Sargent, Daniel J.

PURPOSE Three agents with differing mechanisms of action are available for treatment of advanced colorectal cancer: fluorouracil, irinotecan, and oxaliplatin. In this study, we compared the activity and toxicity of three different two-drug combinations in patients with metastatic colorectal cancer who had not been treated previously for advanced disease. PATIENTS AND METHODS Patients were concurrently randomly assigned to receive irinotecan and bolus fluorouracil plus leucovorin (IFL, control combination), oxaliplatin and infused fluorouracil plus leucovorin (FOLFOX), or irinotecan and oxaliplatin (IROX). The primary end point was time to progression, with secondary end points of response rate, survival time, and toxicity. RESULTS A total of 795 patients were randomly assigned between May 1999 and April 2001. A median time to progression of 8.7 months, response rate of 45%, and median survival time of 19.5 months were observed for FOLFOX. These results were significantly superior to those observed for IFL for all end points (6.9 months, 31%, and 15.0 months, respectively) or for IROX (6.5 months, 35%, and 17.4 months, respectively) for time to progression and response. The FOLFOX regimen had significantly lower rates of severe nausea, vomiting, diarrhea, febrile neutropenia, and dehydration. Sensory neuropathy and neutropenia were more common with the regimens containing oxaliplatin. CONCLUSION The FOLFOX regimen of oxaliplatin and infused fluorouracil plus leucovorin was active and comparatively safe. It should be considered as a standard therapy for patients with advanced colorectal cancer.

Background Previous studies have established that higher baseline quality of life (QOL) scores are associated with improved survival in patients with metastatic colorectal cancer (mCRC). We examined the relationship between overall survival (OS) and baseline QOL. Patients and Methods A total of 1 247 patients with mCRC participating in N9741 (comparing bolus 5-FU/LV, irinotecan [IFL] vs infusional 5-FU/leucovorin [LV]/oxaliplatin [FOLFOX] vs. irinotecan/oxaliplatin [IROX]) provided data at baseline on overall QOL using a single-item linear analogue self-assessment (LASA) 0–100 point scale. The association of OS according to clinically deficient (defined as CD-QOL, score 0–50) vs not clinically deficient (nCD-QOL, score 51–100) baseline QOL scores was tested. A multivariable analysis using Cox proportional hazards modeling was performed to adjust for the effects of multiple baseline factors. An exploratory analysis was performed evaluating OS according to baseline QOL status among patients who did or did not receive second-line therapy. Results Baseline QOL was a strong predictor of OS for the whole cohort (CD-QOL vs nCD-QOL: 11.2 months vs 18.4 months, P < .0001), and in each arm IFL 12.4 vs 15.1 months, FOLFOX 11.1 months vs 20.6 months, and IROX 8.9 months vs 18.1 months. Baseline QOL was associated with baseline performance status (PS) ( P < .0001). After adjusting for PS and treatment arm, baseline QOL was still associated with OS ( P = .017). Conclusions Baseline QOL is an independent prognostic factor for OS in patients with mCRC. The demonstration that patient-assessed QOL and PS are independent prognostic indicators suggests that these assessments provide important complementary prognostic information.

Abstract Context. —Inferences from tumor marker studies are complicated by a variety of statistical concerns, which can make proper interpretation of results difficult. This article focuses on important issues that should be addressed when designing, conducting, and analyzing tumor marker studies. Objective. —To highlight the importance of considering statistical significance, risk ratios, statistical power, reproducibility, multiple testing, confirmatory studies, and missing data in the design of marker studies used for prognosis. Results. —Suggestions are provided for more effectively conducting marker studies. These include more careful attention to adequacy of the number of subjects for a marker study and improved documentation and standardization of assay methods. The importance of complete reporting of study results and description of the statistical analysis methods used is also emphasized. Conclusion. —Cooperation among clinicians, laboratory scientists, and statisticians will be required to conduct statistically sound tumor marker studies and to facilitate prioritizing markers for new and confirmatory studies in an environment of limited patient and specimen resources.

