3.1 Risk factors

There is a significant overlap in the non-modifiable risk factors of EOCRC with LOCRC. For example, in their 2023 systematic review and meta-analysis of 36 studies which included 66,312 participants, Hua et al.⁶ identified that being male (odds ratio [OR] = 1.20, 95% CI 1.08–1.33), Caucasian (OR = 1.44, 95% CI 1.15–1.80), having a family history of CRC (OR = 5.90, 95% CI 3.67-9.48) and having inflammatory bowel disease (IBD) (OR = 4.43, 95% CI 4.05–4.84) increased an individual's risk of developing EOCRC. Interestingly, Agazzi et al.⁷ in their retrospective analysis of 1778 patients aged < 50 who were referred for a colonoscopy, demonstrated that IBD was associated with less adenoma and adenocarcinoma detection (OR = 0.15, 95% CI 0.05–0.51, p = 0.002), attributing this to their intensive surveillance program in IBD patients. Thus, in patients who are not lost to follow-up and are compliant with endoscopic surveillance, IBD may be a protective factor against EOCRC development. Although having a family history was associated with the greatest risk, other studies have shown that on average only 28% of EOCRC patients have a positive family history of CRC and 13% have hereditary cancer syndromes.⁸ Therefore, the majority of EOCRC is sporadic in nature.

There has been a paradigm shift in EOCRC research to focus on childhood or adolescent exposures to modifiable risk factors that alter the gut microbiome (dysbiosis). Li et al.⁹ performed a population-based case-control study of 1,368 German participants aged < 55 (n = 747 cases, n = 621 controls) spanning over a catchment area of two million people. They demonstrated an age-related change in risk profile in patients who were obese (body mass index [BMI] ≥ 30 kg/m²) in comparison to those who were not (BMI < 25 kg/m²) at ages 20 (OR = 2.56, 95% CI 1.2–5.44), 30 (OR = 2.05, 95% CI 1.25–3.50) and 10 years before diagnosis (OR = 1.88, 95% CI 1.30–2.73). In their analysis of the Nurses’ Health Study II (NHSII), which is a database comprised of 116,430 US female nurses aged 25–42 at enrolment in 1989, Zheng et al.¹⁰ demonstrated that those in the highest versus lowest quintile for a Westernised diet (low fiber, high fat, and red meat) had an increased risk of early-onset adenoma (OR = 1.38, 95% CI 1.13–1.68) and a stronger association with adenomas with higher malignant potential (OR = 1.67, 95% CI 1.18–2.37). Hur et al.¹¹ also analyzed the NHSII and found that each serving of a sugar-sweetened beverage/day between the ages 13–18 was associated with a 32% increased EOCRC risk (RR = 1.32, 95% CI 1.00–1.75). Lifetime average alcohol consumption of ≥ 25 g/day has also been correlated with EOCRC versus LOCRC development (OR = 1.3, 95% CI 1.2–2.8, p = 0.011) in a large population-based case-control study of 9235 German participants (n = 5104 cases, n = 4131 controls).¹²

Fewer population-based case-control studies have investigated the role of antibiotic agents in EOCRC development with conflicting results. Whilst Nguyen et al.¹³ found no association between lifetime antibiotic use and EOCRC risk after analyzing 54,804 CRC cases and 201,089 controls in their Swedish cohort, McDowell et al.¹⁴ determined that antibiotics increased EOCRC risk (OR = 1.49, 95% CI 1.07–2.07) after investigating 7,903 CRC cases and 30,418 controls in their US cohort. Further, a study of the UK biobank which is comprised of 113,256 participants (n = 165 EOCRC cases, n = 719 early-onset colorectal adenoma cases), found that long-term or recurrent courses of antibiotics (≥ 3/year) were associated with increased EOCRC risk (OR = 1.48, 95% CI 1.01–2.17, p = 0.046) and adenomas (OR = 1.40, 95% CI 1.17–1.68, p < 0.001).¹⁵ The discrepancies between these studies are likely due to the geographical influence on infection prevalence, local sensitivities and resistances, and antimicrobial stewardship between countries.

