Final stage of liver cancer

Objective: Two recent studies in resectable non-small cell lung cancer by the Lung Cancer Study Group (LCSG) suggested an advantage to adjuvant therapy with cyclophosphamide, doxorubicin (Adriamycin), and cisplatin (CAP). Neither study had a no-treatment control arm. The purpose of this study was to compare the CAP regimen with no treatment in patients with resectable early-stage non-small cell lung cancer.

Methods: After complete resection, eligible patients with T1N1 or T2N0 non-small cell lung cancer were randomly assigned to receive or not to receive four courses of CAP at 3-week intervals beginning on day 30 after surgery after stratification for histology, preoperative white blood cell count, and Karnofsky performance status before surgery. The CAP regimen consisted of 400 mg/[m.sup.2] of cyclophosphamide, 40 mg/[m.sup.2] of doxorubicin, and 60 mg/[m.sup.2] of cisplatin. Of the 269 eligible patients entered in the study, 101 had recurrence and 127 had died at the time of analysis. The mean time since randomization is 6.4 years; mean follow-up is 3.8 years. There were no differences in time to recurrence or overall survival between the two groups even when analyses were adjusted for prognostic variables. Only 53% of the eligible patients received all four courses of CAP, and only 57% of such patients received all four cycles on time. Among the patients who had recurrences, 74% had their initial recurrence at a distant site.

Conclusion: No survival benefit for CAP vs no-treatment control was found in this study. Therefore, adjuvant therapy with CAP should not be recommended for patients with resected early-stage non-small cell lung cancer. Further trials to test adjuvant therapy are indicated, but investigators should use better antiemetics to improve patient compliance as well as more active cisplatin-based chemotherapy regimens.

Most patients with stage I lung cancer who have recurrences have them at systemic sites, including the brain, and less often in the thorax.(1) Therefore, systemic treatments such as chemotherapy are necessary for successful adjuvant therapy even for early-stage non-small cell lung cancer. Two published studies by the Lung Cancer Study Group (LCSG) suggest an advantage to adjuvant therapy with the CAP (cyclophosphamide, doxorubicin, and cisplatin) regimen.(2)(3) In the earlier study, patients with resectable stage II and III adenocarcinoma and large cell carcinoma of the lung were randomly assigned to receive either CAP or intrapleural bacillus Calmette-Guerin (BCG) plus levamisole as an adjuvant to complete surgical resection (protocol 772). There was a significant increase in disease-free interval and overall survival in patients treated with CAP.(2) In the more recent study, patients with resected locally advanced non-small cell lung cancer (LCSG protocol 791) who were randomized to CAP plus thoracic radiotherapy had a superior outcome to those receiving thoracic irradiation alone.(3) In 1980, since neither of the previous studies had an untreated control, LCSG protocol 801 was initiated. It included only patients with T1N1 or T2N0 non-small cell lung cancer. This randomized phase 3 trial compared CAP chemotherapy every 3 weeks for four courses to close follow-up. The detailed results of this study have already been published.(4) This article contains no new data but has a more recent up-to-date review of the literature.

METHODS

Patients with T1N1 or T2N0 non-small cell lung cancer(5) aged 75 years or younger who had sufficiently rapid recovery from surgery and could start treatment by day 30 were included in the trial. Patients with T1N1 were included since they were part of stage I in the staging classification prior to the new classification of Mountain.(5) These patients were not to have received previous cancer treatment or to have had a history of cancer other than nonmelanoma skin cancer or in situ cancer of the cervix. They had to give informed consent and had to be available for appropriate follow-up. Eligible patients were assigned randomly by telephone from the central operations office. Stratification factors included histologic features (squamous vs nonsquamous, preoperative white blood cell count >9,100/m[m.sup.3] vs <9,100/m[m.sup.3]), and Karnofsky performance status before surgery (90% vs 100%). The patients assigned to chemotherapy received cyclophosphamide (400 mg/[m.sup.2]), doxorubicin (40 mg/[m.sup.2]), and cisplatin (60 mg/[m.sup.2]) all at a single visit on day 1; this was then repeated at 3-week intervals for a total of four courses. Dose reduction based on blood cell counts and renal function were mandated. Patients on both arms were followed up every 3 weeks for 3 months, then every 3 months for 2 years and every 6 months thereafter.

Pretreatment Evaluation

Patients were evaluated with a complete history and physical examination, and a chest radiograph, posteroanterior and lateral, as well as routine hematology, chemistry, and an ECG. Only if the alkaline phosphatase or SGOT was >50% above institutional limits were additional tests carried out to further evaluate liver and bone as possible metastatic sites. The chest radiograph and blood work were repeated every visit but scans, when done, were repeated less frequently.

At surgery, stage was determined according to the TNM classification. A lobectomy or pneumonectomy was performed with the highest mediastinal nodes identified and labeled by the surgeon and the pathologist; all mediastinal areas were sampled by surgery plus mediastinoscopy. If the highest node obtainable was positive for tumor, the patient was ineligible for study. Pathologic specimens were reviewed by the pathology reference center at the M. D. Anderson Hospital in Houston.

Statistical Methods

The trial size was determined by the number of patients necessary for 90 events to occur (power 0.9 to detect a twofold difference in median survival with a two-sided log rank test using a significance level of 0.05). The Pearson [x.sup.2] test was used for contingency table analysis; Yates' continuity correction was used in the case of two-by-two tables.(6) Survival and recurrence were estimated by the Kaplan-Meier method(7) and two-sided significance tests were based on log rank statistics as given by Mantel(8) but without continuity correction. Survival analysis stratification variables were used to adjust the data as were other important covariates by using the Cox model(9) or by using stratified log rank score statistics. All end points were defined in the usual way starting on the day of randomization.

RESULTS

Patients were entered in the study from November 1980 to May 1986 and data for the analysis were collected until January 15, 1990. No further update is planned. Among the 283 randomized patients, 14 had major protocol violations and were declared ineligible. Among the 269 eligible patients, 29 did not receive the assigned treatment. Therefore, there were 240 (85%) eligible treated patients. Among the eligible patients, 82% were male, 59% were at least 60 years old, 56% had nonsquamous histologic features, and 84% had tumors classified as T2N0, the remainder as T1N1. There were no statistically significant differences by treatment arm for the potential important prognostic variables that were not included in the stratification (weight loss >10%, neutrophil count >7,000/m[m.sup.3], lobectomy, segmental resection vs pneumonectomy/wedge resection, sex, the presence or absence of heart disease prior to randomization, and complications following surgery).

Relapse and Survival

The mean time from random assignment of the 269 eligible patients was 6.4 years, with a mean follow-up of 3.8 years. One hundred one recurrences occurred along with 31 second primary tumors (all sites, not just lung) and 127 deaths at the time of analysis. We looked for individual potential prognostic variables. Only prior heart disease and a preoperative neutrophil count >7,000/m[m.sup.3] were associated with significantly greater death rates. A number of other parameters were associated with higher recurrence rates.

The time to first recurrence, excluding second primary tumors, is shown in Figure 1. The time to death from any cause is shown in Figure 2. No treatment difference is observed even when adjustment occurred using the Cox model or if the 23 patients who refused the assigned treatments were excluded. When all 283 randomly assigned patients were included, no treatment differences were observed. Among patients with squamous histology, the rates of recurrence (excluding second primary tumors) and death (from all causes) differed in favor of the CAP arm (p = .05, two-sided). Among patients with nonsquamous histology, those on the control arm had somewhat better survival (p = .06, two-sided). No explanation for this finding was forthcoming.

[CHART OMITTED]

Toxicity of Treatment