INTRODUCTION — The majority of patients with metastatic colorectal cancer (mCRC) cannot be cured, although a subset of patients with liver and/or lung-isolated metastatic disease, local recurrence, or limited intraabdominal disease are potentially curable with surgery. For other patients with mCRC, treatment is palliative and generally consists of systemic chemotherapy. (See "Management of potentially resectable colorectal cancer liver metastases" and "Surgical resection of pulmonary metastases: Outcomes by histology" and "Locoregional methods for management and palliation in patients who present with stage IV colorectal cancer", section on 'Aggressive cytoreduction and intraperitoneal chemotherapy for peritoneal metastases'.)

This topic review will cover general principles that underlie chemotherapy treatment of mCRC, including the goals of therapy in patients with potentially resectable metastatic disease versus those with categorically unresectable disease, benefits of treatment compared with supportive care alone, issues related to timing and duration of treatment in patients with unresectable metastatic disease, and assessment during therapy. Specific systemic treatment recommendations for patients with unresectable mCRC, issues relevant to treatment of mCRC in the elderly and those with a poor performance status, and the use of systemic therapy for the purpose of downsizing potentially resectable CRC liver metastases are discussed elsewhere. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach" and "Systemic therapy for nonoperable metastatic colorectal cancer: Approach to later lines of systemic therapy" and "Therapy for metastatic colorectal cancer in older adult patients and those with a poor performance status" and "Management of potentially resectable colorectal cancer liver metastases".)

CHEMOTHERAPY OPTIONS — For decades, fluorouracil (FU) was the sole active agent for advanced colorectal cancer (CRC). This has changed markedly since the year 2000, with the approval of irinotecan; oxaliplatin; three humanized monoclonal antibodies (MoAbs) that target the vascular endothelial growth factor (bevacizumab), the vascular endothelial growth factor receptor (VEGF; ramucirumab), and the epidermal growth factor receptor (cetuximab and panitumumab); intravenous aflibercept, a fully humanized recombinant fusion protein consisting of VEGF binding portions from the human VEGF receptors 1 and 2 fused to the Fc portion of human immunoglobulin G1; regorafenib, an orally active inhibitor of angiogenic (including the VEGF receptors 1 to 3), stromal, and oncogenic kinases; trifluridine-tipiracil (TAS-102), an oral cytotoxic agent that consists of the nucleoside analog trifluridine (a cytotoxic antimetabolite that inhibits thymidylate synthetase and, after modification within tumor cells, is incorporated into DNA causing strand breaks); and tipiracil, a potent thymidine phosphorylase inhibitor, which inhibits trifluridine metabolism and has antiangiogenic properties as well. In addition, other orally active fluoropyrimidines (capecitabine, S-1, tegafur-uracil [UFT]) are also available.

The best way to combine and sequence these agents is still not established. A compilation of commonly used chemotherapy protocols for CRC is available. (See "Treatment protocols for small and large bowel cancer" and "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach".)

Combination versus sequential single agents — First-line therapy with combinations of fluoropyrimidines, oxaliplatin, irinotecan, and biologic agents such as bevacizumab has markedly improved response rates, progression-free survival (PFS), and survival compared with fluoropyrimidines alone or doublet chemotherapy regimens with or without a biologic agent. However, even in patients treated initially with fluoropyrimidine monotherapy, survival is positively impacted by subsequent lines of therapy, and upfront combination therapy (particularly when oxaliplatin and irinotecan are combined) also increases toxicity and cost. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Overview of the therapeutic approach'.)

The question of whether patients should receive initial combination therapy or fluoropyrimidine monotherapy has been addressed in two randomized trials, neither of which showed that survival was adversely impacted by initial single-agent therapy [1,2]. However, the median survival for all groups in both trials (which ranged from 13.9 to 17.4 months) was lower than expected for modern combination chemotherapy. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Overview of the therapeutic approach'.)

One possible reason is the low number of patients who eventually received all three active drugs in both trials. The proportion of patients receiving all three active agents correlates strongly with median survival in all large published phase III trials over the last decade [3]. Furthermore, neither trial used bevacizumab or cetuximab as either first-line or salvage therapy. These agents improve PFS, and bevacizumab also improves overall survival when used in the first-line regimen. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Role of biologics'.)

Some of these issues were addressed in a third trial, the XELAVIRI trial, which randomly assigned 421 patients with untreated metastatic CRC to a fluoropyrimidine plus bevacizumab, followed by the addition of irinotecan at progression, versus initial combined therapy with all three agents [4]. Only 63 percent of patients treated with initial sequential therapy received irinotecan at some point in the course of their treatment, compared with 100 percent in the combination therapy group. Although sequential therapy was shown not to be noninferior to combination therapy for time to failure of strategy (the primary endpoint), survival was not significantly different (median overall survival 23.5 versus 21.1 months). An unplanned subgroup analysis suggested that initial combination therapy particularly benefited patients with wild-type RAS/BRAF tumors.

Thus, the available evidence continues to support initial combination chemotherapy for most patients, particularly for those whose metastases might be potentially resectable after an initial chemotherapy response. (See "Management of potentially resectable colorectal cancer liver metastases", section on 'Neoadjuvant systemic therapy'.)

These trial results and the implications for clinical practice are discussed in detail elsewhere. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Initial doublet combinations versus sequential single agents'.)

