In reviewing the colon cancer literature from the mid-1980s, a recurrent theme could be summed up as, "The overall picture and results of treatment of colorectal cancer have not changed impressively in the last 20 years."[1] During that era, the only drug routinely used to treat colorectal carcinoma was single-agent fluorouracil (5-FU), a drug that had been in use for many years and had shown no proven benefit in the adjuvant setting.

Fortunately, physicians practicing during those years were dedicated to their search for new, more effective therapies for colorectal cancer, especially in the adjuvant setting. Their efforts have led to a significant improvement in the overall picture and results of colorectal cancer treatment over the past 2 decades. In this review, we will compare today's standard of care for adjuvant colorectal carcinoma to that practiced 20 years ago. In addition, we will examine the key questions asked about adjuvant therapy and the answers that ultimately changed clinical practice standards and improved overall survival for patients diagnosed with this disease.

Even with recent advances, colorectal cancer continues to be the third leading cause of cancer deaths in both men and women, accounting for 10% of all cancer-related deaths annually; in 2005, approximately 56,290 deaths associated with the disease were recorded in the United States.[2] Although these statistics are better than those of 1985, when colorectal cancer caused 12% of all cancer-related deaths,[3] more work needs to be done, and research is ongoing. We close this review by considering the questions being asked in current clinical trials and how they might be answered over the next 20 years.

Role of Adjuvant Therapy in 1985

Adjuvant therapy involves the use of chemotherapy following potentially curative surgical resection to destroy residual micrometastatic disease and to prevent relapse. The theory of adjuvant chemotherapy was initially proposed in the 1950s, but in 1985, doctors were still trying to prove its role in the setting of colorectal cancer therapy.

At that time, single-agent, bolus 5-FU was the mainstay chemotherapeutic agent for metastatic colorectal cancer; however, no benefit in the adjuvant setting had yet been proven. Several randomized clinical trials in the 1970s had compared single-agent 5-FU to observation,[4,5] and results demonstrated a trend toward improved overall survival in the chemotherapy arm, but the improvement was not statistically significant.

The finding of a consistent trend toward longer survival after adjuvant treatment with 5-FU, led investigators to develop strategies to enhance the drug's activity using different delivery methods. Many regimens and modifications have been tested over the past 20 years to address a number of clinically relevant questions. Should 5-FU be given as part of a multidrug regimen? Should 5-FU be administered as a continuous infusion or as an oral formulation? Do locally directed therapies to the liver or peritoneum improve outcomes? And perhaps most importantly, who should receive adjuvant chemotherapy?

Do 5-FU Modulators Increase Efficacy?

In the early 1980s, adjuvant chemotherapy trials began to look at the ability of combination regimens, such as 5-FU plus semustine[6,7] or levamisole,[8] to significantly improve outcomes in the adjuvant setting. These trials demonstrated no statistically significant survival benefit among treated patients compared with those in the observation group, although subgroup analyses of several trials demonstrated a significant improvement in overall survival in patients with stage III disease. The largest benefit for this subgroup was found by Wunderlich and colleagues, who reported a 72% overall survival in the adjuvant chemotherapy arm using a regimen of 5-FU, mitomycin, and cytarabine vs 33% in the observation arm (P< .04).[9] Subsequently, this regimen was abandoned in favor of less toxic approaches.

The results of these subgroup analyses were intriguing, but a planned evaluation of adjuvant therapy in stage III patients was still needed to confirm these findings. A randomized controlled trial comparing observation, levamisole alone, and 5-FU/levamisole in stage III colorectal patients was conducted in the 1980s, and the results were published in 1990.[10] At a median follow-up of 3 years, the population of patients treated with the 5-FU/levamisole combination demonstrated a 41% decrease in risk of recurrence (P< .0001) and, more importantly, a reduction in the death rate by 33% (P= .006). These significant benefits were confirmed at reanalysis after a median follow-up of 6.5 years.[11] Based on these results, a National Cancer Institute consensus panel recommended routine adjuvant use of 5-FU/levamisole in stage III colorectal cancer patients.[12]

5-FU/Leucovorin

Despite the proven benefit of the 5-FU/levamisole combination, investigators continued to look for better regimens, often by testing drugs with proven benefit in metastatic disease in the adjuvant setting. One of the first such combinations to show a benefit was 5-FU/leucovorin. Initial trials were combined in a meta-analysis conducted by the International Multicentre Pooled Analysis of Colon Cancer Trials (IMPACT) investigators (Table 1).[13-18] For these studies, patients with stage II/III disease were randomized to observation or six cycles of 5-FU/leucovorin.[13] The adjuvant therapy was well tolerated, and its use yielded a 3-year overall survival of 83% vs 78% in the observation-only arm (P= .02). O’Connell and colleagues used a lower leucovorin dose—20 mg/m2 vs the 200 mg/m2 used in trials included in the IMPACT analysis.[19] Even with the lower dose, the treatment was well tolerated and the 5-FU/leucovorin regimen continued to demonstrate a survival benefit compared with observation.

