Targeted Agents and Oxaliplatin-containing Regimens for the Treatment of Colon Cancer

Evolution of Colorectal Cancer Treatment: The Central Role of Oxaliplatin

As a single agent, oxaliplatin displays only a marginal activity in the clinic (1), but when administered with fluoruracil (5-FU), the combination is highly synergistic. The best example of this synergism comes from the three-arm phase III trial by Rothenberg et al. (2) in which patients with advanced colorectal cancer, progressing following treatment with irinotecan, 5-FU and leucovorin (LV) were randomised to receive 5-FU-alone, oxaliplatin-alone or the fluorouracil, leucovorin, and oxaliplatin (FOLFOX4) combination of the two drugs. Response rate (ORR) and time-to-progression (TTP) were similarly low in patients allocated to receive either 5-FU-alone (0% ORR; TTP=2.7 months) or oxaliplatin-alone (1.3% ORR; TTP=1.6 months) while better results were observed in the combination arm (9.9% ORR; TTP=4.6 months). In vitro data implicate the down-regulation of thymidylate synthase (TS) protein expression by oxaliplatin as a possible molecular mechanism for the observed synergy (3). As front-line treatment, the FOLFOX4 regimen was tested against leucovorin and fluorouracil (LV5FU2) in a randomized study which involved 420 patients, published in 2000 (4). Although patients randomized to receive up-front the combination including oxaliplatin up-front showed a better ORR (50.7% vs. 22.3%) and a longer time-to-progression (9.2 months vs. 6 months), overall survival (OS) was not significantly improved (16.2 vs. 14.7 months). However, the median OS observed in the LV5FU2 arm was about three months longer than that reported in the pre-oxaliplatin era, probably as a result of the introduction of oxaliplatin (or irinotecan) as a second-line treatment. In fact, in the same year, the results of two trials in which patients were randomised to receive a combination of irinotecan and 5-FU or 5-FU-alone were published. 5-FU was given according to different schedules: the weekly bolus infusion according to Roswell Park in the USA (5), weekly prolonged venous infusion (PVI) (the German schedule) or bi-weekly combination of bolus and PVI as in the LV5FU2 in Europe (6). The irinotecan-containing arms performed statistically better regarding all investigated parameters (ORR, TTP and OS). Overall, these data prompted the head-to-head comparison of oxaliplatin- or irinotecan-containing regimens in five trials (7-11). Looking at the most important parameter, OS, in three studies the two drugs tied, while in the remaining two, the oxaliplatin arms were the winners.

Building upon the 5-FU/oxaliplatin cornerstone was the addition of irinotecan in several phase I-II studies (12-23). Although the different schedules make any comparison difficult, dose-limiting toxicities observed in all trials were diarrhea and neutropenia. Response rates ranging from 50% to 72% were observed in advanced colorectal cancer when triplets were administered as first-line treatment, and from 24% to 27% in the studies evaluating triplets in 5-FU-refractory patients (13, 22).

The front-line administration of a triplet, compared to a doublet including 5-FU and irinotecan, has been investigated in two randomised phase III trials. In the trial by the Hellenic Cooperative Group which involved 283 patients, although all the investigated parameters of efficacy were in favour of the triplet arm (ORR 43% vs. 34%; TTP 9.4 vs. 6.9 months; OS 21.5 vs. 19.5 months), none reached statistical significance (23). The second trial included 244 patients and was performed by the Italian Gruppo Oncologico Nord Ovest (GONO) Cooperative Group (24). In this study the administration of the triplet was associated with improved ORR (60% vs. 34%, p<0.001) longer OS (22.6 vs. 16.7 months, p=0.032) and an increased resectability rate (15% vs. 6%, p=0.033). While the 5-FU-irinotecan regimen was the same in both trials, the schedule and the doses of the drugs in the triplet arms were different. This fact, along with differences in the treated populations may account for the discrepant results observed. Alternative combinations such as rapid sequences of doublets [e.g. oxaliplatin, irinotecan, 5FU and LV (FOLFIRINOX)] have shown remarkable activity against pancreatic cancer.

