Search

This article considers the utility of the bounded cumulative hazard model in cure rate estimation, which is an appealing alternative to the widely used two-component mixture model. This approach has the following distinct advantages: (1) It allows for a natural way to extend the proportional hazards regression model, leading to a wide class of extended hazard regression models. (2) In some settings the model can be interpreted in terms of biologically meaningful parameters. (3) The model structure is particularly suitable for semiparametric and Bayesian methods of statistical inference. Notwithstanding the fact that the model has been around for less than a decade, a large body of theoretical results and applications has been reported to date. This review article is intended to give a big picture of these modeling techniques and associated statistical problems. These issues are discussed in the context of survival data in cancer.

This study presents a comparison of the outcomes of patients with epithelial ovarian cancer who did and did not receive care from a gynecologic oncologist. The medical records of 848 women, identified from the Utah CancerRegistry, who were diagnosed with epithelial ovarian cancer between 1992 and 1998 were reviewed for relevant clinical and social information. Nearly 80% of the women lived in one of the four metropolitan areas in Utah. A gynecologic oncologist was involved in the diagnosis or treatment of 39% of the women. This percentage increased over the 6 years of the study, from 33% in the first year to 47.5% in the final year. Women between the ages of 40 and 60 years and urban-dwelling women were more likely to have seen a gynecologic oncologist during the course of their disease (P

In cancer studies the disease natural history process is often observed only at a fixed, random point of diagnosis (a survival time), leading to a current status observation (Sun (2006). The statistical analysis of interval-censored failure time data. Berlin: Springer.) representing a surrogate (a mark) (Jacobsen (2006). Point process theory and applications: marked point and piecewise deterministic processes. Basel: Birkhauser.) attached to the observed survival time. Examples include time to recurrence and stage (local vs. metastatic). We study a simple model that provides insights into the relationship between the observed marked endpoint and the latent disease natural history leading to it. A semiparametric regression model is developed to assess the covariate effects on the observed marked endpoint explained by a latent disease process. The proposed semiparametric regression model can be represented as a transformation model in terms of mark-specific hazards, induced by a process-based mixed effect. Large-sample properties of the proposed estimators are established. The methodology is illustrated by Monte Carlo simulation studies, and an application to a randomized clinical trial of adjuvant therapy for breast cancer.

The shape of the hazard function is of great interest in studies of the efficacy of cancer treatment and post-treatment cancer surveillance. We present estimates of the hazard rates obtained from data on survival of patients with cervical cancer and discuss associated methodological problems. Our study was carried out on survival data for 1826 women with cancer of the cervix uteri stratified by clinical stage and tumor growth pattern. We used nonparametric and various smoothing techniques for estimating the hazard function from the data; these were a nonparametric estimator based on the Nelson-Aalen method and its kernel counterparts, the kernel local likelihood estimator with a data-adaptive bandwidth, and a parametric estimator specifically designed for two-component hazards. For all categories of patients, the estimated hazard functions pass through a clear-cut maximum, tending to zero as the follow-up time becomes sufficiently long. In one stratum of patients we observed a bimodal shape of the hazard function. There are two alternative models that provide equally plausible explanations of this observation; one of them attributes the observed pattern of the hazard function to a certain heterogeneity of tumor cell population, while the competing model refers to a heterogeneity of the subsample of patients under study. Providing the probability of cure is high, as is the case in our setting, there is no way to discriminate between the two models on the basis of survival data.

This paper considers the utility of a stochastic model of carcinogenesis, proposed earlier by Yakovlev and Polig, in quantitative analysis of the incidence of radiation-induced osteosarcomas in beagels injected with various amounts of ^2^3^9Pu. The original version of the model failed to provide a good fit to our experimental data. The model has been generalized by incorporating a simple mechanism of lesion elimination, which is likely to be mediated by the immune system. Two versions of the model were developed: the first version ( Model 1 ) assumed malignant cells to be a target for the immune attack, while in Model 2 initiated cells were assumed to be such a target. Model 2 was rejected by the likelihood ratio test, thereby indicating that the competing model provides a more plausible explanation of the experimental data. Since in experiments with incorporated radionuclides the dose rate varies with time, dose-rate effects cannot be observed directly, and one must rely on mathematical models. The results of our numerical experiments show that, depending on the time of observation, both the direct and the inverse dose-rate effects may manifest themselves even at a fixed total dose level.

In this study, we use a combination of smoothing nonparametric and parametric methods to estimate the hazard function for the disease-free interval from clinical data on patients with carcinoma of the cervix uteri. Three estimation procedures were used for estimating the hazard rate from the data: the life table estimator, a kernel counterpart of the Nelson-Aalen estimator, and a parametric estimator based on a mechanistic model of tumor recurrence. The parametric estimate appears to be in good agreement with its nonparametric counterparts suggesting a bimodal shape of the underlying hazard function. The observed two-component structure of hazard function can be attributed to a certain heterogeneity of ''malignant potential'' remaining in treated tumors. Exophytic and endophytic carcinomas of the cervix uteri show dissimilar patterns of this heterogeneity and its evolution with the time elapsed after treatment. The results of parametric inference suggest that exophytic and endophytic recurrent tumors differ as to quantitative characteristics of a slowly evolving subpopulation of clonogenic tumor cells.

The paper explores methodological and mathematical aspects of a new approach to constructing optimal schedules of cancer screening. This approach consists of systematic use of tumor size at detection, combining stochastic models of tumor latency, tumor growth and tumor detection. and employing a new biologically natural screening efficiency criterion defined as the Kantorovich distance between the tumor size at spontaneous detection in the absence of screening and the tumor size at detection when both spontaneous and screening based mechanisms are in place. An explicit formula for the efficiency functional is obtained. Sample calculations suggests that in the case of exponential tumor growth, the optimal screening schedules with a fixed number of exams have a trend to uniformity.

Inference based on the proportional hazards model is discussed in the presence of long-term survivors. The model is formulated as a cure model yielding an improper survivor function. An algorithm is proposed to fit the proportional hazards model restricted by the fixed survival rates at the end of observation period. A parametric cure model is used to estimate the proportion of long-term survivors. To combine the stability of the parametric method with the flexibility of the nonparametric one, the survival function is estimated nonparametrically conditional on the cure rates provided by the parametric analysis. The methods are applied to the data collected in clinical trials on Hodgkin's disease.

Yakovlev and Polig (1996) developed a mechanistically motivated stochastic model of radiation carcinogenesis allowing for cell death. The key feature of the model is that it allows for radiation-induced cell killing to compete with the process of tumor promotion. This model describes and explains a wide range of experimental findings documented in the radiobiological literature, including the inverse dose-rate effect and radiation hormesis. The model has successfully been applied to various sets of experimental and epidemiological data to gain quantitative insight into the processes of tumorigenesis induced by radiation and chemical carcinogens. In this paper, we discuss the most recent application of the Yakovlev-Polig model to the analysis of epidemiological data on the mortality caused by radiation-induced leukemia (all types) among the atomic bomb survivors (Hiroshima and Nagasaki). Nonparametric estimates of the hazard function for leukemia latency time were obtained for three different dose groups i...

