Colorectal cancer (CRC) is a genetically, epigenetically, transcriptomically and immunologically heterogeneous disease in which such differences make a significant contribution to the differences in disease behavior, treatment responses and patient survival. While traditional histopathological evaluation and tumour–node–metastasis (TNM) staging are the most important prognostic factors, the wide clinical variability seen among patients with the same stage cannot be fully explained by those factors. In fact, due to recent developments in molecular pathology, CRC has moved from a morphologically based disease to a group of molecularly distinct tumor subgroups, which have different biological mechanisms and prognostic implications [1,2]. Multiple genetic and epigenetic alterations in oncogenic signaling, tumor-suppressor pathways, DNA repair mechanisms and epigenetic regulation have been shown through comprehensive genomic studies in CRC.

CRC can have two mechanisms of genomic instability: chromosomal instability (CIN) and microsatellite

instability (MSI). CIN is defined by aneuploidy, loss or gain of chromosomes, and structural abnormalities of the chromosomes and often is associated with mutations in the APC, TP53, and KRAS genes. MSI, on the other hand is mostly caused by defective DNA mismatch repair (MMR) and results in an increased number of mutations in repetitive DNA sequences and higher mutational load [3,4]. MMR deficiency may be due to germline changes linked to Lynch syndrome or to epigenetic changes, such as methylation of the MLH1 promoter. Not only are these molecular pathways relevant for colorectal carcinogenesis, but they also deliver clinically valuable information for prognosis, hereditary cancer risk evaluation and systemic therapy selection [5].

The MMR/MSI molecular phenotype is one of the most clinically significant CRC properties. Immunohistochemistry involves the use of mismatch repair proteins (MLH1, PMS2, MSH2, and MSH6) whereas molecular assays can be used to assess MSI. The current recommendations for diagnostics are supportive of MMR or MSI testing; because tumors that are deficient in MMR (dMMR)/have high MSI (MSI-H) have unique pathological and clinical features and may be highly responsive to immune checkpoint inhibitors [6]. Localized disease is generally associated with favorable prognosis and in the case of colon cancer with dMMR/MSI-H disease, stage II is associated with good prognosis, with possible stage-dependent and/or other molecular alterations. However, in some MSI-H tumours, BRAF V600E mutations can provide significant information on a more aggressive group of patients with worse prognosis, especially at advanced stages [8].

Another big part of CRC molecular pathology is mutations of the RAS – RAF – MAPK signaling pathway. Mutations in KRAS and NRAS drive down stream pathways that activate cells to proliferate and survive in a manner that is independent of the activation of upstream pathway. Since the discovery their role in predicting resistance to epidermal growth factor receptor (EGFR)-targeted therapy is crucial, as patients with activating mutations of the RAS genes do not benefit from anti-EGFR monoclonal antibodies [9]. Another molecular change is BRAF V600E and it has been linked to specific clinicopathological characteristics, such as right-sided primary tumors, less differentiation, and a more aggressive clinical course. Molecular analysis of BRAF is therefore not only important for stratification for prognosis but also to guide selection of targeted treatment strategies in metastatic CRC [8,10].

In addition to the individual mutations, transcriptomic profiling has set up the Consensus Molecular Subtype (CMS) classification as another significant framework understanding CRC heterogeneity. Each of the four main CMS groups (CMS1, MSI immune; CMS2, canonical; CMS3, metabolic; and CMS4, mesenchymal) has specific molecular pathways, tumor microenvironmental components and clinical characteristics [11]. CMS1 tumors are hypermutated and linked to MSI and strong immune activation, while CMS2 tumors show epithelial characteristics with activation of WNT and MYC pathways. CMS3 is defined by metabolic dysregulation; CMS4 shows transforming growth factor-β activation, stromal infiltration, angiogenesis and some characteristics of epithelial–mesenchymal transition [11,12]. These differences in molecules are significant and have prognostic value. Indeed, CMS4 consistently has been linked to worse survival in localized CRC, and stromal and mesenchymal biology also has a role in tumor invasion and metastatic potential [13].