Adding anti-Angiogenic Antibodies to Oxaliplatin-based Regimens

Bevacizumab is a monoclonal antibody that blocks the vascular endothelial growth factor, a critical mediator of angiogenesis (25, 26). Randomized studies in previously untreated patients have shown that the addition of bevacizumab to 5-FU alone (27) or to the combination of 5-FU and irinotecan (28) improves the efficacy of chemotherapy. In patients previously treated with 5-FU and irinotecan, the inclusion of bevacizumab into the FOLFOX4 regimen significantly improved ORR (22.7% vs. 8.6%), TTP (7.3 vs. 4.7 months) and OS (12.9 vs. 10.8 months), when compared to the FOLFOX-only arm (29). The study NO16966 began as a front line study in which patients were randomized to receive either FOLFOX4 or XELOX (oxaliplatin at 130 mg/m² as an intravenous injection on day 1 and capecitabine at 1,000 mg/m² orally b.i.d on days 1-14 every three weeks) (30, 31). In August 2003, after the phase III bevacizumab data became available, the protocol was amended to a 2×2 partially blinded study by adding 7.5 mg/kg of bevacizumab i.v. or placebo on day 1 every three weeks to XELOX and bevacizumab 5 mg/kg i.v. or placebo every two weeks to FOLOFOX4; overall 2,034 patients were included. Main objectives of the study were two-fold: (i) non inferiority of XELOX vs. FOLFOX4; (ii) bevacizumab plus chemotherapy (FOLFOX4 or XELOX) superiority to chemotherapy plus placebo. The primary parameter of efficacy was progression-free survival (PFS): it was similar in both arms (8.5 months in the FOLFOX4 plus placebo or bevacizumab arms, 8 months in the XELOX plus placebo or bevacizumab arms). OS was also very close (19.6 months in the FOLFOX 4 arms, 19.8 months in the XELOX arms). Capecitabine is, therefore, non-inferior to bolus and continuous-infusion 5-FU. Regarding the second main objective, the superiority of bevacizumab to placebo, although the primary objective was met since a longer PFS was observed in the bevacizumab arms (9.4 vs. 8 months, p=0.023), the results were inferior to that observed when bevacizumab was added to 5-FU/LV alone (27) (8.8 months in the FU/LV/bevacizumab group, 5.6 months in the FU/LV group; p<0.001), to irinotecan, leucovorin and fluorouracil (IFL) (28) (10.6 months in the IFL/bevacizumab arm, 6.2 months in the IFL arm; p<0.001) or when combined with FOLFOX in patients progressing following treatment with IFL (see above). Furthermore, in a planned subset analysis, statistical superiority of bevacizumab vs. placebo was evident in the XELOX subgroup Hazard Ratio (HR)=0.77; 97.5% Confidence Interval (CI)=0.63 to 0.94; p=0.0026), but did not reach the significance level in the FOLFOX4 subgroup (HR=0.89; 97.5% CI=0.73 to 1.08; p=0.1871). Finally, ORR was identical (38%) in patients who received bevacizumab or placebo, while OS was only marginally-improved in patients treated with chemotherapy plus bevacizumab (21.3 vs. 19.9 months p=0.07) (31); OS was 20.3 months in the FOLFOX placebo arm vs. 21.2 months in the FOLFOX bevacizumab arm (HR=0.94; 95% CI=0.75-1.16) and 19.2 (XELOX placebo) vs. 21.4 (XELOX bevacizumab) (HR=0.84; 95% CI=0.68-1.04) (31, Appendix online only). NO16966 updated results, showed that median OS was 19.8 months in the pooled XELOX/XELOX-placebo/XELOX-bevacizumab arms vs. 19.5 months in the pooled FOLFOX4/FOLFOX4-placebo/FOLFOX4-bevacizumab arms (HR=0.95, 97.5% CI=0.85-1.06); in the pooled XELOX/XELOX-placebo arms, median OS was 19.0 vs. 18.9 months in the pooled FOLFOX4/FOLFOX4-placebo arms (HR=0.95, 97.5% CI=0.83-1.09) (32), confirming the primary analysis.

Having previously shown the superiority of the three-drug combination with irinotecan, oxaliplatin, leucovorin and fluorouracil (FOLFOXIRI) over the two-drug combination FOLFIRI (24), the GONO group recently reported the conclusion of the TRIBE trial a two-arm protocol in which bevacizumab was added to either FOLFOXIRI or FOLFIRI (33). In this study, the three-drug arm was associated with a statistically significant improvement of PFS (12.2 vs. 9.7 months) and ORR (65% vs. 53%). However, the increase of response rate did not translate into improved resection of liver metastases in the whole population (15% vs. 12%) and in the subgroup of patients with liver-only metastases (32% vs. 28%).