Background and Objectives Research suggests that the addition of dexmedetomidine to local anesthetics can prolong peripheral nerve blocks; however, clinical safety data are limited, and interscalene blocks have not been studied. The present study was designed to test the hypothesis that dexmedetomidine added to ropivacaine would safely enhance the duration of analgesia without adverse effects when compared with ropivacaine alone. Methods We conducted a single-center, prospective, randomized, triple-blind, controlled trial of 62 patients undergoing elective shoulder surgery under general anesthesia with an interscalene block. Patients underwent ultrasound-guided interscalene blocks using either 12 mL of 0.5% ropivacaine or 0.5% ropivacaine plus 150-µg dexmedetomidine. The primary outcomes were self-reported duration of the nerve block and safety assessment (adverse effects and neurological sequelae). Data were analyzed in a blinded fashion. Results The median duration of the nerve block was 18 hours (95% confidence interval, 18–20) in the dexmedetomidine group and 14 hours (95% confidence interval, 14–16) in the ropivacaine group (P = 0.0001). Dexmedetomidine also lowered pain scores for the first 14 hours postoperatively and significantly hastened the time to sensory (P = 0.04) and motor (P = 0.002) block onset. Dexmedetomidine lowered heart rate but blood pressures were stable. Plasma levels of ropivacaine were not different between groups, and plasma dexmedetomidine levels were relatively low. There were no adverse events or neurological sequelae. Conclusions Dexmedetomidine added to ropivacaine for interscalene blocks increased the duration of the nerve block and improved postoperative pain. These additional efficacy and safety data should encourage further study of peripheral perineural dexmedetomidine in humans.

BACKGROUND The question of curability of breast carcinoma remains controversial. Because the probability of cure essentially is an asymptotic notion, the corresponding estimation problems call for special statistical methods. Such methods should account for an intimate connection between the probability of cure and the shape of the hazard function. METHODS The study was performed on survival data for 13,166 women with breast carcinoma identified through the Utah Cancer Registry and stratified by clinical stage and age at diagnosis. For these patients, the follow-up period was 30 years. Three estimation procedures were used for estimating the hazard function from the data: the life table estimator, a kernel counterpart of the Nelson-Aalen estimator, and a parametric estimator specifically designed for two-component hazards. The parametric estimate of the hazard function was used to provide estimates of cure rates for each category of patients. RESULTS For all categories of patients under study, the estimated hazard functions passed through a clear-cut maximum, showing a tendency to decrease as time approached the end of a follow-up period. The hazards appeared to be nonproportional across the strata. The estimated values of the cure rate and the corresponding confidence intervals were determined for each stratum of patients with breast carcinoma. CONCLUSIONS The results of the current study strongly suggest that cure is a possible outcome of breast carcinoma treatment. The condition of proportionality of risks is not met in breast carcinoma survival data. Cancer 1999;85:1789–98. © 1999 American Cancer Society.

Asthma, a family of airway disorders characterized by airway inflammation, has an increasing incidence worldwide. Platelet-activating factor (PAF) may play a role in the pathophysiology of asthma. Its proinflammatory actions are antagonized by PAF acetylhydrolase. A missense mutation (V279F) in the PAF acetylhydrolase gene results in the complete loss of activity, which occurs in 4% of the Japanese population. We asked if PAF acetylhydrolase deficiency correlates with the incidence and severity of asthma in Japan. We found that the prevalence of PAF acetylhydrolase deficiency is higher in Japanese asthmatics than healthy subjects and that the severity of this syndrome is highest in homozygous-deficient subjects. We conclude that the PAF acetylhydrolase gene is a modulating locus for the severity of asthma.

Evaluation of the treatment effect on cytogenetic ordered categorical response is considered in patients treated for chronic myelogenous leukaemia (CML) in a clinical trial initiated by the East German Group for Hematology and Oncology. A simulation model for the cytogenetic response (per cent of Philadelphia chromosome positive metaphases) serially measured in CM L patients was constructed to describe roughly the sparse information available in medical literature. The model was used to construct a summary measure of response and to formulate the treatment effect as a regression with U-shape distributed ordered categorical data. Two simple models (vertical shift model and pooled conditional response model) were specifically designed to model the treatment effect 'observed' in a simulated 'pilot' data set. The powers were contrasted with the traditional proportional odds and binary models. The comparison was based both on repeated sampling from the simulated model and on bootstrap of 'given' pilot data set. We show that the specific models that address the treatment effect directly (as anticipated from pilot data) can gain in power as compared to the traditional proportional odds model when evaluated by bootstrap. However, the proportional odds model appears to be better with repeated sampling from the simulation model. To explain this discrepancy we generated 'pilot data sets' repeatedly from the simulation model and showed that the ordering of the bootstrap power estimates is unstable with reasonably complex models dependent on the random fall of the pilot data sets. This phenomenon clearly limits the usefulness of subtle modelling the form of the treatment difference observed in a small pilot data set.

Nonparametric estimators of survivor function can be used in the presence of long-term survivors to estimate the probability of cure by the last point of a survivor curve. To explain the presence of long-term survivors the cumulative hazard is assumed bounded but otherwise unspecified to yield an improper survivor function. Within this formulation the PH model is asymptotically fully efficient as compared to a parametric model under certain conditions when estimating the ratio of cure rates. However, this is not the case if estimation of cure rates in absolute terms is of interest. The asymptotic relative efficiency tends to zero as the probability of cure decreases. It also decreases with increase of the coefficient of variation of covariates approaching the value computed under homogeneity. Partial likelihood is shown to be inefficient if the baseline hazard is completely known.

A parametric model is used to investigate the latency time of leukaemia observed in patients treated for Hodgkin's disease. In specifying the treatment effect on leukaemia-free survival, account was taken of a fraction of long-term survivors and of time-changing risk associated with the relapse of the primary disease. The model is applied to data collected in the International Database on Hodgkin's Disease. It permits estimation of the contributions of primary and of relapse treatment to the overall risk of induced leukaemia. Baseline hazards appear to be identical after primary and relapse treatments supporting the concept that induced leukaemia have common origin. The probability to induce leukaemia by MOPP chemotherapy is the same, regardless whether used as primary or relapse treatment.

Parametric models are used to investigate the time of leukemia latency in patients treated for Hodgkin's disease. A model of induced carcinogenesis provides a stable projection of the leukemia risk for times beyond the follow-up period. The model is applied to data collected in the International Data Base on Hodgkin's Disease. It permits to estimate the contributions of primary and of relapse treatment to the overall risk of induced leukemia. Latency periods appear to be identical after both treatments supporting the concept of induced leukemogenesis.