TREATMENT GOALS — The goals of chemotherapy for mCRC differ according to the clinical scenario. For most patients, treatment will be palliative and not curative (a fact that may not be understood by patients [5]), and the treatment goals are to prolong overall survival and maintain quality of life (QOL) for as long as possible.

Potentially resectable disease — However, some patients with stage IV disease (particularly those with liver-limited metastases) can be surgically cured of their disease. Even selected patients with initially unresectable liver metastases may become eligible for resection if the response to chemotherapy is sufficient.

This approach has been termed "conversion therapy" [6] to distinguish it from "neoadjuvant therapy," which applies to preoperative chemotherapy given to patients who present upfront with apparently resectable disease. The key parameter for selecting the specific regimen in this scenario is not survival or improved QOL, but instead, response rate (ie, the ability of the regimen to shrink metastases) [7]. (See "Management of potentially resectable colorectal cancer liver metastases".)

Nonresectable disease — The following general principles guide the use of palliative chemotherapy in the setting of nonoperable disease:

●In general, for patients without symptomatic disease (ie, the majority of patients), induction of a tumor response is not as important as is delaying tumor progression for as long as possible. In the palliative setting, objective response rate is not the best indicator of treatment benefit (prolonged survival and/or progression-free survival [PFS]) [8-10]. Thus, achieving stable disease as the best response to therapy might be considered a treatment success. (See 'Assessment during therapy' below.)

●Patients benefit more from access to all active agents than from a particular treatment sequence of specific regimens used as individual "lines" of therapy. In all large published phase III trials testing various combinations of cytotoxic agents and targeted agents conducted over the last decade, the proportion of patients receiving all active agents has correlated strongly with median survival [3,11]. Although no such analysis has yet been performed after the introduction of biologic agents, it is conceivable that the overall principle of optimizing outcomes through exposure to all active agents is still valid.

●Despite these findings, the available evidence suggests that only a minority of American patients with mCRC are exposed to all active agents in the course of their therapy for mCRC [12].

●Because of the survival benefit from second- and even third-line chemotherapy, the routine practice of crossover in clinical trials severely limits the ability to detect an overall survival advantage of one treatment regimen over another. Therefore, the actual activity of a new agent or combination regimen is better captured by the endpoint PFS, in particular in the first-line setting. Improvements in PFS correlate with longer survival [13-15] and are not affected by crossover or subsequent therapy.

These concepts can be illustrated by results from the EPIC trial, in which patients failing initial oxaliplatin-based therapy were randomly assigned to irinotecan with or without cetuximab [16]. There were significant differences in PFS, objective response, and disease control rates that favored combined therapy, but no overall survival advantage. This was attributed to the fact that 50 percent of the patients in the irinotecan arm crossed over to cetuximab at progression. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Approach to later lines of systemic therapy", section on 'Patients not initially treated with cetuximab/panitumumab'.)

●Endpoints other than PFS (eg, duration of disease control, time to failure of strategy) have been proposed, but none are widely used [17,18].

●The model of distinct "lines" of chemotherapy (in which regimens containing non-cross-resistant drugs are each used in succession until disease progression) is being abandoned in incurable metastatic mCRC in favor of a "continuum of care" approach [19]. This implies an individualized treatment strategy that may include phases of maintenance chemotherapy interspersed with more aggressive treatment protocols, as well as reutilization of previously administered chemotherapy agents in combination with other active drugs.

The following sections will emphasize the practical issues that arise when choosing the appropriate treatment strategy for individual patients with inoperable mCRC. Specific recommendations for therapy as well as management of patients with potentially resectable liver metastases are discussed elsewhere. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach" and "Systemic therapy for nonoperable metastatic colorectal cancer: Approach to later lines of systemic therapy" and "Therapy for metastatic colorectal cancer in older adult patients and those with a poor performance status" and "Management of potentially resectable colorectal cancer liver metastases".)

CHEMOTHERAPY VERSUS SUPPORTIVE CARE — Systemic fluorouracil (FU)-based chemotherapy produces meaningful improvements in median survival and progression-free survival (PFS) compared with best supportive care (BSC) alone [20-22]. These benefits are most pronounced with regimens containing irinotecan or oxaliplatin in combination with FU. Although no trial has compared these regimens with BSC alone, median survival durations in clinical trials of oxaliplatin- and irinotecan-containing chemotherapy now consistently exceed two years; by contrast, for patients with unresectable mCRC who receive best supportive care (BSC) alone, median survival is approximately five to six months [20-22].

Long-term survival is improving with the availability of more active anticancer agents [23-25]. As an example, in a report of pooled data from North Center Cancer Treatment Group (NCCTG) trials conducted in the FU plus leucovorin (LV) era, only 1.1 percent of patients were alive at five years [26]. By contrast, in a report from the phase III FIRE-3 trial (first-line irinotecan with short-term infusional FU plus LV [FOLFIRI] plus either bevacizumab or cetuximab), the five-year survival rate for patients with RAS wild-type tumors treated with FOLFIRI plus cetuximab was approximately 20 percent [27]. Although many of the survival gains are attributable to advances in chemotherapy treatment, more aggressive use of surgical resection of metastatic disease has also contributed [25]. (See "Management of potentially resectable colorectal cancer liver metastases" and "Surgical resection of pulmonary metastases: Outcomes by histology", section on 'Colorectal cancer'.)

