Mixture modeling for joint analysis of survival, discrete, and continuous data

Mixture modeling is commonly used to model categorical latent variables that represent subpopulations in which population membership is unknown but can be inferred from the data. In relatively recent years, the potential of finite mixture models has been applied in time-to-event data. However, the commonly used survival mixture model assumes that the effects of the covariates involved in failure times differ across latent classes, but the covariate distribution is homogeneous. The aim of this dissertation is to develop a method to examine time-to-event data in the presence of unobserved heterogeneity under a framework of mixture modeling. A joint model is developed to incorporate the latent survival trajectory along with the observed information for the joint analysis of a time-to-event variable, its discrete and continuous covariates, and a latent class variable. It is assumed that the effects of covariates on survival times and the distribution of covariates vary across different latent classes. The unobservable survival trajectories are identified through estimating the probability that a subject belongs to a particular class based on observed information. We applied this method to a Hodgkin lymphoma study with long-term follow-up and observed four distinct latent classes in terms of long-term survival and distributions of prognostic factors. Our results from simulation studies and from the Hodgkin lymphoma study demonstrated the superiority of our joint model compared with the conventional survival model. This flexible inference method provides more accurate estimation and accommodates unobservable heterogeneity among individuals while taking involved interactions between covariates into consideration.^

Scanned Ion Beam Therapy for Thoracic Tumors

Although frequently cured of Hodgkin lymphoma, adolescents and young adults can develop radiation induced second cancers. These patients could potentially benefit from scanned ion radiotherapy yet likely would require motion mitigation strategies. In theory, four-dimensional (4D) optimization of ion beam fields for individual motion states of respiration can enable superior sparing of healthy tissue near moving targets, compared to other motion mitigation strategies. Furthermore, carbon-ion therapy can sometimes provide greater relative biological effectiveness (RBE) for cell sterilization in a target but nearly equivalent RBE in tissue upstream of the target, compared to proton therapy. Thus, we expected that for some patients with Hodgkin lymphoma, carbon-ion therapy would reduce the predicted risk of second cancer incidence in the breast compared with proton therapy. The purpose of this work was to determine whether 4D-optimized carbon-ion therapy would significantly reduce the predicted risk of radiation induced second cancers in the breast for female Hodgkin lymphoma patients while preserving tumor control compared with proton therapy. To achieve our goals, we first investigated whether 4D-optimized carbon beam tracking could reduce dose to volumes outside a moving target compared with 3D-optimized carbon beam tracking while preserving target dose coverage. To understand the reliability of scanned carbon beam tracking, we studied the robustness of dose distributions in thoracic targets to uncertainties in patient motion. Finally, we investigated whether using carbon-ion therapy instead of proton therapy would significantly reduce the predicted risk of second cancer in the breast for a sample of Hodgkin lymphoma patients. We found that 4D-optimized ion beam tracking therapy can reduce the maximum dose to critical structures near a moving target by as much as 53%, compared to 3D-optimized ion beam tracking therapy. We validated these findings experimentally using a scanned carbon ion synchrotron and a motion phantom. We found scanned carbon beam tracking to be sensitive to a number of motion uncertainties, most notably phase delays in tracking, systematic spatial errors, and interfractional motion changes. Our findings indicate that a lower risk of second cancer in the breast might be expected for some Hodgkin lymphoma patients using carbon-ion therapy instead of proton therapy. For our reference scenario, we found the ratio of risk to be 0.77 ± 0.35 for radiogenic breast cancer after carbon-ion therapy versus proton therapy. Our findings were dependent on the RBE values for tumor induction and the radiosensitivity of breast tissue, as well as the physical dose distribution.