Although there are some high-quality, large-scale population-based case-control studies investigating the risk factor profile of EOCRC, these are predominantly in German and US cohorts limiting the translatability of their results into an Australasian setting. Further, analyses of the NHSII are limited by a lack of analysis of males which could result in either under- or over-estimation of odds ratios. Future research should address these limitations and focus on establishing nomograms that predict EOCRC risk such that high-risk patients can be identified and educated regarding the importance of early screening and red-flag symptoms (such as rectal bleeding and changes to bowel habits).

3.2 Histopathological profile

In comparison to LOCRC, EOCRCs harbor more aggressive features on histopathological analysis. In two large-scale comparative analyses of the Surveillance, Epidemiology, and End Results (SEER) database consisting of 1,334 patients aged 20–40 years versus 46,457 patients aged 60–80 years and North American National Cancer databases consisting of 64,068 patients < 50 years versus 524,801 patients > 50 years, EOCRCs had significantly higher proportions of cancers with poor differentiation (20 and 27% vs 17. and 18%, both p < 0.001) as well as mucinous and signet ring cell components (12.6 and 15.7% vs. 10.8 and 11.5%, both p < 0.001).^{16, 17} The presence of other adverse histological features is also more common. In a recent single-institution study of 34,067 CRC patients (n = 6369 < 50 years vs. n = 27,698 > 50 years) in China, EOCRC patients were more likely to have perineural (29.7% vs. 22.4%, p < 0.001) and vascular (22.1% vs. 17.5%, p < 0.001) invasion. The same study also found EOCRC patients had the more advanced nodal disease (N2) (19.3% vs 12.8%, p < 0.001) and T stage (pT4) (2.5% vs. 1.7%, p = 0.004) at the time of presentation.¹⁸ Whilst no current Australian study directly compares EOCRC to LOCRC, a recent study of EOCRC across four tertiary referral centers in New South Wales showed lower rates of perineural invasion (22.6%) but a higher percentage of pT4 disease (30.4%).¹⁹

We now recognize three pathways for general CRC carcinogenesis: chromosomal instability (CIN), microsatellite instability (MSI), and CpG island methylator phenotype (CIMP).²⁰ Pathogenic germline mutations associated with hereditary cancer syndromes are present in only 2%–5% of LOCRC, with CIN being the most implicated pathway. Although sporadic cancers remain the most common, the proportion of EOCRC patients who demonstrate germline mutations is higher than in LOCRC. MSI-high has been demonstrated in 10%–30% of EOCRC patients, which has clinical relevance because of the benefit of programmed cell death protein 1 receptor inhibitors in patients with MSI-high cancers, especially when EOCRC patients tend to have more advanced disease and are more likely to have metastatic disease.²¹ You et al.²² demonstrated that of 25 EOCRC patients harboring germline mutations, 48% (n = 12) and 20% (n = 5) had mutations in mismatch-repair (MSI pathway) and APC/MUTYH (CIN pathway) genes respectively. Although germline mutations are increased in EOCRC, most EOCRCs are sporadic and hence there has been a focus on identifying somatic mutations unique to EOCRC in recent literature. Stoffel and Murphy²³ showed a lower rate of APC/BRAF and a higher rate of CTNNB1 mutations in their EOCRC versus LOCRC group. This was corroborated by Lieu et al.²⁴ who demonstrated, via next-generation tumor sequencing, BRAF mutations in 5% and 48% of their MSI-high EOCRC and LOCRC, respectively. Other studies have linked EOCRC to mutations in genes resulting in DNA-repair deficiencies, histone modification, and increased cell turnover (Table 1).^25-28 This higher tumor mutational burden is important as it indicates the potential role and favorable response of EOCRCs to immune checkpoint inhibitors.²⁹ Knowledge of molecular characterization has not yet translated into therapeutic improvements.