Abstract Background. —Under the auspices of the College of American Pathologists, the current state of knowledge regarding pathologic prognostic factors (factors linked to outcome) and predictive factors (factors predicting response to therapy) in colorectal carcinoma was evaluated. A multidisciplinary group of clinical (including the disciplines of medical oncology, surgical oncology, and radiation oncology), pathologic, and statistical experts in colorectal cancer reviewed all relevant medical literature and stratified the reported prognostic factors into categories that reflected the strength of the published evidence demonstrating their prognostic value. Accordingly, the following categories of prognostic factors were defined. Category I includes factors definitively proven to be of prognostic import based on evidence from multiple statistically robust published trials and generally used in patient management. Category IIA includes factors extensively studied biologically and/or clinically and repeatedly shown to have prognostic value for outcome and/or predictive value for therapy that is of sufficient import to be included in the pathology report but that remains to be validated in statistically robust studies. Category IIB includes factors shown to be promising in multiple studies but lacking sufficient data for inclusion in category I or IIA. Category III includes factors not yet sufficiently studied to determine their prognostic value. Category IV includes factors well studied and shown to have no prognostic significance. Materials and Methods. —The medical literature was critically reviewed, and the analysis revealed specific points of variability in approach that prevented direct comparisons among published studies and compromised the quality of the collective data. Categories of variability recognized included the following: (1) methods of analysis, (2) interpretation of findings, (3) reporting of data, and (4) statistical evaluation. Additional points of variability within these categories were defined from the collective experience of the group. Reasons for the assignment of an individual prognostic factor to category I, II, III, or IV (categories defined by the level of scientific validation) were outlined with reference to the specific types of variability associated with the supportive data. For each factor and category of variability related to that factor, detailed recommendations for improvement were made. The recommendations were based on the following aims: (1) to increase the uniformity and completeness of pathologic evaluation of tumor specimens, (2) to enhance the quality of the data needed for definitive evaluation of the prognostic value of individual prognostic factors, and (3) ultimately, to improve patient care. Results and Conclusions. —Factors that were determined to merit inclusion in category I were as follows: the local extent of tumor assessed pathologically (the pT category of the TNM staging system of the American Joint Committee on Cancer and the Union Internationale Contre le Cancer [AJCC/UICC]); regional lymph node metastasis (the pN category of the TNM staging system); blood or lymphatic vessel invasion; residual tumor following surgery with curative intent (the R classification of the AJCC/UICC staging system), especially as it relates to positive surgical margins; and preoperative elevation of carcinoembryonic antigen elevation (a factor established by laboratory medicine methods rather than anatomic pathology). Factors in category IIA included the following: tumor grade, radial margin status (for resection specimens with nonperitonealized surfaces), and residual tumor in the resection specimen following neoadjuvant therapy (the ypTNM category of the TNM staging system of the AJCC/UICC). Factors in category IIB included the following: histologic type, histologic features associated with microsatellite instability (MSI) (ie, host lymphoid response to tumor and medullary or mucinous histologic type), high degree of MSI (MSI-H), loss of heterozygosity at 18q ( DCC gene allelic loss), and tumor border configuration (infiltrating vs pushing border). Factors grouped in category III included the following: DNA content, all other molecular markers except loss of heterozygosity 18q/DCC and MSI-H, perineural invasion, microvessel density, tumor cell–associated proteins or carbohydrates, peritumoral fibrosis, peritumoral inflammatory response, focal neuroendocrine differentiation, nuclear organizing regions, and proliferation indices. Category IV factors included tumor size and gross tumor configuration. This report records findings and recommendations of the consensus conference group, organized according to structural guidelines defined herein.

11093 Background: Defective DNA mismatch repair (MMR) results in microsatellite instability (MSI) and is detected in approximately 15% of sporadic colon cancers. MMR status has been shown to provide prognostic and predictive information in primary colon cancers. We sought to develop a model to predict MMR deficiency using clinically available data, and thereby facilitate patient selection for MMR or MSI testing. Methods: TNM stage II and III colon carcinomas (n= 982) were studied from six 5- fluorouracil-based adjuvant therapy trials conducted by the North Central Cancer Treatment Group. MMR status in tumors had been analyzed by MSI (using mono- and dinucleotide markers) or by immunohistochemistry for MMR proteins (hMLH1 and hMSH2). Logistic regression and a recursive partitioning and amalgamation (RPA) analysis was used to identify important predictive factors of MMR status. Factors explored included age, gender, histologic grade, tumor site, stage, lymph node metastases, and T-stage. Results: Defective MMR was found in147 (15%) cancers. Tumor site was the most important predictor of MMR status followed by histologic grade. Distal tumors had a low likelihood of defective MMR (3% rate overall; 13/468), whereas proximal tumors had a greater likelihood of defective MMR (26%; 130/506). For patients with proximal tumors, the addition of histologic grade and gender increased the prediction of defective MMR ( Table ). Using tumor site, histologic grade, and gender, the logistic regression model showed excellent discrimination (c- statistic = 0.81). Conclusions: Tumor site is an important predictor of defective MMR that is rare in distal and increased in proximal tumors. The combination of proximal site, poor differentiation, and female gender resulted in a 51% likelihood of defective MMR. Therefore, this model can facilitate the selection of sporadic colon cancers for MMR or MSI testing to enable its use in clinical decision-making. [Table: see text] No significant financial relationships to disclose.