Additionally, there is significant intra-tumoral and inter-tumoral heterogeneity, increasing the complexity of the clinical relevance of molecular classification. A single colorectal tumour can have several molecular subpopulations and the molecular makeup of a primary tumour can be different to that of the metastases [14]. This variability may affect drug sensitivity and lead to treatment resistance. These molecular profiling studies, in turn, have highlighted the importance of combining genomic changes, MSI/MMR status, transcriptomic signatures, immune traits and features of the tumor microenvironment instead of using just one of these [2,14,15].

Molecular pathology, therefore, is becoming an increasingly important tool in the evaluation of prognosis in CRC. Molecular biomarkers should not be used as a substitute for conventional histopathology and TNM staging; they should be used in combination with them to provide additional (and complementary) information regarding the biology of the tumor that is not available from the morphology. Combining MMR/MSI status with RAS and BRAF mutations, CpG island methylator phenotype and molecular subtype classification may provide further risk stratification and aid personalized therapeutic decisions [5,9,15]. The future of molecular risk prediction is likely to be enhanced and further developed through continued advances in next generation sequencing, digital pathology, liquid biopsy, and transcriptomic technologies. Therefore, a detailed knowledge of the molecular profile of CRC is vital to establishing more accurate prognostic models, identifying those patients with a higher risk of recurrence or death, and hence the development of more personalized treatment approaches.

The aim of this study was to explore the molecular pathological features of CRC and its correlation with clinicopathological characteristics and prognosis in a retrospective observational study in the tertiary care hospital. The study cohort consisted of 80 patients who were diagnosed with colorectal adenocarcinoma from 1 January 2026 to 30 June 2026 and were confirmed by histopathologic examination. Patients were screened using a consecutive sampling approach, based on hospital medical records, histopathology databases, pathology archives and available molecular diagnostic records. The patients were included if they were ≥18 years of age, had a confirmed colorectal adenocarcinoma diagnosis, and had sufficient clinical, pathological, and follow-up data. Patients whose colorectal malignancy recurs, whose histological type was non-adenocarcinoma, or who had incomplete medical or pathological records, inadequate tissue samples, or inadequate outcome information were not included. Data relating to age, sex, symptoms, family history, site of tumor, and the treatment were collected from the existing medical records. Histopathological parameters included tumour histological type and grade, depth of invasion, lymphovascular invasion, perineural invasion, lymph-node status and pathological TNM stage, and were retrieved from pathology reports. The formalin-fixed paraffin embedded tumour blocks were retrieved and examined for molecular pathological results. Mismatch repair (MMR) status was determined by immunohistochemical expression of MLH1, PMS2, MSH2, and MSH6. When one or more protein of the MMR system was not expressed, it was considered to be deficient mismatch repair (dMMR), and when all four proteins were expressed, it was considered to be proficient mismatch repair (pMMR). In addition, whenever possible, the molecular pathology records were used to obtain microsatellite instability (MSI) status and molecular alterations involving KRAS, NRAS, and BRAF. prognosis was obtained retroactively from hospital records, oncology notes, radiological reports, pathology records and documented follow-up visits. Disease recurrence, local or distant progression, metastatic disease, treatment response and survival status were all evaluated outcomes. Overall survival (OS) was measured from the date of initial diagnosis until death by any cause or last documented follow up date. The time to progression-free survival (PFS) was measured from diagnosis or treatment initiation to documented progression or recurrence, death, or last follow-up. Assessment of short term prognosis was performed at least 6 months post diagnosis (where follow-up information was available). For patients without documented progression or death, these were censored at the date of last follow-up. The IBM SPSS Statistics version 27 software was used for data analysis. The Shapiro–Wilk test was used to determine if continuous variables were normally distributed and then they were presented as mean ± standard deviation or as a median with interquartile range as appropriate. Categorical variables were displayed in frequency and percentage. The associations between molecular markers and categorical clinicopathological variables were assessed by chi-square test or Fisher's exact test, and the associations between molecular markers and continuous variables were assessed by independent-samples t-test and Mann–Whitney U test when the data were normally distributed and non-normally distributed, respectively. Logistic regression was used to determine independent clinicopathological and prognostic factors with odds ratios (ORs) and 95% confidence intervals (CIs). OS and PFS were estimated using the Kaplan–Meier analysis and survival differences between molecular subgroups were evaluated using the log-rank test. The two-sided p-value < 0.05 was regarded as statistically significant. Prior to data extraction ethical approval was obtained from the institutional research and ethical review committee of the participating tertiary care hospital. This study is retrospective and no intervention was done with the participants as only information recorded from the clinical and pathological history was used. Patient confidentiality was ensured by coding and the anonymisation of the data used for analyses. The study was carried out, following the Declaration of Helsinki guidelines.