Environmental exposures resulting in gut dysbiosis and EOCRC development have been well established in current EOCRC literature. Inflammation has been hypothesized to be the link between exposure, the breakdown of intestinal barriers, and increased accessibility of food-borne mutagens.³⁰ Subsequently, EOCRC studies have focused on identifying genes and tumor microbes associated with a pro-inflammatory state. Gardner et al.³¹ used NanoString immune profiling to analyze mRNA expression of immune genes in EOCRC (n = 40) and LOCRC (n = 39). They identified increased gene expression of C7, CFD (elements of complement), and SAA1 (a gene overexpressed in inflammatory bowel disease) in EOCRC patients. This is significant as a dysregulated innate immune response has been found to be the main driver behind “inflammaging”, a phenomenon characterized by persistent low levels of systemic and tissue microenvironment inflammation.³² Other studies investigating the relationship between certain environmental exposures and EOCRC-related gene and metabolomic modifications have established links between inflammation, obesity, and the NOTCH1-GATA4-IRG1 axis, alcohol consumption and overexpression of NOTUM/GDF11, and red meat consumption and choline/tryptophan/bile acid accumulation and upregulation of phosphatidylcholine and tryptophan biosynthesis genes in EOCRC patients.^33-35 A Westernised diet has also been correlated with an EOCRC fecal taxa signature characterized by an abundance of Falvonifractor plautii, B. vulgatus, Alistipes putredinis, and Intestinomonas butyriciproducens species, resulting in decreased flavonoid degradation, increased proteolytic and elastase deficiency and thus inflammation.^{34, 36} Although studies have demonstrated the correlation between numerous genes, inflammation and EOCRC, a reproducibly implicated gene remains to be identified and future research is needed.

3.3 Diagnostic biomarkers for early detection

EOCRC patients are generally younger than the minimum age for CRC screening in Australia. Despite increasing EOCRC incidence, as a proportion of all CRC cases, EOCRC is still small with a reduction of screening initiation age to 45 having been found to have a less favorable benefits-to-harms ratio in an Australian simulation study.³⁷ As such, serological diagnostic biomarkers represent a promising minimally invasive alternative. Loomans-Kropp et al.³⁸ found that the Epi proColon V2.0, a commercially available mSEPT9 cell-free DNA detection kit for those > 50 years in the US, had good statistical performance characteristics to predict EOCRC in comparison to all healthy controls. Such a test might in the future be cost-effective if used to screen the 35-50-year-old population. Nakamura et al.³⁹ established a combination signature of four circulating micro RNA (miRNA) (miR-193aa-5p, miR-210, miR-513a-5p, and miR-628-3p) via genome-wide transcriptomic profiling as a diagnostic biomarker for EOCRC in their training cohort with good performance characteristics which were reproduced in two independent validation cohorts. It is important to note that EOCRC diagnostic biomarker research is evolving with studies investigating their cost feasibility as well as performance in diagnosing colorectal adenomas still yet to be published.

3.4 Treatment patterns and short-term outcomes

EOCRC patients tend to undergo more radical chemotherapy regimens and surgeries due to their high-performance status and generally fewer comorbidities.^40-43 Kneuertz et al.⁴² performed a US cohort study of 13,102 EOCRC and 37,007 LOCRC patients and found that younger patients were more likely to receive adjuvant chemotherapy at all stages (despite not being standard of care in stage I and stage IIa disease) in comparison to LOCRC with odds ratios of 2.88 (95% CI 2.21–3.77) for stage I, 3.93 (95% CI 2.44–3.07) for stage II, 2.42 (95% CI 2.18–2.68) for stage III and 2.74 (95% CI 2.44–3.07) for stage IV. After performing matching for stage, EOCRC patients did not demonstrate any appreciable survival gains. Thus, at present, the 2023 DIRECt guidelines recommend that neoadjuvant and adjuvant systemic therapy should not differ between EOCRC and LOCRC.¹ However, given the high proportion of MSI-high EOCRC and evolving research implicating the role of inflammation in the carcinogenesis of EOCRC, further research into the efficacy of immune checkpoint inhibitors in the treatment of EOCRC is necessary.