11088 Background: Obesity is associated with an increased risk of colon cancer. However, the influence of body mass index (BMI) upon the prognosis of patients with established colon cancer remains unknown. Methods: We conducted a retrospective study of 1,803 patients with surgically resected stage III colon cancer who were enrolled in five randomized trials of 5-fluorouracil-based adjuvant chemotherapy conducted by the North Central Cancer Treatment Group. Patient height and weight were recorded at study entry and BMI (kg/m2) was calculated and categorized. Cancer recurrence or death were monitored during 5 years of follow-up. The score and likelihood ratio p-values were determined from univariate and multivariate Cox regression models respectively, after stratifying by study. Results: Among stage III colon cancer patients, 19% were obese (BMI 30 kg/m2), 37% were overweight (BMI, 25 to 29.9 kg/m2), 38% were of normal-weight (BMI, 20 to 24.9 kg/m2), and 6% were underweight (BMI < 20 kg/m2). Obese versus normal-weight patients showed higher rates of lymph node (LN) metastasis (>3 LNs; 38% vs. 29%, p <0.01) and tumor site was more likely to be distal versus proximal (52% vs. 45%, p= 0.03). No differences in age, gender, or histologic grade were found. In a univariate analysis, obese patients had significantly worse disease-free survival (DFS) compared with normal-weight patients (hazard ratio 1.25 (95% CI: 1.04 -1.51; p= 0.02). The 5 year DFS rates were 49% in obese patients versus 57% in normal weight subjects. Furthermore, poorer DFS was observed for obese patients after adjusting for age, sex, histologic grade, and tumor site (p= 0.03). Neither overweight nor underweight patients (vs. normal-weight) had significantly different DFS. Analysis of the predictive impact of BMI for 5-FU-based adjuvant therapy is in progress. Conclusions: Obesity (BMI 30 kg/m2) was associated with a greater number of metastatic lymph nodes and poorer disease-free survival in patients with stage III colon cancer, suggesting that obesity influences tumor progression. No significant financial relationships to disclose.

9599 Background: We have previously identified overall a single-item measure for baseline quality of life (QOL) as a strong prognostic factor for survival (Tan, ASCO 2008), and that fatigue was an important component of patient QOL (Sloan, 2007). To explore whether patient-reported fatigue was supplemental or redundant to the prognostic information of overall QOL, we performed a patient-level pooled analysis of 43 NCCTG and MCCC oncology clinical trials of the effect of baseline fatigue on OS. Methods: 3,915 patients participating in 43 trials provided data at baseline for fatigue on a single-item 0–100 point scale. OS was tested for association with clinically deficient fatigue (CDF, score 0–50, n=1,497) vs not clinically deficient fatigue (nCDF, score 51–100, n=2,418). Cox proportional hazards models adjusted for the effects of overall QOL, performance score, race, disease site, age and gender. Results: Baseline fatigue was a strong predictor of OS for the entire patient cohort (CDF vs. nCDF: 31.5 mos vs >83.9 mos, p<0.0001). The effect sizes were consistent across different disease sites (GI, esophageal, head and neck, prostate, lung, breast and others). After controlling for covariates, including performance status and overall QOL, baseline fatigue remained a strong prognostic factor in multivariate models (CDF vs. nCDF: HR=1.23, p=0.02). Conclusions: Fatigue is a strong prognostic factor for OS independent of overall QOL and PS in a wide variety of oncology patient populations. Single-item measures of overall QOL and fatigue can help to identify vulnerable subpopulations among cancer patients. No significant financial relationships to disclose.