TIMING OF CHEMOTHERAPY — Although the value of early chemotherapy versus treatment deferral until symptoms develop is controversial, we suggest instituting chemotherapy at diagnosis for patients with categorically unresectable mCRC, and when possible, before patients become symptomatic.

Many patients with mCRC are asymptomatic. Data are limited on optimal timing of chemotherapy, and the only randomized trials directly addressing this issue studied older regimens like fluorouracil (FU) and leucovorin (LV):

●In an early trial in which 182 patients with asymptomatic mCRC were randomly assigned to initial or deferred chemotherapy with sequential methotrexate, FU, and LV, earlier treatment was associated with improvements in median survival (14 versus 9 months), symptom-free interval, and progression-free survival (PFS) [20].

●In a combined analysis of 168 asymptomatic patients who were enrolled in two trials randomly testing early versus delayed FU-based chemotherapy, there was a non-statistically significant two-month benefit in median survival with early treatment (13 versus 11 months) [28].

Whether these results can be extrapolated to patients treated with irinotecan, oxaliplatin, or biologic therapies, especially in the era of modern diagnostic procedures that can detect lower volume metastatic disease, is unclear. Regimens such as these are associated with clear-cut survival benefits, particularly if patients are serially exposed to all active agents. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Initial doublet combinations versus sequential single agents'.)

The only available data from the era of modern chemotherapy come from a retrospective report of 736 patients with mCRC diagnosed between January 2003 and December 2010 at a single Australian center; 377 (51 percent) received immediate chemotherapy, 167 (23 percent) did not because they were deemed inappropriate for therapy or refused, and 192 (26 percent) adopted a "watch and wait" policy initially, 168 of whom eventually received chemotherapy (at a median of 3.7 months from diagnosis) [29]. Compared with immediate treatment, the fraction of patients in the delayed chemotherapy group who eventually received treatment with all active agents was slightly less (30 versus 39 percent), but the median survival was superior (27 versus 17 months).

Importantly, these data are not from a randomized trial, and interpretation is limited by the potential for selection bias (ie, patients who had treatment deferred were likely to be those with favorable biology [asymptomatic, lower volume metastatic disease, better performance status]), all of which could have contributed to the longer survival in this group. At least in the United States, most patients institute treatment at a time when they are still asymptomatic from their cancer. An alternate approach, which may be particularly appropriate for asymptomatic elderly patients, is an initial period of observation to judge the tempo of disease progression.

CHEMOTHERAPY DOSING IN OBESE PATIENTS — For cancer patients with a large body surface area (BSA), chemotherapy drug doses are often reduced because of concern for excess toxicity. However, there is no evidence that fully dosed obese patients experience greater toxicity from chemotherapy for mCRC; furthermore, obese patients who are given reduced doses may have inferior outcomes [30]. Although limited, the available data do not support the policy of routine dose reduction (or capping the maximal BSA to 2.0 m²) for obese patients with mCRC. Guidelines from the American Society of Clinical Oncology (ASCO) recommend that full weight-based cytotoxic chemotherapy doses be used to treat obese patients with cancer [31]. (See "Dosing of anticancer agents in adults", section on 'Overweight/obese patients'.)

CONTINUOUS VERSUS INTERMITTENT THERAPY — The optimal duration of initial chemotherapy for unresectable disease that does not progress is controversial. In general, the decision to permit treatment breaks during therapy must be individualized and based upon several factors, including tolerance of and response to chemotherapy, disease bulk and location, and symptomatology:

●There are many patients with small volume but multiple sites of disease who respond to chemotherapy or have a prolonged period of disease stability. Even if their disease triples in volume off therapy, they will not likely be symptomatic or develop organ dysfunction. Patients with favorable characteristics may be able to tolerate chemotherapy-free intervals of multiple months per year and go on to respond favorably to drugs for many years.

●On the other end of the spectrum are patients with retained primary tumors, bulky disease, poor performance scores due to tumor related symptoms, peritoneal disease that may lead to unsalvageable bowel obstruction as the first sign of progression, and those with extensive symptomatic disease who progress through treatment regimens in quick succession with either short-lived responses or no response. These patients may be better approached with continuous chemotherapy. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Duration of initial chemotherapy'.)

When fluorouracil (FU) was the only treatment alternative, patients generally stayed on treatment until their disease progressed or they developed unacceptable toxicity. This typically meant that patients were treated for four to six months (the median progression-free survival [PFS] duration) and then were placed on supportive care alone until they died (median duration of survival approximately one year).

Compared with FU alone, newer combinations are more effective (median survival durations now consistently approach two years), but they are also more toxic. This is particularly true for oxaliplatin-containing regimens, which cause cumulative neurotoxicity; several studies have shown that more patients come off of therapy because of toxic effects than because of progressive disease [32,33]. This is particularly problematic for regimens that include antiangiogenic therapy; the available data suggest that patients who receive treatment to tumor progression derive the greatest benefit from the addition of bevacizumab to the chemotherapy backbone [32]. Intermittent rather than continuous chemotherapy has the potential to improve outcomes and reduce toxicity as well as cost.

Whether continued chemotherapy provides better outcomes than intermittent therapy to best response followed by a chemotherapy "holiday" has been addressed in several trials, most of which have studied chemotherapy regimens that contain oxaliplatin, a drug that is associated with dose-limiting neurotoxicity. Intermittent oxaliplatin-free therapy can be achieved through a complete break in therapy or the use of a non-oxaliplatin-containing "maintenance regimen." (See "Overview of neurologic complications of platinum-based chemotherapy", section on 'Cumulative sensory neuropathy'.)