Computer simulations have been conducted to provide a realistic model of tumor recurrence in a cancer patient, following treatment. The simulation model incorporates description of the temporal organization of various biological processes underlying tumor development at the cellular level: proliferation, differentiation, death of tumor cells, growth control in neoplastic tissues along with the tumor treatment effect. The prime object of our concern is whether the simple parametric model of tumor recurrence proposed by Hoang et al. [6] allows estimation of actual value of the tumor growth potential. A good fit has been demonstrated of the parametric model when applied to the samples of simulated tumor recurrence times as well as to real data samples of tumor recurrence in breast cancer patients with various regimen of radiotherapy.

The current gold standard for diagnosis of hepatic fibrosis and cirrhosis is the traditional invasive liver biopsy. It is desirable to assess hepatic fibrosis with noninvasive means. Targeted proteomic techniques allow an unbiased assessment of proteins and might be useful to identify proteins related to hepatic fibrosis. We utilized selected reaction monitoring (SRM) targeted proteomics combined with an organ-specific blood protein strategy to identify and quantify 38 liver-specific proteins. A combination of protein C and retinol-binding protein 4 in serum gave promising preliminary results as candidate biomarkers to distinguish patients at different stages of hepatic fibrosis due to chronic infection with hepatitis C virus (HCV). Also, alpha-1-B glycoprotein, complement factor H and insulin-like growth factor binding protein acid labile subunit performed well in distinguishing patients from healthy controls.

Rationale: To compare cancers detected during use of 5α-reductase inhibitors (5αRI) with cancers detected in untreated controls stratified for tumor size. Methods: Prostate biopsies were performed on 235 consecutive patients “for cause” (elevated or rising PSA, positive digital rectal examination, or focal hypoechoic lesion). Fifty patients were excluded for a prior diagnosis of cancer, leaving 185 as the study group (5αRI=41, control=144). Patients in the 5αRI group had been treated for a mean of 3.5 years. Cancer was ultimately diagnosed in 114/185 patients. Results: Cancer was diagnosed in 31/41 (76%) of patients treated with 5αRI and 83/144 (58%) of the control group (p=0.04). Control tumors were larger (14.3 mm) than those in 5αRI treated patients (9.4 mm, p=0.0007). No differences in mean PSA or PSA kinetics were detected between groups. For tumors less than 1.0 cm, the proportion of high grade cancers (Gleason 7-10 and Gleason 4+3-10) was higher in 5αRI subjects than in controls (p

BACKGROUND: Prostate cancer mortality rates in the United States declined by >40% between 1991 and 2005. The impact of changes in primary treatment and adjuvant and neoadjuvant hormone therapy on this decline is unknown. METHODS: The authors applied 3 independently developed models of prostate cancer natural history and disease detection under common assumptions about treatment patterns, treatment efficacy, and survival in the population. Primary treatment patterns were derived from the Surveillance, Epidemiology, and End Results registry; data on the frequency of hormone therapy were obtained from the CaPSURE (Cancer of the Prostate Strategic Urologic Research Endeavor) database; and treatment efficacy was based on estimates from randomized trials and comparative effectiveness studies of treatment alternatives. The models projected prostate cancer mortality without prostate-specific antigen screening and in the presence and absence of treatment benefit. The impact of primary treatment was expressed as a fraction of the difference between observed mortality and projected mortality in the absence of treatment benefit. RESULTS: The 3 models projected that changes in treatment explained 22% to 33% of the mortality decline by 2005. These contributions were accounted for mostly by surgery and radiation therapy, which increased in frequency until the 1990s, whereas hormone therapies contributed little to the mortality decline by 2005. Assuming that treatment benefit was less for older men, changes in treatment explained only 16% to 23% of the mortality decline by 2005. CONCLUSIONS: Changes in primary treatment explained a minority of the observed decline in prostate cancer mortality. The remainder of the decline probably was because of other interventions, such as prostate-specific antigen screening and advances in the treatment of recurrent and progressive disease. Cancer 2012;118:5955-63. V C 2012 American Cancer Society.

In order to describe the life-prolonging effect of some agents that are harmful at higher doses, ionizing radiations in particular, a stochastic model is developed in terms of accumulation and progression of intracellular lesions caused by the environment and by the agent itself. The processes of lesion repair, operating at the molecular and cellular level, are assumed to be responsible for this hormesis effect within the framework of the proposed model. Properties of lifetime distributions, derived for analysis of animal experiments with prolonged and acute irradiation, are given special attention. The model provides efficient means of interpreting experimental findings, as evidenced by its application to analysis of some published data on the hormetic effects of prolonged irradiation and of procaine on animal longevity.

SummaryIn semiparametric models, the dimension d of the maximum likelihood problem is potentially unlimited. Conventional estimation methods generally behave like O(d3). A new O(d) estimation procedure is proposed for a large class of semiparametric models. Potentially unlimited dimension is handled in a numerically efficient way through a Nelson–Aalen-like estimator. Discussion of the new method is put in the context of recently developed minorization–maximization algorithms based on surrogate objective functions. The procedure for semiparametric models is used to demonstrate three methods to construct a surrogate objective function: using the difference of two concave functions, the EM way and the new quasi-EM (QEM) approach. The QEM approach is based on a generalization of the EM-like construction of the surrogate objective function so it does not depend on the missing data representation of the model. Like the EM algorithm, the QEM method has a dual interpretation, a result of merging the idea of surrogate maximization with the idea of imputation and self-consistency. The new approach is compared with other possible approaches by using simulations and analysis of real data. The proportional odds model is used as an example throughout the paper.

Several authors have indicated that incorrectly classified cause of death for prostate cancer survivors may have played a role in the observed recent peak and decline of prostate cancer mortality. Motivated by the suggestion we studied a competing risks model where other cause of death may be misattributed as a death of interest. We first consider a naive approach using unconstrained nonparametric maximum likelihood estimation (NPMLE), and then present the constrained NPMLE where the survival function is forced to be monotonic. Surprising observations were made as we studied their small-sample and asymptotic properties in continuous and discrete situations. Contrary to the common belief that the non-monotonicity of a survival function NPMLE is a small-sample problem, the constrained NPMLE is asymptotically biased in the continuous setting. Other isotonic approaches, the supremum (SUP) method and the Pooled-Adjacent-Violators (PAV) algorithm, and the EM algorithm are also considered. We found that the EM algorithm is equivalent to the constrained NPMLE. Both SUP method and PAV algorithm deliver consistent and asymptotically unbiased estimator. All methods behave well asymptotically in the discrete time setting. Data from the Surveillance, Epidemiology and End Results (SEER) database are used to illustrate the proposed estimators.

