Crowd-sourcing inspiration: Data science contest asks entrants to diagnose lung cancer

William Doss | | Radiology Business | Imaging Informatics

A $1 million data science contest organized by Booz Allen Hamilton asked entrants to design deep-learning algorithms to identify lung cancer using just 2,000 images—a small data set in the machine learning world. While the winning entry won’t necessarily be used in clinical settings, the contest highlights the potential for crowd-sourcing inspiration.  

The winning team taught a neural network to first identify nodules in low-dose CT images from the National Cancer Institute before diagnosing cancer.

“We think that explicitly dividing this problem into two stages is critical, which seems also to be what human experts would do,” said Zhe Li, a member of the winning team and a student at China’s Tsinghua University, in an interview with MIT Technology Review.

The winning algorithms are available for free online, allowing medical researchers and data scientists to draw from minds outside the industry—what better way to think outside of the (black) box?

Read the full story here:

Million-Dollar Prize Hints at How Machine Learning May Someday Spot Cancer

William Doss

As a Senior Writer for TriMed Media Group, Will covers radiology practice improvement, policy, and finance. He lives in Chicago and holds a bachelor’s degree in Life Science Communication and Global Health from the University of Wisconsin-Madison. He previously worked as a media specialist for the UW School of Medicine and Public Health. Outside of work you might see him at one of the many live music venues in Chicago or walking his dog Holly around Lakeview.

Related Content

ChatGPT's medical advice may be deterring women from necessary imaging
'RadGPT' system improves radiology report readability and is ready for 'immediate' use
What radiologists think about using ChatGPT and AI in breast imaging
Newly proposed cybersecurity rule could cripple private radiology practices, RBMA says
AI is revolutionizing radiology workflow and patient care
Konica Minolta and ImagineSoftware partner to expand revenue cycle management offerings

Around the web

Cardiovascular Business
FDA-cleared software increases throughput for flurpiridaz back-to-back rest-stress imaging

The new F-18 flurpiridaz radiotracer is expected to help drive cardiac PET growth, but it requires waiting between rest and stress scans. Software from MultiFunctional Imaging can help care teams combat that problem.

Health Imaging
MRI projectile incident likely to result in a 'seven-figure sum' repair bill for hospital

News of an incident is a stark reminder that healthcare workers and patients aren’t the only ones who need to be aware around MRI suites.

Health Imaging
American College of Radiology announces significant expansion of Lung Cancer Screening Registry

The ACR hopes these changes, including the addition of diagnostic performance feedback, will help reduce the number of patients with incidental nodules lost to follow-up each year.