CircRNAs are generated during transcription through the formation of a 3’-5’ phosphodiester linkage between an upstream 3’splice site (splice acceptor) and a downstream 5’splice site (splice donor) (14, 15). Predominantly, circRNAs are primarily categorized into three principal categories, contingent upon their originating sequences: Exon-intro circRNAs (ElciRNAs), intron-derived circRNAs (ciRNAs), and exonic circRNAs (ecRNAs) (16–18). EIciRNAs and ciRNAs typically reside in the nucleus and are implicated in transcriptional regulation through interactions with the Pol II transcription machinery and U1 snRNP, respectively (19, 20). Conversely, EcRNAs predominantly localize in the cytoplasm, where they can modulate gene expression by acting as miRNA sponges or interacting with RNA-binding proteins (21, 22). A fourth category of circRNAs, known as intergenic circRNAs, has recently been identified. These circRNAs primarily originate from the intergenic regions of the genome and consist of two intronic circRNA fragments formed by the reverse variable splicing of the flanking integrated circRNA. Intergenic circRNAs are non-exonic and have gained significant attention due to advancements in RNA sequencing technology and bioinformatics. Unlike linear RNAs, intergenic circRNAs contain two intronic fragments that are flanked by GT-AC splicing signals, which act as loop-linked splice donors and acceptors. Sequence analyses have indicated that intergenic circRNAs generally exhibit weak conservation. Despite this, knowledge about these circular transcripts and the specific mechanisms regulating circRNA biogenesis is still limited. Some research has provided insight into their potential roles. For instance, a study identified an intergenic circRNA, hsa_circ_0007379, which is downregulated and acts as a novel tumor suppressor in CRC (23). The diverse localization and functionalities of these circRNA classes underscore their potential multifaceted roles in cellular processes and disease states, including CRC.

Originally deemed as inconsequential, low-abundance splicing byproducts (18), circRNAs have been catapulted into scientific prominence with advancements in sequencing technologies and analytical methodologies, revealing that myriad genes are capable of generating circRNAs, with a single gene often producing diverse circRNAs through alternative splicing cycles. The biogenesis of circRNAs often hinges on non-canonical splicing patterns within pre-mRNAs.

ElciRNAs interact with U1 snRNP, predominantly localize within the nucleus, and facilitate the transcription of their parental genes. Exon-intron circRNAs, or ElciRNAs are characterized by the retention of introns between circularized exons. The majority of eukaryotic circRNAs, known as ecircRNAs, are formed by exonic sequences. EIciRNAs ( Figure 1 ), which are made up of both intronic and exonic sequences, and ecRNAs ( Figure 1 ), consisting of exonic sequences, are produced through back splicing in the nucleus. Additionally, since ciRNAs ( Figure 1 ) are formed from introns, a typical intron may be modified to become one. The prevailing theory posits that eukaryotic tRNA splicing endonuclease (TSEN) splicing is orchestrated by a ruler mechanism, whereby splice sites are selected according to their proximity to the recognized tRNA body (24). Subsequent to TSEN cleavage, tRNA halves are conjoined to form a mature tRNA. The RtcB enzyme, essential for tricRNA ligation, ligates tRNAs, underscoring the requisite nature of RtcB’s ligase activity. During pre-tRNA maturation, TSEN components cleave pre-intron-containing pro-tRNAs in the conventional bulge-helix-bump (BHB) motif. The resultant intron ends are ligated by RtcB ligase, culminating in the formation of tricRNA, a stable circRNA (25).

Compared to mRNA, circRNA exhibits a notably resilient structure. Presently, three principal processes are posited to orchestrate circRNA degradation (26). Firstly, the circRNA-miRNA axis, which dismantles certain circRNAs by leveraging highly complementary binding sites for specific microRNAs (miRNAs), for instance, miR-671 directs the degradation of CDR1as. Secondly, circRNA degradation mediated by N6-methyladenosine (m6A) modification emerges as another pivotal mechanism (27). YTHDF2 (27) interacts with RNase P/MRP (mitochondrial RNA processing) to recognize circRNAs containing the m6A gene through heat-reactive protein 12 (HRSP12) bridging. Additionally, exosomes or extracellular vesicles may be used to remove circRNAs. For instance, exosomes encompass molecules like circSCMH1 and MeCP2 that exhibit persistent presence. Thirdly, MDM2, instrumental in quelling apoptosis by inhibiting p53, Foxo3, and their downstream molecule Puma, plays a crucial role. During cancer cell apoptosis, the circ-Foxo3 expression has been observed to markedly escalate. Consequently, the down-regulation of circ-Foxo3 impedes cancer cell apoptosis (28, 29). For example, the formation of the circ-Foxo3-p21-CDK2 ternary complex arrested the function of CDK2 and blocked cell cycle progression (30). Deng et al. noted a diminution of circ-FOXO3 in tumor tissues, and the overexpression of LATS543 precipitated apoptosis in CRC cells (29).

