Publications by authors named "F A Minhas"
In computational pathology, labels are typically available only at the whole slide image (WSI) or patient level, necessitating weakly supervised learning methods that aggregate patch-level features or predictions to produce WSI-level scores for clinically significant tasks such as cancer subtype classification or survival analysis. However, existing approaches lack a theoretically grounded framework to capture the holistic distributional differences between the patch sets within WSIs, limiting their ability to accurately and comprehensively model the underlying pathology. To address this limitation, we introduce HistoKernel, a novel WSI-level Maximum Mean Discrepancy (MMD) kernel designed to quantify distributional similarity between WSIs using their local feature representation.
View Article and Find Full Text PDFArticle Synopsis
- * This study utilized deep learning to analyze ITH in a large sample of early-stage luminal breast cancer by assessing morphological features from whole slide images of tissue samples.
- * Findings showed that higher ITH correlates with more aggressive tumor traits (like larger size and low estrogen receptor expression) and can independently predict worse patient outcomes.
Despite the existence of established standards and guidelines for pathology reporting, many pathology reports are still written in unstructured free text. Extracting information from these reports and formatting it according to a standard is crucial for consistent interpretation. Automated information extraction from unstructured pathology reports is a challenging task, as it requires accurately interpreting medical terminologies and context-dependent details.
View Article and Find Full Text PDFArticle Synopsis
- The study introduces a method to predict protein expression in tissues using standard H&E stained images, aiming to enhance understanding of diseases like cancer and improve precision medicine outcomes.
- The authors develop a framework called Ouroboros that generates H&E images from protein profiles and vice versa, highlighting spatial changes in glioblastoma samples.
- Validation of this method with extensive data shows significant improvements over previous techniques in predicting protein expression and generating virtual images, suggesting potential for better diagnostic and therapeutic decisions.
Background: Emotional problems in adolescents living in low-income and middle-income countries (LMICs) remain largely unaddressed; key reasons include a scarcity of trained mental health professionals and unavailability of evidence-based, scalable psychological interventions. We aimed to evaluate the effectiveness of a non-specialist-delivered, group psychological intervention to reduce psychosocial distress in school-going adolescents in Pakistan.
Methods: In a two-arm, single-blind, cluster randomised controlled trial, eligible public school clusters from a rural subdistrict of Gujar Khan, Rawalpindi, Pakistan, were randomised (1:1, stratified by sex) using permuted block randomisation into intervention (n=20) and wait-list control (n=20) groups.