Mike Bradburn

Survival Analysis Part I: Basic concepts and first analyses

Survival analysis is a collection of statistical procedures for data analysis where the outcome variable of interest is time until an event occurs. Because of censoring–the nonobservation of the event of interest after a period of follow-up–a proportion of the survival times of interest will often be unknown. It is assumed that those patients who are censored have the same survival prospects as those who continue to be followed, that is, the censoring is uninformative. Survival data are generally described and modelled in terms of two related functions, the survivor function and the hazard function. The survivor function represents the probability that an individual survives from the time of origin to some time beyond time t. It directly describes the survival experience of a study cohort, and is usually estimated by the KM method. The logrank test may be used to test for differences between survival curves for groups, such as treatment arms. The hazard function gives the instantaneous potential of having an event at a time, given survival up to that time. It is used primarily as a diagnostic tool or for specifying a mathematical model for survival analysis. In comparing treatments or prognostic groups in terms of survival, it is often necessary to adjust for patient-related factors that could potentially affect the survival time of a patient. Failure to adjust for confounders may result in spurious effects. Multivariate survival analysis, a form of multiple regression, provides a way of doing this adjustment, and is the subject the next paper in this series.

Survival Analysis Part II: Multivariate data analysis – an introduction to concepts and methods

Survival Analysis Part II: Multivariate data analysis – an introduction to concepts and methods

Survival Analysis Part III: Multivariate data analysis - choosing a model and assessing its adequacy and fit

Survival Analysis Part III: Multivariate data analysis – choosing a model and assessing its adequacy and fit

Carcinosarcoma of the Ovary 19 Years of Prospective Data from a Single Center

A review of clinicopathologic features and outcome in women with carcinosarcoma of the ovary (also known as malignant mixed mesodermal tumor [MMMT]) compared with a group of women with serous adenocarcinoma (SAC) of the ovary was conducted.

The Randomised Assessment of Treatment using Panel Assay of Cardiac Markers (RATPAC) trial: a randomised controlled trial of point-of-care cardiac markers in the emergency department

Objectives To determine whether using a point-of-care cardiac biomarker panel would increase the rate of successful discharge home after emergency department assessment, and affect the use of cardiac tests and treatments, subsequent attendance at or admission to hospital and major adverse events. Design and setting Pragmatic multicentre randomised controlled trial in six acute hospitals in the UK. Participants Patients attending with acute chest pain due to suspected myocardial infarction (N=2243). Interventions Diagnostic assessment using a point-of-care biomarker panel consisting of creatine kinase, myocardial type, myoglobin and troponin I measured at baseline and 90 min compared with standard care without the point-of-care panel. Main outcome measures The primary outcome was successful discharge home, defined as having left hospital or awaiting transport home by 4 h after attendance and no major adverse events up to 3 months. Secondary outcome measures included length of stay, use of coronary care, cardiac interventions and inpatient beds, emergency department attendances, subsequent admissions, outpatient visits and major adverse events. Results Point-of-care panel assessment was associated with an increased rate of successful discharge (358/1125 (32%) vs 146/1118 (13%); OR 3.81, 95% CI 3.01 to 4.82; p<0.001), reduced median length of initial hospital stay (8.8 vs 14.2 h; p<0.001) and greater use of coronary care (50/1125 (4.0%) vs 31/1118 (3.0%); p=0.041), but no difference in mean length of initial stay (29.6 vs 31.7 h; p=0.462), mean inpatient days over follow-up (1.8 vs 1.7; p=0.815) or major adverse events (36 (3%) vs 26 (2%); OR 1.31, 95% CI 0.78 to 2.20; p=0.313). Conclusions Point-of-care panel assessment increases successful discharge home and reduces median length of stay, but does not alter overall hospital bed use. Trial registration Current controlled trials ISRCTN37823923.

Survival analysis part IV: further concepts and methods in survival analysis.