The stability of circRNA is mainly reflected in its unique circular structure and its ability to resist degradation. Circular RNA forms a closed loop structure through the reverse splicing mechanism, which makes it unaffected by RNA exonuclease, thus having high stability and not susceptible to degradation. (31)Specifically, due to the lack of a 5 ′ cap and 3 ′ poly (A) tail, circRNA is generally resistant to exbonuclease attack and shows higher stability than homologous linear mRNA with the same nucleotide sequence (32). This property provides the circular RNA with a longer cellular half-life and higher stability, which is crucial for the durable effect of the drug. Moreover, the stability of circular RNA is also manifested in its resistance to nucleases (33). Due to its circular structure, the circular RNA is resistant to degradation by nucleases, maintaining its integrity. This stability allows circular RNA to exist in living organisms for long periods, participate in various biological processes, and has potential applications in disease diagnosis and treatment. (21) Because the internal translation of circRNA is generally inefficient compared to classical cap-dependent translation of linear mRNA, detection of protein and peptide products produced by circRNA is challenging due to technical limitations.

CircRNAs are now recognized to affect the expression of parental genes, function as miRNA sponges, interact with proteins, serve as miRNA sponges, and contribute to the onset and progression of diseases (34).

The disruption of the canonical cell cycle and the presence of enhanced proliferative capabilities are considered the mainstays of tumorigenic pathways (48). Tumorigenesis is predominantly characterized by the activation of oncogene and the inactivation of tumor suppressor genes, which result in aberrant cell cycle. Recently, several circRNAs have been found to support the biological action of cyclins or CDKs in CRC. For instance, Circ_001621 attenuated the restraint of matrix metallopeptidase 9 (MMP9) and cycling-dependent kinase 4 (CDK4) by miR-578, thereby increasing osteosarcoma proliferation and migration, respectively (49). By regulating the balance of glycine/serine metabolism and redox in a p53-dependent manner, circMYH9 overexpression encourages CRC growth (50). CircPPP1R12A influences the Hippo-YAP pathway in CRC cells, enhancing the cells’ ability to proliferate, migrate, and invade (43). Conversely, overexpression of circPIP5K1A promotes CRC cell invasion and migration, while its diminution curtails cell viability and motility (51).

Certain circRNAs modulate CDK or cyclin activity and regulate cell cycle-specific genes to promote cellular growth (52–54). Abnormally produced circRNAs in malignancies may influence cancer cell proliferation by controlling key cell cycle regulators such as cyclins, p53, CDKs, CDK inhibitors (CKIs), and pRB. CircRNAs also affect cell proliferation by forming complexes with factors involved in the cell cycle, controlling CCN through RNA-binding proteins, and regulating major Wnt/β-catenin, phosphatidylinositol-3 kinase (PI3K)/AKT/FOXO, and other signaling pathways (55).

Moreover, circRNAs can affect CRC’s invasion, tumor lymph node metastasis stage, malignant phenotypes, histological grade, and chemotherapy treatment susceptibility by influencing the aforementioned cell cycle variables.