The study included a total of 80 patients who have histopathologically confirmed colorectal adenocarcinoma. The average age of the subjects in the study was 56.8 ± 12.4 years and ranged from 29 to 82 years. There were 47 (58.8%) male and 33 (41.2%) female patients. Rectal bleeding (56.3%), change in bowel function (51.3%), abdominal pain (46.3), and loss of weight (32.5) were the most common presenting symptoms. In 52 (65.0%) patients the primary tumor was in the colon and in 28 (35.0%) patients in the rectum. Moderately differentiated adenocarcinoma was the most frequent tumor grade (49 [61.3%]) followed by poorly differentiated (19 [23.8%]) and well differentiated (12 [15.0%]). Lymphovascular invasion was seen in 31 (38.8%) cases and perineural invasion in 24 (30.0%) cases. The pathological TNM classification of the disease showed that stage III was the most common (37.5%) followed by stage II (32.5%), stage IV (21.3%) and stage I disease (8.8%) (Table 1).

Table 1. Demographic and clinicopathological characteristics of patients with colorectal cancer (n=80)

61 patients (76.3%) were proficient mismatch repair (pMMR) and 19 patients (23.8%) were deficient mismatch repair (dMMR) on MMR immunohistochemistry. Loss of MLH1/PMS2 expression was the most common of the dMMR tumors. MSI status was available in 64 patients of which 15 (23.4%) had an MSI-high tumor and 49 (76.6%) had an MSI-stable or MSI-low tumor. Of those with molecular testing, 27 (39.7%) were found to have a KRAS mutation and 7 (10.3%) had a BRAF V600E mutation. Four (5.9%) patients were found to have NRAS mutation. Selected clinicopathological features were significantly associated with molecular abnormalities. dMMR/MSI high tumors were more likely to be in the right colon, and were associated with poorer histological differentiation, while KRAS and BRAF alterations were more common among patients with advanced stage disease. The distribution of major molecular markers based on clinicopathological features is shown in Table 2.

Table 2. Molecular pathological characteristics of colorectal cancer

In the follow up period available, 22 (27.5%) patients had disease progression or recurrence, and 18 (22.5%) had distant metastases documented. During the follow-up, 8 patients (10.0%) died. Progression was more common in patients with pMMR tumors compared to patients with dMMR tumors (31.1% vs. 15.8%), but not statistically significant (p=0.18). The incidence of disease progression was higher in patients with KRAS-mutated tumor than in patients with KRAS wild-type tumor (37.0% vs. 19.5%, p=0.08). There were a higher proportion of adverse outcome (progression or recurrence) in the BRAF-mutated tumors (4/7, 57.1%) than in the non-BRAF-mutated tumors (18/61, 29.5%), p=0.14. LVI, perineural invasion, lymph-node positivity and advanced TNM stage were significant factors for poor prognosis (Table 3).

