Introduction
According to global cancer statistics, colorectal cancer (CRC) is the third most common type of cancer diagnosed worldwide, accounting for just over 10% of all diagnosed cancers. This is surpassed only by cancers of the lung and breast (the latter in women). CRC is the second leading cause of cancer-related mortality, accounting for just over 9% of cancer-related deaths (1). Approximately 20% of patients with CRC demonstrate metastases at initial diagnosis (metastatic CRC; mCRC), whether detected on imaging or confirmed during biopsy, with up to 80% of such patients deemed unresectable at initial diagnosis (2,3).
Systemic therapy is the mainstay of treatment for patients with inoperable CRC. Localized therapy in the palliative setting (namely surgical resection and irradiation of the primary tumour) is commonly limited to patients suffering from primary tumour complications. These can include bowel obstruction, perforation, localized pain and bleeding from the tumour (4,5).
Chemotherapeutic agents typically employed in mCRC consist of the antimetabolite 5-fluorouracil (5FU) (6,7), the pyrimidine analogue capecitabine (8), the topoisomerase inhibitor irinotecan (9) and the alkylating agent oxaliplatin (10). EGFR-targeting monoclonal antibodies include cetuximab (11) and panitumumab (12) for confirmed KRAS/NRAS wild-type tumours and VEGF-targeting bevacizumab (13) and ramucirumab (14), the latter typically in combination with chemotherapy using the FOLFIRI regimen (15).
Oxaliplatin-containing regimens, such as FOLFOX (16,17) or XELOX (18,19) and irinotecan-containing regimens, such as FOLFIRI (20-22) and XELIRI (23-25), have both been established equally effective in terms of progression-free and overall survival as first-line palliative systemic therapy in mCRC (26). Oxaliplatin-containing regimens may be preferred over irinotecan-based regimens as first-line therapy in mCRC due to their slightly more favourable median overall survival and toxicity profiles (21,27,28), including in elderly patients (29).
This may vary in patients for whom either oxaliplatin or irinotecan are contraindicated as first-line therapy due to their varying toxicity profiles. In patients deemed unsuitable for additional oxaliplatin and irinotecan with 5FU or capecitabine (such as patients of advanced age, with co-morbidities or poor performance status), single-agent 5FU or capecitabine may be preferred. Such single-agent 5FU regimens include the Roswell Park (30,31) and QUASAR (32,33) regimens and capecitabine (34,35) in either the adjuvant (36) or palliative (37) setting.
Combined nucleic acid analogue/thymidine phosphorylase inhibitor TAS-102 (tipiracil hydrochloride) (38) and the multikinase inhibitor regorafenib (39) are typically reserved for patients in whom 5FU-based chemotherapy in combination with oxaliplatin or irinotecan has failed (or if the patient is deemed clinically unsuitable to receive these treatments). Depending on the patient, irinotecan- or oxaliplatin-based chemotherapy may be rechallenged in the palliative setting if a significant interval has elapsed since completing a previous course of treatment, provided cumulative toxicity allows for this approach (40,41).
The targeted tyrosine kinase inhibitors encorafenib and binimetinib may be considered in BRAF V600E mutant tumours in combination with other systemic therapies (42). Immune checkpoint inhibitors are reserved for microsatellite instability-high (MSI-H) or deficient mismatch repair (dMMR) tumours (43); these include the monoclonal antibody inhibitors nivolumab (44,45) and pembrolizumab (46), which inhibit programmed death receptor 1, and ipilimumab (47), which inhibits cytotoxic T-lymphocyte-associated protein 4.
Tumour heterogeneity exists in mCRC, whether within the primary tumour (intra-tumoural heterogeneity) or between the primary and metastatic tumours (inter-tumoural heterogeneity) (48-50). Intra- or inter-tumoural heterogeneity has been implicated in the mechanisms underlying resistance to systemic therapy (51). Tumour heterogeneity in mCRC can vary during the course of the disease (52).