In spite of the negligible benefits observed when bevacizumab was added to oxaliplatin-based regimens in the front-line setting, bevacizumab is commonly included in these regimens following the hypothesis that it is fundamental in order to normalize the abnormal structure and function of tumor vasculature to make it more efficient for oxygen and drug delivery to cancer cells (34). Further corollaries from this hypothesis have lent support to the use of bevacizumab as maintenance following front-line treatment and also upon progression, changing the chemotherapy partners since resistance to chemotherapy results from changes in tumour cell biology and is often agent-specific while bevacizumab resistance, if it occurs, can result from the development of alternative angiogenesis pathways (35, 36).

The role of bevacizumab as a single-maintenance agent was investigated in only one randomised non-inferiority design aimed at demonstrating the non-inferiority of a no-treatment strategy, compared to the administration of bevacizumab (37). Eligibility criteria for this trial requested the inclusion of bevacizumab in an up-front regimen with an oral or intravenous fluoropyrimidine (alone or in combination with irinotecan or oxaliplatin). The primary end-point of the study, which included 262 patients, was TTP (from randomisation) while secondary end-points were PFS (from the beginning of front-line chemotherapy), time-to-second line treatment, OS, adverse events and costs. TTPs was 4.1 months among 124 patients randomised to receive bevacizumab maintenance and 2.9 months among 123 patients randomised to observation (p=0.47, i.e. non-inferiority was not demonstrated). With 168 events, OS results were 25.1 months among bevacicumab-maintained and 22.8 months among observed patients (p=0.218). As expected, the costs of the two strategies are astonishingly different: 37,596 USD for each bevacizumab-maintained patients and 8,180 USD for each observed patient.

Bevacizumab beyond progression was investigated in the TML; ML18147, a randomised phase III trial in which 820 patients with advanced colorectal cancer, progressing on first-line chemotherapy plus bevacizumab, were randomised to chemotherapy-alone (switching to fluoropyrimidine plus irinotecan or oxaliplatin, depending on the type of front-line chemotherapy administered) or to chemotherapy and bevacizumab (38). The primary end-point was OS from initiation of second-line chemotherapy. The primary end-point was met with and observed OS of 11.2 months among patients who continued bevacizumab and 9.8 months among patients who received chemotherapy alone (HR=0.81; 95% CI=0.69-0.95, p=0.0062). Bevacizumab beyond progression was tested in another smaller phase III study (BEBYP), in 185 patients (39). This study had planned to enroll 262 patients but was closed prematurely following the presentation of the TML results. The primary end-point was PFS that was longer in the bevacizumab compared to the placebo arm (5.2 vs. 6.7 months; p=0.007). ORR was 18% in the placebo arm and 21% in the bevacizumab arm (p=0.71); survival data are still immature but similar in the two arms (16 vs. 16.5 months; p=0.34).

Finally the concept of keeping the anti-vascular agent upon progression was studied in another randomised phase III trial, the VELOUR trial, in which patients with metastatic colorectal cancer progressing on first-line oxaliplatin-based chemotherapy with or without bevacizumab (n=1226) received either FOLFIRI alone or FOLFIRI plus the anti-vascular agent aflibercept (40). The observed gain in OS (primary end-point) was similar of that observed in the TML trial: 13.5 months for patients receiving FOLFIRI plus aflibercept and 12.1 months for patients receiving FOLFIRI plus placebo (HR=0.82, p=0.0032), this last trial supporting the US FDA approval of aflibercept after failure of second-line treatments.

Adding anti-Epidermal Growth Factor Receptor Antibodies (EGFR) to Oxaliplatin-based Regimens

Cetuximab and panitumumab are monoclonal antibodies against the EGFR, a molecular target that is expressed in the majorty of colorectal tumors (41). Panitumumab is a fully human immunoglobulin (Ig) G2 monoclonal antibody (42) while Cetuximab is a chimeric human-mouse IgG1 monoclonal antibody (43); both antibodies bind the EGFR extracellular domain with high affinity and inhibit ligand-induced EGFR tyrosine phosphorylation, tumor cell activation, and tumor cell proliferation.

