Introduction. Most patients with early-stage endometrial carcinoma are treated with hysterectomy. The knowledge of the stage of disease could impact greatly on the eventual therapeutic decision. Numerous studies have investigated the use of endovaginal ultrasonography (US) and magnetic resonance (MR) imaging in the preoperative staging of endometrial carcinoma. Although a number of studies have demonstrated no significant difference in the accuracy of MR imaging and endovaginal US in assessing myometrial invasion, more recently, dynamic contrast-enhanced MR imaging has been found to be significantly superior to the non-enhanced T2-weighted MR imaging in the preoperative evaluation of endometrial carcinoma. / Purpose. (1) To evaluate prospectively the relative accuracy of endovaginal US (including Doppler) and MR imaging using phased array pelvic multicoil in the preoperative staging of endometrial carcinoma with histopathological correlation on the same patient population, (2) To determine whether the addition of contrast-enhanced MR sequences to high resolution fast spinecho (FSE) T2-weighted sequences, can improve the diagnostic accuracy of staging endometrial carcinoma. (Abstract shortened by UMI.)

Relatively little is known about psychological and social factors as potential determinants for developing cervical cancer. Health related beliefs and detailed sociodemographic information were compared among women with different stages of cervical cancer precursor lesions, and among women with HPV infections from a high-risk area for cervical cancer. Both cytology and cervicography were used to obtain cervical lesion status of women, and HPV DNA detection was done by a polymerase chain reaction-based protocol (MY09/11). For cervical cancer precursor lesions, three different groups of women were compared to women remaining cytologically normal (N = 841) throughout four years of follow-up: (1) women with any abnormal cytological diagnosis, atypical squamous cells of undetermined significance, low-grade, and high-grade squamous intraepithelial lesions (ASCUS, LSILs, and HSILs respectively) (N = 487); (2) women with either LSILs or HSILs (N = 420); and (3) women with HSILs only (N = 37). (Abstract shortened by UMI.)

<p> Cytotoxic T lymphocytes (CTLs) are the primary component of the adaptive immune system responsible for clearance of virally infected and tumorigenic cells. In cancer however, this tumor-specific immune response is often impaired. The impairment is multifactorial; some cancers utilize mechanisms to evade the immune system through downregulation of Major Histocompatability Complex I or lack of tumor-specific antigens, while others use methods to actively inhibit local function of tumor-induced immune responses via production of immunosuppressive cytokines, Fas-mediated apoptosis, or recruitment of T regulatory cells (Tregs). These Tregs function to further immune regulate and inhibit CTLs, using methods such as suppressive cytokines, and cytotoxic killing. All of these components lead to an &ldquo;on/off&rdquo; phenotype, where CTL effector function is shut down within the Tumor Immunosuppressive Microenvironment (TIM), but can be recovered quickly upon removal of CTLs from the TIM. The transient impairment of Tumor Infiltrating Lymphocytes (TIL) has been described in mouse models, but is poorly characterized in humans. </p><p> In this dissertation, we examined infiltration of CTLs across several types of human pediatric cancers, taken from patients who had not undergone prior treatment. We found tumors associated with favorable prognoses, including Wilms&rsquo; Tumor and Neuroblastoma (NB), had higher levels of CTL infiltration than those with less favorable prognoses, e.g. Ependymoma, which possessed no observable infiltration. Additionally, we demonstrate the TIL &ldquo;on/off&rdquo; phenotype in a case of Pilocytic Astrocytoma, demonstrating significant recovery of TIL effector function. </p><p> We proposed that the poor infiltration and impaired effector function in these pediatric tumors was a direct result of the TIM, and sought to improve this immune response by developing an attenuated live cell vaccine, utilizing a murine NB model, Neuro2a, to create a NB line with knock down (KD) of Inhibitor of Differentiation 2 (Id2), which impaired their ability to form tumors in vivo. In prophylactic and therapeutic models, introduction of Id2-KD cells in combination with the immune checkpoint blockade inhibitor anti-CTLA-4, induced an increase in CTLs capable of homing to the tumor, that were also able to employ effector function within the TIM, resulting in clearance of wild-type Neuro2a tumors. </p><p> A separate emerging immunotherapeutic approach is to express a Chimeric Antigen Receptor (CAR) on CTLs that allows them to be activated to kill cells expressing the CAR-specific protein, bypassing MHC presentation. Using a murine Rhabdomyosarcoma model, we demonstrate that tumor infiltrating Tregs express lytic molecules, encouraging us to develop a method of successfully transducing Tregs with a CAR (DC101), rather than CTLs, thereby exploiting characteristics of the Treg in the TIM, specifically their cytotoxic capability and their unique recruitment and ability to thrive in that environment. We demonstrate in vitro CAR-mediated redirection of lytic effector function using DC101-expressing CTLs against tumor cell lines, though attempting to increase Treg cytotoxicity in vitro via known inducers of CTL cytotoxicity (IFN&alpha; or IL-12) or known inducers of Tregs within the TIM (TGF-&beta;1) showed no increase in Treg cytotoxicity.</p>