TYPE I INSULIN-LIKE GROWTH FACTOR RECEPTOR TYROSINE KINASE AS A MOLECULAR TARGET IN MANTLE CELL LYMPHOMA

Mantle cell lymphoma (MCL) is an aggressive B-cell lymphoid malignancy representing 5-10% of all non-Hodgkin’s lymphomas. It is distinguished by the t(11;14)(q13;q32) chromosomal translocation that juxtaposes the proto-oncogene CCND1, which encodes cyclin D1 at 11q13 to the IgH gene at 14q32. MCL patients represent about 6% of all new cases of Non-Hodgkin’s lymphomas per year or about 3,500 new cases per year. MCL occurs more frequently in older adults – the average age at diagnosis is the mid-60s with a male-to-female ratio of 2-3:1. It is typically characterized by the proliferation of neoplastic B-lymphocytes in the mantle zone of the lymph node follicle that have a prominent inclination to disseminate to other lymphoid tissues, bone marrow, peripheral blood and other organs. MCL patients have a poor prognosis because they develop resistance/relapse to current non-specific therapeutic regimens. It is of note that the exact molecular mechanisms underlying the pathogenesis of MCL are not completely known. It is reasonable to anticipate that better characterization of these mechanisms could lead to the development of specific and likely more effective therapeutics to treat this aggressive disease. The type I insulin-like growth factor receptor (IGF-IR) is thought to be a key player in several different solid malignancies such as those of the prostate, breast, lung, ovary, skin and soft tissue. In addition, recent studies in our lab showed evidence to support a pathogenic role of IGF-IR in some types of T-cell lymphomas and chronic myeloid leukemia. Constitutively active IGF-IR induces its oncogenic effects through the inhibition of apoptosis and induction of transformation, metastasis, and angiogenesis. Previous studies have shown that signaling through IGF-IR leads to the vi activation of multiple signaling transduction pathways mediated by the receptor-associated tyrosine kinase domain. These pathways include PI3K/Akt, MAP kinase, and Jak/Stat. In the present study, we tested the possible role of IGF-IR in MCL. Our results demonstrate that IGF-IR is over-expressed in mantle cell lymphoma cell lines compared with normal peripheral blood B- lymphocytes. Furthermore, inhibition of IGF-IR by the cyclolignan picropodophyllin (PPP) decreased cell viability and cell proliferation in addition to induction of apoptosis and G2/M cell cycle arrest. Screening of downstream oncogenes and apoptotic proteins that are involved in both IGF-IR and MCL signaling after treatment with PPP or IGF-IR siRNA showed significant alterations that are consistent with the cellular changes observed after PPP treatment. Therefore, our findings suggest that IGF-IR signaling contributes to the survival of MCL and thus may prove to be a legitimate therapeutic target in the future.

Estimated current patterns of food consumption and perceived changes among patients with lymphoma

Usual food choices during the past year, self-reported changes in consumption of three important food groups, and weight changes or stability were the questions addressed in this cross-sectional survey and retrospective review. The subjects were 141 patients with Hodgkin's disease or other B-cell types of lymphoma within their first three years following completion of initial treatments for lymphoma at the University of Texas M. D. Anderson Cancer Center in Houston, Texas. ^ The previously validated Block-98 Food Frequency Questionnaire was used to estimate usual food choices during the past year. Supplementary questions asked about changes breads and cereals (white or whole grain) and relative amounts of fruits and vegetables compared with before diagnosis and treatment. Over half of the subjects reported consuming more whole grains, fruits, and/or vegetables and almost three quarters of those not reporting such changes had been consuming whole grains before diagnosis and treatment. ^ Various dietary patterns were defined in order to learn whether proportionately more patients who changed in healthy directions fulfilled recognized nutritional guidelines such as 5-A-day fruits and vegetables and Dietary Reference Intakes (DRIB) for selected nutrients. ^ Small sizes of dietary pattern sub-groups limited the power of this study to detect differences in meeting recommended dietary guidelines. Nevertheless, insufficient and excessive intakes were detected among individuals with respect to fruits and vegetables, fats, calcium, selenium, iron, folate, and Vitamin A. The prevalence of inadequate or excess intakes of foods or nutrients even among those who perceived that they had increased or continued to eat whole grains and/or fruits and vegetables is of concern because of recognized effects upon general health and potential cancer related effects. ^ Over half of the subjects were overweight or obese (by BMI category) on their first visit to this cancer center and that proportion increased to almost three-quarters by their last follow-up visits. Men were significantly heavier than women, but no other significant differences in BMI measures were found even after accounting for prescribed steroids and dietary patterns. ^