Survival analysis for lung cancer patients: A comparison of Cox regression and machine learning models.

INTRODUCTION: Survival analysis based on cancer registry data is of paramount importance for monitoring the effectiveness of health care. As new methods arise, the compendium of statistical tools applicable to cancer registry data grows. In recent years, machine learning approaches for survival analysis were developed. The aim of this study is to compare the model performance of the well established Cox regression and novel machine learning approaches on a previously unused dataset. MATERIAL AND METHODS: The study is based on lung cancer data from the Schleswig-Holstein Cancer Registry. Four survival analysis models are compared: Cox Proportional Hazard Regression (CoxPH) as the most commonly used statistical model, as well as Random Survival Forests (RSF) and two neural network architectures based on the DeepSurv and TabNet approaches. The models are evaluated using the concordance index (C-I), the Brier score and the AUC-ROC score. In addition, to gain more insight in the decision process of the models, we identified the features that have an higher impact on patient survival using permutation feature importance scores and SHAP values. RESULTS: Using a dataset including the cancer stage established by the Union for International Cancer Control (UICC), the best performing model is the CoxPH (C-I: 0.698±0.005), while using a dataset which includes the tumor size, lymph node and metastasis status (TNM) leads to the RSF as best performing model (C-I: 0.703±0.004). The explainability metrics show that the models rely on the combined UICC stage and the metastasis status in the first place, which corresponds to other studies. DISCUSSION: The studied methods are highly relevant for epidemiological researchers to create more accurate survival models, which can help physicians make informed decisions about appropriate therapies and management of patients with lung cancer, ultimately improving survival and quality of life.

Longer Survival From Melanoma in Germany.

BACKGROUND: New treatment options for cutaneous melanomas with a poor prognosis have been available since 2011, including immune therapies and targeted drugs. Randomized controlled trials have demonstrated that these treatments improve survival, but no population- level studies have been available to date. METHODS: All patients in the database of the Center for Cancer Registry Data (Zentrum für Krebsregisterdaten) who had a diagnosis of melanoma (ICD10: C43) in the years 2000 to 2019 were included in the study. The relative five-year survival (5YRS) was calculated for four 5-year periods (2000-04, 2005-09, 2010-14, 2015-19). The data were standardized/stratified according to sex, age group, and UICC stage to correct for differences between regions and over time. Regression models were used to detect statistically significant secular trends. RESULTS: 301 486 individuals were included in the study. The overall 5YRS rose from 93% (2000-04) to 95% (2015-19). The 5YRS in 2015-19 was similar to or greater than that in 2000-04 for all subgroups. The largest rises in 5YRS were between 2010-14 and 2015-19, and specifically in advanced stages: for UICC stage IV tumors, the 5YRS rose from 31% to 36%. There was a significant rising trend across the four time periods (p < 0.001). CONCLUSION: The survival of melanoma patients has improved over the past 20 years. From 2010-14 to the most recent period, the largest changes were seen in advanced tumor stages. This favorable development coincided with the introduction of new therapies.

Reduced Psychosocial Well-Being among the Children of Women with Early-Onset Breast Cancer.

BACKGROUND: Approximately 27% of female breast cancer patients are diagnosed before the age of 55, a group often comprising mothers with young children. Maternal psychosocial well-being significantly impacts these children's psychosocial well-being. This study assesses the well-being of children with mothers who have early-onset breast cancer. METHODS: We examined the eldest child (up to 15 years old) of women with nonmetastatic breast cancer (<55 years old, mean age: 40) enrolled in the mother-child rehab program 'get well together'. Using maternal reports on children's well-being (the Strengths and Difficulties Questionnaire; SDQ), we describe the prevalence of abnormally high SDQ scores and identify protective and risk factors via linear regression. RESULTS: The mean SDQ scores of 496 children (4-15 years old, mean age: 8) fell below the thresholds, indicating psychosocial deficits. However, most SDQ scores deviated negatively from the general population, especially for emotional problems, with one in ten children displaying high and one in five displaying very high deficits. Female sex, more siblings, a positive family environment and maternal psychosocial well-being were protective factors for children's psychosocial well-being. CONCLUSIONS: Children of mothers with breast cancer may benefit from improved maternal well-being and family support. Further research is needed to identify appropriate interventions.

Young Adult German Breast Cancer Patients Participating in a Three-Week Inpatient Mother-Child Rehab Program Have High Needs for Supportive Care.

A known cut-off problem hampers the interpretation of quality of life (QOL) scores. The purpose of this study was to apply a novel approach for the EORTC QLQ-C30 instrument to identify the proportion of breast cancer (BC) patients in need of supportive care. Changes in QOL during the COVID-19 pandemic were evaluated, as well as changes over time (after treatment termination and up to 4 years later). Data were obtained from a cohort study on young adult BC patients with minor children participating in a mother-child rehab program. Cross-sectional QOL data were collected from 2015 to 2021 (baseline). Follow-up data were available for up to 4 years after diagnosis for a subgroup. The baseline cohort included 853 women (mean age 35 years). More than 50% had a need for supportive care. In the subgroup with follow-up, this proportion remained at a high level up to several years after diagnosis. During the COVID-19 pandemic, changes regarding the proportion with this need were not as high as expected-with the exception of changes on the QLQ-C30 scale 'role functioning' (+15%). Even several years after diagnosis, every second BC patient with minor children had a need for supportive care, which is much higher than previously found. Healthcare staff should be aware of this potential need and should address this issue.