<p> Mathematical modeling of tumor response to radiation therapy (RT) has great potential for designing therapy plans that are more personalized, more adaptive, and more reliable for outcome predictions. A preexisting model of tumor response to radiation therapy for cervical cancer has been shown to generate model parameters that correlate strongly with both tumor local control and disease-specific survival. This model is further developed through incorporation of another effect of RT not previously accounted for: the oxygen effect. An easily obtainable form of input data, hemoglobin level, enables simulation of the oxygen effect simultaneously with the other major model effects. For the Local Control (LC) patient group, the changes in the model parameters caused by incorporation of the oxygen effect are found to significantly improve the agreement of those parameters with actual patient data. For the Local Failure (LF) group and the overall patient group, the oxygen effect is incorporated without significant change to the agreement between the model-simulated output parameters and the actual patient data. Also, a strategy is presented for solving the main model equations to obtain analytic expressions for surviving cell fraction and regression volume ratio as functions of time.</p>

<p> Although clinical trials are essential for the development of cancer treatments, only approximately 3% of cancer patients in the U.S. participate in them. While 55% of these patients are enrolled in cancer clinical trials through community-based practices and around 80% of all cancer patients are seen at this type of practice, there is a lack of knowledge about the enrollment barriers at these sites. This study evaluates enrollment barriers at a community-based cancer clinic at the levels of the investigative site, healthcare provider, and patient. Barriers to enrollment and strategies to increase enrollment are evaluated through historical data analyses and results from a survey assessing the opinions of healthcare providers on enrollment and research practices. </p>

<p> This dissertation examines the role of retinoic acid receptor activation in the post-transcriptional regulation of a fibroblast growth factor binding protein. Previous work showed that all-trans retinoic acid (ATRA) reduces mRNA expression of the angiogenic switch molecule, Fibroblast Growth Factor Binding Protein 1 (FGFBP1 or FGF-BP), independent of an effect on transcription of the FGFBP1 mRNA. I hypothesized that a retinoid-induced microRNA was involved in FGFBP1 mRNA loss. MicroRNAs (miRs) are 19-22 nucleotide (nt) single stranded non-coding RNAs that post-transcriptionally repress mature mRNA function, thereby reducing expression of their target proteins. The current dogma suggests that miRs canonically bind to the 3' untranslated region (UTR) of mRNA through a 7-nt seed-matched site. However, recent data indicate that miRs may also bind the open reading frame (ORF) of mRNAs. In this dissertation, I show that miR-27b-3p and miR-125a-5p are induced by ATRA and target FGFBP1. Overexpression of miR-27b-3p and miR-125a-5p rapidly reduced FGFBP1 mRNA levels through a target site in the open reading frame of the FGFBP1 mRNA. Both microRNAs showed specificity for regions within the ORF of FGFBP1, suggesting that these microRNAs may also be involved in inhibiting translation of the FGFBP1 protein. Next generation sequencing data from The Cancer Genome Atlas shows that loss of these microRNAs is characteristic of several epithelial cancers, including head and neck, lung, and cervical squamous cell carcinomas, suggesting a tumor suppressor role for miRs 27a-3p and 125a-5p. In total, these data suggest an important regulatory role for miRs 27b-3p and 125a-5p in the oncogenesis of squamous cell carcinomas, through modulation of FGFBP1 expression.</p>