The computation in the multinomial logit mixed effects model is costly especially when the response variable has a large number of categories, since it involves high-dimensional integration and maximization. Tsodikov and Chefo (2008) developed a stable MLE approach to problems with independent observations, based on generalized self-consistency and quasi-EM algorithm developed in Tsodikov (2003). In this paper, we apply the idea to clustered multinomial response to simplify the maximization step. The method transforms the complex multinomial likelihood to Poisson-type likelihood and hence allows for the estimates to be obtained iteratively solving a set of independent low-dimensional problems. The methodology is applied to real data and studied by simulations. While maximization is simplified, numerical integration remains the dominant challenge to computational efficiency.

Hepatocellular carcinoma (HCC) is a major health problem worldwide, including Egypt. In the recent past, HCC has become the second most prevalent cancer among men in Egypt. Since HCC has not been well studied in the rural population of Egypt, this case-control study was conducted to investigate the epidemiologic risk factors of HCC in the predominantly rural region of Gharbiah, Egypt.A total of 150 cases and 150 controls matched to cases on age (±5 years) and sex were recruited from the Gharbiah Cancer Society and Tanta Cancer Center. Exposure data were collected by an interviewer-administered standardized questionnaire about epidemiologic, occupational, medical and family history of HCC. Conditional logistic regression was utilized to calculate unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (CI). The effect modification of HCC risk between viral infection and environmental risk factors was also assessed.Being an industrial worker was an independent risk for developing HCC (OR 3.54, 95% CI 1.18, 10.63) after adjusting for viral infection, schistosomiasis and tobacco smoking. High relative risk of HCC was observed among HCV-infected individuals who were farmers (OR 9.60, 95% CI 3.72, 24.76), industrial workers (OR 12.90, 95% CI 4.33, 38.43) or active smokers (OR 5.95, 95% CI 2.20, 16.08).Occupational exposure may play an important role in the development of HCC. Farming, industrial exposures and cigarette smoking may increase the risk of HCC among HCV-seropositive individuals. Future research focusing on mechanisms of occupational exposures among HCV patients in this population is needed.

In bladder cancer, clinical grade and stage fail to capture outcome. We developed a clinically applicable quantitative PCR (QPCR) gene signature to predict progression in non–muscle-invasive bladder cancer. Comparative metaprofiling of 12 DNA microarray data sets (comprising 631 samples and 241,298 probe sets) identified 96 genes, which showed differential expression in seven clinical outcome categories, or were identified as outliers, historic markers, or housekeeping genes. QPCR was done to determine mRNA expression from 96 bladder tumors. Fifty-seven genes differentiated T2 from non-T2 tumors (P < 0.05). Principal components analysis and Cox regression models were used to predict probability of T2 progression for non-T2 patients, placing them into high- and low-risk groups based on their gene expression. At 2 years, high-risk patients exhibited greater T2 progression (45% for high-risk patients versus 12% for low-risk patients; P = 0.003, log-rank test). This difference remained significant within T1 tumors (61% for high-risk patients versus 22% for low-risk patients; P = 0.02) and Ta tumors (29% for high-risk patients versus 0% for low-risk patients; P = 0.03). The best multivariate Cox model included stage and gender, and this signature provided predictive improvement over both (P = 0.002, likelihood ratio test). Immunohistochemistry was done for two genes in the signature not previously described in bladder cancer, ACTN1 and CDC25B, corroborating their up-regulation at the protein level with disease progression. Thus, we identified a 57-gene QPCR panel to help predict progression of non–muscle-invasive bladder cancers and delineate a systematic, generalizable approach to converting microarray data into a multiplex assay for cancer progression. [Cancer Res 2009;69(9):3810–8]

Background The time by which prostate-specific antigen (PSA) screening advances prostate cancer diagnosis, called the lead time, has been reported by several studies, but results have varied widely, with mean lead times ranging from 3 to 12 years. A quantity that is closely linked with the lead time is the overdiagnosis frequency, which is the fraction of screen-detected cancers that would not have been diagnosed in the absence of screening. Reported overdiagnosis estimates have also been variable, ranging from 25% to greater than 80% of screen-detected cancers. Methods We used three independently developed mathematical models of prostate cancer progression and detection that were calibrated to incidence data from the Surveillance, Epidemiology, and End Results program to estimate lead times and the fraction of overdiagnosed cancers due to PSA screening among US men aged 54-80 years in 1985-2000. Lead times were estimated by use of three definitions. We also compared US and earlier estimates from the Rotterdam section of the European Randomized Study of Screening for Prostate Cancer (ERSPC) that were calculated by use of a microsimulation screening analysis (MISCAN) model. Results The models yielded similar estimates for each definition of lead time, but estimates differed across definitions. Among screen-detected cancers that would have been diagnosed in the patients' lifetimes, the estimated mean lead time ranged from 5.4 to 6.9 years across models, and overdiagnosis ranged from 23% to 42% of all screen-detected cancers. The original MISCAN model fitted to ERSPC Rotterdam data predicted a mean lead time of 7.9 years and an overdiagnosis estimate of 66%; in the model that was calibrated to the US data, these were 6.9 years and 42%, respectively. Conclusion The precise definition and the population used to estimate lead time and overdiagnosis can be important drivers of study results and should be clearly specified.