Exosomes, diminutive small extracellular vesicles, play a pivotal role in tumor spreading by carrying specific information between cells. They have been found to contribute to various processes involved in cancer progression, such as the epithelial-to-mesenchymal transition (EMT), extracellular matrix (ECM) remodeling, tumor angiogenesis, immune evasion, and the establishment of pre-metastatic niches. Exosomes can also influence organotropic metastasis, which refers to the preference of cancer cells to metastasize to specific organs (39). For instance, in CRC, the circRNAs Hsa_circ_0001666 has been found to decelerate disease progression by modulating the miR-576-5p/PCDH10 axis (56). Moreover, it diminishes EMT and stemness in CRC cells, while also affecting the Wnt/β-catenin signaling pathway ( Figure 2 ). Another circRNA, circRNA-ACAP2, is expressed in elevated levels in CRC tissues compared to normal tissues. The knockdown of circRNA-ACAP2 inhibits the proliferation, migration, and invasion of SW480 cells by upregulating miR-21-5p and inhibiting Tiam1, a downstream target of miR-21-5p (57) ( Figure 2 ). Additionally, circPIP5K1A has been found to promote colon oncogenesis by inhibiting miR-1273a (51) and amplifying AP-1 activity ( Figure 2 ). Furthermore, hsa_circ_0006732 may act as an oncogene in CRC by altering the miR-127-5p/RAB3D axis, influencing CRC cell proliferation, migration, invasion, and EMT (58) ( Figure 2 ).

CircFoxo3 is a well-studied circRNA derived from the tumor suppressor gene Foxo3 (28), has been implicated in various roles across the development of numerous cancer types. For instance, by boosting FOXO3a expression, ANGPTL1 significantly reduces the expression of the stem cell transcription factor SOX2 in CRC cells, thereby attenuating the cells’ migratory, invasive capabilities, expression of cancer stem cell (CSC) markers, and sphere-forming ability (59, 60). Additionally, ALKBH5, modulating the FOXO3/miR21/SPRY2 axis in CRC, acting as an anticancer agent (61–63). In CRC tissue, circZNF609 expression is downregulated, and it fosters apoptosis by upregulating p53 in CRC cells (64). Furthermore, the disruption of circAPLP2 slows glycolysis in vitro and mitigates tumor growth in vivo by elevating miR-485-5p and diminishing FOXK1 (65).

Neovascularization is pivotal in the growth of malignancies, enabling distant metastasis and facilitating nutrient supply and waste removal as tumors attain substantial size (66). Certain circRNAs have been found to promote angiogenesis by modulating angiogenic regulators such as vascular endothelial growth factor (VEGF). For instance, VEGFs interact with tyrosine kinase cell receptors (VEGFR1-3) to activate VEGF signaling in endothelial cells, thereby influencing cell proliferation, migration, and vascular permeability during angiogenesis (67). In CRC, CircCCT3 has been shown to promote CRC metastases through the miR-613/WNT3 or miR-613/VEGFA axis (68, 69) ( Figure 3 ). Additionally, Circ_001971 acts as a competing endogenous RNA (ceRNA) to counteract miR-29C-3p’s suppression of VEGFA, thereby enhancing CRC growth, invasion, and angiogenesis (70) ( Figure 3 ). Moreover, Circ_0056618 upregulates CXCR4 and VEGF-A in CRC, promoting cell proliferation, migration, and angiogenesis by sponging miR-206 and eliminating its repressive effects (70, 71) ( Figure 3 ).

Glycolysis, pivotal for ensuring an efficient energy supply, particularly in low-oxygen environments, involves key glycolytic enzyme called HK2, which catalyzes the conversion of glucose to glucose-6-phosphate (72, 73). Overexpression of HK2 promoted cell invasion and migration by increasing glycolysis. In CRC, the downregulation of Circ-PITHD1 inhibited CRC invasion and proliferation by regulating miR-590-5p/HK2 signaling-mediated aerobic glycolysis. In certain cancers, hypoxia-induced circRNF13, mediated by HIF-1 and EIF4A3, promotes tumor development and glycolysis (74–76). Hsa_circ_0000231, which suppresses cell glycolysis and promotes cell death, exhibited diminished when hsa-502-5p inhibitor was utilized (77, 78). However, further study is urgently required to elucidate the role of circRNA in cellular energetics of CRC.

The PD-1/PD-L1 pathway, an immunological checkpoint, assists tumor cells in evading the immune system, thereby emerging as a compelling target for antitumor immunity (79). Delivery of circRERE-AAV causes significant anticancer effects in preclinical models of CRC, and a combination therapy utilizing circRERE-AAV and an anti-PD-1 antibody exhibits synergistic effects on tumor formation (80). By increasing the expression of PD-L1, hsa_circ_ 0136666 (81), hsa_circ_ 0020397 (82), CDR1-AS (83), and circEIF3K (84) promotes the carcinogenesis and immune evasion of CRC, potentially providing new insights into the molecular immunopathogenesis of CRC. Further studies are needed to better understand the role of noncoding RNAs in tumor immunotherapy and the cancer immune response.