The two drugs were initially tested as single agents, and in combination with anticancer drugs, in patients with advanced colorectal cancer patients selected on the basis of the expression of EGFR, in different lines of chemotherapy. Only later it did become evident that the status of KRAS was a better biomarker for selecting patients for a treatment with antibodies to EGFR. In fact, it has been shown in retrospective studies and in subgroup analyses from randomised trials that clinical responses to antibodies against EGFR are restricted to patients (approximately 55 to 70%) whose tumors are wild type for KRAS (44-52).

The most relevant randomised trials in which antibodies to EGFR are investigated in combination with oxaliplatin-based regimens are reported in Table I.

The OPUS (Oxaliplatin and Cetuximab in First-Line Treatment of mCRC) was a randomised phase II study that compared the use of FOLFOX4 plus cetuximab (n=169) vs. FOLFOX4 alone (n=168) in patients with previously untreated advanced colorectal cancer (53, 54). The primary objective was to determine whether the addition of the antibody improved the ORR compared to chemotherapy alone. Secondary objectives included the determination of rates of potentially curative metastasectomy, duration of response, PFS, OS, and safety. A retrospective subgroup analysis done on regulatory request investigated the potential relationship between PFS/ORR and tumor KRAS mutational status (n=315, 93% of the whole population; KRAS wild-type population n=179, 57%). In the original patient population, adding cetuximab improved the ORR by an absolute increase of 10% (36% to 46%), which did not reach the original goal [the odds ratio (OR) was 1.52]. Median PFS was identical for both arms at 7.2 months. However, cetuximab was effective in the population with tumors that demonstrated wild-type KRAS. The ORR increased from 34% to 57% (OR 2.551; p=0.0027), and PFS alone improved from 7.2 to 8.3 months (OR=0.567; p=0.0064). The observed OS did not differ significantly between the two groups: 18.5 months with FOLFOX alone; 22.8 months with FOLFOX plus cetuximab (p=0.39). Remarkably, in the KRAS-mutant population, the ORR, median PFS, and overall PFS were markedly worse for patients who received cetuximab.

The COIN trial enrolled 1,630 patients with advanced colorectal cancer, randomised to receive a combination of fluoropyrimidine and oxaliplatin with (arm B) or without (arm A) cetuximab (55). Tumour samples from 1,316 (81%) patients were used for somatic molecular analyses; 565 (43%) had KRAS mutations. In patients with KRAS wild-type tumours (arm A, n=367; arm B, n=362), OS did not differ between treatment groups (median survival 17.9 months in the control group vs. 17.0 months in the cetuximab group; HR=1.04, 95% CI=0.87-1.23, p=0.67). Similarly, there was no effect on progression-free survival (8.6 months in the control group vs. 8.6 months in the cetuximab group; HR=0.96, 0.82-1.12, p=0.60). ORR increased from 57% (n=209) with chemotherapy alone to 64% (n=232) with addition of cetuximab (p=0.049).

In the first line phase III NORDIC VII study patients were randomly assigned to receive either standard Nordic FLOX (arm A), cetuximab and FLOX (arm B), or cetuximab combined with intermittent FLOX (arm C). The primary end-point was PFS. OS, response rate (RR), R0 resection rate, and safety were secondary end-points (56). KRAS status was assessed during conduction of the trial, but no adjustment of the sample size was performed. As a result, the number of patients in the two main comparison arms, FLOX with cetuximab and FLOX without cetuximab, with KRAS wild-type tumours was 97 per arm, which clearly diminishes the power of the trial. OS was almost identical for the three groups (20.4, 19.7 and 20.3 months, respectively), and confirmed response rates were 41%, 49%, and 47%, respectively. In patients with KRAS wild-type tumours, cetuximab did not provide any additional benefit compared with FLOX-alone. In patients with KRAS mutations, no significant difference was detected, although a trend toward improved PFS was observed in arm B.

The Panitumumab Randomized Trial in Combination With Chemotherapy for Metastatic Colorectal Cancer to Determine Efficacy (PRIME) a randomized phase III trial designed to evaluate panitumumab at 6 mg/kg every two weeks plus FOLFOX4 (n=593) vs. FOLFOX4 alone (n=590) as a first-line treatment in patients with wild-type or mutant KRAS tumours (57). Among patients with wild-type KRAS tumours, the addition of panitumumab to first-line FOLFOX4 significantly improved PFS (the primary endpoint; 9.6 vs. 8.0 months, respectively; HR=0.80; 95% CI=0.66-0.97; p=0.02) compared with FOLFOX4 alone. Among patients with wild-type KRAS tumours the addition of panitumumab to FOLFOX4 vs. FOLFOX4 alone was not associated with an improvement of median OS (23.9 vs. 19.7 months, respectively; HR=0.83; 95% CI=0.67-1.02; p=0.072) or with a better overall response rate (55% with FOLFOX plus panitumumab, 48% with FOLFOX alone; OR=1.35; p=0.068). Patients with KRAS mutant tumours who received panitumumab plus FOLFOX4 had significantly shorter PFS than patients treated with FOLFOX4 alone (7.3 vs. 8.8 months, respectively; HR=1.29; 95% CI=1.04-1.62; p=0.02), but median OS and the ORR were not significantly different.

