Outcome and presentation of heart failure in breast cancer patients: findings from a Swedish register-based study

Theory and Practice of Integrating Machine Learning and Conventional Statistics in Medical Data Analysis

The practice of medical decision making is changing rapidly with the development of innovative computing technologies. The growing interest of data analysis with improvements in big data computer processing methods raises the question of whether machine learning can be integrated with conventional statistics in health research. To help address this knowledge gap, this paper presents a review on the conceptual integration between conventional statistics and machine learning, focusing on the health research. The similarities and differences between the two are compared using mathematical concepts and algorithms. The comparison between conventional statistics and machine learning methods indicates that conventional statistics are the fundamental basis of machine learning, where the black box algorithms are derived from basic mathematics, but are advanced in terms of automated analysis, handling big data and providing interactive visualizations. While the nature of both these methods are different, they are conceptually similar. Based on our review, we conclude that conventional statistics and machine learning are best to be integrated to develop automated data analysis tools. We also strongly believe that machine learning could be explored by health researchers to enhance conventional statistics in decision making for added reliable validation measures.

Relative Risk of Cardiovascular Mortality in Breast Cancer Patients: A Population-Based Study

Aims

To investigate the risk of cardiovascular disease (CVD) mortality in breast cancer patients compared with the general female population.

Methods

Data was retrieved from the Surveillance, Epidemiology, and End Results database. 924,439 female breast cancer patients who were at the age of follow-up ≥ 30 years and diagnosed during 1990-2016 as well as the aggregated general female population in the US were included. Using multivariable Poisson regression, we calculated incidence rate ratios (IRRs) of CVD mortality among female breast cancer patients compared with the female population.

Results

The risk of CVD mortality was mildly increased among breast cancer patients at the age of follow-up 30-64 years (IRR 1.06, 95% confidence interval [CI] 1.03-1.10) compared with the general population. This growth of risk reached its peak within the first month after diagnosis (IRR 3.33, 95% CI 2.84-3.91) and was mainly activated by diseases of the heart (IRR 1.11, 95% CI 1.07-1.15). The elevation was greatest in survivors at the age of follow up 30-34 years (IRR 3.50, 95% CI 1.75-7.01).

Conclusions

Clinicians should provide risk mitigation strategies with early monitoring of CVD mortality for breast cancer survivors, especially those who were young or with aggressive tumor stage.

Cardiovascular mortality among cancer survivors who developed breast cancer as a second primary malignancy

BACKGROUND

To assess the risk of cardiovascular mortality among cancer survivors who developed breast cancer as a second malignancy (BCa-2) compared with patients with first primary breast cancer (BCa-1) and the general population.

METHODS

Using the Surveillance, Epidemiology, and End Results database, we conducted a population-based cohort study including 1,024,047 BCa-1 and 41,744 BCa-2 patients diagnosed from the age 30 between 1975 and 2016, and the corresponding US female population (994,415,911 person-years; 5,403,551 cardiovascular deaths). Compared with the general population and BCa-1 patients, we calculated incidence rate ratios (IRRs) of cardiovascular deaths among BCa-2 patients using Poisson regression. To adjust for unmeasured confounders, we performed a nested, case-crossover analysis among BCa-2 patients who died from cardiovascular disease.

RESULTS

Although BCa-2 patients had a mildly increased risk of cardiovascular mortality compared with the population (IRR 1.08) and BCa-1 patients (IRR 1.15), the association was pronounced among individuals aged 30-49 years (BCa-2 vs. population: IRR 6.61; BCa-2 vs. BCa-1: IRR 3.03). The risk elevation was greatest within the first month after diagnosis, compared with the population, but comparable with BCa-1 patients. The case-crossover analysis confirmed these results.

CONCLUSION

Our findings suggest that patients with BCa-2 are at increased risk of cardiovascular mortality.