The prevalence of drug resistance is notably high, significantly diminishing the chemotherapeutic efficacy once resistance is established. In CRC tissues and cells, an elevation in miR-31-5p and a downregulation of CircDDX17 and KANK1 have been observed. The upregulation of CircDDX17 increased 5-Fu sensitivity and slowed CRC progression (85). CircDDX17, by sponging miR-31-5p, impedes CRC development and 5-Fu resistance. Conversely, the knockdown of FOXO3 via FOXO3 shRNA restored the doxorubicin sensitivity of doxorubicin-resistant HCT116 DR cells. Conversely, FOXO3 overexpression decreased the doxorubicin sensitivity of HCT116 cells (86). In CRC cells, CircLHFPL2 successfully overcame PIK3CA-mediated MEK inhibitor resistance (87). Additionally, CircRNA_101277 has been found to augment cisplatin resistance by suppressing miR-370, elevating the miR-370 target gene IL-6, and activating the miR-370/IL-6 axis (88, 89).

The TME encompasses various components, including Cancer-associated endothelial/fibroblast cells (CAEs/CAFs), tumor-related macrophages (TAMs), NK cells, and TEXs (90–93). CAFs, subjected to either normoxic or hypoxic treatment, have been observed to release exosomes under hypoxic conditions (94). CRC cells treated with exosomes from sh-circEIF3K CAFs formed fewer cell colonies compared to those treated with exosomes from sh-NC CAFs. Moreover, patients with low circEIF3K levels exhibited longer median survival times compared to their high-level counterparts. TAMs are key players in tumor immunity and escape within the TME. The results of the ISH experiment revealed that circMERTK was substantially more abundant in CRC tissues than in comparable normal tissues, with its expressions largely overlapping with the macrophage marker CD68. The M2 polarization of macrophages is facilitated by exosomal circRNAs (95, 96). Exosomal circEIF3K may influence PD-L1 expression in CRC (84), while exosomal circUHRF1 promotes TIM3 expression in NK cells, diminishing NK cell functionality. Blocking MALAT-1 in the CCL5/MALAT-1/Snail axis prevented Snail overexpression from causing CCL5-induced migration and invasion, demonstrating the vital of this axis in the DC-dependent cancer development. Cancer cells can mediate immune evasion by using exosomal PD-L1. The function of T cells may be inhibited if TEXs carrying PD-L1 reach the lymphatic system from the blood, preventing immune cells from identifying and eliminating tumor cells. Patients with CRC exhibited elevated expression of hsa-circ0004771 in serum exosomes, suggesting its potential future utility as a CRC biomarker (97).

The early detection of CRC is pivotal for effective treatment and is often associated with successful outcomes (84, 101). Given their distinct expression patterns in CRC tissues and blood, compared to healthy controls, circRNAs emerge as promising candidates for tissue or liquid biopsies aimed at diagnosing CRC. For instance, hsa_circ_001978, hsa_circ_105039, and hsa_circ_103627 have been identified as potential novel biomarkers for CRC prediction (102). CircRNA_105039, crucial for augmenting CCND2 expression and promoting the formation of cardiomyocytes formation through miR17 absorption, along with circ-FMN2, circ-LMNB1 (103), and circ-ZNF609 may serve as viable diagnostic and prognostic markers for CRC identification. Hsa_circ_0026416 may act as a ceRNA by upregulating NFIB expression by competitively absorbing miR-346 (104). Circ_PVT1 and circ_001569, exhibiting significant upregulation in individuals with CRC, suggest their potential utility as diagnostic and prognostic marker for CRC (105). Hsa_circ_0007534 may serve as a potential cancer marker in CRC patients and might be associated with adverse prognostic outcomes (106). Furthermore, the diminished expression of hsa_circ_0001649 in CRC suggests its potential utility as a biomarker for targeted and sensitive CRC screening (107).