Oxaliplatin — For patients who are responding to an oxaliplatin-based initial regimen, we suggest discontinuing oxaliplatin before the onset of severe neurotoxicity (usually after three to four months of therapy). Continuation of oxaliplatin is an alternative for responding patients who have no clinically significant neuropathy.

For patients who choose to discontinue oxaliplatin, we suggest the following approach:

●For most patients, we suggest maintenance therapy over a complete break in therapy, with reintroduction of oxaliplatin at disease progression. A switch to an irinotecan-based regimen is also an option at disease progression for those with persistent neuropathy.

●For patients initially treated without a biologic agent, we suggest maintenance therapy with a fluoropyrimidine with or without leucovorin (LV). (See 'OPTIMOX and CONcePT trials' below.)

●For patients initially treated with oxaliplatin plus bevacizumab, we suggest maintenance therapy with a fluoropyrimidine with or without bevacizumab, rather than bevacizumab alone. (See 'Maintenance bevacizumab alone' below.)

●For responding patients who were initially treated with oxaliplatin plus an agent targeting the epidermal growth factor receptor (EGFR), we suggest combining fluorouracil with the anti-EGFR agent rather than the EGFR agent alone. (See 'Patients initially treated with an EGFR inhibitor' below.)

●A complete break in therapy is also a valid option, particularly if a complete clinical response is observed or for those with small-volume metastatic disease who have a partial response or stable disease to the initial course of chemotherapy. Decision-making should also consider patient preference. In such cases, close follow-up with tumor assessment at two-month intervals and early resumption of chemotherapy at the first sign of progression is recommended. (See 'Complete break in therapy' below.)

Oxaliplatin-based regimens (eg, FOLFOX [oxaliplatin plus LV and short-term infusional FU]) are commonly used for first-line chemotherapy in mCRC [12]. However, oxaliplatin is associated with a cumulative sensory neuropathy, which may be dose limiting. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'FOLFOX versus FOLFIRI' and "Treatment protocols for small and large bowel cancer".)

Whether long-term neurotoxicity can by mitigated by intermittent oxaliplatin-free intervals has been addressed in several trials. The following represents an overview of the most important findings.

OPTIMOX and CONcePT trials

●The OPTIMOX1 trial randomly assigned 620 previously untreated patients to FOLFOX4, administered every two weeks until disease progression (arm A), or FOLFOX7 (table 1) for six cycles only, followed by reintroduction of oxaliplatin at the time of progression after 12 cycles of a non-oxaliplatin-containing maintenance regimen (leucovorin-modulated FU) [17]. Individuals in arm B had a significantly lower risk of developing grade 3 or 4 toxicity during cycles 6 to 18 (but not overall). The authors concluded that a survival benefit for maintenance non-oxaliplatin chemotherapy could have been masked by the large number of patients in this cohort (60 percent) who did not receive planned oxaliplatin reintroduction [34].

●The subsequent OPTIMOX-2 trial was initially designed as a 600 patient phase III trial, but when bevacizumab became available, accrual was halted with 202 patients enrolled [35]. The primary endpoint was the duration of disease control, calculated as the sum of the duration of PFS both following the initial three-month course of modified FOLFOX7 (mFOLFOX7) (table 1), as well as after the subsequent reintroduction of oxaliplatin. An important characteristic of OPTIMOX-2 was that randomization occurred after six cycles of therapy regardless of response, and metastases were allowed to progress back to baseline levels before FOLFOX was reintroduced.

Complete discontinuation of therapy had an adverse impact on prognosis; the group receiving maintenance therapy had significantly longer median duration of disease control and median PFS from the time of randomization; there was also a trend toward improved median overall survival (24 versus 20 months, p = 0.42). These data mandate caution and both careful patient selection and vigilant patient monitoring so that therapy can be reinstated promptly at progression when considering chemotherapy-free intervals.

●Another multicenter trial, the CONcePT trial, in which patients were randomly assigned to continuous versus intermittent oxaliplatin (alternating every eight cycles with and without oxaliplatin) also confirmed the benefit of intermittent rather than continuous oxaliplatin for increasing time on first-line therapy for oxaliplatin/bevacizumab-based combinations [36]. Rates of peripheral sensory neuropathy were significantly lower in the intermittent therapy group.

NO-16966 and MRC COIN trials — Two other trials have also concluded that early discontinuation of systemic chemotherapy before disease progression might be detrimental:

●In the NO16966 trial, a randomized trial of two oxaliplatin-based regimens (XELOX and FOLFOX4 (table 1)) that was later modified to allow for further randomization to bevacizumab versus no bevacizumab [32]. Significantly more patients in the chemotherapy/bevacizumab arms discontinued therapy for reasons other than disease progression, typically adverse effects. After reanalyzing the results according to whether patients were on- or off-treatment at the time of disease progression, the authors concluded that treatment continuation until disease progression was necessary to optimize benefit from the addition of bevacizumab.