Inter-tumoural heterogeneity accounts for identifiable discordances in scientifically validated tests of advanced CRC. These discordances may include differentiation of adenocarcinoma (53,54), in which cancer stem cells play a role (55), mutation status (56-59), including KRAS/NRAS (60-62) and BRAF (63,64) status, MSI status (65) and dMMR status (66,67). This may lead to discordances in biomarker profiles between the primary tumour and metastases. Inter-tumoural heterogeneity appears to be more prevalent in mCRC with certain pathological and molecular features, such as confirmed MSI (68,69).
Among diagnosed CRCs, 30-40% of carry a pathogenic somatic KRAS mutation (70), 10% carry a NRAS mutation (71) and 10% carry a BRAF mutation (72). dMMR is identified in 10-20% (73-75) and MSI is identified in 10-20% of diagnosed colon cancers (76,77). Approximately 3% of colon cancers arise from germline mutations leading to MSI (78), a condition known as Lynch syndrome or hereditary non-polyposis CRC. Approximately 1% of cases result from germline defects in the α-fetoprotein gene (79), a condition known as familial adenomatous polyposis. Hypermethylation of the MLH1 gene promoter occurring in tumours with the CpG island methylator phenotype appears to be the predominant somatic mechanism of action of MSI-H in colorectal tumours (80).
Discordant responses between the primary tumour and metastatic sites in mCRC may arise from this underlying heterogeneity between tumour sites. Tumour cells at the metastatic sites may harbour clones that have gained (or lost) mutations advantageous to their survival compared to those residing at the primary site, or vice versa (81,82). These molecular discrepancies in mCRC are not yet fully understood; clonal evolution, cancer stem cells and ‘The Big Bang’ model have all been hypothesized to play a role (83-85).
The incidence of these biomarker discordances in mCRC varies depending on the resources consulted. Part of the literature, including meta-analysis studies, suggests that the rate of biomarker concordance is high between primary tumours and metastases in mCRC (86,87). These studies further suggest that tissues from either the primary tumour or metastatic site are sufficient for confirming the biomarker status of mCRC to help guide the systemic therapy approach (88). However, more recent studies suggest that the rate of these discordances increases if next-generation sequencing is used (89,90). There is high concordance (>90%) between immunohistochemical analysis and molecular testing of dMMR (91).
The rates of molecular discordances between the primary tumour and metastatic sites in mCRC may be as high as 10-15%, depending on the study (86,92). The rate of discrepancies in MSI or MMR status tends to be low (<5%) upon comparison (93,94). The expression of other specific biomarkers, such as programmed death ligand 1 (PD-L1), may vary more markedly between primary tumour and metastatic sites in up to one-third of patients (95).
Evaluating molecular characteristics between primary and metastatic tumours separately in a single patient with mCRC can demonstrate variable biological behaviour and response to systemic therapy due to these identified subclones (96). Non-genetic factors, such as post-translational modification, epigenetics and the tumour microenvironment, also contribute to this phenomenon. Comparing such predictive or prognostic molecular signatures between the primary and metastatic tumours in mCRC has yielded different results. For example, mutant KRAS status exhibits concordance between the primary tumour and distant organ metastases in up to 90% of patients with mCRC (97). Conversely, comparisons between the primary tumour and lymph node metastases demonstrate lower concordance rates with a KRAS mutant status of ~37% (98).
The role of pre-emptive localized therapies in patients with relatively asymptomatic primary tumours remains controversial and has demonstrated an inconsistent clinical benefit (99-101). Radiation to rectal and rectosigmoid cancer primary tumours has been associated with a reduced risk of death in one retrospective study (102). Previous findings have demonstrated no additional benefit, reduced risk of complications or death when radiation treatment is administered before systemic therapy (103), while other findings suggest that prior resection of the colorectal primary in mCRC offers benefit in selected patients (104). However, this should not be routinely considered in asymptomatic patients, as it offers no additional benefit (105), taking into account the currently available systemic therapies (106).
Discussion
The present study demonstrated that up-front palliative systemic therapy can effectively control primary tumours in patients with de novo mCRC with primary tumours in situ that are deemed inoperable/non-curable at initial diagnosis. However, while up-front systemic therapy with palliative intent is predominantly effective in mCRC with primary tumour in situ, 24% of patients in our study required localized intervention due to complications arising from the primary tumour (on first-line systemic therapy).
