{Reference Type}: Journal Article
{Title}: CNN-based deep learning method for predicting the disease response to the Neoadjuvant Chemotherapy (NAC) treatment in breast cancer.
{Author}: Kirelli Y;Arslankaya S;Koçer HB;Harmantepe T;
{Journal}: Heliyon
{Volume}: 9
{Issue}: 6
{Year}: 2023 Jun
{Factor}: 3.776
{DOI}: 10.1016/j.heliyon.2023.e16812
{Abstract}: UNASSIGNED: The objective of the study is to evaluate the performance of CNN-based proposed models for predicting patients' response to NAC treatment and the disease development process in the pathological area. The study aims to determine the main criteria that affect the model's success during training, such as the number of convolutional layers, dataset quality and depended variable.
UNASSIGNED: The study uses pathological data frequently used in the healthcare industry to evaluate the proposed CNN-based models. The researchers analyze the classification performances of the models and evaluate their success during training.
UNASSIGNED: The study shows that using deep learning methods, particularly CNN models, can offer strong feature representation and lead to accurate predictions of patients' response to NAC treatment and the disease development process in the pathological area. A model that predicts 'miller coefficient', 'tumor lymph node value', 'complete response in both tumor and axilla' values with high accuracy, which is considered to be effective in achieving complete response to treatment, has been created. Estimation performance metrics have been obtained as 87%, 77% and 91%, respectively.
UNASSIGNED: The study concludes that interpreting pathological test results with deep learning methods is an effective way of determining the correct diagnosis and treatment method, as well as the prognosis follow-up of the patient. It provides clinicians with a solution to a large extent, particularly in the case of large, heterogeneous datasets that can be challenging to manage with traditional methods. The study suggests that using machine learning and deep learning methods can significantly improve the performance of interpreting and managing healthcare data.