-based prediction model for colorectal cancer detection in symptomatic patients

BACKGROUND

Risk prediction models for colorectal cancer (CRC) detection in symptomatic patients based on available biomarkers may improve CRC diagnosis. Our aim was to develop, compare with the NICE referral criteria and externally validate a CRC prediction model, COLONPREDICT, based on clinical and laboratory variables.

METHODS

This prospective cross-sectional study included consecutive patients with gastrointestinal symptoms referred for colonoscopy between March 2012 and September 2013 in a derivation cohort and between March 2014 and March 2015 in a validation cohort. In the derivation cohort, we assessed symptoms and the NICE referral criteria, and determined levels of faecal haemoglobin and calprotectin, blood haemoglobin, and serum carcinoembryonic antigen before performing an anorectal examination and a colonoscopy. A multivariate logistic regression analysis was used to develop the model with diagnostic accuracy with CRC detection as the main outcome.

RESULTS

We included 1572 patients in the derivation cohort and 1481 in the validation cohorts, with a 13.6 % and 9.1 % CRC prevalence respectively. The final prediction model included 11 variables: age (years) (odds ratio [OR] 1.04, 95 % confidence interval [CI] 1.02-1.06), male gender (OR 2.2, 95 % CI 1.5-3.4), faecal haemoglobin ≥20 μg/g (OR 17.0, 95 % CI 10.0-28.6), blood haemoglobin <10 g/dL (OR 4.8, 95 % CI 2.2-10.3), blood haemoglobin 10-12 g/dL (OR 1.8, 95 % CI 1.1-3.0), carcinoembryonic antigen ≥3 ng/mL (OR 4.5, 95 % CI 3.0-6.8), acetylsalicylic acid treatment (OR 0.4, 95 % CI 0.2-0.7), previous colonoscopy (OR 0.1, 95 % CI 0.06-0.2), rectal mass (OR 14.8, 95 % CI 5.3-41.0), benign anorectal lesion (OR 0.3, 95 % CI 0.2-0.4), rectal bleeding (OR 2.2, 95 % CI 1.4-3.4) and change in bowel habit (OR 1.7, 95 % CI 1.1-2.5). The area under the curve (AUC) was 0.92 (95 % CI 0.91-0.94), higher than the NICE referral criteria (AUC 0.59, 95 % CI 0.55-0.63; p < 0.001). On the basis of the thresholds with 90 % (5.6) and 99 % (3.5) sensitivity, we divided the derivation cohort into three risk groups for CRC detection: high (30.9 % of the cohort, positive predictive value [PPV] 40.7 %, 95 % CI 36.7-45.9 %), intermediate (29.5 %, PPV 4.4 %, 95 % CI 2.8-6.8 %) and low (39.5 %, PPV 0.2 %, 95 % CI 0.0-1.1 %). The discriminatory ability was equivalent in the validation cohort (AUC 0.92, 95 % CI 0.90-0.94; p = 0.7).

CONCLUSIONS

COLONPREDICT is a highly accurate prediction model for CRC detection.