Using Social Media to Understand Primary Discussions in Gastrointestinal Cancers: Machine Learning Approach.

BACKGROUND: The incidence and mortality rates of gastrointestinal (GI) cancers are high in the United States as well as worldwide. The widespread use of social media provides unique opportunities to facilitate the dissemination of information, especially in the context of health. OBJECTIVE: We aim to characterize the public's primary discussions, including perceptions, concerns, and interests toward GI cancers, from prevention, diagnosis, and treatment to survivorship care through the social media platform Twitter, using tweets posted by Twitter users. METHODS: We analyzed 87 860 Twitter posts related to GI cancers. We used machine learning with natural language processing to identify salient topics and themes in the collected tweets. RESULTS: The most common themes across all GI cancer types included cancer risk prevention and awareness outreach programs, risk factors including lifestyles (primarily diet), and cancer survivorship-related discussions (primarily GI symptoms and quality of life). GI symptom-related tweets were prevalent in patients with colorectal and stomach cancers, whereas themes of newer clinical trials, end-of-life trials, palliative care trials, and disease prognosis were common in tweets related to liver/biliary and pancreatic cancers. CONCLUSIONS: Our research emphasizes the importance of individualized approaches in managing GI cancers, considering lifestyle and diet, the need for comprehensive survivorship care, raising awareness, delivering information, and improving targeted interventions related to GI cancers. IMPLICATIONS FOR PRACTICE: Our study suggests utilizing Twitter data to better understand the real-world interest and concerns about GI cancers among the public, which can guide future patient-centered research in this field.

Racial/Ethnic Disparities in HRQoL and Associated Risk Factors in Colorectal Cancer Survivors: With a Focus on Social Determinants of Health (SDOH).

PURPOSE: This study aimed to understand how health-related quality of life (HRQoL) differs by race/ethnicity in colorectal (CRC) survivors. We aimed to 1) examine racial/ethnic disparities in HRQoL, and 2) explore the roles of social determinants of health (SDOH) risk factors for HRQoL differ by racial/ethnic groups. METHODS: In 2,492 adult CRC survivors using Behavioral Risk Factor Surveillance System (BRFSS) survey data (from 2014 to 2021, excluding 2015 due to the absence of CRC data), we used the Centers for Disease Control and Prevention (CDC) HRQoL measure, categorized into "better" and "poor." Multivariate logistic regressions with prevalence risk (PR) were employed for our primary analyses. RESULTS: Compared with non-Hispanic Whites (NHW), non-Hispanic Blacks (NHB) (PR = 0.61, p = .045) and Hispanics (PR = 0.32, p < .001) reported worse HRQoL in adjusted models. In adjusted models, unemployed/retired and low-income levels were common risk factors for worse HRQoL across all comparison groups (NHW, NHB, non-Hispanic other races, and Hispanics). Other SDOH associated with worse HRQoL include divorced/widowed/never married marital status (non-Hispanic other races and Hispanics), living in rural areas (NHW and NHB), and low education levels (NHB and Hispanics). Marital status, education, and employment status significantly interacted with race/ethnicity, with the strongest interaction between Hispanics and education (PR = 2.45, p = .045) in adjusted models. CONCLUSION: These findings highlight the need for culturally tailored interventions targeting modifiable factors (e.g., social and financial supports, health literacy), specifically for socially vulnerable CRC survivors, to address the disparities in HRQoL among different racial/ethnic groups.

Analysis of Changes in Plasma Cytokine Levels in Response to IL12 Therapy in Three Clinical Trials.

The ability of IL12 to stimulate natural killer (NK) cell and T-cell antitumor activity makes it an attractive candidate for the immune therapy of cancer. Our group has demonstrated that IL12 enhances the NK cell response to antibody-coated tumor cells and conducted three clinical trials utilizing IL12 with mAbs (OSU-9968, OSU-0167, and OSU-11010). To better characterize IL12-induced immunity, plasma cytokine levels were measured in 21 patients from these trials with favorable and unfavorable responses. t-statistics and linear modeling were used to test for differences within and between response groups by examining levels at baseline and post-IL12 administration. Patients exhibited significant increases in 11 cytokines post-IL12 administration when analyzed collectively. However, several cytokines were differentially induced by IL12 depending on response. GMCSF was significantly increased in complete/partially responding patients, while stable disease patients had significant increases in IL10 and decreases in VEGF-C. Patients who experienced progressive disease had significant increases in CCL3, CCL4, IL18, TNFα, CXCL10, CCL8, CCL2, IL6, and IFNγ. The increases in CCL3, CCL4, and IL6 in progressive disease patients were significantly higher than in clinically benefitting patients and most prominent within the first two cycles of IL12 therapy. This correlative pilot study has identified changes that occur in levels of circulating cytokines following IL12 administration to patients with cancer, but this report must be viewed as exploratory in nature. It is meant to spark further inquiry into the topic via the analysis of additional cohorts of patients with similar characteristics who have received IL12 in a uniform fashion. SIGNIFICANCE: IL12 activates immune cells and is used to treat cancer. The profile of circulating cytokines was measured in an exploratory fashion in patients with cancer that received IL12 in combination with mAbs. This correlative pilot study could serve as the basis for additional studies of IL12 effects on the production of immune cytokines.

Risk Factors of Health-Related Quality of Life among Gastrointestinal Cancer Survivors in the U.S.: With a Focus on Social and Behavioral Determinants of Health (SBDH).