Table 3. Molecular and clinicopathological factors associated with adverse prognostic outcomes

Differences in progression-free survival were shown using Kaplan–Meier analysis, by molecular subgroup. The progression-free survival (PFS) was relatively longer for patients with dMMR/MSI-high tumors compared with patients with pMMR/MSS tumors (Figure 1).

Figure 1. Kaplan–Meier curve comparing progression-free survival between dMMR/MSI-high and pMMR/MSS colorectal cancer patients.

 In the same way, the patients with KRAS wild-type tumors had a higher PFS rate compared to those with KRAS-mutant tumors, but the result was not statistically significant. The overall survival was significantly affected by TNM stage, with patients with stage I–II disease having a significantly better overall survival than patients with stage III–IV disease (Figure 2). Advanced stage, TNM and lymph-node involvement were the most powerful independent predictors on multivariable analysis for unfavorable six month outcome, while molecular markers were weaker, independent of other clinicopathological factors.

Figure 2. Kaplan–Meier curve demonstrating overall survival according to TNM stage (I–II versus III–IV).

The present study analyzed the molecular pathological data of colorectal adenocarcinoma and investigated the correlations between these data and clinicopathological factors and short-term prognosis. Of these 80 patients, proficient mismatch repair (pMMR) was found in a higher proportion of patients than deficient mismatch repair (dMMR) and KRAS mutations were the most common of the tested RAS/RAF mutations. The study also showed that traditional pathological features, such as advanced TNM stage, lymphovascular invasion, lymph-node involvement, and perineural invasion were still highly correlated with poor prognosis. The results highlight the role of molecular biomarkers as a source of biological information, but also they must be used in conjunction with known histopathological and staging parameters. Among the current cohort, 23.8% of patients had been identified as dMMR and 23.4% of patients in whom MSI testing was possible were identified as MSI-high. This ratio is of clinical importance as mismatch repair deficiency is one of the most significant molecular phenotypes of colorectal cancer. The MMR status is a reflection of the underlying genomic instability of the tumour and is relevant to prognosis and therapeutic decision making. The molecular pathology classification currently classifies colorectal cancer as a disease characterized by the presence or absence of MMR and MSI, depending on the molecular alterations present, although MMR alone may have different prognostic values in different stages of the disease [16]. The relatively good progression profile that was seen in all dMMR/MSI-high tumors in the current study, therefore, is biologically plausible and consistent with the known behavior of this group of tumors in the clinic. In the present study, it was also observed that the dMMR/MSI-high tumors were more common in right-sided colon tumor location and poor differentiation. These features are typical of the MSI-associated colorectal cancer phenotype, which is characterized by tumor location in the proximal colon, immune-cell infiltration, and pathological features [16,17]. However, caution must be used in interpreting the lack of statistical significance between the differences in progression between dMMR and pMMR tumors in our cohort. There were only a limited number of patients and follow-up, which may have limited the sensitivity of the study to identify clinically significant differences. Additionally, MSI/MMR status is just one of many aspects of the molecular picture, and the prognostic impact may be influenced by BRAF status, tumour stage, immune microenvironment and treatment exposure. In the present study, about 40% of the molecularly examined tumors contained KRAS mutations. It is important to note that patients with tumors bearing the KRAS mutation had a numerically higher rate of progression and/or recurrence than those with KRAS wild-type tumors, but the difference was not statistically significant at the 95% level. The results are consistent with previous studies that suggest that KRAS mutation may be linked to poorer prognosis in CRC. Systematic review and meta-analysis of stage II and III colon cancer revealed that analysis of KRAS and BRAF alterations has prognostic significance but their independent effects were dependent upon disease setting and therapeutic context [17]. Likewise, a meta-analysis of studies on metastatic colorectal cancer patients showed that KRAS mutations are linked to poorer survival, suggesting that it could be a predictive and possibly a prognostic marker [18]. Therefore, the nonsignificant results in the current study might be due to a small number of subjects and to the relatively short follow-up period, and not to a lack of a biological relationship. BRAF V600E mutation was detected in 10.3% of patients who had testing and was found to be numerically more common in patients who progressed or recurred. The association was not statistically significant, but the direction of the association is congruent with previous data suggesting that BRAF