The National Surgical Adjuvant Breast and Bowel Project (NSABP C-03) took the next step and compared 5-FU/leucovorin with MOF (semustine [methyl-CCNU], vincristine [Oncovin], and 5-FU) in 1,081 stage II /III colorectal cancer patients. The data from that trial showed the leucovorin arm to have a 3-year overall survival rate of 84%, compared with 77% in the MOF group (P= .03).[14]

Leucovorin vs Levamisole

The NSABP then compared levamisole to leucovorin in 2,151 stage II/III colorectal cancer patients who received either 5-FU/leucovorin, 5-FU/levamisole, or a combination of all three drugs. [15] When compared with those in the 5-FU/levamisole arm, the patients who received 5-FU/leucovorin had a longer 5-year disease-free survival rate (65% vs 60%, P= .04) but no significant overall survival benefit (overall survival: 74% vs 70%; P= .07). The three-drug combination offered no additional benefit over the two-drug regimen.

Another trial of 680 patients also demonstrated the superiority of leucovorin over levamisole in patients with stage III disease.[16] After a median follow-up of 46.5 months, the 5-FU/leucovorin arm had a lower mortality rate (P= .0089) and a 4-year overall survival of 63 months, compared with 55 months in the 5-FU/levamisole arm. These results, published in 2001, led to the adoption of 5-FU/leucovorin as the standard of care in the adjuvant setting and in future adjuvant trials.

In December 2005, publication of 10-year Intergroup 0089 follow-up data brought another important update to the leucovorin-vs-levamisole debate.[17] In this trial, 3,794 patients with stage II/III colon cancer were randomized to 1 year of levamisole plus 5-FU or to one of three experimental treatment regimens: low-dose leucovorin/5-FU (Mayo Clinic regimen) given over 6 to 8 months, high-dose leucovorin/5-FU (Roswell Park regimen) given over 8 months, or a low dose of leucovorin/5-FU plus levamisole given for 6 months. No difference in overall survival was noted among the four arms, with overall survival at 10 years ranging from 50% to 59%. However, the leucovorin toxicity profile, especially at the higher dose, was superior to levamisole, with more grade 3/4 diarrhea but less neutropenia and stomatitis recorded. Additionally, the superiority of leucovorin was reaffirmed, as investigators reported that leucovorin given for 6 to 8 months was as effective as levamisole given for 12 months.

The high- and low-dose leucovorin regimens are both considered acceptable alternatives. However, the high-dose Roswell Park regimen is more frequently employed by oncologists using 5-FU/leucovorin, because it is associated with less neutropenia and alopecia.

How Should Systemic 5-FU Be Administered?

When it first came into use, 5-FU was delivered as a bolus injection. However, because the drug is metabolized rapidly and has a short serum half-life (< 20 minutes), researchers became interested in the potential for increasing its benefits when administering it as a continuous infusion rather than as a bolus injection.

In the 1970s, Seifert et al reported that continuous-infusion 5-FU in advanced-stage colorectal cancer yielded an increased response rate (44%) over bolus injections (22%).[20] The two routes also showed a different side-effect profile, with myelotoxicity more severe in the bolus arm—31% of patients reached a total white blood cell count (WBC) of less than 2,000 cells/mm³, whereas no patients in the continuous-infusion arm had a WBC that low. The improved response rate found with continuous-infusion 5-FU in the metastatic setting led to several trials of the regimen in the adjuvant setting.

In 2003, Saini et al reported on a randomized control trial of 716 patients with stage II/III disease who were given either six monthly bolus injections of 5-FU/leucovorin (425 mg/m2 of 5-FU, 20 mg/m2 of leucovorin on days 1-5) or continuous-infusion 5‑FU (300 mg/m2/d for 12 weeks).[21] At a median follow-up of almost 20 months, no difference in overall survival was noted. However, more patients in the bolus-injection arm than in the continuous-infusion arm were alive and free of relapse at follow-up (80% vs 68.6%, P = .023). The continuous-infusion arm also had an improved quality-of-life score and significantly fewer episodes of grade 3/4 neutropenia, alopecia, diarrhea, and stomatitis (P< .0001).

The decrease in toxicity seen with continuous-infusion 5-FU was supported by subsequent trials. Intergroup 0153 compared bolus 5-FU/leucovorin/levamisole for six cycles to continuous-infusion 5-FU/levamisole given in three 8-week cycles.[22] The bolus regimen consisted of the Mayo Clinic regimen of 425 mg/m2 of 5-FU and 20 mg/m2 of leucovorin given for 5 days every 4 to 5 weeks, and the continuous-infusion regimen was 250 mg/m2/d of 5-FU given for three cycles of 56 days. As with previous trials, the most significant difference between the two regimens was the toxicity profile. In the continuous-infusion arm, only 5% of patients experienced a grade 4 toxicity, compared with 39% of patients in the bolus-infusion arm.

Despite the overall decreased toxicity, patients in the continuous-infusion arm were almost twice as likely to withdraw from the protocol before completion. Early discontinuation seemed most related to inconvenience of the pump, chronic hand-foot syndrome, and clotting episodes, all of which were more common in the infusional arm. Unlike Saini's results, this trial demonstrated no significant difference in disease-free survival; like the Saini study, it showed no difference in overall survival between the two arms.