Intervening with prognostic variables is imperative to enhance clinical outcomes and potentially facilitate curative approaches in CRC (108–110). CircRNA_0001178 and circRNA_0000826 have emerged as potential diagnostic biomarkers for liver metastases originating from CRC (111, 112). MiR-7, considered a tumor suppressor in CRC, inhibits the expression of numerous oncogenes and has been implicated in tumor suppression through combined circHIPK3 silencing and miR-7 overexpression, as evidenced by in vitro and in vivo studies (113, 114). CircLONP2, acting as a pivotal initiator of metastasis during CRC progression, influences miR-17 intracellular maturation and intercellular transfer (115). Circ_382390 inhibits miR-30c-5p, which increases MYC and CCND1 in the target gene TCF7 (116). CircHUWE1, potentially sponging miR-486, may counteract the tumor-suppressive effects of silence when miR-486 is downregulated (117). Additionally, both hsa_circ_0052184 (118) and CircVAPA have been implicated in promoting CRC proliferation and migration (119). Circ_0021977, acting through the circ_0021977/miR-10b-5p/p21&p53 axis, presents itself as a potential target for CRC therapy, influencing invasion (120).

CircRNAs have emerged as potential therapeutic targets in CRC, with over 70 elevated circRNAs exerting a tangible impact on CRC carcinogenesis. The silencing of these circRNAs yields antithetical effects both in vitro and in vivo, positioning these oncogenic circRNAs as viable therapeutic targets. Specially created targeted interference RNAs can precisely target special post-splicing ligations of cancer-causing circRNAs for anti-tumor effects. For instance, circ-RNF121 absorbs miR-1224-5p and associates with FOXM1, thereby inhibiting tumor development in vivo (121). Additionally, circ-133 (122) and hsa_circ_0005963 (74) have been identified as therapeutic biomarkers in CRC. CircIFT80, by sponging miR-634, miR-568, and miR-142, and its circIFT80’s positive connection with CTNNB1 (β-catenin), augments gene expression (123). The circFNDC3B-miR-97-5p-TIMP3 pathway has demonstrated tumor-suppressive activity, and circFNDC3B-enriched exosomes have been shown to inhibit angiogenesis and CRC progression (124). CircCOG2, when upregulated, accelerates CRC proliferation and invasion via the miR-1305/TGF-2/SMAD3 pathway (125). Exosomes derived from CRC cells with high metastatic potential could spread this effect to CRC cells with lower metastatic potential. Circ_0006174, an oncogenic circRNA, modulates the miR-1205/CCBE1/Wnt pathway, facilitating the progression of CRC (126) ( Table 1 ). It is ensured that all referenced studies have been approved by the respective ethics committees and adhere to all necessary ethical standards, all participants signed an informed consent form.

The decision to use targeted therapies has traditionally been based on histopathological features and genotyping. However, only a small proportion of patients benefit from these approaches, highlighting the need for more refined biomarkers, particularly those that indicate specific sensitivity to targeted drugs. Next-generation sequencing and high-throughput technologies have significantly advanced biomarker discovery and enhanced our understanding of the underlying mechanisms of targeted therapies.

Due to its short half-life and significant fluctuations in expression under different physiological conditions, mRNA is not an ideal predictor for targeted therapy responses. In contrast, circRNAs offer advantages over linear mRNA due to their stable, circular structure, which confers significant resistance to degradation and greater stability (127).

Anlotinib (128) is a common targeted agent used in cancer treatment. It inhibits tumor growth in vitro and in vivo by downregulating circHAS2. These findings suggest a promising strategy for targeting patients with advanced colorectal cancer (CRC) who exhibit higher levels of circHAS2, potentially benefiting approximately 52.9% of such patients. In parallel, a prospective phase II clinical study has been initiated to evaluate the drug’s efficacy in 14 patients with advanced CRC (ClinicalTrials.gov identifier: NCT05262335). Although recent guidelines do not yet include anlotinib, circRNAs may offer valuable guidance in future studies of targeted drugs.

Cetuximab (129) is another targeted therapy with excellent efficacy in treating CRC, particularly in metastatic cases. It is widely used as a first-line treatment for CRC. Despite its promising clinical benefits, including improved progression-free survival (PFS), overall survival (OS), and quality of life with minimal side effects, the effectiveness of cetuximab is often limited by chemoresistance. Dysregulation of circRNA expression in CRC, breast cancer, and non-small cell lung cancer (NSCLC) has been linked to resistance to chemotherapy and targeted therapies. However, this resistance can often be reversed by modulating circRNA expression in tumor cells.