BACKGROUND: Increasing numbers of long-term gastrointestinal (GI) cancer survivors highlight the importance of understanding the factors contributing to their health-related quality of life (HRQoL). We investigated the risk factors of HRQoL, including demographics, clinical characteristics, and social and behavioral determinants of health (SBDH). METHODS: Data on adult GI cancer survivors (n = 3201) from the Behavioral Risk Factors Surveillance System (BRFSS) surveys from 2014-2021 (except for 2015) were analyzed. Unadjusted/adjusted logistic regression was used. RESULTS: The majority were women (54%) and white (78%), with a median age of 67. Survivors who were 65 years or older, diagnosed with colorectal cancer, or who had fewer comorbidities were more likely to report significantly better HRQoL. Significant social factors of poor HRQoL included unmarried, racial and ethnic minorities, poor socioeconomic status, and poor healthcare access. Significant behavioral factors of poor HRQoL were lack of physical activity, heavy alcohol consumption, and current smoking, with lack of physical activity being the most significant factor. CONCLUSIONS: The SBDH has a critical role in HRQoL. Future studies are warranted to develop a tailored survivorship intervention, such as physical rehabilitation, and to explore machine learning/artificial intelligence-based predictive models to identify cancer survivors at a high risk of developing poor HRQoL.

A flexible copula-based approach for the analysis of secondary phenotypes in ascertained samples.

Data collected for a genome-wide association study of a primary phenotype are often used for additional genome-wide association analyses of secondary phenotypes. However, when the primary and secondary traits are dependent, naïve analyses of secondary phenotypes may induce spurious associations in non-randomly ascertained samples. Previously, retrospective likelihood-based methods have been proposed to correct for sampling biases arising in secondary trait association analyses. However, most methods have been introduced to handle studies featuring a case-control design based on a binary primary phenotype. As such, these methods are not directly applicable to more complicated study designs such as multiple-trait studies, where the sampling mechanism also depends on the secondary phenotype, or extreme-trait studies, where individuals with extreme primary phenotype values are selected. To accommodate these more complicated sampling mechanisms, only a few prospective likelihood approaches have been proposed. These approaches assume a normal distribution for the secondary phenotype (or the latent secondary phenotype) and a bivariate normal distribution for the primary-secondary phenotype dependence. In this paper, we propose a unified copula-based approach to appropriately detect genetic variant/secondary phenotype association in the presence of selected samples. Primary phenotype is either binary or continuous and the secondary phenotype is continuous although not necessary normal. We use both prospective and retrospective likelihoods to account for the sampling mechanism and use a copula model to allow for potentially different dependence structures between the primary and secondary phenotypes. We demonstrate the effectiveness of our approach through simulation studies and by analyzing data from the Avon Longitudinal Study of Parents and Children cohort.

Use of spectral gamma ray as a lithology guide for fault rocks: A case study from the Wenchuan Earthquake Fault Scientific Drilling project Borehole 4 (WFSD-4).

The main purpose of the Wenchuan Earthquake Fault Scientific drilling project (WFSD) was to produce an in-depth borehole into the Yingxiu-Beichuan (YBF) and Anxian-Guanxian faults in order to gain a much better understanding of the physical and chemical properties as well as the mechanical faulting involved. Five boreholes, namely WFSD-1, WFSD-2, WFSD-3P, WFSD-3 and WFSD-4, were drilled during the project entirety. This study, therefore, presents first-hand WFSD-4 data on the lithology (original rocks) and fault rocks that have been obtained from the WFSD project. In an attempt to determine the physical properties and the clay minerals of the lithology and fault rocks, this study analyzed the spectral gamma ray logs (Total gamma ray, Potassium, Thorium and Uranium) recorded in WFSD-4 borehole on the Northern segment of the YBF. The obtained results are presented as cross-plots and statistical multi log analysis. Both lithology and fault rocks show a variability of spectral gamma ray (SGR) logs responses and clay minerals. This study has shown the capabilities of the SGR logs for well-logging of earthquake faults and proves that SGR logs together with others logs in combination with drill hole core description is a useful method of lithology and fault rocks characterization.

Mixture regression models for closed population capture-recapture data.

In capture-recapture studies, the use of individual covariates has been recommended to get stable population estimates. However, some residual heterogeneity might still exist and ignoring such heterogeneity could lead to underestimating the population size (N). In this work, we explore two new models with capture probabilities depending on both covariates and unobserved random effects, to estimate the size of a population. Inference techniques including Horvitz-Thompson estimate and confidence intervals for the population size, are derived. The selection of a particular model is carried out using the Akaike information criterion (AIC). First, we extend the random effect model of Darroch et al. (1993, Journal of American Statistical Association 88, 1137-1148) to handle unit level covariates and discuss its limitations. The second approach is a generalization of the traditional zero-truncated binomial model that includes a random effect to account for an unobserved heterogeneity. This approach provides useful tools for inference about N, since key quantities such as moments, likelihood functions and estimates of N and their standard errors have closed form expressions. Several models for the unobserved heterogeneity are available and the marginal capture probability is expressed using the Logit and the complementary Log-Log link functions. The sensitivity of the inference to the specification of a model is also investigated through simulations. A numerical example is presented. We compare the performance of the proposed estimator with that obtained under model Mh of Huggins (1989 Biometrika 76, 130-140).