On February 27, 2008 the world's scientific community lost one of its most brilliant and innovative members – Dr. Andrei Yakovlev. He was 63, and at the zenith of his intellectual powers. With his uncanny ability to persuade and to ignite interest in his ideas, he challenged and stimulated everyone around him through his keen scientific vision and unparalleled intellectual depth and rigor. He collaborated with dozens of researchers – mathematicians, statisticians, biologists, epidemiologists, medical doctors and engineers. For many of them, these collaborations were a life changing experience. His unflinching belief in the power of mathematics and his tireless pursuit of new biomedical knowledge were irresistibly infectious and inspiring to two generations of his students and hundreds of his colleagues. For more than three decades he passionately advocated his vision for the future of Biology and Medicine as disciplines built upon a solid foundation of mathematical theory and statistical methodology. He has always enjoyed and cherished those wonderful "moments of truth" when predictions based on his models were confirmed by biological experiments or provided a perfect fit to known biomedical data, and he shared his joy with his co-workers, students and friends. He was a natural leader, caring mentor, and paternal figure for everyone privileged to work under his tutelage. He led by his wisdom, honesty, dignity and championship for scientific truth. He was a towering moral authority for many of his friends and coworkers, always ready to help, offer advice and even, for better or for worse, make critical decisions for them. Andrei Yakovlev was born on November 30, 1944 in Leningrad (now St. Petersburg), Russia. The period of his adolescence and youth fell upon the era of the post-WWII revival of mathematics and natural sciences in the Soviet Union. The combination of good education, professional opportunity and relative freedom from the oppressive communist ideology attracted thousands of talented and daring young people to mathematics, theoretical and experimental physics, chemistry, biology and medicine. Andrei took full advantage of this unique historical chance and received an M.D. from the 1st Medical Institute in Leningrad (1967), a Ph.D. in cell biology from the Pavlov Institute of Physiology of the Academy of Sciences of the USSR (1973) and a Doctor of Science degree (which is the highest scientific degree in Russia) in Mathematics from Moscow State University (1981). With such a formidable background he was uniquely positioned to transcend the boundaries of the four different sciences, cultures and intellectual worlds of Mathematics, Statistics, Biology and Medicine. Dr. Yakovlev made definitive and lasting contributions to a stunning number of fields: cell biology; cell population dynamics; branching stochastic processes; stochastic models in radiation biology; optimization of radiation cancer treatment; carcinogenic risk assessment; stochastic models of carcinogenesis, cancer growth, progression and detection; optimal schedules of cancer surveillance and screening; survival analysis; cancer epidemiology; statistical inference from microarray gene expression data; genetic regulatory networks; biostatistics; bioinformatics; and computational biology. His pioneering ideas were embodied in four research monographs and about 200 papers. He was the principal investigator on 12 major research grants, and since 1985 he gave more than 150 invited talks in the US and Europe alone. Yet he thought that his main tasks lay ahead of him as more and more groundbreaking ideas were taking shape in his mind, waiting to be tested, elaborated and implemented. Dr. Yakovlev had an incredible intuition as to which biomedical problems are amenable to mathematical approaches and what is the right mathematical model or statistical test to employ. As one example, following A.N. Kolmogorov's groundbreaking works of the 1930–1940s on branching stochastic processes, Dr. Yakovlev made this approach a powerful tool in the study of cell proliferation kinetics and produced new results on stem cell differentiation that were theretofore too difficult even to contemplate. Furthermore, his work led to formulation of new concepts and discovery of new problems and directions of research in the mathematical theory of branching stochastic processes. As yet another example of the interdisciplinary impact of his ideas, his contributions to biomathematical modeling of carcinogenesis and tumor recurrence have sparked the development of new frequentist and Bayesian biostatistical methodology in cure models based on their hierarchical mixture origin, a field of biostatistics of general significance far beyond the initial focus on cancer. Over the last seven years of his life Dr. Yakovlev developed a strong interest in microarray data analysis. In this very active area of research he designed a novel statistical methodology that allowed him to uncover unknown patterns of gene expression and interaction of great biological importance. In particular, this led him to the discovery of the major impact that the correlation structure of gene expression data has on the stability of multiple testing procedures – a fact that has never been noticed before. His scientific legacy does not consist only of ideas, methods and results. Most importantly, he left behind scientific principles and research paradigms that will guide his co-workers and students for decades to come. He forged them through incisive analysis, trial and error, and excruciating, sometimes heated and fierce, debates with his colleagues and disciples. He has never articulated them explicitly, perhaps leaving this endeavor for a future time, but followed them religiously in his own work. Dr. Yakovlev has held numerous leading positions throughout his distinguished scientific career. In 1978 he founded the department of Biomathematics in the Central Research Institute of Roentgenology and Radiology (Leningrad, Russia) and headed it until 1988. From 1988–1992 he chaired the Department of Applied Mathematics of Leningrad (later St. Petersburg) State Polytechnical University. In 1992–96 he held visiting professorships at several universities and research institutions in the US, France, Germany and Australia. In 1996 he was appointed to the position of Director of Biostatistics at the Huntsman Cancer Institute of the University of Utah and settled in Salt Lake City. His last academic post was at the University of Rochester Medical Center, which recruited him as a founding chairman of the Department of Biostatistics and Computational Biology, a labor of love for the last six years of his life. As a result of his tireless effort, the department has seen a three-fold expansion and a six-fold increase in external research funding, which placed it among the world's top departments in the field. Dr. Yakovlev's academic life spanned several countries (Russia, Germany, France and the US) leaving in his wake numerous admirers, disciples, followers and co-workers as well as those apprehensive of his fierce pursuit of scientific truth, relentless questioning of conventional wisdom, zero tolerance for junk science, and little patience for scientific orthodoxy. Many of his ideas, especially in the fields of biostatistics, cancer epidemiology, cell biology and microarray data analysis, appeared decades ahead of their time. Due to their revolutionary nature, mathematical sophistication and broad interdisciplinary scope, they were found inaccessible by some of his peers. With the system of anonymous peer refereeing currently in place, which Dr. Yakovlev considered fundamentally broken, some of his best and most innovative papers and grant proposals were undervalued by narrowly specialized reviewers. That is why Dr. Yakovlev embraced with great enthusiasm the new framework of open peer reviewed publication pioneered by BMC-Biology Direct and became the founding editor of its Mathematical Biology section. It is comforting to us that his exceptional research accomplishments were recognized by many prominent national and international organizations. He was a recipient of the Alexander von Humboldt Award (1994) and a John Simon Guggenheim Fellowship (1999) and was named elected member of six academies and societies. In recognition of his important contributions to statistics, he was named Honorary Fellow of the Institute of Mathematical Statistics (1998), Fellow of the American Statistical Association (2000) and Elected Member of the International Statistical Institute (2002). Among many honors bestowed on Dr. Yakovlev are Honorary Doctor of Science of Idaho State University (2002), Distinguished Scholarly and Creative Research Award of the University of Utah (2002) and the title of Distinguished Lecturer of Creighton University (2006). We believe that with the passage of time the broad scientific community will come to realize that in Dr. Yakovlev (figure ​(figure1)1) we have lost one of the greatest scientific visionaries of our time. It is heart wrenching that he did not live long enough to witness this recognition. Figure 1 A recent photograph of Dr. Yakovlev. He is survived by his wife Nina and their 11 year old son Yuri.

Objective. The emergency department (ED) can be a particularly challenging environment in which to offer care for chronic pain. This study tried to determine if beliefs held by patients and providers about noncancer-related chronic pain affect evaluation and management of pain in ED. Intervention. We surveyed 103 patients presenting to the ED with chronic pain, 34 ED physicians, and 44 ED nurses to assess the influence of 15 possible barriers to managing chronic pain in the ED. Results. Patients were significantly more likely than providers to believe that their pain had to have a diagnosed physical component to be treated. Providers were significantly more likely than patients to believe that patients came to the ED because they lacked a primary care physician. All agreed that chronic pain treatment was not a priority in the ED and the potential for addiction, dependence, diversion, and forged prescriptions was low. Conclusions. Patients in chronic pain may need to be reassured that their pain will be treated, even in the absence of objective signs or magnified symptoms. Providers may wrongly believe that lack of a primary care physician brings these patients to the ED. Providers and patients appear to believe that treating chronic pain in the ED has a low priority. Both groups may underestimate the problems inherent with prescribing opioids in this setting.

