Background

This trial was designed to further study active immunotherapy in patients with stage II or III colon cancer, using an enhanced vaccination schedule (with four instead of three doses of vaccine) and with the vaccine prepared by a central laboratory to assure quality control.

Methods

Patients were eligible if they had undergone curative resection of stage II or III adenocarcinoma of the colon. 

Patients were excluded if there was intestinal perforation or peritoneal seeding, if patients had inflammatory bowel  disease, if the cancer was rectal, if the CEA level did not normalize within 21 days, if there was a prior history of cancer or if they were being treated with steroids or cytotoxic agents. 

Vaccine preparation

Surgery was performed at one of 12 hospitals. Tumor samples were processed at one central site. The tumor was minced, dissociated with collagenase and DNase and frozen. About 60% of vaccines prepared in this fashion contained normal GI flora, but none were declared unsafe for use.  Prior to administration, the vaccine was thawed and irradiated with 200 Gy. Each vaccine contained approximately 10⁷ cells. The first two vaccinations administered to each patient also contained 10⁷ BCG microorganisms. 

Randomization and vaccine administration

The specifics of randomization are not described in the article. Patients were randomized to either receive or not receive the vaccine, and were stratified by institution, tumor location and pathological stage. 

Patients who were randomized to receive the vaccine were given a series of four intradermal injections, starting 4-5 weeks after surgery.  The first two vaccinations were given one week apart, were injected in the anterior thighs and included BCG organisms. Vaccines three and four were administered at 3 weeks and 6 months (booster), were injected into the upper arms and did not include the BCG. 

Follow-up and analysis

Patients were followed up every three months for the first two years, every 6 months for the next three years and annually thereafter. CEA was measured at each follow-up, chest x-rays were performed regularly and CT scan and colonoscopy were performed annually. 

The primary endpoints were recurrence-free interval and recurrence-free survival (time to first occurrence of any malignant disease and time to occurrence of malignancy or death from any cause).  Subgroup analysis by pathological stage was also prespecified. Secondary endpoints included deaths from recurrence of colon cancer and all-cause mortality. 
 

Results

There were 480 samples with adequate tumor cells. 161 patients declined to participate and 65 were ineligible, leaving 254 patients randomized (128 to vaccine, 126 to no vaccine). 

Patient and disease characteristics were well matched, and included (for both groups combined, unless otherwise indicated): 

• Median age: 65-66 years; male: 68%
• Right colon 40%; left colon 54%
• Greater than 4 positive lymph nodes: 25 in control group, 20 in vaccine group
• Stage II disease: 62%; stage III disease 32%

The mean time from tumor resection to the start of vaccine preparation was 2.98 hours.  Based on average specimen size and number of cells per gram of specimen, the authors estimated that their approach would yield enough tumor cells to prepare four vaccine doses in 88% of patients overall. 

Local reactions to the vaccines were more intense with the first two vaccinations, which contained BCG.  These included induration in all patients and some degree of ulceration in 93%.  Eight patients did not receive BCG in the second vaccination because of local reaction to the first.  Mild systemic reactions occurred in 22% of patients, severe reactions in 5%. 

Local reactions to the third and fourth vaccinations, which did not contain BCG, were less intense but did occur in most patients.  Induration greater than 10 mm occurred in 87% of patients with the third vaccination and 92% with the fourth.  Ulceration only occurred in 7% and 4% of patients with these vaccinations. Mild and severe systemic reactions were only present in 7% and 0.5% of patients. 
 

Endpoints

Median follow-up time was 5 years 4 months. 

Among all patients, recurrence of (any) malignant disease occurred in 31.7% of unvaccinated patients and in 19.5% of vaccinated patients. This was statistically significant.  This effect was not seen in stage III patients (41.4% vs. 34.9%). Thus, the reduced overall recurrence rate was due to a particularly significant effect among stage II patients (27.1% vs. 11.8%). 

There was a trend towards reduction in the combined endpoint of  malignancy or death, which reached significance for stage II patients (36.5% vs. 23.5%).  Overall mortality was not affected, but there were trends in reduction of disease-related mortality, particularly for stage II patients (16.5% vs. 8.2%). 

Author's discussion

The authors argue that there are two reasons why this study showed a benefit, whereas two other trials did not. The first is the addition of a fourth vaccination at 6 months, and the second is greater quality control in the preparation of the vaccine by the centralized laboratory.  Adequacy of vaccine preparation facilities is thus an important issue. 

The authors note that adjuvant chemotherapy, which is effective for stage III patients, has not been shown to benefit stage II patients. Thus, immunotherapy would be a logical choice for stage II patients. Whether or not it would benefit stage III patients in addition to chemotherapy is an issue that they intend to assess in future trials.