RESUMO
Objective: Patients with breast cancer who have indeterminate extra-mammary lesions, for example in lung, liver or bone, without other metastatic lesions pose a clinical dilemma regarding subsequent management. This study aimed to investigate the prevalence, characteristics and outcomes of such lesions detected on initial staging imaging, and address the clinical significance of these incidental findings. Materials and Methods: Medical records of patients with newly diagnosed breast cancer who underwent computed tomography scans and bone scintigraphy between January 1, 2015 and June 30, 2021 were reviewed. Patients with indeterminate extra-mammary lesions on imaging were included. Patients with obvious metastatic disease were excluded. Lesion characteristics, breast cancer staging, duration of follow-up and natural history of disease progression were analysed. Results: The study included 52 patients with indeterminate lesions on pre-operative imaging. The median follow-up duration was 14 (range: 6-41) months. The most common site of occurrence of indeterminate lesions was the lung (60.9%) followed by the liver (26.1%). Forty-six had lesions that remained stable (88.5%), while six (11.5%) had progression to metastatic disease. Out of these six, only two (3.8%) developed metastasis in the same site as the original indeterminate lesion, whereas the remaining four developed metastases in other sites. Conclusion: Patients with breast malignancy found to have indeterminate extra-mammary lesions without obvious distant metastasis on initial staging scans are associated with a small risk of subsequently developing metastatic disease. Although most of these lesions remain quiescent, surveillance imaging is recommended because a small but significant proportion of patients with such lesions eventually harbour actual metastatic disease.
RESUMO
Credit scoring is an important tool used by financial institutions to correctly identify defaulters and non-defaulters. Support Vector Machines (SVM) and Random Forest (RF) are the Artificial Intelligence techniques that have been attracting interest due to their flexibility to account for various data patterns. Both are black-box models which are sensitive to hyperparameter settings. Feature selection can be performed on SVM to enable explanation with the reduced features, whereas feature importance computed by RF can be used for model explanation. The benefits of accuracy and interpretation allow for significant improvement in the area of credit risk and credit scoring. This paper proposes the use of Harmony Search (HS), to form a hybrid HS-SVM to perform feature selection and hyperparameter tuning simultaneously, and a hybrid HS-RF to tune the hyperparameters. A Modified HS (MHS) is also proposed with the main objective to achieve comparable results as the standard HS with a shorter computational time. MHS consists of four main modifications in the standard HS: (i) Elitism selection during memory consideration instead of random selection, (ii) dynamic exploration and exploitation operators in place of the original static operators, (iii) a self-adjusted bandwidth operator, and (iv) inclusion of additional termination criteria to reach faster convergence. Along with parallel computing, MHS effectively reduces the computational time of the proposed hybrid models. The proposed hybrid models are compared with standard statistical models across three different datasets commonly used in credit scoring studies. The computational results show that MHS-RF is most robust in terms of model performance, model explainability and computational time.
