Abstract:
Background: Breast cancer remains a leading cause of cancer-related death in women worldwide. Since 2019, female breast cancer has surpassed lung cancer as the most commonly diagnosed cancer, with an estimated 2.3 million new cases (11.7%), followed by lung cancer (11.4%). The presence of metastasis plays an essential role in affecting the prognosis of patients with breast cancer, but there is no clear understanding when it becomes metastatic. We want to examine what factors are associated with metastasis. Plenty of researches focus on cancer-specific factors to predict its metastasis along with patient demographics such as race and age. However, few researches consider patient-centered health factors, which also play a significant role in the likelihood of developing metastasis in breast cancer. The National Inpatient Sample (NIS) database, which provides de-identified patient admission data collected annually from a randomly selected 20% sample of hospital admissions across the United States, enables this analysis.
Hypothesis: We hypothesize that, compared to tumor-related factors, patient-centered factors in breast cancer, such as family cancer history, previous cancer treatments, overall health condition, and comorbidities, play an indispensable role in the development of metastasis. Ignoring these factors could lead to inaccurate identification of metastasis-associated factors and impact the accuracy of metastasis prediction.
Methods: The study included female patients diagnosed with breast cancer, with no exclusions applied. A total of 4296 female breast cancer patients were identified from the 2021 NIS, out of which 1691(39.36%) had metastasis and 2605 (60.64%) did not. A thorough review of 130 factors in the NIS was conducted, leading to the selection of 21 key factors for analysis. A logistic regression model integrating all these factors was developed to identify significant factors of metastasis and quantify their impact. The model was further refined using backward elimination and validated through cross-validation.
Results: The analysis revealed that APRDRG severity, malignant pleural effusion, cancer treatment, and anemia were significant risk factors for metastasis. Specifically, APRDRG severity was a strong predictor, with moderate (OR = 1.31), major (OR = 3.60), and extreme (OR = 5.16) levels vs. minor level (baseline) significantly increasing the risk (p < 0.05). Anti-neoplastic treatment (OR = 1.24, p < 0.05), anemia (OR = 1.32, p < 0.05), and malignant pleural effusion (OR = 1.86, p < 0.001) were also independently associated with a higher likelihood of metastasis. In contrast, factors associated with a reduced risk of metastasis included elective admission (OR = 0.29, p < 0.001), history of cancer (OR = 0.69, p < 0.05), and obesity (OR = 0.76, p < 0.05). The model demonstrated strong performance, achieving an AUC of 0.7556 and high sensitivity (0.8663). Cross-validation indicated consistent performance, with standard deviations < 0.05 and narrow 95% confidence intervals, confirming model reliability.
Conclusions: Our study offers valuable insights for physicians to guide personalized treatment strategies, allowing for intensified therapy in high-risk patients and maintenance treatment for those at lower risk, ultimately improving breast cancer patient outcomes.
