China
Africa
Explore chapters and articles related to this topic
Radiomics analysis for gynecologic cancers
Published in Ruijiang Li, Lei Xing, Sandy Napel, Daniel L. Rubin, Radiomics and Radiogenomics, 2019
Majority of the works in gynecologic cancers have studied the feasibility of using subjectively assessed and semi-quantitative imaging parameters to predict outcomes, tumor grade, and to correlate such features with underlying genetic expression. Subjective radiologic assessments from computed tomography (CT) imaging have been shown to be associated with aggressive mesenchymal transcriptomic profile (Classification of Ovarian Cancer [CLOVAR]) of high-grade serous ovarian cancer (HGSOC) [10,11], and with BRCA mutation status in HGSOC [12] and survival [11,12]. Tumor volume combined with apparent diffusion coefficient (ADC) computed from diffusion-weighted magnetic resonance (MR) images have been shown to be correlated with outcomes in endometrial [13] cancers. Halle [14] identified a dynamic contrast enhanced (DCE)-MR parameter A(Brix), where low A(Brix) was significantly associated with upregulation of genes related to hypoxia. Semi-quantitative DCE-MRI parameters have been shown to predict chemoradiotherapy [15], radiation response from pre-treatment imaging for cervical cancers [16,17]. ADC computed from diffusion-weighted (DW-MRI) quantifies the diffusion in the tissues. ADC has been shown to be associated with clinical prognostic factors [18] and predictive of outcomes to therapy in cervical [19,20] and ovarian cancers [21].
Health
Published in Diane P. Michelfelder, Neelke Doorn, The Routledge Handbook of the Philosophy of Engineering, 2020
The physiological model of health and disease, in contrast to the ontological model, conceives of health and disease as located on a continuous scale. They are only quantitatively different, and there can be a substantial grey zone between them. Health and disease are matters of gradation. What counts (literally!) as a disease is determined by cut-off points that seem epistemologically arbitrary. This way of thinking often leads to a ‘cascade model of disease’ (Boenink 2010). The image of the cascade is often used to label hypothesized disease mechanisms, like the ‘rheumatoid arthritis cascade’ and the ‘amyloid cascade hypothesis’ (supposedly reconstructing the process leading to Alzheimer’s disease). Such a model has important implications for the way we approach health. Cascades are usually conceived of as something that you hardly notice at first: they start really small (on the molecular level). The process that follows is thought of in a mechanistic or even deterministic way: each step tends to trigger the next one, unless an external factor intervenes. Moreover, the process is usually conceived of in a unidirectional way—there is no turning back, although blocking the cascade may be possible. It is not hard to understand that such a cascade model invites proposals for early detection and early treatment. The assumption is that the earlier you intervene, the easier it should be to counter the disease process. Doing so requires that you are able to identify the disease process in its early stages (‘downstream’). This way of thinking underlies, among others, screening programmes for breast and colon cancer, but also current proposals to diagnose Alzheimer’s disease at an earlier stage by way of molecular biomarkers. An extreme but striking illustration of the ‘early detection logic’ driven by the cascade model is the response of some women who carry a BRCA mutation. The mutation means these women are at substantially increased risk of hereditary breast and ovarian cancer, which is why they sometimes decide to have their healthy breasts and/or ovaries removed via preventive surgery. Some of these women call themselves ‘pre-vivors’ (Force 2018): they did not survive disease and suffering, but proactively intervened in the supposed cascade to prevent the disease from reaching the stage in which it causes bodily complaints and suffering.
On the Use of a Logistic Regression Model in the Gene-Environment Problem: A Bayesian Approach
Published in American Journal of Mathematical and Management Sciences, 2019
Akanksha Gupta, S. K. Upadhyay
The list of covariates includes BRCA1 and/or BRCA2 mutation (genetic factor) that are supposed to be important components in the occurrence of ovarian cancer. There are two environmental components OC use and parity. It has been established that the long duration of OC use and increasing parity reduce the risk of ovarian cancer (see Gupta & Upadhyay, 2014; Mukherjee & Chatterjee, 2008; etc.). The covariate age of a woman is divided into 5 categories with 1 representing the age less than 40 years, 2 for the ages in the range 40-49 years, 3 for the ages in the range 50-59 years, 4 for the ages in the range 60-69 years and 5 is used for representing the ages greater than or equal to 70 years. In the discussion that follows, agei is used to denote the age category taking value i, . Similarly, ethnicity is classified into 3 categories; namely, Ashkenazi (As), Non-Ashkenazi (NAs), and mixed ancestry (MAs) coded as 1, 2, and 3, respectively. BRCA history represents the presence of BRCA mutation in the past, if any. BRCA history also takes binary values with value 1 representing availability of BRCA mutation in the past, and 0 representing no such history. GS history takes value 1 if there has been any gynecological surgery in the past; otherwise, it takes value 0. Family history (FH) denotes the history of breast and ovarian cancer in the first degree relative of the patient. It is coded as 0 for no such disease, 1 for the existence of breast cancer, and 2 for the existence of both breast and ovarian cancers. Obviously, the logistic regression model (2) with these covariates can be rewritten as, where is the vector of regression coefficients associated with the corresponding components of predictor variables (influential factors) Z. As such, the term γTZ can be written as, where PHB denotes BRCA history and GSh denotes the GS history. is an identity function that takes value unity if the subject lies under the corresponding category δ, and takes value zero otherwise. Finally, is used to denote the associated regression coefficient with the corresponding category of predictor variable, whereas is the associated coefficient with the corresponding interaction terms. Obviously, is the coefficient which gives the odds ratio of the category δ as compared to the baseline category. Say, for example, for the variables corresponding to age, the baseline category is the age group less than 40 years. Similarly, for ethnicity the baseline category is Ashkenazi, and for family history of breast and ovarian cancer, the baseline category is group of patients without breast or ovarian cancer, i.e. no disease group.