For example, cetuximab resistance in CRC has been associated with mutations in RAS, BRAF, PIK3CA, and PTEN. A study by Zhang et al. (130) demonstrated that circIFNGR2 regulates the target gene KRAS through a miR-30b sponge mechanism at the post-transcriptional level, thereby contributing to cetuximab resistance in CRC cells. In contrast, inhibiting circHIPK3 (113) expression enhances the responsiveness of CRC cells to cetuximab. Moreover, a retrospective analysis revealed that circHIF1A-positive patients had lower tumor regression rates and poorer long-term prognosis compared to circHIF1A-negative patients. The expression of circHIF1A in tumor tissues may serve as a predictive marker for the efficacy and outcomes of cetuximab treatment in RAS/BRAF wild-type metastatic CRC (mCRC) patients (129).

Programmed cell death ligand 1 (PD-L1) is a critical immune checkpoint protein that regulates immune responses through its interaction with PD-1. Knockdown of circ_0007422 (131) has been shown to suppress tumor progression and immune escape in CRC by modulating the miR-1256/PD-L1 pathway, offering a novel therapeutic approach for CRC. Additionally, circ_0089761 (132) competes with PD-L1 for binding to miR-27b-3p, relieving miR-27b-3p’s inhibitory effect on PD-L1 expression. This results in increased PD-L1 levels, promoting metastatic capacity in CRC cells. Therefore, the circ_0089761/miR-27b-3p/PD-L1 axis may serve as an important therapeutic target for CRC treatment. Elevated levels of circ_0089761 in CRC cells contribute to the upregulation of PD-L1, further enhancing tumor progression.

An increasing body of research has highlighted the dysregulation of circRNA expression in the progression of CRC, underscoring its critical role in the development of these cancers. Previous studies have shown that circRNAs not only act as miRNA sponges in the progression of intestinal cancers but also modulate drug sensitivity. Furthermore, many circRNAs have been identified as potential prognostic markers for the early diagnosis and prognosis of CRC, owing to the stability provided by their unique circular structure.

Current blood-based biomarkers, such as CEA and CA19-9, suffer from low sensitivity and specificity. Additionally, other diagnostic methods are either invasive or costly. As a result, recent research has shifted toward minimally invasive testing, particularly liquid biopsy. Exosomes, which are abundant, stable in circulation, and rich in genetic information and other biomolecules, are emerging as key molecules for biomarker discovery. These exosomal non-coding RNAs, including miRNAs, lncRNAs, and circRNAs, have shown promise in early CRC detection, offering higher sensitivity and specificity than traditional biomarkers like CEA and CA19-9. Moreover, exosomal circRNAs have demonstrated significant prognostic potential for TNM staging and could serve as predictive biomarkers for chemotherapy regimens such as 5-FU and FOLFOX.

CircLPHN3 has been shown to be closely correlated with various clinical features of CRC. The expression level of circLPHN3 in cancer tissues negatively correlated with CRC prognosis in Kaplan-Meier survival analysis. MiR-142-5p, which is highly expressed in CRC, is negatively regulated by circLPHN3. Overexpression of circLPHN3 suppresses CRC cell growth, migration, and invasion, mediated through miR-142-5p (133). In a study involving 100 paired CRC and adjacent normal tissue samples, circXRN2 was found to promote CRC cell proliferation and metastasis via the miR-149-5p/ENC1/EMT axis, suggesting that circXRN2 may serve as a potential therapeutic target and a novel biomarker for CRC progression (134).

In a study involving 157 CRC patients and 100 healthy controls, the expression of hsa_circ_0008621 was quantified in both serum and tissue samples by qRT-PCR. The levels of hsa_circ_0008621 were found to be significantly higher in the serum and tissues of CRC patients compared to healthy controls. Kaplan-Meier analysis and multivariate Cox regression indicated that high serum levels of hsa_circ_0008621 are associated with poor survival in CRC patients, promoting tumor progression. These findings suggest that hsa_circ_0008621 could serve as a promising non-invasive prognostic biomarker and therapeutic target for CRC (135).

A recent study also demonstrated that circCOL1A1 was significantly upregulated in the plasma of CRC patients and positively correlated with TNM stage (p = 0.001). Moreover, dynamic monitoring of circCOL1A1 in CRC patients undergoing surgical resection revealed a significant decrease in circCOL1A1 levels post-surgery, while levels increased in patients with recurrent or metastatic CRC. These findings suggest that circCOL1A1 may serve as a prognostic biomarker for postoperative recurrence in CRC (136).