mutation is a poor prognostic feature. However, the prognostic impact of BRAF cannot be considered independent from the MMR/MSI background, as BRAF-mutated tumors have been found to be associated with adverse outcome even in the dMMR/MSI-high population [19]. The relatively small number of BRAF-mutated tumors in our study makes it difficult to draw conclusions for an independent prognostic effect. More patients and a longer follow-up are needed to see if BRAF status gives us any prognostic information over and above TNM stage and MMR status. The present study yielded an important result, in that the pathological variables were more strongly associated with adverse outcomes than the conventional pathological variables. Hazardous lymphovascular invasion and perineural invasion were significantly related to progression or recurrence. The results corroborate the idea that pathological features of invasive tumor activity are also of great relevance in the age of molecular oncology. In particular, there is a strong correlation between perineural invasion and aggressive disease and survival in colorectal cancer. According to a recent meta-analysis, PNI is a poor prognostic factor, and suggested that the prognosis of stage II patients with PNI might be comparable with some stage III patients [20]. Hence, PNI and LVI, if documented routinely, continue to be relevant for the development of personalized risk profiles. The biological role of tumor–stroma interaction and epithelial–mesenchymal transition is also in favor of the tumor invasion–adverse outcome relationship. Tumor budding was not assessed in the present study, but is another histopathological parameter associated with invasive potential. Tumor budding has been linked to metastasis and poor prognosis, and is believed to represent a shift toward a more aggressive tumor type [21,22]. The use of such pathological features in conjunction with molecular markers could give a better estimate of the aggressiveness of a tumour than either marker alone. The majority of our patients had advanced disease, and TNM stage was strongly correlated with outcome, further underscoring the continued importance of the conventional TNM classification system. Frequent progression or recurrence were seen in patients with stage III–IV disease compared to stage I–II patients. Anatomic staging should therefore be used in conjunction with molecular staging. At present, there is some evidence that molecular subtyping can offer further prognostic information, and that it is most valuable when combined with existing pathological and clinical markers [23]. Consensus molecular subtypes have shown different prognosis, in particular, mesenchymal/CMS4 has a worse prognosis in localized disease [23]. The potential of molecular biomarkers to predict outcome is growing beyond the confines of mutation analysis of tissue. With the advent of new techniques like circulating tumor DNA (ctDNA), there is the potential to get dynamic information on the presence of residual disease or early recurrence, which cannot be obtained by routine histopathological interpretation in all cases. A recent systematic review and meta-analysis showed that there was a statistically significant correlation between ctDNA detection and shorter recurrence-free and overall survival in colorectal cancer patients [24]. The use of ctDNA in the current study was not evaluated; however, combining tissue-based MMR, RAS, BRAF and other molecular markers may help facilitate postoperative risk stratification and surveillance in future studies. There are some limitations of the present study. For one, it has a retrospective design, which means that it is prone to selection bias and information bias due to the fact that the analysis was based on the completeness and accuracy of existing medical records. Second, the number of patients (80) was small, reducing the statistical power of the analyses of subgroups, including those with relatively rare alterations, such as BRAF mutations and NRAS mutations. Third, not all patients were tested for all biomarkers, which created uneven denominators across the individual biomarkers. Fourth, the relatively short follow-up time restricts the evaluation of long-term overall and disease-free survival. Lastly, comprehensive molecular profiling (CMS classification, tumor mutational burden, alterations of POLE and POLD1, and ctDNA) was not available. In spite of these shortcomings the study offers clinically relevant evidence and supports the need for molecular pathology to be used in conjunction with traditional histopathological evaluation. Overall, these results suggest that CRC is highly molecularly and pathologically heterogeneous and that alterations in MMR/MSI, KRAS and BRAF may yield complementary information about the biology and prognosis of CRC. But, among these, TNM stage, LVI, LN involvement and perineural invasion were more strongly associated with adverse outcome than single molecular markers were in this group. Therefore, molecular pathology should be considered in addition to and not instead of conventional pathological staging. More extensive and longer follow-up studies are warranted to develop comprehensive molecular–pathological prognostic models for CRC.