To determine if pain perception is affected by an extreme bout of exercise that causes ongoing exercise-related pain.Repeated-measures design.Pre-race registration area and finish area of an endurance race.Twenty-one competitors in the 2005 Western States 100 Mile Endurance Run and 11 control subjects who were assisting at the race but not running.Not applicable.Overall pain and pain ratings on a pressure pain test before and after the event.Mean overall pain +/- standard deviation on a 100-mm scale increased (P.05) from 3+/-6mm before the run to 39+/-28mm after the run among the runners. The faster runners showed a mean reduction (P.05) in pain ratings after the race of 15+/-20mm (on a 100-mm scale), whereas there was no change for the slower runners and controls. Findings were confirmed by model-based analysis.The faster runners in a 100-mile (161-km) running race experience a modest temporary reduction in pressure pain perception that does not appear to be augmented by ongoing pain related to the exercise. The lack of a reduction in pain perception among the slower runners may be because an extreme bout of exercise of this nature can "exhaust" the systems responsible for exercise-induced analgesia in all but the most well-trained of runners, or that these systems were not activated because the slower runners were unable to maintain a high enough exercise intensity during the later stages of the race.

The Food and Drug Administration (FDA), Substance Abuse and Mental Health Services Administration (SAMHSA), and the National Institute for Drug Abuse (NIDA) report that no sound scientific studies support the medicinal use of cannabis. Despite this lack of scientific validation, many patients routinely use "medical marijuana," and in many cases this use is for pain related to nerve injury. We conducted a double-blinded, placebo-controlled, crossover study evaluating the analgesic efficacy of smoking cannabis for neuropathic pain. Thirty-eight patients with central and peripheral neuropathic pain underwent a standardized procedure for smoking either high-dose (7%), low-dose (3.5%), or placebo cannabis. In addition to the primary outcome of pain intensity, secondary outcome measures included evoked pain using heat-pain threshold, sensitivity to light touch, psychoactive side effects, and neuropsychological performance. A mixed linear model demonstrated an analgesic response to smoking cannabis. No effect on evoked pain was seen. Psychoactive effects were minimal and well-tolerated, with some acute cognitive effects, particularly with memory, at higher doses. Perspective This study adds to a growing body of evidence that cannabis may be effective at ameliorating neuropathic pain, and may be an alternative for patients who do not respond to, or cannot tolerate, other drugs. However, the use of marijuana as medicine may be limited by its method of administration (smoking) and modest acute cognitive effects, particularly at higher doses.

Insulin resistance, infertility, and hirsutism, common characteristics of polycystic ovary syndrome (PCOS), improve with even modest weight loss. Optimal dietary treatment for PCOS is not known.We compared the effects of acute protein administration with those of glucose challenges on hormones related to obesity and insulin resistance (ie, cortisol and insulin), hirsutism [ie, dehydroepiandosterone (DHEA) and androstenedione], and hunger (ie, ghrelin).Patients with PCOS (n = 28; aged 26 +/- 2 y) were tested with a 5-h oral-glucose-tolerance test (OGTT) and a euvolemic, euenergetic protein challenge.Glucose ingestion caused larger fluctuations in blood glucose and more hyperinsulinemia than did protein (P0.01, overall treatment-by-time interaction). During the protein challenge, cortisol and DHEA declined over 5 h. During OGTT, cortisol and DHEA increased after the third hour and began to show significant divergence from protein from the fourth hour (Por= 0.01). During OGTT, 18 patients who had a blood glucose nadir of69 mg/dL had elevated cortisol (baseline: 10.4 +/- 0.4; nadir: 5.9 +/- 0.1; peak: 12.7 +/- 0.9 microg/dL) and DHEA (baseline: 15.6 +/- 1.3; nadir: 11.2 +/- 1.0; peak: 24.6 +/- 1.6 ng/mL) (P0.01), whereas the remaining 10 patients with a glucose nadir of 76 +/- 2 mg/dL had no increase in adrenal steroids. Both glucose and protein suppressed ghrelin (from 935 +/- 57 to 777 +/- 51 pg/mL and from 948 +/- 60 to 816 +/- 61 pg/mL, respectively). After glucose ingestion, ghrelin returned to baseline by 4 h and increased to 1094 +/- 135 pg/mL at 5 h. After the protein challenge, ghrelin remained below the baseline (872 +/- 60 pg/mL) even at 5 h. The overall treatment effect was highly significant (P0.0001).Glucose ingestion caused significantly more hyperinsulinemia than did protein, and it stimulated cortisol and DHEA. Protein intake suppressed ghrelin significantly longer than did glucose, which suggested a prolonged satietogenic effect. These findings provide mechanistic support for increasing protein intake and restricting the simple sugar intake in a PCOS diet.

Hepatic steatosis is a recognized problem in patients after orthotopic liver transplant (OLT). However, de novo development of nonalcoholic fatty liver disease (NAFLD) has not been well described. The aim of this study was to determine the prevalence and predictors of de novo NAFLD after OLT. A retrospective analysis was performed on 68 OLT patients with donor liver biopsies and posttransplantation liver biopsies. Individual medical charts were reviewed for demographics, indication for OLT, serial histology reports, genotypes for hepatitis C, comorbid conditions, and medications. Liver biopsies were reviewed blindly and graded according to the Brunt Scoring System. Multivariate logistic regression analysis was used to study the risk factors for developing NAFLD. The interval time from OLT to subsequent follow-up liver biopsy was 28 ± 18 months. A total of 12 patients (18%) developed de novo NAFLD, and 6 (9%) developed de novo NASH. The regression model indicated that the use of angiotensin-converting enzyme inhibitors (ACE-I) was associated with a reduced risk of developing NAFLD after OLT (odds ratio, 0.09; 95% confidence interval, 0.010-0.92; P = 0.042). Increase in body mass index (BMI) of greater than 10% after OLT was associated with a higher risk of developing NAFLD (odds ratio, 19.38; 95% confidence interval, 3.50-107.40; P = 0.001). In conclusion, de novo NAFLD is common in the post-OLT setting, with a significant association with weight gain after transplant. The use of an ACE-I may reduce the risk of developing post-OLT NAFLD. Liver Transpl, 2006. © 2006 AASLD.

Estimates that smoking contributes 38-72% of the United States (US) Black male cancer death rate leave a wide range of uncertainty. This paper uses additional and regional data, and refined methods, to reassess that range.This study uses lung cancer rates as an exposure index, linear regression, age adjusted US 1950-2001 and US regional 1969-2001 death rates (rates), and the formula: smoking-attributable fraction (SAF)=(1-((rate in the unexposed) / (rate in the exposed))). Estimated lung cancer rates in the unexposed range between rates predicted for a population with no smoking-attributable lung cancers to rates seen in "nonsmokers."Lung cancer death rates predicted 99.9% and 99.8% of the variances in non-lung non-stomach cancer death rates from 1950-1980 and 1950-1988, respectively (each P0.0001). That suggests 2001 all-sites cancer death SAFs of 63% (sensitivity range 60-66%) nationally and from 43% in the Northeast to 67% in the South.Smoking may cause most premature cancer deaths and temporal and regional cancer death rate disparities in Black men.