Other studies have demonstrated a reduced risk of primary tumour-related complications and the need for emergent surgical intervention with up-front localized interventions prior to undertaking palliative systemic therapy (3,101). Most international guidelines currently recommend combination chemotherapy as the initial treatment for unresectable mCRC with primary in situ (109,110). Local interventions for primary tumours are typically reserved for when complications arise from the primary tumour after palliative systemic therapy has already been employed (including bowel obstruction, perforation, significant pain, or bleeding from the primary tumour) (111,112).
In the present study, higher rates of radiological response were observed in primary tumours compared with metastatic tumour sites. Nonetheless, the rate of complications arising from the primary tumour requiring intervention during systemic therapy remained high (up to 25% of cases in this patient cohort).
In a retrospective study involving 233 patients with mCRC receiving combined chemotherapy with or without bevacizumab as up-front first-line systemic therapy (113), only 7% of these patients required emergent surgical intervention, and 4% required emergent non-surgical intervention (radiation and endoscopic stenting) while on systemic therapy. On the other hand, the remaining 213 patients (89%) never required local intervention for their primary tumour. Another study observed that, among 83 asymptomatic patients with non-curable mCRC treated with first-line chemotherapy (114), only 5% required surgery, while 4% required colonic stenting to manage complications arising from the primary tumour.
Conversely, other studies support prophylactic surgical resection of the primary tumours in non-curable mCRC before undertaking palliative-intent systemic therapy to reduce the future risk of primary tumour complications (2,3,101). One meta-analysis reviewed eight retrospective studies including 1,062 patients (101) and observed that up-front primary tumour resection was associated with reduced rates of primary tumour-associated complications requiring emergent localized intervention and increased overall survival rate. This was compared to patients receiving up-front palliative systemic therapy alone, who were 7.3 times more likely to suffer acute complications requiring localized interventions while on palliative systemic therapy.
Current randomized control trials, such as the SYNCHRONOUS trial (ISRCTN30964555) and the iPACS study (JCOC1007) are comparing up-front palliative chemotherapy alone with up-front primary tumour resection followed by palliative chemotherapy in patients with non-curative mCRC with asymptomatic primary tumours at diagnosis (115,116).
Multiple studies have observed conflicting results when addressing the concordance rates of KRAS, NRAS and BRAF mutation status between the primary colorectal tumour and metastatic sites. While some results showed no significant difference in mutation status (namely KRAS) between the primary tumours and corresponding metastases, others showed discordant results in 4-32% of the patients (16). One study including 305 patients demonstrated a high concordance rate of KRAS mutation status (96.4%) between primary colorectal tumours and corresponding liver metastases (23,24). Mutation status discordance rates of ≤25% between the primary tumour and the lymph node metastases were also observed (117).
There were certain limitations to the present study. Certain patient variables (such as past medical history, ethnicity, dietary history, smoking history, and whether patients did or did not attend a colorectal screening program) were not assessed as part of the analysis, as they were considered to be outside the scope of this study, and due to relatively small patient number. This is further taking into account the small number of patients accrued in this data analysis, which limits the validity of statistical associations observed. Further research (ideally a meta-analysis) is required to assess and, ultimately, validate the associations observed in this study.
In the present study, statistically appreciable rates of discordant radiological responses to up-front palliative systemic therapy were observed between the primary tumour and metastatic tumour sites in patients with inoperable/non-curative de novo mCRC (in up to 60% of our patient cohort). Approximately one-third of the patients demonstrating radiological control of the primary tumour otherwise demonstrated progression at metastatic sites while on first-line up-front single-modality palliative systemic therapy at the first interval restaging imaging.
This has implications for molecular analyses of the tissues obtained from patients diagnosed with mCRC. Our analysis suggests that standard molecular panels performed in mCRC (including KRAS, NRAS and BRAF status with MMR and MSI analyses) should preferentially be performed on tissue from metastatic sites rather than on tissue from the primary tumour.
Up-front localized management strategies, such as palliative radiation to the primary tumour, surgical interventions (including stoma formation) and endoscopic procedures (such as colonic stenting) should be considered in certain patients with inoperable mCRC.