Bayesian variable selection and survival modeling: assessing the Most important comorbidities that impact lung and colorectal cancer survival in Spain

Sine-G family of distributions in Bayesian survival modeling: A baseline hazard approach for proportional hazard regression with application to right-censored oncology datasets using R and STAN

In medical research and clinical practice, Bayesian survival modeling is a powerful technique for assessing time-to-event data. It allows for the incorporation of prior knowledge about the model's parameters and provides a more comprehensive understanding of the underlying hazard rate function. In this paper, we propose a Bayesian survival modeling strategy for proportional hazards regression models that employs the Sine-G family of distributions as baseline hazards. The Sine-G family contains flexible distributions that can capture a wide range of hazard forms, including increasing, decreasing, and bathtub-shaped hazards. In order to capture the underlying hazard rate function, we examine the flexibility and effectiveness of several distributions within the Sine-G family, such as the Gompertz, Lomax, Weibull, and exponentiated exponential distributions. The proposed approach is implemented using the R programming language and the STAN probabilistic programming framework. To evaluate the proposed approach, we use a right-censored survival dataset of gastric cancer patients, which allows for precise determination of the hazard rate function while accounting for censoring. The Watanabe Akaike information criterion and the leave-one-out information criterion are employed to evaluate the performance of various baseline hazards.

Prognostic factors and survival disparities in right-sided versus left-sided colon cancer

Right-sided colon cancer (RCC) and left-sided colon cancer (LCC) differ in features and outcomes because of variations in embryology, epidemiology, pathology, and prognosis. This study sought to identify significant factors impacting patient survival through Bayesian modelling. Data was retrospectively analysed from a colorectal neoplasia database. Data on demographics, perioperative risks, treatment, mortality, and survival was analysed from patients who underwent colon cancer surgery from January 2010 to December 2021. This study involved 2475 patients, with 58.7% having RCC and 41.3% having LCC. RCC patients had a notably higher mortality rate, and their overall survival (OS) rates were slightly lower than those with LCC (P < 0.05). RCC stages I-IV consistently exhibited worse OS and relapse-free survival (RFS) than LCC (P < 0.05). Factors like age, BMI, ASA score, cancer stage, and comorbidities had significant associations with OS and RFS. Poor and moderate differentiation, lower lymph node yield, and organ resection were linked to lower survival while receiving chemotherapy; higher BMI levels and elective surgery were associated with better survival (all P < 0.05). Our study reveals key differences between RCC and LCC, emphasising the impact of age, BMI, ASA score, cancer stage, and comorbidities on patient survival. These findings could inform personalised treatment strategies for colon cancer patients.

Seven preoperative factors have strong predictive value for postoperative pneumonia in patients undergoing thoracoscopic lung cancer surgery

Background

Postoperative pneumonia (POP) is a hospital acquired pneumonia that occurs >48 hours after tracheal intubation. The diagnosis of POP should be based on clinical and radiological findings within 30 days after surgery. It is a common complication after thoracoscopic surgery for lung cancer patients. However, the specific impact of preoperative comorbidities on the incidence of POP remains unclear. This study aimed to analyze the preoperative data of patients with lung cancer to help surgeons predict the risk of incidence of POP after thoracoscopic lung resection.

Methods

This study is a prospective study that included patients with lung cancer who were scheduled for thoracoscopic surgery in 1 year. All cases came from two medical centers. Preoperative demographic information, tumor information, preoperative comorbidities, quality of life scores, and incidence of POP were collected. Variables were screened by univariate analysis and multivariate regression. Finally, a prediction model was constructed. A total of 53 preoperative factors were included as candidate predictors. The binary outcome variable was defined as the presence or absence of POP. The incidence of POP was the primary outcome variable. The predictive performance of the model was verified internally through 1,000 iterations of bootstrap resampling.

Results

A total of 1,229 patients with lung cancer who underwent thoracoscopic surgery were enrolled. In addition, 196 cases (15.95%) had POP; 1,025 (83.40%) patients had comorbid conditions. The total number of comorbidity diagnosed in all samples was 2,929. The prediction model suggested that patients with advanced age, high body mass index (BMI), smoking, poor physical condition, respiratory diseases, diabetes, and neurological diseases were more likely to develop POP. The area under the curve (AUC) and Brier scores were 0.851 and 0.091, respectively. The expected and observed results were in strong agreement, according to the likelihood of POP calibration curve.

Conclusions

The constructed model is capable of evaluating the probability of POP occurrence in patients with lung cancer. Seven preoperative factors in patients with lung cancer were found to be associated with increased probability of having pneumonia after thoracoscopic lung resection. This model can help predict the incidence of POP after surgery.