Finally, the addition of cetuximab to a fluoropyrimidine-oxalipatin combination was investigated in 272 patients randomised into the NewEPOC study (58). Eligible patients were required to be KRAS wild-type and to have operable liver metastases. The study was prematurely closed following a recommendation from the Independent Data Monitoring Committe since PFS was significantly worse in the cetuximab arm (14.8 vs. 24.2 months, p<0.048).

In conclusion, the addition of EGFR-targeting antibodies to oxaliplatin-containing regimens for the front-line therapy of patients with KRAS wild-type advanced colorectal cancer, is associated with results which were not consistently observed in the reported trials. In fact, ORR increased in the OPUS and COIN trials; PFS increased in the OPUS and PRIME trials. No differences in OS were observed while in the new EPOC trials the addition of cetuximab is clearly detrimental.

The findings of NORDIC are particularly intriguing since it is the only trial in which patients with KRAS-mutated cancer experienced a trend towards improvement in PFS with an antibody to EGFR. It is also the only trial in which the observed response rate in patients with KRAS-mutated tumors was numerically higher than in these with KRAS wild-type cancer. The possible explanations for the different results by adding cetuximab or panitumumab to oxaliplatin-(or irinotecan)-containing regimens (not presented here) are extensively discussed elsewhere (59).

Use of Both anti-Angiogenic and anti-EGFR Antibodies with Chemotherapy

Since combining cytotoxic drugs that act through different mechanisms improved outcomes in patients with metastatic colorectal cancer, it seemed logical that further progress would result from combining bevacizumab with either cetuximab or panitumumab and administering these monoclonal antibodies together with chemotherapy.

The CAIRO 2 study was a randomised phase III trial in which 732 patients with previously untreated metastatic colorectal cancer were randomly assigned to receive capecitabine, oxaliplatin, and bevacizumab or the same three drugs accompanied by cetuximab in cycles administered every three weeks (60). After a median follow-up of 23 months, the addition of cetuximab to the combination of capecitabine, oxaliplatin, and bevacizumab significantly reduced the median PFS, from 10.7 months to 9.4 months; the addition of cetuximab was also associated with a trend towards reducing the median OS from 20.3 months to 19.4 months. Tumor tissue was assessed for the status of the KRAS gene in 71% of patients; mutations were found in 40% of the specimens. The addition of cetuximab did not improve the outcome in patients whose tumors contained wild-type KRAS but was deleterious for those with tumors bearing a mutant KRAS gene. The negative effect of adding an antibody to EGFR to a chemotherapy–bevacizumab combination was also observed by Hecht et al. in the Panitumumab Advanced Colorectal Cancer Evaluation (PACCE) trial (61), in which 823 patients who had not received previous treatment for metastatic colorectal cancer were randomly assigned to receive FOLFOX and bevacizumab, either alone or accompanied by panitumumab. The addition of panitumumab reduced both the median PFS and the median OS.

Finally, the concept of combining both antibodies with FOLFOX or FOLFIRI was investigated in the CALGB Intergroup study C80405 (62). This was a three-arm study powered for survival. The physician selected either FOLFOX or FOLFIRI, and then patients were randomly assigned to bevacizumab alone, cetuximab alone or the combination of the two. The arm in which cetuximab and bevacizumab were combined was closed in September 2009. Therefore, this trial will compare cetuximab and bevacizumab head-to-head (and FOLFOX vs. FOLFIRI). The results of this study are awaited with interest since thus far the two monoclonal antibodies have been compared in combination with the FOLFIRI-only backbone, in the FIRE trial (63).

The Adjuvant Setting: Where Are We in 2013?