●The inferiority of intermittent as compared with continuous oxaliplatin and fluoropyrimidine-based first-line chemotherapy was also shown in the MRC COIN trial, in which 1630 patients were randomly assigned to continuous therapy until disease progression, toxicity, or withdrawal of patient consent, versus 12 weeks of therapy followed by a chemotherapy-free interval until disease progression [37]. The trial was powered for non-inferiority. Median survival was not significantly better in the continuous arm (19.6 versus 18 months; hazard ratio [HR] 1.087, 95% CI 0.986-1.198); however, the upper limit of the confidence interval exceeded the predefined non-inferiority boundary interval for survival (1.162), and the noninferiority of no maintenance treatment could not be established. Rates of grade 3 or worse peripheral neuropathy were 27 versus 5 percent in the continuous and intermittent therapy arms.

Maintenance bevacizumab — Several trials have explored the benefit of maintenance bevacizumab in patients initially treated with a bevacizumab-containing regimen, both alone and in combination with a fluoropyrimidine.

Maintenance bevacizumab plus a fluoropyrimidine

●CAIRO3 – The utility of maintenance treatment with capecitabine plus bevacizumab was addressed in the Dutch CAIRO3 trial, which randomly assigned 558 patients with stable disease or better after six cycles of XELOX plus bevacizumab who were not eligible for potentially curative metastasectomy to continued capecitabine (625 mg/m² twice daily every day) plus bevacizumab (7.5 mg/kg every three weeks) or observation alone [38]. Upon first progression (PFS1), patients in both arms were supposed to be treated with XELOX plus bevacizumab until the second progression (PFS2) per protocol. The primary endpoint was PFS2, which was calculated from the time of randomization. Maintenance therapy was associated with a significantly longer PFS2 (11.7 versus 8.5 months, HR 0.67, p<0.0001), and there was a trend toward improved overall survival, as well (median 21.6 versus 18.1 months, HR 0.89, p = 0.22).

●German AIO KRK 0207 trial – Similarly, a benefit for continued fluoropyrimidine plus bevacizumab as compared with observation alone was also shown in the German AIO KRK 0207 trial, in which patients without progressive disease after six months of oxaliplatin plus a fluoropyrimidine and bevacizumab were randomly assigned to maintenance with the same fluoropyrimidine plus bevacizumab, bevacizumab alone, or observation only [39]. The primary endpoint was the "time to failure of strategy" or TFS, which included the duration of maintenance plus the time from reinduction after first progression to a second disease progression. The trial was powered to demonstrate noninferiority with a noninferiority margin set at 3.5 months, corresponding to an HR of 1.42. The median TFS in the fluoropyrimidine plus bevacizumab and observations arms was not significantly different (6.9 and 6.4 months, respectively; HR 1.26, 95% CI 0.99-1.60). However, the observation arm was not non-inferior to fluoropyrimidine plus bevacizumab because the upper limit of the 95 percent confidence interval exceeded the threshold set for non-inferiority (1.43). Notably, few patients in either arm were exposed to reinduction treatment (19 percent with combined therapy, and 46 percent of those undergoing observation), rendering the primary endpoint, TFS, non-informative and clinically irrelevant.

Results from the comparison of bevacizumab alone versus no maintenance or bevacizumab plus fluoropyrimidine are discussed below. (See 'Maintenance bevacizumab alone' below.)

●STOP and GO trial – A slightly different approach was tested in the Turkish STOP and GO trial, in which, following six cycles of bevacizumab plus XELOX, 123 patients were randomly assigned to continued therapy or discontinuation of oxaliplatin and maintenance with bevacizumab plus capecitabine until progression [40]. The median PFS was significantly better in the group receiving maintenance therapy with bevacizumab plus capecitabine (11 versus 8.3 months), with less grade 3 or 4 diarrhea (3.3 versus 11.3 percent), hand-foot syndrome (1.6 versus 3.2 percent), and neuropathy (1.6 versus 8.1 percent).

Maintenance bevacizumab alone — For patients who have no disease progression after an initial course of bevacizumab plus oxaliplatin-containing chemotherapy, we suggest not pursuing bevacizumab alone for maintenance therapy; this approach is also not recommended in consensus-based guidelines for the treatment of mCRC from NCCN [41] and ESMO [42].

The role of maintenance bevacizumab alone has been studied in three trials, all of which used different comparator arms, and all of which came to different conclusions:

●In the Spanish MACRO trial, patients received six cycles of first-line XELOX plus bevacizumab followed by a randomization to continued therapy or bevacizumab maintenance therapy alone until progression or treatment intolerance [43]. There was no arm in which patients received no maintenance therapy. The median PFS and overall survival durations in patients treated with maintenance bevacizumab alone were not significantly worse, and rates of severe neurotoxicity, hand-foot syndrome, and fatigue were significantly lower. However, the trial failed to achieve its primary endpoint of non-inferiority for PFS, because the projected upper limit of the 95 percent confidence interval for PFS exceeded the preset limit.

●In the Swiss SAKK 41-06 trial, 262 patients with mCRC were randomly assigned to bevacizumab continuation versus no maintenance after four to six months of first-line bevacizumab-containing chemotherapy (62 percent oxaliplatin-containing, 31 percent irinotecan-containing, and the rest fluoropyrimidine alone) [44]. Like the MACRO trial, the trial failed to achieve its primary endpoint of non-inferiority for TTP with the projected upper limit of the 95 percent confidence interval for TTP exceeding the preset limit. The median TTP was 4.1 for bevacizumab continuation versus 2.9 months for no continuation (HR 0.74, 95% CI 0.57-0.95). However, in our view, this study has significant limitations; it includes trials conducted over almost two decades, contains a very heterogenous patient population, and it is heavily influenced by the COIN trial due to its size. As a result, it should not be used to justify use of bevacizumab alone as effective maintenance therapy.

