The histological and molecular characteristics of early-onset colorectal cancer: a systematic review and meta-analysis

Abstract

Background: Early-onset colorectal cancer (CRC), defined as diagnosis before age 50, has increased in recent decades. Although more often diagnosed at advanced stage, associations with other histological and molecular markers that impact prognosis and treatment remain to be clarified. We conducted a systematic review and meta-analysis concerning the prevalence of prognostic and predictive tumor markers for early- vs. late-onset CRC, including oncogene mutations, microsatellite instability (MSI), and emerging markers including immune cells and the consensus molecular subtypes.

Methods: We systematically searched PubMed for original research articles published between April 2013-January 2024. Included studies compared the prevalence of tumor markers in early- vs. late-onset CRC. A meta-analysis was completed and summary odds ratios (ORs) with 95% confidence intervals (CIs) were obtained from a random effects model via inverse variance weighting. A sensitivity analysis was completed to restrict the meta-analysis to studies that excluded individuals with Lynch syndrome, a hereditary condition that influences the distribution of tumor markers for early-onset CRC.

Results: In total, 149 articles were identified. Tumors from early-onset CRC are less likely to include mutations in KRAS (OR, 95% CI: 0.91, 0.85-0.98), BRAF (0.63, 0.51-0.78), APC (0.70, 0.58-0.84), and NRAS (0.88, 0.78-1.00) but more likely to include mutations in PTEN (1.68, 1.04-2.73) and TP53 (1.34, 1.24-1.45). After limiting to studies that excluded Lynch syndrome, the associations between early-onset CRC and BRAF (0.77, 0.64-0.92) and APC mutation (0.81, 0.67-0.97) were attenuated, while an inverse association with PIK3CA mutation was also observed (0.88, 0.78-0.99). Early-onset tumors are less likely to develop along the CpG Island Methylator Phenotype pathway (0.24, 0.10-0.57), but more likely to possess adverse histological features including high tumor grade (1.20, 1.15-1.25), and mucinous (1.22, 1.16-1.27) or signet ring histology (2.32, 2.08-2.57). A positive association with MSI status (1.31, 1.11-1.56) was also identified. Associations with immune markers and the consensus molecular subtypes are inconsistent.

Discussion: A lower prevalence of mutations in KRAS and BRAF is consistent with extended survival and superior response to targeted therapies for metastatic disease. Conversely, early-onset CRC is associated with aggressive histological subtypes and TP53 and PTEN mutations, which may serve as therapeutic targets.

Keywords: colon cancer; colorectal cancer; early-onset; molecular characteristics; oncogenes; prognosis; rectal cancer.

Figure 1

Literature review flowchart. ^a Inappropriate study design includes studies concerning colorectal cancer incidence, colonoscopy or other colorectal cancer screening, population level summary statistics for colorectal cancer, and studies of colorectal cancer in model organisms or in vitro studies. ^b Markers of interest include oncogene mutations in KRAS, NRAS, BRAF, PIK3CA, PTEN, TP53, APC, and HER2; histological phenotypes including high-grade tumors and mucinous or signet ring histology; molecular carcinogenesis pathways including microsatellite instability and the CpG island methylator phenotype (CIMP); and novel tumor prognostic phenotypes including immune markers in the tumor microenvironment and the consensus molecular subtypes.^c Studies where late-onset colorectal cancer was defined as ≥ 40 years at diagnosis (or younger), or early-onset CRC was defined as ≤ 60 years at diagnosis (or older).

Figure 2

Odds ratios for KRAS mutation in early-onset CRC. Data presented as odds ratios (95% confidence interval) for KRAS mutation in early-onset relative to late-onset colorectal cancer. The pooled odds ratio is obtained via a random effects model using inverse variance weighting. AACR, American Association for Cancer Research; MDACC, MD Anderson Cancer Center; MSI, microsatellite instability; MSS, microsatellite stable.

Figure 3

Odds ratios for BRAF mutation in early-onset CRC. Data presented as odds ratios (95% confidence interval) for BRAF mutation in early-onset relative to late-onset colorectal cancer. The pooled odds ratio is obtained via a random effects model using inverse variance weighting. AACR, American Association for Cancer Research; MDACC, MD Anderson Cancer Center; MSI, microsatellite instability; MSS, microsatellite stable.

Figure 4

Odds ratios for APC mutation in early-onset colorectal cancer. Data presented as odds ratios (95% confidence interval) for APC mutation in early-onset relative to late-onset colorectal cancer. The pooled odds ratio is obtained via a random effects model using inverse variance weighting. AACR, American Association for Cancer Research; COH, City of Hope National Medical Center; CI, confidence interval; EO-CRC, early-onset colorectal cancer; MDACC, MD Anderson Cancer Center; MSKCC, Memorial Sloan Kettering Cancer Center; UCD, University of California, Davis.

Figure 5

Odds ratios for NRAS mutation in early-onset colorectal cancer. Data presented as odds ratios (95% confidence interval) for NRAS mutation in early-onset relative to late-onset colorectal cancer. The pooled odds ratio is obtained via a random effects model using inverse variance weighting. AACR, American Association for Cancer Research; CI, confidence interval; EO-CRC, early-onset colorectal cancer; MDACC, MD Anderson Cancer Center.

Figure 6

Odds ratios for TP53 mutation in early-onset colorectal cancer. Data presented as odds ratios (95% confidence interval) for TP53 mutation in early-onset relative to late-onset colorectal cancer. The pooled odds ratio is obtained via a random effects model using inverse variance weighting. AACR, American Association for Cancer Research; CI, confidence interval; EO-CRC, early-onset colorectal cancer; MDACC, MD Anderson Cancer Center; MSI, microsatellite instability; MSS, microsatellite stability.

Figure 7

Odds ratios for PTEN mutation in early-onset colorectal cancer. Data presented as odds ratios (95% confidence interval) for PTEN mutation in early-onset relative to late-onset colorectal cancer. The pooled odds ratio is obtained via a random effects model using inverse variance weighting. AACR, American Association for Cancer Research; CI, confidence interval; EO-CRC, early-onset colorectal cancer; MDACC, MD Anderson Cancer Center.