The present study revealed significant molecular and clinicopathological heterogeneity of colorectal cancer. TNM stage and other conventional pathological parameters, such as lymphovascular invasion, lymph-node invasion, and perineural invasion, were also highly correlated with poor prognosis, while mismatch repair status, MSI, KRAS and BRAF alterations gave valuable insights into tumour biology and prognosis. The results highlight the importance of the combination of molecular pathological evaluation with routine histopathological examination and TNM staging to optimize the prognosis stratification and consequently more individualized treatment of colorectal cancer patients.

Recommendations

Appropriate molecular assessment, such as MMR/MSI testing and clinically relevant RAS/ BRAF mutations, should be included in the diagnosis, prognosis, and management of colorectal cancer. This should not be used alone, and should be combined with TNM stage and known histopathological risk factors. These associations should be confirmed in larger multicenter prospective studies with longer follow-up and in-depth genomic profiling, and molecular–pathological prognostic models should be developed. The combination of new biomarkers, like circulating tumor DNA, could further enhance the ability to detect minimal residual disease, early recurrence and individualized postoperative monitoring.

1. Woischke C, Michl M, Neumann J. Molecular pathology of colorectal cancer. Pathologie (Heidelb). 2023;44(5):279-286. doi:10.1007/s00292-023-01201-9.
2. Guo L, Wang Y, Yang W, et al. Molecular profiling provides clinical insights into targeted and immunotherapies as well as colorectal cancer prognosis. Gastroenterology. 2023;165(2):414-428.e7. doi:10.1053/j.gastro.2023.04.029. Active DOI (PubMed)
3. Cancer Genome Atlas Network. Comprehensive molecular characterization of human colon and rectal cancer. Nature. 2012;487(7407):330-337. doi:10.1038/nature11252.
4. Taieb J, Svrcek M, Cohen R, et al. Deficient mismatch repair/microsatellite unstable colorectal cancer: diagnosis, prognosis and treatment. Eur J Cancer. 2022;175:136-157. doi:10.1016/j.ejca.2022.07.020.
5. Imyanitov E, Kuligina E. Molecular testing for colorectal cancer: clinical applications. World J Gastrointest Oncol. 2021;13(10):1288-1301. doi:10.4251/wjgo.v13.i10.1288.
6. Bartley AN, Mills AM, Konnick E, et al. Mismatch repair and microsatellite instability testing for immune checkpoint inhibitor therapy: guideline from the College of American Pathologists in collaboration with the Association for Molecular Pathology and Fight Colorectal Cancer. Arch Pathol Lab Med. 2022;146(10):1194-1210. doi:10.5858/arpa.2021-0632-CP.
7. Zhang C, Zhan Y, Ni K, et al. Effects of deficient mismatch repair on the prognosis of patients with stage II and stage III colon cancer during different postoperative periods. BMC Cancer. 2022;22:1156. doi:10.1186/s12885-022-10266-3.
8. Tan E, Whiting J, Xie H, et al. BRAF mutations are associated with poor survival outcomes in advanced-stage mismatch repair-deficient/microsatellite high colorectal cancer. Oncologist. 2022;27(3):191-197. doi:10.1093/oncolo/oyab055.
9. Clinical significance of somatic mutations in RAS/RAF/MAPK signaling pathway in Moroccan and North African colorectal cancer patients. Asian Pac J Cancer Prev. 2022;23(11):3725-3732. doi:10.31557/APJCP.2022.23.11.3725.