Most analyses of survival data use primarily Kaplan–Meier plots, logrank tests and Cox models. We have described the rationale and interpretation of each method in previous papers of this series, but here we have sought to highlight some of their limitations. We have also suggested alternative methods that can be applied when either the data or a given model is deficient, or when more difficult or specific problems are to be addressed. For example, analysis of recurrent events can make an important contribution to the understanding of the survival process, and so investigating repeat cancer relapses may be more informative than concentrating only on the time until the first. More fundamentally, missing data are a common issue in data collection that in some cases can seriously flaw a proposed analysis. Such considerations may be highly relevant to the analysis of a data set, but are rarely mentioned in the analysis of survival data. One possible reason for this is a perceived lack of computer software, but many of the approaches discussed here are currently incorporated into existing commercial statistical packages (e.g. SAS, S-Plus, Stata) and freeware (e.g. R). On the other hand, the desire may be to ‘keep things simple for the readership’. This view is reasonable, but is valid only where a simple analysis adequately represents the survival experience of patients in the study. Ensuring the analyses are appropriate is therefore crucial. More advanced survival methods can derive more information from the collected data; their use may admittedly convey a less straightforward message, but at the same time could allow a better understanding of the survival process. The aim of this series has been to aid awareness, understanding and interpretation of the many and varied methods that constitute the analysis of survival data. It is paramount that analyses are performed in the knowledge of the assumptions that are made therein, and the more complex methods, in particular, are best applied by a statistician.

Subcellular localisation of cyclin B, Cdc2 and p21(WAF1/CIP1) in breast cancer. association with prognosis.

The heterodimeric cyclin B/Cdc2 protein kinase governs entry into mitosis, and can be negatively regulated through p53-mediated transcriptional induction of the cyclin-dependent kinase inhibitor p21(WAF1/CIP1). Ectopic expression of p21(WAF1/CIP1) in cultured cells has been shown previously to influence the subcellular distribution of the cyclin-dependent kinases (CDKs) including Cdc2. In this study, we have examined the subcellular localisation of Cdc2, cyclin B and p21(WAF1/CIP1) by immunohistochemistry in a well characterised series of primary breast cancers. Surprisingly, p21(WAF1/CIP1) was predominantly cytoplasmic in many of the tumours, where it was associated with high p53 levels; cytoplasmic p21(WAF1/CIP1) and high cyclin B levels were also significant predictors of poor prognosis. We conclude that breast tumorigenesis may be characterised by abnormalities in pathways determining not only levels of expression of key regulatory molecules, but also their subcellular localisation. Investigation of the subcellular distribution of cell cycle regulatory proteins, particularly p21(WAF1/CIP1), could provide valuable prognostic markers in breast cancer.

Factors which predict unsuccessful mobilisation of peripheral blood progenitor cells following G-CSF alone in patients with non-Hodgkin's lymphoma.

INTRODUCTION High-dose therapy with haematopoietic progenitor cell support has increasingly been utilised for patients with haematological malignancies. Peripheral blood is the stem cell source of choice, however, various mobilisation strategies are used by different centres. PATIENTS AND METHODS Over a 2-year period, 52 patients with non-Hodgkins lymphoma (median age 47 years, range 16-64 years) underwent peripheral blood progenitor cell mobilisation using G-CSF alone (16 microg/kg/day). The harvest was considered successful if > or =1 x 10(6) CD34(+) cells/kg were collected by leukapheresis. The histological subtypes of non-Hodgkins lymphoma comprised: follicular (24 patients), diffuse large B-cell (14 patients), lymphoplasmacytoid (four patients), mantle cell (three patients), lymphoblastic lymphoma (one patient) and small lymphocytic lymphoma/chronic lymphocytic leukaemia (six patients). The median interval from diagnosis of non-Hodgkins lymphoma to mobilisation was 27 months (range 2 months to 17 years). The median number of prior treatment episodes was 2 (range 1-5); 26 patients had received fludarabine alone or in combination. At the time of peripheral blood progenitor cell mobilisation, 20 patients were in 1st remission and 32 were in > or =2nd remission; 30 patients were in partial remission and 22 were in complete remission; the bone marrow was involved in nine patients. RESULTS Peripheral blood progenitor cell mobilisation/harvest was unsuccessful in 19 out of 52 (37%) patients (mobilisation: 18, harvest: 1). The factors associated with unsuccessful mobilisation or harvest were: prior fludarabine therapy (P=0.002), bone marrow involvement at diagnosis (P=0.002), bone marrow involvement anytime prior to mobilisation (P=0.02), histological diagnosis of follicular, mantle cell, or lymphoplasmacytoid lymphoma, or small lymphocytic lymphoma/chronic lymphocytic leukaemia (P=0.03) and female gender (P=0.04). CONCLUSION Although peripheral blood progenitor cells can be successfully mobilised and harvested from the majority of patients with non-Hodgkins lymphoma after treatment with G-CSF alone, the latter is unsuccessful in approximately one-third of patients. These factors should be taken into account when patients are being considered for high-dose treatment.