The bounded cumulative hazard model is a generalization of the proportional hazard model with several distinct advantages: (1) It has superior flexibility, allowing a fit to data where the proportional hazards assumption does not apply; (2) It has a form that is suitable for semiparametric inference; and (3) It may offer an interpretation in terms of biologically meaningful parameters. In this paper the bounded cumulative hazard model is discussed. Several versions of the model are applied to breast cancer recurrence data from the Curie Institute.

The ultimate success of microarray technology in basic and applied biological sciences depends critically on the development of statistical methods for gene expression data analysis. The most widely used tests for differential expression of genes are essentially univariate. Such tests disregard the multidimensional structure of microarray data. Multivariate methods are needed to utilize the information hidden in gene interactions and hence to provide more powerful and biologically meaningful methods for finding subsets of differentially expressed genes. The objective of this paper is to develop methods of multidimensional search for biologically significant genes, considering expression signals as mutually dependent random variables. To attain these ends, we consider the utility of a pertinent distance between random vectors and its empirical counterpart constructed from gene expression data. The distance furnishes exploratory procedures aimed at finding a target subset of differentially expressed genes. To determine the size of the target subset, we resort to successive elimination of smaller subsets resulting from each step of a random search algorithm based on maximization of the proposed distance. Different stopping rules associated with this procedure are evaluated. The usefulness of the proposed approach is illustrated with an application to the analysis of two sets of gene expression data.

Objectives. To evaluate HER2neu and epidermal growth factor receptor (EGFR) expression with respect to overall survival and disease-free survival (DFS), and correlate expression with pretreatment factors. Comparative evaluations of manual and automated immunohistochemical imaging systems for HER2neu and EGFR expression were made. Methods. Fifty-five patients with stages I–IVA carcinoma of the cervix were treated with definitive radiation therapy. Immunohistochemistry was performed for HER2neu and EGFR, and scored by both manual and automated methods. Univariate and multivariate analyses were performed with disease-free survival (DFS) and overall survival (OS) as primary endpoints, and biomarkers were evaluated for correlation between prognostic factors. Results. Strong correlations in HER2neu and EGFR protein expression were observed between digitally and manually analyzed staining ( P ≤ 0.0001). Increased FIGO stage and decreased HER2neu expression were significant for reduced DFS on univariate analysis ( P ≤ 0.001 and P = 0.03, respectively). Increased FIGO stage, decreased HER2neu expression, and increased membranous staining of EGFR were significant for diminished OS on univariate analysis ( P ≤ 0.0001, P = 0.002, and P = 0.043, respectively). Multivariate analysis revealed only increased membranous staining of EGFR associated with diminished DFS and OS ( P = 0.046 and P = 0.012, respectively). Overexpression of HER2neu correlated significantly with adenocarcinoma, and overexpression of EGFR correlated significantly with squamous cell carcinoma histology ( P = 0.038 and P = 0.035). Inverse correlations were observed between HER2neu expression and clinical stage, EGFR membranous staining, and EGFR distribution ( P = 0.007, P = 0.006, and P = 0.034, respectively). Conclusions. Increased expression of HER2neu and decreased EGFR membranous staining identified patients with improved DFS and OS on univariate analysis, although only decreased EGFR membranous staining was significant on multivariate analysis. We also found strong correlation of results between manually and automated imaging methods.

Optimization of probe design for array-based experiments requires improved predictability of oligonucleotide hybridization behavior. Currently, designing oligonucleotides capable of interacting efficiently and specifically with the relevant target is not a routine procedure. Multiple examples demonstrate that oligonucleotides targeting different regions of the same RNA differ in their hybridization ability. The present work shows how thermodynamic evaluations of oligo-target duplex or oligo self-structure stabilities can facilitate probe design. Statistical analysis of large sets of hybridization data reveals that thermodynamic evaluation of oligonucleotide properties can be used to avoid poor RNA binders. Thermodynamic criteria for the selection of 20 and 21mers, which, with high probability, interact efficiently and specifically with their targets, are suggested. The design of longer oligonucleotides can also be facilitated by the same calculations of DeltaG(o) (T) values for oligo-target duplex or oligo self-structure stabilities and similar selection schemes.

The purpose of this study was to examine a variety of biomarkers in carcinoma of the cervix to better characterize (1). the natural history of the disease, (2). response to radiotherapy (RT), and (3). potential for new therapeutic strategies.Fifty-five patients with Stage IB-IVA carcinoma of the cervix, treated with definitive intent RT, and on whom tumor tissue blocks were available were included in this study. Charts were reviewed for clinical parameters and disease status. Immunohistochemistry was performed for epidermal growth factor receptor (EGFR), vascular endothelial growth factor (VEGF), CD34, topoisomerase II alpha (topo-II), and cyclooxygenase-2 (COX-2). Univariate and multivariate Cox proportional hazards modeling was performed with disease-free survival (DFS) and overall survival (OS) as the end points. Biomarkers were evaluated for correlation between various prognostic factors.In this series of 55 patients with carcinoma of the cervix treated with definitive RT, only stage was significant on univariate analysis for DFS (p0.0001). On univariate analysis, increasing FIGO stage (p0.0001) and membranous staining of EGFR (p0.037) indicated diminished OS. On multivariate analysis for DFS, COX-2, VEGF, and stage were significant (p = 0.012, p = 0.014, and p = 0.03, respectively), with increased expression indicating a worse prognosis. For OS, multivariate analysis revealed that VEGF, EGFR, and FIGO stage were significant (p = 0.005, p = 0.011, and p0.0001, respectively). Significant direct correlations were identified between VEGF and CD34 (p = 0.04), COX-2 and topo-II (p = 0.04), COX-2 and grade (p = 0.04), and tumor size and clinical stage (p = 0.04).Multivariate analysis revealed that increased staining for VEGF and COX-2 indicated diminished DFS, and VEGF and EGFR identified patients at increased risk of death. A significant direct correlation between VEGF and CD34 implicates the process of angiogenesis. Topo-II is a proliferative marker and it correlated directly with COX-2, indicating that expression of COX-2 may be greater in more proliferative tumors. Increased expression of EGFR, VEGF, and COX-2 has identified patients with a worse prognosis in cancer of the cervix. These data support the investigation of therapeutics that target these proteins in carcinoma of the cervix.