10. Cohen R, Liu H, Fiskum J, et al. BRAF V600E mutation in first-line metastatic colorectal cancer: an analysis of individual patient data from the ARCAD database. J Natl Cancer Inst. 2021;113(10):1386-1395. doi:10.1093/jnci/djab042.
11. Guinney J, Dienstmann R, Wang X, et al. The consensus molecular subtypes of colorectal cancer. Nat Med. 2015;21(11):1350-1356. doi:10.1038/nm.3967.
12. Ros J, Baraibar I, Martini G, et al. The evolving role of consensus molecular subtypes: a step beyond inpatient selection for treatment of colorectal cancer. Curr Treat Options Oncol. 2021;22(12):113. doi:10.1007/s11864-021-00913-5. Pietrantonio F, et al. Clinical value of consensus molecular subtypes in colorectal cancer: a systematic review and meta-analysis. J Natl Cancer Inst. 2021. doi:10.1093/jnci/djab106.
13. Joanito I, Wirapati P, Zhao N, et al. Single-cell and bulk transcriptome sequencing identifies two epithelial tumor cell states and refines the consensus molecular classification of colorectal cancer. Nat Genet. 2022;54(7):963-975. doi:10.1038/s41588-022-01100-4.
14. Eide PW, Moosavi SH, Eilertsen IA, et al. Metastatic heterogeneity of the consensus molecular subtypes of colorectal cancer. npj Genom Med. 2021;6:59. doi:10.1038/s41525-021-00223-7.
15. Noack P, Langer R. Molecular pathology of colorectal cancer. Memo. 2023;16:116–121. doi: 10.1007/s12254-023-00893-2.
16. Formica V, Sera F, Cremolini C, Riondino S, Morelli C, Arkenau HT, et al. KRAS and BRAF mutations in stage II and III colon cancer: a systematic review and meta-analysis. J Natl Cancer Inst. 2022;114(4):517–527. doi: 10.1093/jnci/djab190.
17. Yoshino T, Van Cutsem E, Li J, Shen L, Kim TW, Sriuranpong V, et al. Effect of KRAS codon 12 or 13 mutations on survival with trifluridine/tipiracil in pretreated metastatic colorectal cancer: a meta-analysis. ESMO Open. 2022;7(3):100511. doi: 10.1016/j.esmoop.2022.100511.
18. Tan E, Whiting J, Xie H, Imanirad I, Carballido E, Felder S, et al. BRAF mutations are associated with poor survival outcomes in advanced-stage mismatch repair-deficient/microsatellite high colorectal cancer. Oncologist. 2022;27(3):191–197. doi: 10.1093/oncolo/oyab055.
19. Prognostic value of perineural invasion in colorectal cancer: a meta-analysis. Int J Colorectal Dis. [Meta-analysis]. doi: 10.1016/S1091-255X(23)00348-7
20. Shah AH, Gami AJ, Desai NH, Gandhi JS, et al. Tumor budding as a prognostic indicator in colorectal carcinoma: a retrospective study of primary colorectal carcinoma cases in a tertiary care center. J Cancer Res Ther. 2022;13(3):459–467. doi: 10.1007/s13193-022-01498-7.
21. Hatthakarnkul P, Quinn JA, Ammar A, Lynch G, Van Wyk H, McMillan DC, et al. Molecular mechanisms of tumour budding and its association with microenvironment in colorectal cancer. Clin Sci (Lond). 2022;136(8):521–535. doi: 10.1042/CS20210886.
22. Ten Hoorn S, de Back TR, Sommeijer DW, Vermeulen L. Clinical value of consensus molecular subtypes in colorectal cancer: a systematic review and meta-analysis. J Natl Cancer Inst. 2022;114(4):503–516. doi: 10.1093/jnci/djab106.
23. Zhou Q, Chen X, Zeng B, Zhang M, Guo N, Wu S, et al. Circulating tumor DNA as a biomarker of prognosis prediction in colorectal cancer: a systematic review and meta-analysis. J Natl Cancer Cent. 2025;5(2):167–178. doi: 10.1016/j.jncc.2024.05.007.