9563 Background: Fatigue is a prevalent and debilitating symptom reported by cancer patients (pts) which compromises a pt's quality of life (QOL). This study examined the relationship between PR fatigue and QOL as well as cancer-related symptoms (CRS) in 43 North Central Cancer Treatment Group and Mayo Clinic Cancer Center clinical trials. Methods: 3,915 pts from 43 oncology clinical trials provided baseline fatigue data on a single-item 0–100 point scale. Pts' QOL assessment included a single-item overall QOL and associated QOL domains measured by numerical analogues, the Profile of Mood States (POMS), and PR symptom assessment measures. Associations between fatigue and QOL domains were assessed by Spearman correlation coefficients. Wilcoxon rank sum test compared QOL scores between pts with clinically deficient fatigue(CDF, score ≤50) vs. no clinically deficient fatigue (nCDF, score>50). Changes from baseline in fatigue and QOL were compared by Wilcoxon rank sum test with a 20-point change defined as clinically meaningful. Results: 38% of pts reported CDF at baseline and 45% of pts reported CDF at last assessment. Fatigue was only moderately correlated at best with overall QOL, pain, POMS, social and physical function (Spearman rho's of .27,.40, .56, .38 and .38 respectively). Pts with CDF averaged over 10 points lower overall QOL, pain, POMS, social, and physical function (see table below, all p<.0001) as well as worsening CRS including sleepiness, nausea, headache, abnormal sweating, trouble sleeping, dry mouth, and sexual dysfunction (all p<.001). Pts with 20+ points worsening in fatigue declined in overall QOL, physical function, pain and POMS (all p<.0001). Conclusions: Patients with CDF suffer greater deficits in QOL and CRS. Patients report fatigue as distinctly different from overall QOL, pain, physical, social, mood status and CRS. Fatigue appears with a broad spectrum of CRS clusters. Routine measurement and management of fatigue could impact QOL and treatment-related symptoms. [Table: see text] No significant financial relationships to disclose.

e19031 Background: Useful predictors of survival in cancer patients have been found including such as gender, race, age, marital status, state of disease and tumor location. We set to explore the prognostic value of baseline QOL assessments on overall survival in lung cancer patients. Methods: A total of 1,319 lung cancer patients were followed at Mayo Clinic between 1997 and 2007. Overall QOL was assessed by either Lung Cancer Symptom Scale (LCSS) or a single-item Linear Analogue Self Assessment (LASA) with comparable wording. QOL scores were transformed to a 0–100 scale with higher scores representing better status. Patients QOL scores assessed within the first 6 months of follow-up were dichotomized using a priori definition for clinically deficiency score (CDS, ≤50 vs. >50). The Kaplan-Meier method was used to estimate the distribution of survival times by QOL clinical deficient categorization. Cox Proportional Hazards Models were used to evaluate the prognostic importance of QOL on OS alone and in the presence of age, smoking category, gender, treatment, stage of disease, smoking cessation and years of consuming a pack every day. Results: Baseline QOL was significantly associated with OS univariately (HR=0.46 with p<0.0001 based on the CDS categorization). After controlling for all these other factors, the indication of a clinically deficient baseline QOL contributed significantly to the prediction of patient survival (p=0.0002; HR=0.67). Age, treatment, disease stage, smoking cessation and smoking category were also significant predictors of OS multivariately. The median survival for patients that had an overall QOL score less than the CDS (≤50) was 1.6 versus 4.6 years for patients with QOL >50. Conclusions: Baseline overall QOL measured by LCSS or LASA is a significant and independent prognostic factor for OS in lung cancer patients. A meaningful next step in this research is to investigate interventions to improve lung cancer patient QOL and thereby increase their likelihood of survival as well. No significant financial relationships to disclose.

6512 Background: The use of randomized phase II designs with an experimental arm and a standard-treatment control arm (R2PII) instead of a conventional single-arm design is clearly increasing in oncology. In practice, sample size, related cost issues, the belief that historical controls are adequate, and the use of a standard-treatment control arm in a phase II setting are frequently raised objections to R2PII trials. As the expense and complexity of definitive phase III trials increase, the ability of phase II trials to provide reliable and accurate results is critical. Methods: We investigated the ability of single arm vs R2PII trials to provide accurate conclusions by modeling variability in historical controls, patient outcome drifts independent of the tested therapy, and patient selection effects. Simulations compared R2PII and single-arm designs with binary endpoints under realistic parameters (e.g. alpha = beta = 0.10, historical control success rate = 20%, target success rate = 40%). Results: In the absence of variability in historical controls, estimated false positive and negative rates in both designs mirror the designated specifications. However, even in the presence of a modest drift effect in the population (mean 5% absolute shift in true control success rate), the false positive rate in single-arm designs is inflated two to three fold (to 20%-30%), while the R2PII retains the desired error rates. Greater confidence in historical controls in the single-arm design corrects only a small portion of the deviations. Increasing the sample size in each trial inflates the false positive error rate further to as much as 50%. Varying several sets of parameters gave similar results. Conclusions: In the presence of variability in historical controls, patient drift and/or selection effects, the false positive error rate of a single arm design is unacceptably high. In contrast, the R2PII design is reliable and robust despite the complexities in patient outcome drift and selection effects, and variability in historical control success rates. Given the rapid improvements in outcomes of many tumor types, the R2PII design should be the preferred method to evaluate novel agents in oncology in spite of the associated costs and the use of a reference control arm. No significant financial relationships to disclose.