Exploration of sialic acid receptors as a potential target for cancer treatment: A comprehensive review.

The potential to target anticancer drugs directly to cancer cells is the most difficult challenge in the current scenario. Progressive works are being done on multifarious receptors and are on the horizon, expected to facilitate tailored treatment for cancer. Among several receptors, one is the sialic acid (SA) receptor by which cancer cells can be targeted directly as hyper sialylation is one of the most distinguishing characteristics of cancer cells. SA receptors have shown tremendous potential for tumor targeting because of their elevated expression in a range of human malignancies including prostate, breast, gastric cells, myeloid leukemia, liver, etc. This article reviews the overexpression of SA receptors in various tumors and diverse strategies for targeting these receptors to deliver drugs, enzymes, and genes for therapeutic applications. It also summarizes the diagnostic applications of SA-grafted nanoparticles for imaging various SA-overexpressing cancer cells and technological advances that are propelling sialic acid to the forefront of cancer therapy.

RadBERT-CL: Factually-Aware Contrastive Learning For Radiology Report Classification.

Radiology reports are unstructured and contain the imaging findings and corresponding diagnoses transcribed by radiologists which include clinical facts and negated and/or uncertain statements. Extracting pathologic findings and diagnoses from radiology reports is important for quality control, population health, and monitoring of disease progress. Existing works, primarily rely either on rule-based systems or transformer-based pre-trained model fine-tuning, but could not take the factual and uncertain information into consideration, and therefore generate false positive outputs. In this work, we introduce three sedulous augmentation techniques which retain factual and critical information while generating augmentations for contrastive learning. We introduce RadBERT-CL, which fuses these information into BlueBert via a self-supervised contrastive loss. Our experiments on MIMIC-CXR show superior performance of RadBERT-CL on fine-tuning for multi-class, multi-label report classification. We illustrate that when few labeled data are available, RadBERT-CL outperforms conventional SOTA transformers (BERT/BlueBert) by significantly larger margins (6-11%). We also show that the representations learned by RadBERT-CL can capture critical medical information in the latent space.