Publications by authors named "Maham Siddique"
Breast cancer remains the leading cause of cancer-related deaths in women worldwide. Current screening regimens and clinical breast cancer risk assessment models use risk factors such as demographics and patient history to guide policy and assess risk. Applications of artificial intelligence methods (AI) such as deep learning (DL) and convolutional neural networks (CNNs) to evaluate individual patient information and imaging showed promise as personalized risk models.
View Article and Find Full Text PDFObjective: To investigate the ability of our convolutional neural network (CNN) to predict axillary lymph node metastasis using primary breast cancer ultrasound (US) images.
Methods: In this IRB-approved study, 338 US images (two orthogonal images) from 169 patients from 1/2014-12/2016 were used. Suspicious lymph nodes were seen on US and patients subsequently underwent core-biopsy.
Article Synopsis
- A study evaluated a weakly supervised deep learning model for classifying breast MRI lesions, aiming to improve specificity without requiring detailed pixel-level segmentation.
- The dataset included 278,685 image slices from 438 patients, and the Resnet-101 architecture was utilized for training, resulting in a classifier with high accuracy.
- The model achieved an AUC of 0.92 and a classification accuracy of 94.2%, demonstrating that this approach effectively distinguishes between benign and malignant MRI images.
Objective: To determine the relationship between two documented indicators of tumor aggressiveness, SUV and volume doubling time (VDT) for stage I non-small cell lung cancer (NSCLC).
Methods: 116 pathology proven solid NSCLC patients with 2 pretreatment CT and 1 PET/CT scan were retrospectively identified. The 2 CT scans were at least 85 days apart.
Article Synopsis
- * The CNN consisted of multiple convolutional and residual layers, implemented with dropout and L2 normalization, and trained using the Adam optimizer, achieving a diagnostic accuracy of 72.5% in classifying patients based on their chemotherapy response.
- * Results showed that the model provided a reasonable prediction capability, making it feasible to utilize CNN algorithms for predicting NAC response across multiple institutions in breast cancer patients.
Article Synopsis
- A convolutional neural network (CNN)-based algorithm developed to distinguish atypical ductal hyperplasia (ADH) from ductal carcinoma in situ (DCIS) is validated using a new dataset of 280 mammographic images from 140 patients.
- The study involved a rigorous analysis of these images, utilizing advanced CNN techniques and standard metrics to assess diagnostic performance, focusing on sensitivity, specificity, and accuracy.
- Results showed the algorithm achieved a high area under the curve (0.90), with diagnostic accuracy at 80.7%, sensitivity at 63.9%, and specificity at 93.7%, confirming its effectiveness on unseen data.