ABSTRACT
Artificial Intelligence (AI)-based image analysis has immense potential to support diagnostic histopathology, including cancer diagnostics. However, developing supervised AI methods requires large-scale annotated datasets. A potentially powerful solution is to augment training data with synthetic data. Latent diffusion models, which can generate high-quality, diverse synthetic images, are promising. However, the most common implementations rely on detailed textual descriptions, which are not generally available in this domain. This work proposes a method that constructs structured textual prompts from automatically extracted image features. We experiment with the PCam dataset, composed of tissue patches only loosely annotated as healthy or cancerous. We show that including image-derived features in the prompt, as opposed to only healthy and cancerous labels, improves the Fréchet Inception Distance (FID) by 88.6. We also show that pathologists find it challenging to detect synthetic images, with a median sensitivity/specificity of 0.55/0.55. Finally, we show that synthetic data effectively train AI models.
ABSTRACT
COVID-19 pandemic situation has affected millions of people with tens of thousands of deaths worldwide. Despite all efforts for finding drugs or vaccines, the key role for the survival of patients is still related to the immune system. Therefore, improving the efficacy and the functionality of the immune system of COVID-19 patients is very crucial. The potential new, non-invasive, FDA-approved biophysical technology that could be considered in this regard is tumor treating fields (TTFields) based on an alternating electric field has great biological effects. TTFields have significant effects in improving the functionality of dendritic cell, and cytotoxic T-cells, and these cells have a major role in defense against viral infection. Hence, applying TTFields could help COVID-19 patients against infection. Additionally, TTFields can reduce viral genomic replication, by reducing the expressions of some of the vital members of DNA replication complex genes from the minichromosome maintenance family (MCMs). These genes not only are involved in DNA replication but it has also been proven that they have a crucial role in viral replication. Also, TTFields suppress the formation of the network of tunneling nanotubes (TNTs) which is knows as filamentous (F)-actin-rich tubular structures. TNTs have a critical role in promoting the spread of viruses through improving viral entry and acting as a protective agent for viral components from immune cells and even pharmaceuticals. Moreover, TTFields enhance autophagy which leads to apoptosis of virally infected cells. Thus, it can be speculated that using TTFields may prove to be a promising approach as a subsidiary treatment of COVID-19.