The results obtained with the addition of oxaliplatin (or irinotecan) to the 5-FU backbone in advanced-stage disease, prompted the launch of a series of randomised phase III studies in stage II and III colon cancer. In these studies patients were randomised to receive either a combination of 5-FU and LV or the same combination plus oxaliplatin (or irinotecan). The role of oxaliplatin was investigated in two studies (64, 65) while the impact of irinotecan was investigated in three (66-68). These trials show that while the addition of oxaliplatin to the regimens including 5-FU and LV is associated with a better outcome for patients, the introduction of irinotecan is either marginally effective or detrimental. Furthermore, the activity of oxaliplatin is independent of the 5-FU-LV schedule since the results initially observed in the MOSAIC trial in which oxaliplatin is given bi-weekly with bolus and continuous infusion 5-FU were confirmed in the NSABP C07 trial in which bi-weekly oxaliplatin is combined with weekly 5-FU and LV. However, the schedule issue matters in terms of toxicity, especially neurotoxicity, enteropathy and toxic death. The lower incidence of grade III neurotoxicity reported in the NSABP C07 trial (8%) compared to the MOSAIC trial (12.4%) is easily explained by the lower cumulative doses of oxaliplatin in the former study (765 mg/m² vs. 1,025 mg/m²). The GI toxicity in the NSABP C07 was described in detail (69). A syndrome of bowel wall injury characterised by hospitalisation for the management of severe diarrhoea or dehydration and radiographic or endoscopic evidence of bowel wall thickening or ulceration occurred in 79 out of 1,857 patients (4.3%). The majority of these patients (65%) had received bolus 5-FU-LV and oxaliplatin. Besides this syndrome, not observed in the MOSAIC trial, the incidence of grade III and IV diarrhoea was also higher in the oxaliplatin containing arm of the NSABP C07 trial (38% vs. 10.8%). Finally treatment related deaths were 0.5% in the MOSAIC trial and 1.2% in the NSABP C07 trial. In conclusion, FOLFOX should be the preferred regimen in the adjuvant setting, although the lower incidence of grade III neurotoxicity, the ease of administration of bolus delivery and the avoidance of a central catheter could make FLOX preferable in some cases, with the warning that enteric toxicity should be carefully managed. The positive results of the MOSAIC trial, which enrolled 40% of patients with stage II disease, led the regulatory Authorities FDA in the USA and the European Medicine Agency to approve this protocol for the adjuvant treatment of only stage III disease. These decisions were criticised in a commentary by Grothey and Sargent (70) published in the Journal of Clinical Oncology, for the main reasons that were based upon a sub-group analysis. The recently reported improved OS observed only in stage III in both trials (71, 72), with an absolute difference of 4.4% (p=0.029) in the MOSAIC trial and of 2.5% in the NSABP C07 (p-value not reported) in favour of the oxaliplatin-containing arm could, however, lend support to the conservative approach applied by the regulatory agencies.

Furthermore, the observed interaction between age and treatment suggest to use caution recommending oxaliplatin-containing regimens for the elderly. In fact, updated OS results of MOSAIC and NSABP trials indicate that the gain is limited to patients younger than 70 years old. In particular, in NSABP C07 OS significantly improved with FLOX compared with FULV for patients younger than age 70 (HR=0.80; 95% CI=0.68 to 0.95; p<0.013) (72). The 5-year OS estimates were 78.8% for FULV and 81.8% for FLOX, a 3.1% improvement in patients younger than age 70 years. OS did not vary significantly by treatment in patients aged 70 years (p=0.30), but nominal OS at 5 years was 4.7% worse for patients treated with FLOX (71.6%) compared with that for patients treated with FULV (76.3%) among these older patients. In the MOSAIC trial, 5-year OS was not significantly different in the elderly (n=315): 75.8% with FOLFOX vs. 76.1% with LV5FU2 (p=0.661) (73).

