Deep Learning V.S. Cancer

Hello all!

Welcome to the new Lazarus website. As some of might have seen, we’ve launched Prophet, our cervical cancer detection app. Prophet uses a deceptively simple 37 question survey to detect an individual’s risk of cervical cancer: all without any physical input. If you haven’t checked it out, you can try it here. It’s free until December 10th, so try it out until then.

Meanwhile, let’s kick off this shiny new site with a shiny new blog post. Today’s topic revolves around different applications of deep learning towards the fight against cancer.

 
image1.png
 

More Effective Imaging

Google’s team has recently been hard at work using deep neural networks to improve the accuracy of metastatic breast cancer detection. By leveraging LYNA, they’ve been able to create a system that has significantly more precision than the average pathologist, at over 99% accuracy. Google’s team has made the gold standard in imaging tech.

 
 
 

Quicker Drug Discovery

Two months ago, Atomwise entered into an evaluation agreement with Pfizer to test its product in their supply chain. Atomwise is a startup that uses deep learning to improve the speed and efficacy of drug discovery. By improving the speed of drug discovery, we can find more effective medications to combat tumours more quickly.

 
 
 
Screen Shot 2018-11-19 at 1.32.42 AM.png
 
 

Predicting at Risk Patients

Last week, the Prophet beta formally began. By leveraging over 47,000 different deep learning models, Prophet uses a patient’s data to predict the onset of cervical cancer. A current private beta for breast cancer is being internally run at select hospitals. Early detection increases the survival rate of cervical cancer to over 93%, from 16% at IV.

 
 

This week we’ll hopefully release something in addition to this post. A reminder for those of you who haven’t signed up for Prophet, go and do it here. It’s free and it might save your life or the lives of your patients.

Thanks for coming with us on this ride,

-Ariel