Patient-reported outcomes assessment in cancer trials: taking stock, moving forward.

To evaluate and improve the use of cancer trial end points that reflect the patient's own perspective, the National Cancer Institute organized an international conference, Patient-Reported Outcomes Assessment in Cancer Trials (PROACT), in 2006. The 13 preceding articles in this special issue of the Journal were commissioned in preparation for or in response to the PROACT conference, which was cosponsored by the American Cancer Society. Drawing from these articles and also commentary from the conference itself, this concluding report takes stock of what has been learned to date about the successes and challenges in patient-reported outcome (PRO) assessment in phase III, phase II, and symptom management trials in cancer and identifies ways to improve the scientific soundness, feasibility, and policy relevance of PROs in trials. Building on this synthesis of lessons learned, this article discusses specific administrative policies and management procedures to improve PRO data collection, analysis, and dissemination of findings; opportunities afforded by recent methodologic and technologic advances in PRO data collection and analysis to enhance the scientific soundness and cost efficiency of PRO use in trials; and the importance of better understanding the usefulness of PRO data to the full spectrum of cancer decision makers, including patients and families, health providers, public and private payers, regulatory agencies, and standards-setting organizations.

Challenges and recommendations for advancing the state-of-the-science of quality of life assessment in symptom management trials.

Major findings are presented from a workshop on Quality of Life Assessment in Cancer Symptom Management Trials, sponsored by the National Cancer Institute. Data-driven research reports focused on 3 topics, 1) the rationale and utility of health-related quality of life (HRQOL) assessment, 2) conceptual models, and 3) measurement and design issues. Recommendations for including HRQOL assessment cited the potential value of: capturing additional treatment effects (eg, fatigue + depression); describing the patient experience; predicting patient prognosis; identifying potential adverse effects; observing interactions among symptoms; calculating quality adjusted survival and cost-effectiveness; and generating new hypotheses. Recommendations for developing more fully developed conceptual models focused on maintaining clear distinctions among symptoms, function, summary measures of HRQOL, and global HRQOL assessments; identifying symptom clusters; pursuing hypotheses about whether clustering is better explained as symptom-related or as patient-related (genetic predispositions); and gaining a better understanding of the dynamic and reciprocal influences of symptoms on each other. With respect to measurement and design issues, because different HRQOL measures cover different domains with various degrees of sensitivity, there is a need to select measures that are carefully tailored to the study's hypotheses. Finally, there is a growing appreciation that trials must be powered to test for effects on secondary endpoints.

A genome scan of 18 families with chronic lymphocytic leukaemia.

Chronic lymphocytic leukaemia (CLL) accounts for about 30% of all leukaemias and is most prevalent in older individuals. Significant familial aggregation has been demonstrated but the mode of inheritance is unknown. Recurrent cytogenetic abnormalities are frequently found in CLL tumour cells but no susceptibility genes have been confirmed. We have collected clinical data and biospecimens on families ascertained for having at least two living patients with CLL. The current study included DNA samples from 94 individuals (38 affected patients) in 18 families. We have carried out a genome scan using the ABI 28-panel medium density linkage mapping set (average spacing of 10 cM and average heterozygosity of 80%). Genotypes for 359 markers were scored. Multipoint limit of detection (lod) scores were calculated, assuming both dominant and recessive inheritance and allowing for increased penetrance with age and genetic heterogeneity. Non-parametric linkage scores were also calculated. Lod scores of 1.0 or greater were found on regions of chromosomes 1, 3, 6, 12, 13 and 17, but none of these loci achieved statistical significance. Four of these six regions (6q, 13q, 12 and 17p) coincide with areas where cytogenetic abnormalities are frequently observed in CLL tumour cells and are, therefore, strong candidate regions for containing germ line changes.

B-cell monoclonal lymphocytosis and B-cell abnormalities in the setting of familial B-cell chronic lymphocytic leukemia.

BACKGROUND: Among all hematologic malignancies, B-cell chronic lymphocytic leukemia (BCLL) has the highest familial clustering (three- to sevenfold increase), strongly suggesting a genetic component to its etiology. Familial BCLL can be used as a model to study the early pathogenesis of this disease. METHODS: We examined nine kindreds from the National Cancer Institute's Familial BCLL Registry, consisting of 19 affected members with BCLL and 33 clinically unaffected first-degree relatives. Flow cytometric immunophenotyping to detect a B-cell monoclonal lymphocytosis (BCML) was performed. Monoclonality was confirmed by polymerase chain reaction analysis of whole blood DNA. Cell cycle analysis for aneuploidy was conducted. RESULTS: In all affected individuals, we observed the classic BCLL CD5/CD19/CD20/CD23 immunophenotypic patterns. Six of the 33 unaffected individuals (18%) had evidence of BCML. Additional individuals (13/33, 39%) showed some other abnormality, whereas 14 individuals (42%) were normal. Based on an estimated prevalence of 0.7% for BCML in the general population, the finding of six subjects (18%) with clonal abnormalities in this relatively modest sample was significantly greater than expected (i.e., 18% vs. 0.7%, P < 5.7 x 10(-9)). CONCLUSIONS: Individual components of BCML and other B-cell abnormalities were observed in almost half of the apparently unaffected individuals. Our findings suggested that BCML may be an early detectable abnormality in BCLL. The spectrum of some of these observed abnormalities suggested the involvement of different B-cell subpopulations or different pathways in clonal evolution. Population-based, longitudinal studies will be required to determine the incidence of BCML and other B-cell abnormalities and their relation to disease progression in BCLL and other closely related B-cell lymphoproliferative disorders.