More recently were reported the results of NO16968 (XELOX in Adjuvant Colon Cancer Treatment [XELOXA]), a randomised phase III study designed to investigate whether the addition of oxaliplatin to an oral fluoropyrimidine, would maintain the efficacy of treatment compared to earlier trials offering at the same time a potentially more convenient regimen (74). In this study, oxaliplatin was combined with capecitabine (XELOX) and compared with bolus FU/FA (Mayo Clinic or Roswell Park regimens) in patients with stage III colon cancer. The intention-to-treat population comprised 1,886 patients; 944 patients were randomly assigned to XELOX and 942 to FU/FA (Mayo Clinic, n=664; Roswell Park, n=278). After 57 months of follow-up for the primary analysis, 295 patients (31.3%) in the XELOX group had relapsed, developed new primary colon cancer, or died compared to 353 patients (37.5%) in the FU/FA group (HR for DFS=0.80; 95% CI=0.69 to 0.93; p=0.0045). The 3-year DFS rate was 70.9% with XELOX and 66.5% with FU/FA. The HR for OS for XELOX compared to FU/FA was 0.87 (95% CI=0.72 to 1.05; p=0.1486). The 5-year OS for XELOX and FU/FA were 77.6% and 74.2%, respectively. Although direct comparisons of oxaliplatin-based regimens in the adjuvant setting have not been performed, data from large phase III trials evaluating FOLFOX-4 and XELOX in the first- or second-line treatment of metastatic colorectal cancer have demonstrated that XELOX is non-inferior to FOLFOX4 in terms of PFS, OS, and objective response rate (75, 76).

Furthermore, a systematic review and meta-analysis of 40 trials comparing oxaliplatin-containing regimens to fluoropyrimidine regimens did not show any significant difference in DFS at a median follow-up of three-years (or closest reported analysis) for XELOX vs. FLOX (HR=0.99, 95% CI=0.80-1.22) or FOLFOX (HR=1.00, 95% CI=0.72-1.41) (77). There was also no significant difference in OS at a median follow-up of at least 5 years. In this study no interaction was observed between age and treatment arm and DFS benefit was the same in the populations of patients <65 or ≥65 years old.

Molecularly-targeted agents demonstrated some degrees of activity in patients with metastatic colorectal cancer and other advanced malignant diseases. In fact, several phase III studies have resulted in the approval of bevacizumab in combination with a fluoropyrimidine-, oxaliplatin-, or irinotecan-based chemotherapy regimen for metastatic colorectal cancer. It was therefore anticipated that bevacizumab would also improve the outcome of patients with high-risk stage II and stage III colon cancer when treated with the standard infusional FOLFOX chemotherapy. This expectation formed the basis of the National Surgical Adjuvant Breast and Bowel Project (NSABP) C-08 and of the AVANT (AVastin adjuvANT) trials (78, 79).

In the NSABP C08 study conducted in the USA stage II and III colon cancer 2,672 patients were randomised to either FOLFOX6 alone (an evolution of FOLFOX4 in which the bolus of 5-FU on day 2 is taken off while 5-FUPVI is given at the dose of 2,400 mg/m² over 46 h starting on day 1), or to the same combination plus bevacizumab 5 mg/kg for 12 cycles followed by bevacizumab alone at the same dose for 12 additional cycles (78).

The AVANT trial is a three-arm European trial in which 3,451 stage III and high risk stage II patients were randomised to receive: (i) FOLFOX4 alone for 12 cycles, (ii) FOLFOX4 plus bevacizumab at 5 mg/kg for 12 cycles followed by bevacizumab alone at the dose of 7.5 mg/kg every three weeks for eight additional cycles or (iii) XELOX (capecitabine oxaliplatin) plus bevacizumab at 7.5 mg/kg every three weeks for 8 cycles followed by bevacizumab alone at the same dose and interval for eight additional cycles (79). This study addresses not only the value of bevacizumab in this population but also the possibility of substituting capecitabine to bolus and PVI 5-FU, thus avoiding the need for a central catheter. The results show that neither NSABP C08 nor the AVANT met their primary end-point of prolonging DFS after three years. Furthermore, in the AVANT trial the administration of bevacizumab was associated with a detrimental effect on OS in the FOLFOX arm (HR=1.27; 95% CI=1.03-1.57; p=0.02) but not in the XELOX arm (HR=1.15; 95% CI=0.93-1.42; p=0.21).

The role of cetuximab as an addition to FOLFOX6 was investigated in the North Central Cancer Treatment Group (NCCTG) N0147 trial, a randomised phase III trial which enrolled 2686 patients resected for stage III wild-type KRAS colon cancer (80). The primary randomized comparison was 12 bi-weekly cycles of mFOLFOX6 with and without cetuximab. KRAS mutation status was centrally determined. The trial was halted after a planned interim analysis of 48% of predicted events (246/515) occurring in 1,863 (out of 2,070 planned) patients with tumours having wild-type KRAS. DFS in patients with wild-type KRAS mutations was the main end-point of the study. Secondary end points included OS and toxicity.

