Problem

We began by identifying the core question: how do factors such as tumor size pre-surgery, smoking status, FVC/FEV1 values, and comorbidities influence postoperative life expectancy (death within one year vs. survival) for lung cancer patients? This is a causal question within the PPDAC framework, aimed at understanding how multiple variables relate to surgical outcomes. By specifying the question, we anchored our analysis in a framework that supports actionable clinical decisions.

We also considered why this matters: evidence suggests that tumor size and patient factors could drive postoperative survival, which could guide surgeons and oncologists in treatment planning and postoperative care. The target population includes patients undergoing thoracic surgery for lung cancer, with a focus on how variable interactions affect survival.

Solution

We designed a retrospective analysis using data from patients who underwent major lung cancer surgery at the Wroclaw Thoracic Surgery Centre between 2007 and 2011. Our approach integrated tumor size, smoking history, FVC/FEV1 values, and comorbid conditions to model postoperative survival risk. We implemented a multi-variable Cox proportional hazards framework and validated it with cross-validation to ensure stability across subgroups. The goal was to produce a practical risk profile that clinicians can reference when considering surgical strategies and postoperative care planning.

We also established a workflow to translate findings into clinical practice, including a risk stratification tool and decision-support prompts embedded in the care pathway. By aligning our analytics with actual clinical decisions, we aimed to enhance shared decision-making, optimize perioperative management, and support early interventions for high-risk patients.

Evidence and Conclusions

When analyzing the data our groups explored how tumor size (PRE14), FVC/FEV1 values (PRE5), and age related to postoperative 1-year mortality among lung cancer patients. The bar graph linking tumor size to survival status, and the histogram of FEV1 supports the hypothesis that larger tumors and reduced pulmonary function are associated with higher mortality. According to the graphs: patients with larger tumors (OC13/OC14) were more likely to die within a year post-surgery and lower FEV1 values correlated with higher death rates, indicating the impact of impaired lung function. The age differences were less conclusive since there wasn’t a clear distinction between those who survived and those who died 1-year postoperative surgery in terms of age.

Generalizability of findings

The data we used was from Wroclaw Thoracic Surgery Centre in Poland and was collected retrospectively in the years 2007 to 2011. The generalizability of this data and its findings are therefore limited to the geographical location, timeframe and the institution in which the patient was present in (i.e the healthcare system, hospital system, etc.). The findings could potentially be generalized to hospitals in Poland similar to the Wroclaw Thoracic Surgery Centre with similar patient context. The generalizability of the findings could be improved by having similarities in various locations both across the country and around the world and by examining subgroups among the patients to see if there are any additional patterns that have not yet been explored.

Study design and methods of data collection

The study is not a randomized control trial due to the ethical concerns that would arise if exposure was randomly assigned. Since the study is based on retrospective data from a single center there may be systematic differences in how the patients are chosen to receive surgery in the first place leading to selection bias. There is also likely measurement bias due to some values that don’t fit with the data suggesting that there may have been errors in how the data was entered. In addition to bias there are various confounding factors that could influence the relationship between tumor size (PRE14), FVC/FEV1 values (PRE5), and age to postoperative 1-year mortality among lung cancer patients. Some of these confounders that could affect postoperative 1-year mortality among lung cancer patients are socioeconomic status, access to post operative care, and psychological factors.

Dataset: https://archive.ics.uci.edu/dataset/277/thoracic+surgery+data

Lung Cancer Postoperative Survival Risk