EOCRC patients are more likely to present urgently due to symptoms of obstruction or perforation requiring more open approaches, non-restorative procedures, and extended radical resections in an emergency setting.^{18, 44} Research examining the short-term postoperative outcomes of EOCRC patients is limited in number and by the preponderance of single institutional, small retrospective cohort studies. Of the larger-scale studies, Ewongwo et al.⁴⁵ and Hanna et al.⁴⁶ utilized the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database to analyze 7538 rectal cancer and 15,957 CRC patients, respectively. They found that early-onset rectal and EOCRC patients had significantly reduced 30-day mortality (0.4% vs. 1.8%, p = 0.04 and 0.3% vs. 1.3%, p = 0.04, respectively) and 30-day postoperative complications (18% vs. 22%, p = 0.02 and 25% vs. 29%, p = 0.02, respectively) on univariate analysis in comparison to their older counterparts. These results, however, did not remain statistically significant after accounting for confounding factors (comorbidities, stage, and type of surgery) on multivariate analyses.

3.5 Long-term outcomes

At present, conflicting data exists regarding EOCRC survival due to heterogenous EOCRC definitions and survival measures. Franklyn et al.⁴⁰ found that 5-year overall survival (OS), even when adjusted for age and gender, was superior for younger-onset CRC patients (< 40 years) in comparison to CRC patients aged ≥ 40 years (71.6% vs 47.2%, p < 0.001). Other studies have shown improved survival outcomes after adjusting for stage.^{49, 50} Saraste et al.⁴⁷ determined that EOCRC patients had a superior 5-year stage-adjusted disease-free survival (DFS) in comparison to those aged 50–74 and ≥ 75 (stage I: 0.96 vs. 0.88 vs. 0.69, p < 0.001; stage II: 0.90 vs. 0.82 vs. 0.62, p < 0.001 and stage III: 0.77 vs. 0.68 vs. 0.49, p < 0.001). O'Sullivan et al.⁴³ found a survival benefit for those aged 40–49 with stage IV disease (hazard ratio = 0.79, 95% CI: 67%−95%) compared to those aged < 40, 50–59, and 60–70 years. The superior survival of EOCRC patients has been credited to their fewer comorbidities and tendency to undergo multimodal systemic therapies and aggressive surgery. In contrast to studies examining US, Canadian, English, and Japanese cohorts, the only Australian survival study to date found poorer median DFS (4.6 vs. 16 months, p = 0.023) after neoadjuvant radiotherapy and surgery and poorer progression-free survival (2.6 vs. 9.7 months, p = 0.006) and OS (40.5 vs. 583 months, p = 0.036) after relapse in early-onset versus late-onset rectal cancer. This was believed to be due to adverse histopathological features and potentially treatment-resistant environment of early-onset tumors.⁴⁸

As it has been recognized that EOCRC patients are likely to receive more aggressive treatments and live longer, studies investigating their quality of life (QoL) and long-term functional outcomes are important. However, the literature to date is scarce. A systematic review of 15 studies including a total of 18,146 CRC patients demonstrated that mean global QoL scores and psychosocial aspects of health relating to career, finances, relationships, and emotional/social/family wellbeing were significantly impaired in EOCRC patients in comparison to LOCRC patients. This systematic review was limited by the large proportion of cross-sectional studies prohibiting the evaluation of temporal trends in QoL and by the significant heterogeneity across studies.⁴⁹ Recently, the REACCT Collaborative⁵⁰ reported on the functional outcomes of 1429 early-onset rectal cancer patients with a median follow-up time of 42 months. Of the 29% of patients with persistent functional impairment, the majority had bowel dysfunction (16%), followed by bladder (7%) and sexual (4.5%) dysfunction and infertility (1%). Further, Sandberg et al.⁵¹ in their longitudinal follow-up of the QoLiRECT study, found that low anterior resection syndrome improved only minimally over time for younger versus older patients. This highlights the desirability for thorough preoperative counseling to ensure informed consent, the implementation of preoperative nomograms to predict postoperative functional impairment, and careful consideration of sphincter-preserving and nerve-sparing techniques where possible.⁵⁰ To ascertain the long-term feasibility of current EOCRC treatment patterns, large prospective studies comparing the QoL of EOCRC and LOCRC patients in a quantitative manner with questionnaires tailored to CRC-specific functional outcomes are warranted.