●On the other hand, noninferiority of bevacizumab alone compared with bevacizumab plus a fluoropyrimidine was shown in the German AIO KRK 0207 trial, described above [39]. The primary endpoint (the median time to failure of strategy, TFS) in the fluoropyrimidine plus bevacizumab and bevacizumab alone arms was 6.9 and 6.1 months, respectively. Compared with fluoropyrimidine plus bevacizumab, the bevacizumab only arm was non-inferior (HR 1.08, 95% CI 0.85-1.37). However, the upper boundary of the noninferiority margin was very generous (HR 1.43). Notably, few patients in either arm were exposed to reinduction treatment (19 percent with combined therapy, and 43 percent of those receiving bevacizumab alone), rendering the primary endpoint, TFS, non-informative and clinically irrelevant. Results from the comparison of no maintenance versus fluoropyrimidine plus bevacizumab are discussed above. (See 'Maintenance bevacizumab plus a fluoropyrimidine' above.)

Patients initially treated with an EGFR inhibitor — For patients initially treated with an agent targeting the epidermal growth factor receptor (EGFR), we suggest maintenance therapy using fluorouracil plus the anti-EGFR agent rather than an anti-EGFR agent or fluoropyrimidine alone.

Benefit from anti-EGFR therapies is limited to patients whose tumors lack mutations in one of the RAS oncogenes (ie, wild-type [WT] RAS). (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Agents targeting the EGFR'.)

Three trials have addressed the benefit of maintenance therapy with an EGFR inhibitor after initial treatment with FOLFOX plus an EGFR inhibitor:

●The phase II MACRO-2 trial randomly assigned 193 patients with KRAS (exon 2 only) WT tumors to receive FOLFOX plus cetuximab for four months (eight courses) followed by either continued therapy with the same regimen or cetuximab monotherapy alone (250 mg/m² weekly) [45]. Cetuximab monotherapy was noninferior to the combination of continued FOLFOX plus cetuximab, as judged by the primary endpoint, the proportion of patients who were progression free at nine months (60 versus 72 percent, HR 0.60, 95% CI 0.31-1.15).

●On the other hand, results with panitumumab alone were inferior to maintenance treatment with FU/LV plus panitumumab following four months of induction therapy with FOLFOX plus panitumumab in the phase II noninferiority VALENTINO trial [46]. Ten-month PFS was inferior with panitumumab alone (49 versus 60 percent).

●A slightly different question, the benefit of adding panitumumab to leucovorin (LV) modulated FU versus FU/LV alone after six cycles of induction therapy with FOLFOX plus panitumumab in RAS wild-type advanced CRC was addressed in the phase III PANAMA trial [47]. In a preliminary report, presented at the 2021 annual ASCO meeting, median PFS, the primary endpoint, was significantly better with combined therapy as compared with leucovorin-modulated FU alone (8.8 versus 5.7 months, HR 0.72, 95% CI 0.60-0.85), and there was also a trend to better overall survival that also favored maintenance panitumumab.

Irinotecan — While intermittent treatment approaches appear to be almost mandatory for the majority of patients receiving oxaliplatin because of cumulative neurotoxicity, the advantages of intermittent treatment with irinotecan-based regimens are less clear. For most patients, we treat for as long as tumor shrinkage continues and treatment is tolerable. Nevertheless, intermittent treatment does not appear to compromise outcomes and could be considered in responding patients, with or without maintenance bevacizumab.

The benefits/risks of intermittent chemotherapy with an irinotecan-containing regimen were addressed in an Italian trial, which demonstrated that patients started on FOLFIRI (irinotecan with short-term infusional FU plus LV (table 2)) as first-line therapy had similar overall outcome (PFS and overall survival) whether or not the regimen was administered continuously until progression or toxicity or in "two months on/two months off" intervals [48]. The mean chemotherapy-free period in the intermittent treatment group was three months. However, there were no demonstrable differences in treatment-related toxicity between the continuous versus intermittent treatment groups. Of note, further second- and third-line therapy did not follow a "stop-and-go" approach, so that for overall survival, any potential differences obtained in first-line therapy could have been obscured by subsequent treatment. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Irinotecan-based regimens'.)

A lack of benefit for maintenance bevacizumab versus no treatment until progression following six months of induction FOLFIRI plus bevacizumab was shown in the randomized phase III PRODIGE 9 trial [49].

Complete break in therapy — The above data have led to the general conclusion that some form of maintenance therapy is preferred rather than a complete break in therapy in patients who are responding to or have stable disease after induction chemotherapy therapy. However, the main benefit for maintenance therapy is in prolonging PFS and none of the trials have shown that this approach is associated with better overall survival compared with a complete break in therapy. In our view, either maintenance therapy or a complete break in therapy represents a valid option, particularly for patients who have had a complete clinical response or for those with small-volume metastatic disease who have a partial response or stable disease to the initial course of chemotherapy.

