Evan Calabrese, MD PhD

@ecalabr

Neuroradiologist at Duke University Medical Center. Views are my own.

Joined June 2009

224 Following 2.8K Followers 1K Posts

Latest posts 5

Evan Calabrese, MD PhD @ecalabr Nov 3

When you use the 18G Quinke for an LP:

Oren Gottfried, MD Verified account @OGdukeneurosurg Nov 3

Cauda equina in the flesh

View quoted post on X

Evan Calabrese, MD PhD @ecalabr Oct 29

I find it so interesting that AI finds the finding I thought was tricky but completely fabricates a finding that is not there and could seriously harm the patient.

Some commenters made the same mistake, though they may also be using AI.

Grok Verified account @grok Oct 29

The nearly overlooked finding is the large T1-hyperintense mass in the pelvis (bottom of the FOV). Differential includes dermoid/teratoma (fat) or endometrioma (blood products).

Also present: low-lying conus/tethered cord (ends near L5-S1). Classic “check the four corners” case.
View quoted post on X

Evan Calabrese, MD PhD @ecalabr Oct 29

I have to admit that I nearly overlooked the major abnormality here...

Evan Calabrese, MD PhD @ecalabr Oct 21

Curious case. Have you seen one? According to some papers its impossible.

Evan Calabrese, MD PhD @ecalabr Oct 19

Classifying MRI series is essential but tricky, thanks to inconsistent protocols and naming. Ryan Fishman and I built an open-source AI GUI that uses DICOM headers and pixel data for semi- or fully-automated classification: github.com/ecalabr/mrid