<p> Introduction: Pancreatic adenocarcinoma is anticipated to become the most fatal gastrointestinal neoplasm by 2017. Because patients who present with symptoms uniformly have advanced disease, 5-year survival is poor. Surveillance programs in the highest risk patients have produced disappointing results and use invasive tests that would not be safe or practical at the population level. Molecular markers, specifically aberrant DNA methylation, show promise in early studies. We aimed to: 1) measure the sensitivity and specificity of aberrant DNA methylation in stools as a minimally invasive way to detect pancreatic cancer; and, 2) identify novel methylation candidate markers highly sensitive and specific for pancreatic cancer. </p><p> Methods: To select candidate markers among those reported in the literature, DNA was extracted from unmatched tissue samples of pancreatic cancers and of normal colonic epithelia and assayed for aberrant methylation by methylation specific PCR. The most discriminant candidates and mutant <i>KRAS</i> were then assayed from matched archival stools of patients with pancreatic cancer in comparison to stools from healthy controls. Sensitivity and specificity for pancreatic cancer were determined from multivariate logistic regression models. To identify novel candidate markers, DNA was extracted from matched, archival tissues of pancreas cancer and two control groups (normal pancreas and normal colon) and sequenced using the reduced representation bisulfite technique. Among all mapped regions, those with the highest variance in methylation differences between cases and controls were filtered prior to analysis. Significant regions were then blindly assayed by methylation specific PCR in an independent, matched sample set, where sensitivity and specificity were measured using univariate logistic regression. </p><p> Results: In tissues, methylated <i>BMP3, EYA4, UCHL1</i> and <i> MDFI</i> were highly discriminant for pancreatic cancer in comparison to normal colon samples. However, when assayed in stools, only <i>BMP3 </i> remained significant. In combination with mutant <i>KRAS</i>, the area under the receiver operating characteristics curve for <i>BMP3 </i> in detection of pancreatic cancer, was 0.85, indicating strong association. Results were not significantly influenced by tumor location or stage. Reduced representation bisulfite sequencing identified over 500 differentially methylated regions which met a priori significance thresholds. The top 25 novel candidates were validated in independent samples, showing both strong association with cases, compared to controls and high signal to noise ratio. </p><p> Conclusions: We report the first demonstration of feasibility for the detection of pancreatic cancer using assay of aberrantly methylated DNA markers from stool. This is a critical first step in the long-term goal of developing a minimally invasive screening tool to curb the mortality rate of this devastating disease. Because the discriminant candidate markers, methylated <i>BMP3 </i> and mutant <i>KRAS</i>, are not unique to pancreatic cancer, we developed a marker discovery strategy which yielded dozens of highly discriminant, validated, novel candidates, many of which have never before been reported in association with cancer. Further studies are indicated to measure the site-specificity of these markers for pancreatic cancer, compared to other gastrointestinal neoplasms and to study the clinical utility of these novel candidates in distant biologic media, such as blood, stool or pancreatic juice.</p>