At least two meta-analyses and a more recent trial have specifically addressed the role of observation (ie, a complete break in therapy) versus maintenance treatment in patients initially treated with either oxaliplatin or irinotecan-based initial systemic therapy for mCRC:

●An early meta-analysis of continuous versus intermittent strategies of delivering systemic chemotherapy to previously untreated mCRC included eight randomized trials, four of which did not employ maintenance therapy, one of which used maintenance therapy with a fluoropyrimidine alone, two of which used biologic therapy alone, and one of which used a fluoropyrimidine plus a biologic agent; seven of the eight trials examined oxaliplatin, and only one [48] involved irinotecan [50]. Overall survival was similar in patients treated with continuous or intermittent strategies (HR 1.03, 95% CI 0.96-1.10), regardless of whether maintenance treatment was included. Quality of life was the same or better with intermittent therapy.

●A more recent network meta-analysis included 12 randomized trials comparing the different treatment strategies of continued chemotherapy, observation, and maintenance therapy (including fluoropyrimidine alone, bevacizumab alone, or fluoropyrimidine plus bevacizumab) [51]. Different induction regimens were used in the different trials, including an oxaliplatin-based regimen in nine [17,35-40,43,44,52], an irinotecan-based regimen in two [49,50], and mixed regimens in one trial [44].

Comparisons of any maintenance therapy versus observation demonstrated that maintenance therapy was associated with improved PFS in both direct (HR 0.63, 95% CI 0.45-0.86) and indirect analyses (HR 0.58, 95% CI 0.43-0.77), but the effect on overall survival was not significant (HR for the indirect analysis 0.91, 95% CI 0.83-1.01). Analyses of each specific maintenance strategy (fluoropyrimidine alone, bevacizumab alone, fluoropyrimidine plus bevacizumab) versus observation also found improved PFS but not overall survival for all comparisons.

●Additional data are available from the randomized FOCUS4-N trial, in which 254 patients with stable or responding disease after 16 weeks of induction therapy with a variety of regimens were randomly assigned to a complete break in therapy with active monitoring versus single-agent capecitabine (1250 mg/m² twice daily on days 1 through 14 of each 21-day cycle), until progression [53]. Maintenance therapy with capecitabine doubled the time to progression and return to full-dose chemotherapy (median PFS 3.88 versus 1.87 months, HR 0.40, 95% CI 0.21-0.75), but had no impact on median overall survival (14.8 versus 15.2 months, adjusted HR 0.93, 95% CI 0.69-1.27). Furthermore, those assigned to maintenance capecitabine had significant higher rates of cumulative toxicity, especially diarrhea, fatigue, nausea, and palmar plantar erythrodysesthesia, although these were primarily low grade.

ASSESSMENT DURING THERAPY — During chemotherapy, response is typically assessed by periodic assay (every one to three months) of serum carcinoembryonic antigen (CEA) levels, if initially elevated, and interval radiographic evaluation (typically every 8 to 12 weeks, or as prompted by a rising CEA level). Although persistently rising CEA levels are highly correlated with disease progression, confirmatory radiologic confirmatory studies should be obtained prior to a change in therapeutic strategy.

Radiographic response — Radiographic tumor response is usually quantified using Response Evaluation Criteria In Solid Tumors (RECIST) (table 3) [54,55].

Immunotherapy using immune checkpoint inhibitors is increasingly being integrated into the care of patients with mismatch repair-deficient/microsatellite instability-high (dMMR/MSI-H) mCRC. (See "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Patients with deficient DNA mismatch repair/microsatellite unstable tumors' and "Systemic therapy for nonoperable metastatic colorectal cancer: Approach to later lines of systemic therapy", section on 'Microsatellite unstable/deficient mismatch repair tumors'.)

Individuals treated with immune checkpoint inhibitors for dMMR/MSI-H mCRC can have pseudoprogression [56], and objective response criteria specifically developed for these drugs should be used (eg, immune-modified RECIST [imRECIST] (table 4)). (See "Principles of cancer immunotherapy", section on 'Immunotherapy response criteria'.)

Serum tumor markers — If initially elevated, a 50 percent or greater declines in CEA from baseline to first restaging can predict disease nonprogression and correlate with favorable long-term outcomes [57]. On the other hand, persistently rising CEA levels (particularly rapidly rising levels [58]) are highly correlated with disease progression [59,60]. However, confirmatory radiologic studies are generally recommended in both settings, particularly if a change in therapeutic strategy is being considered because of a rising CEA. Caution should be used when interpreting a rising CEA level during the first four to six weeks of a new therapy, since spurious early elevation in serum CEA may occur, especially after oxaliplatin [61-63].

Circulating tumor DNA (ctDNA) is the fraction of circulating DNA that is derived from a patient's cancer. Colorectal cancers shed DNA into the blood, and interest in using ctDNA as a surrogate indicator of treatment response has grown as techniques to detect and quantify such DNA have improved. At least one early report suggests that a decrease in levels of ctDNA after chemotherapy is associated with higher objective response rates and longer median progression-free survival and may represent a sensitive and specific marker of therapeutic efficacy [64]. However, there is not yet enough known about the mechanisms controlling ctDNA change and how well radiologic responses and ctDNA markers correlate with each other to understand whether ctDNA can replace or supplement periodic assay of CEA or radiologic assessment.

A year 2018 joint review of the utility of ctDNA analysis in patients with cancer by ASCO and the College of American Pathologists (CAP) concluded that there is insufficient evidence of clinical validity and utility for the majority of ctDNA assays in advanced cancer [65]. Few large prospective validation studies have been performed on ctDNA-based treatment monitoring. At least in advanced breast cancer, there are some data that suggest that ctDNA responses do not always parallel imaging-based responses [66], and no studies convincingly demonstrate improved patient outcomes or any cost savings when compared with standard of care monitoring approaches.