Three-year DFS for mFOLFOX6 alone was 74.6% vs. 71.5% with the addition of cetuximab (HR=1.21; 95% CI=0.98-1.49; p=0.08) in patients with wild-type KRAS, and 67.1% vs. 65.0% (HR=1.12; 95% CI=0.86-1.46; p=0.38) in patients with mutated KRAS, with no significant benefit in any subgroups assessed. Among all patients, grade 3 or higher adverse events (72.5% vs. 52.3%; OR]=2.4; 95% CI=2.1-2.8; p<0.001) and failure to complete 12 cycles (33% vs. 23%; OR=1.6; 95% CI=1.4-1.9; p<0.001) were significantly higher with cetuximab.

Increased toxicity and greater detrimental differences in all outcomes were observed in patients aged 70 years or older with wild-type KRAS. In particular patients who were treated with mFOLFOX6 alone demonstrated significantly better OS (HR=2.00; 95% CI=1.05-3.78; p=0.03), with a three-year estimate of 86.2% (95% CI=78.9%-94.1%) for the mFOLFOX6 alone group vs. 72.5% (95% CI=64.0%-82.0%) for the mFOLFOX6 with cetuximab group.

Similar negative results were reported in the Pan-European Trials in Alimentary Tract Cancer (PETACC) 8 trial in which 2,559 patients with completely-resected colon cancer (1,602 with wild-type KRAS) were randomised to receive either FOLFOX4 alone (arm A) or the same chemotherapy plus cetuximab (arm B) (81). Among KRAS wild-type patients, no difference in the primary end-point, three-year DFS, was observed between the two arms: 78% in arm A, 75.1% in arm B (HR=1.05; 95% CI=0.85-1.29; p=0.66). Grade 3 or higher adverse events were significantly increased in arm B (80.9%) vs. arm A (66.2%; HR=1.09; 95% CI=0.81-1.47; p=0.56).

The results of all these studies clearly demonstrated that the addition of molecularly targeted agents to oxaliplatin-containing regimens is not associated with an improvement of DFS/OS. Therefore, oxalipatin-fluoropyrimidine regimens are the only ones to have consistently shown a favorable impact in terms of DFS (72-74) and OS (72, 73) compared to fluoropyrimidine alone. However, the mature results of the MOSAIC and C-07 studies, performed on patients selected only on the basis of the clinical and pathological stage, provided outcome data that were considerably less positive than had been anticipated, suggesting the need to look for predictive biomarkers. Although most cases of colorectal cancer develop a chromosomal instability pathway, some 15% of cases are characterized by microsatellite instability, a molecular marker of defective DNA mismatch repair (dMMR) (82). Preliminary retrospective data in 303 patients with stage III completely resected colon cancer who received adjuvant FOLFOX were recently reported and demonstrated a significant impact of the MMR status on DFS (83). The 3-year DFS rate was 90.5% (95% CI=73.2%–96.9%) in patients with dMMR tumours (n=269) and 73.8% (95% CI=67.9%–78.8%) in patients with pMMR tumours (n=34) (log-rank test, p=0.027). MMR status was the only parameter to cross the level of the statistical significance in the multivariate analysis (p=0.015). From this retrospective study it is possible to conclude that MMR status is an independent prognostic biomarker in patients with stage III colon cancer treated with adjuvant FOLFOX chemotherapy. The predictive effect of MMR status in the response to FOLFOX adjuvant chemotherapy should be assessed by the undergoing analyses of tissue samples from previously completed randomized trials comparing 5-FU plus oxaliplatin with 5-FU alone, such as the MOSAIC and NASBP-C07 studies. The underlying biological mechanisms of MMR effect on adjuvant chemotherapy remain to be identified.

To summarise, in less than two decades, the three-year DFS for patients with stage III colon cancer increased from 44% with surgery alone in 1990, to 62% with the six month combination of 5-FU and LV in 1995, to 65% observed in the LV5-FU2 arm in the MOSAIC trial and finally to the 72% obtained in both oxaliplatin-containing arms of the MOSAIC and NSABP C07 trials. Improvements in the imaging which allowed a better selection of patients contributed to these results but this success story is mostly attributable to oxaliplatin-based adjuvant chemotherapy. Therefore, the USA Intergroup and the Italian TOSCA studies seeking a reduction in toxicity by randomizing three versus six cycles of FOLFOX following stage III curative resection is a key matter for the future.