<p>Technological advances in the recent decades have enabled cancer researchers to probe the disease at multiple resolutions. This wealth of experimental data combined with computational systems biology methods is now leading to predictive models of cancer progression and response to therapy. We begin by presenting our research group's multi-scale in silico framework for modeling cancer, whose core is a tissue-scale computational model capable of tracking the progression of tumors from a diffusion-limited avascular phase through angiogenesis, and into invasive lesions with realistic, complex morphologies. We adapt this core model to consider the delivery of systemically-administered anticancer agents and their effect on lesions once they reach their intended nuclear target. We calibrate the model parameters using in vitro data from the literature, and demonstrate through simulation that transport limitations affecting drug and oxygen distributions play a significant role in hampering the efficacy of chemotherapy; a result that has since been validated by in vitro experimentation. While this study demonstrates the capability of our adapted core model to predict distributions (e.g., cell density, pressure, oxygen, nutrient, drug) within lesions and consequent tumor morphology, nevertheless, the underlying factors driving tumor-scale behavior occur at finer scales. What is needed in our multi-scale approach is to parallel reality, where molecular signaling models predict cellular behavior, and ultimately drive what is seen at the tumor level. Models of signaling pathways linked to cell models are already beginning to surface in the literature. We next transition our research to the molecular level, where we employ data mining and bioinformatics methods to infer signaling relationships underlying a subset of breast cancer that might benefit from targeted therapy of Androgen Receptor and associated pathways. Defining the architecture of signaling pathways is a critical first step towards development of pathways models underlying tumor models, while also providing valuable insight for drug discovery. Finally, we develop an agent-based, cell-scale model focused on predicting motility in response to chemical signals in the microenvironment, generally accepted to be a necessary feature of cancer invasion and metastasis. This research demonstrates the use of signaling models to predict emergent cell behavior, such as motility. The research studies presented in this dissertation are critical steps towards developing a predictive, in silico computational model for cancer progression and response to therapy. Our Laboratory for Computational & Predictive Oncology, in collaboration with research groups throughout in the United States and Europe are following a computational systems biology paradigm where model development is fueled by biological knowledge, and model predictions are refining experimental focus. The ultimate objective is a virtual cancer simulator capable of accurately simulating cancer progression and response to therapy on a patient-specific basis.

Telomerase is the ribonucleoprotein enzyme whose principal function is to catalyze the de novo addition of telomeric repeats to the ends of linear chromosomes. Telomerase activity is predominantly observed in cancerous tissues and rarely in normal somatic cells making telomerase an attractive anticancer target. BIBR1532 is a highly selective pharmacological inhibitor of telomerase catalytic activity and induces telomere shortening and eventual growth arrest. We treated drug-sensitive and drug-resistant human leukemia and breast cancer cells lines with BIBR1532 and observed telomere shortening and a progressive decrease in proliferative capacity or colony forming ability. This effect was increased with the co-treatment of traditional chemotherapeutics, suggesting that pharmacological telomerase inhibition in combination with chemotherapeutics may be a valid strategy for the treatment of both drug-resistant and drug-sensitive cancers. Finally, our results support further investigation and development of pharmacological inhibitors of telomerase catalytic activity.

Squamous intraepithelial lesions (SIL) are believed to be precursors of in situ cervical neoplasia and of invasive cervical cancer. Though the burden of this disease has been greatly reduced thanks to the availability of an effective screening test, the worldwide morbidity and mortality remain high. / On the basis of a classic approach to analysis of data from a prospective cohort study, I correlated HPV status at enrollment with subsequent risk of incident SIL during up to three years of follow-up among women of Brazil enrolled to the Ludwig-McGill cohort. / The risk of occurrence of a first instance of SIL among women was strongly associated with HPV infection. After adjustment for a number of socio-demographic factors that have previously been established as risk factors for cervical cancer, the magnitude of association with high-risk HPVs remained unchanged while the association with low risk types was dampened, furthering the evidence for the role of oncogenic HPV types in the development of cervical cancer. / Parity was found to have an independent positive association with cervical precursor lesions, after adjustment for HPV status, age, previous history of Pap screening, number of sexual partners, age at first intercourse, and income. Income was also positively associated with the development of lesions.