Cytoplasmic p21WAF1/CIP1 expression is correlated with HER-2/ neu in breast cancer and is an independent predictor of prognosis

BackgroundHER-2 (c-erbB2/Neu) predicts the prognosis of and may influence treatment responses in breast cancer. HER-2 activity induces the cytoplasmic location of p21WAFI/CIPI in cell culture, accompanied by resistance to apoptosis. p21WAFI/CIPI is a cyclin-dependent kinase inhibitor activated by p53 to produce cell cycle arrest in association with nuclear localisation of p21WAFI/CIPI . We previously showed that higher levels of cytoplasmic p21WAFI/CIPI in breast cancers predicted reduced survival at 5 years. The present study examined HER-2 and p21WAFI/CIPI expression in a series of breast cancers with up to 9 years of follow-up, to evaluate whether in vitro findings were related to clinical data and the effect on outcome.MethodsThe CB11 anti-HER2 monoclonal antibody and the DAKO Envision Plus system were used to evaluate HER-2 expression in 73 patients. p21WAFI/CIPI staining was performed as described previously using the mouse monoclonal antibody Ab-1 (Calbiochem, Cambridge, MA, USA).ResultsHER-2 was evaluable in 67 patients and was expressed in 19% of cases, predicting reduced overall survival (P = 0.02) and reduced relapse-free survival (P = 0.004; Cox regression model). HER-2-positive tumours showed proportionately higher cytoplasmic p21WAFI/CIPI staining using an intensity distribution score (median, 95) compared with HER-2-negative cancers (median, 47) (P = 0.005). There was a much weaker association between nuclear p21WAFI/CIPI and HER-2 expression (P = 0.05), suggesting an inverse relationship between nuclear p21WAF1/CIP1 and HER-2.ConclusionThis study highlights a new pathway by which HER-2 may modify cancer behaviour. HER-2 as a predictor of poor prognosis may partly relate to its ability to influence the relocalisation of p21WAFI/CIPI from the nucleus to the cytoplasm, resulting in a loss of p21WAFI/CIPI tumour suppressor functions. Cytoplasmic p21WAFI/CIPI may be a surrogate marker of functional HER-2 in vivo.

Randomised trial of educational visits to enhance use of systematic reviews in 25 obstetric units.

Abstract Objective To evaluate the effectiveness of an educational visit to help obstetricians and midwives select and use evidence from a Cochrane database containing 600 systematic reviews. Design Randomised single blind controlled trial with obstetric units allocated to an educational visit or control group. Setting 25 of the 26 district general obstetric units in two former NHS regions. Subjects The senior obstetrician and midwife from each intervention unit participated in educational visits. Clinical practices of all staff were assessed in 4508 pregnancies. Intervention Single informal educational visit by a respected obstetrician including discussion of evidence based obstetrics, guidance on implementation, and donation of Cochrane database and other materials. Main outcome measures Rates of perineal suturing with polyglycolic acid, ventouse delivery, prophylactic antibiotics in caesarean section, and steroids in preterm delivery, before and 9 months after visits, and concordance of guidelines with review evidence for same marker practices before and after visits. Results Rates varied greatly, but the overall baseline mean of 43% (986/2312) increased to 54% (1189/2196) 9 months later. Rates of ventouse delivery increased significantly in intervention units but not in control units; there was no difference between the two types of units in uptake of other practices. Pooling rates from all 25 units, use of antibiotics in caesarean section and use of polyglycolic acid sutures increased significantly over the period, but use of steroids in preterm delivery was unchanged. Labour ward guidelines seldom agreed with evidence at baseline; this hardly improved after visits. Educational visits cost 860