A flexible class of semi-parametric survival models is proposed that takes account of long- and short-term covariate effects in cancer survival. The diversity of responses described by the models include non-proportional and crossing survival curves as well as a fraction of long-term survivors. Restricted non-parametric maximum likelihood estimation procedures (RNPMLE) are developed to provide point estimates, confidence intervals and tests for the models. Numerical algorithms to fit semi-parametric survival models are emphasized. The methods are applied to analyse post-treatment survival of breast cancer patients diagnosed in Utah by age and stage.

Lack of adequate statistical methods for the analysis of microarray data remains the most critical deterrent to uncovering the true potential of these promising techniques in basic and translational biological studies. The popular practice of drawing important biological conclusions from just one replicate (slide) should be discouraged. In this paper, we discuss some modern trends in statistical analysis of microarray data with a special focus on statistical classification (pattern recognition) and variable selection. In addressing these issues we consider the utility of some distances between random vectors and their nonparametric estimates obtained from gene expression data. Performance of the proposed distances is tested by computer simulations and analysis of gene expression data on two different types of human leukemia. In experimental settings, the error rate is estimated by cross-validation, while a control sample is generated in computer simulation experiments aimed at testing the proposed gene selection procedures and associated classification rules.

Motivation: Existing analyses of microarray data often incorporate an obscure data normalization procedure applied prior to data analysis. For example, ratios of microarray channels intensities are normalized to have common mean over the set of genes. We made an attempt to understand the meaning of such procedures from the modeling point of view, and to formulate the model assumptions that underlie them. Given a considerable diversity of data adjustment procedures, the question of their performance, comparison and ranking for various microarray experiments was of interest. Results: A two-step statistical procedure is proposed: data transformation (adjustment for slide-specific effect) followed by a statistical test applied to transformed data. Various methods of analysis for differential expression are compared using simulations and real data on colon cancer cell lines. We found that robust categorical adjustments outperform the ones based on a precisely defined stochastic model, including some commonly used procedures. Availability: A program implementing the proposed adjustment and test procedures is available at http://www.hci.utah.edu/groups/biostat/szabo.html. Contact: alexander.tsodikov@hci.utah.edu * To whom correspondence should be addressed.

Since 1980, electron arc irradiation of the postmastectomy chest wall has been the preferred radiotherapy technique at the University of Utah for patients with advanced breast cancer. We report the results of this technique in 156 consecutive Stage IIA-IIIB patients treated from 1980 to 1998.CT treatment planning was used in all patients to identify chest wall thickness and internal mammary lymph node depth. Computerized dosimetry was used to deliver total doses of 50 Gy in 5-1/2 weeks to the chest wall and the internal mammary lymph nodes with electron arc therapy. Patients were assessed for local, regional, and distant control of disease and for survival. Univariate and multivariate proportional hazards were modeled using a hierarchical nonproportional semiparametric model testing the following prognostic factors: age, stage, tumor size, number of positive lymph nodes, estrogen receptor status, and dose. End points evaluated included disease-free survival, cause-specific survival, and overall survival.Eighty-one percent of patients were at high risk for local-regional failure because ofT2 primary tumor or3 positive axillary lymph nodes. The median number of positive lymph nodes was 5, and the median tumor size was 3.5 cm. Actuarial 10-year local-regional control and overall survival were 95% and 52%, respectively. In multivariate analysis, the only factor prognostic for disease-free survival, cause-specific survival, and overall survival was the number of positive lymph nodes (p0.001). The 10-year rates of local-regional control for patients with 0, 1-3, 4-9, andor = 10 involved lymph nodes were 100%, 98%, 93%, and 89%, respectively. The only rates of acute and chronic radiotherapy toxicityor = 2 by RTOG/EORTC criteria were skin related and observed in 44% and 10% for acute and late reactions, respectively.These data demonstrate excellent local-regional control rates with electron arc therapy of the postmastectomy chest wall in patients with advanced breast cancer. Our 20-year experience with electron arc radiotherapy has demonstrated the safety and efficacy of this technique. The advantage of this technique is that the internal mammary lymph node chain can be easily encompassed while the dose to heart and lung is minimized; it also obviates match lines in areas of high risk.

This paper explores the applicability of a mechanistic survival model, based on the distribution of clonogens surviving a course of fractionated radiation therapy, to clinical data on patients with prostate cancer. The study was carried out using data on 1,100 patients with clinically localized prostate cancer who were treated with three-dimensional conformal radiation therapy. The patients were stratified by radiation dose (group 1

This paper is concerned with modern approaches to mechanistic modeling of the process of cancer detection. Measurements of tumor size at diagnosis represent a valuable source of information to enrich statistical inference on the processes underlying tumor latency. One possible way of utilizing this information is to model cancer detection as a quantal response variable. In doing so, one relates the chance of detecting a tumor to its current size. We present various theoretical results emerging from this approach and illustrate their usefulness with numerical examples and analyses of epidemiological data. An alternative approach based on a threshold type mechanism of tumor detection is briefly described.

This final report is concerned with stochastic modeling of breast cancer detection and estimation problems associated with the two-variate distribution of age and tumor size at diagnosis. Our methodological approach is designed to accommodate generally-structured data. Another research avenue was related to optimal scheduling of breast cancer screening by maximizing the expected reduction of tumor size at detection. We developed a Monte- Carlo EM algorithm for estimation of biologically meaningful parameters incorporated into the joint distribution of age and tumor size at detection. The proposed estimation techniques were tested by computer simulations and applied to epidemiological data on individuals identified through the Utah Population Database (UPDB) and Utah Cancer Registry. We studied various indicators of family history and used one of them to stratify the data on breast cancer obtained from the UPDB. An optimal schedule has been constructed for low- and high-risk groups of individuals identified through the UPDB. While the efficacy of the optimal schedule tends to be higher in high-risk families, its structure appears to be robust to variations in breast cancer risk.

Design of antisense oligonucleotides targeting any mRNA can be much more efficient when several activity-enhancing motifs are included and activity-decreasing motifs are avoided. This conclusion was made after statistical analysis of data collected from >1000 experiments with phosphorothioate-modified oligonucleotides. Highly significant positive correlation between the presence of motifs CCAC, TCCC, ACTC, GCCA and CTCT in the oligonucleotide and its antisense efficiency was demonstrated. In addition, negative correlation was revealed for the motifs GGGG, ACTG, AAA and TAA. It was found that the likelihood of activity of an oligonucleotide against a desired mRNA target is sequence motif content dependent.

A branching stochastic process proposed earlier to model oligodendrocyte generation by O-2A progenitor cells under in vitro conditions does not allow invoking the maximum likelihood techniques for estimation purposes. To overcome this difficulty, we propose a partial likelihood function based on an embedded random walk model of clonal growth and differentiation of O-2A progenitor cells. Under certain conditions, the partial likelihood function yields consistent estimates of model parameters. The usefulness of this approach is illustrated with computer simulations and data analyses.