SPECIAL CONSIDERATIONS DURING THE COVID-19 PANDEMIC — The COVID-19 pandemic has increased the complexity of cancer care. Important issues in areas where viral transmission rates are high include balancing the risk from delaying cancer treatment versus harm from COVID-19, minimizing the number of clinic and hospital visits to reduce exposure whenever possible, mitigating the negative impacts of social distancing on delivery of care, and appropriately and fairly allocating limited health care resources. Specific considerations for patients undergoing palliative chemotherapy for stage IV colorectal cancer include establishing goals of care and discussing advance care planning, utilizing oral rather than intravenous therapy, where appropriate, transitioning outpatient care (eg, pump disconnection) to home whenever possible, and using intermittent rather than continuous therapy (with or without maintenance therapy), where feasible. (See 'Continuous versus intermittent therapy' above and "Systemic therapy for nonoperable metastatic colorectal cancer: Selecting the initial therapeutic approach", section on 'Not candidates for intensive therapy'.)

These and other recommendations for cancer care during active phases of the COVID-19 pandemic are discussed separately. (See "COVID-19: Considerations in patients with cancer".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Colorectal cancer".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Colon and rectal cancer (The Basics)")

●Beyond the Basics topics (see "Patient education: Colon and rectal cancer (Beyond the Basics)" and "Patient education: Colorectal cancer treatment; metastatic cancer (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●General considerations – Systemic chemotherapy produces meaningful improvements in survival that are most pronounced in patients who are exposed to all active drugs. Understanding of how to combine and sequence drugs for metastatic colorectal cancer (mCRC) is evolving. (See 'Chemotherapy versus supportive care' above and 'Chemotherapy options' above.)

Some patients with stage IV disease can be surgically cured of their disease, and the goal of initial chemotherapy is maximal reduction in tumor burden. For most, treatment is palliative, and the goals are to prolong overall survival and maintain quality of life (QOL) for as long as possible. (See 'Treatment goals' above.)

For most patients, we suggest early rather than deferred initiation of chemotherapy, and when possible, before patients become symptomatic (Grade 2C). (See 'Timing of chemotherapy' above.)

Full weight-based cytotoxic chemotherapy doses should be used for the treatment of obese patients with mCRC. (See 'Chemotherapy dosing in obese patients' above.)

●Duration of initial therapy – The optimal duration of initial chemotherapy for unresectable mCRC is controversial. The decision to permit treatment breaks for responding patients must be individualized and based upon the regimen being used, tolerance of and response to chemotherapy, disease bulk and location, symptomatology, and patient preference. (See 'Continuous versus intermittent therapy' above.)

•Oxaliplatin – For patients who are responding to an oxaliplatin-based initial regimen, we suggest discontinuing oxaliplatin before the onset of severe neurotoxicity (usually after three to four months of therapy) rather than continuous therapy (Grade 2B). Continuation of oxaliplatin is another alternative for responding patients who have no clinically significant neuropathy. (See 'Oxaliplatin' above.)

For patients who choose to discontinue oxaliplatin, the following represents our approach:

-For most patients, we suggest maintenance therapy over a complete break in therapy, with reintroduction of oxaliplatin at disease progression (Grade 2B). A switch to an irinotecan-based regimen is also an option at disease progression for those with persistent neuropathy.

-For patients initially treated without a biologic agent, we suggest maintenance therapy with a fluoropyrimidine with or without leucovorin (Grade 2B). (See 'OPTIMOX and CONcePT trials' above.)

-For patients initially treated with oxaliplatin plus bevacizumab, we suggest maintenance therapy with a fluoropyrimidine with or without bevacizumab, rather than bevacizumab alone (Grade 2C). (See 'Oxaliplatin' above and 'Maintenance bevacizumab alone' above.)

-For responding patients who were initially treated with oxaliplatin plus an agent targeting the epidermal growth factor receptor (EGFR), we suggest maintenance therapy with fluorouracil plus the anti-EGFR agent rather than an anti-EGFR agent or a fluoropyrimidine alone (Grade 2C). (See 'Patients initially treated with an EGFR inhibitor' above.)

-A complete break in therapy is also a valid option, particularly if a complete clinical response is observed or for those with small-volume metastatic disease who have a partial response or stable disease to the initial course of chemotherapy. Decision-making should also consider patient preference. In such cases, close follow-up with tumor assessment at two-month intervals and early resumption of chemotherapy at the first sign of progression is recommended. (See 'Complete break in therapy' above.).

•Irinotecan – The advantages of intermittent treatment with irinotecan-based regimens are less clear, and for most patients, we continue treatment for as long as tolerability and tumor shrinkage continue. Intermittent treatment with or without maintenance bevacizumab is an option for responding patients who desire a break in therapy. (See 'Irinotecan' above.)

●Response assessment – Response to chemotherapy is typically assessed by periodic assay of serum carcinoembryonic antigen (CEA) levels, if initially elevated, and interval radiographic evaluation. Although persistently rising CEA levels are highly correlated with disease progression, confirmatory radiologic confirmatory studies should be obtained prior to a change in therapeutic strategy. (See 'Assessment during therapy' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Axel Grothey, MD, who contributed to an earlier version of this topic review.