A radiologist’s battle with a taunting pixel: I think I won, for now.
I stared at the gray dot on my monitor. Actually, I was staring at a number of dots around it, too, but this particular pixel was the troublemaker. It was a data-point which tipped the balance from one diagnostic consideration to another. Fluid simple versus complex, structure-wall thin or thick, cortex or blood, etc. More generally speaking, interpretative confidence rendered uncertain.
Somehow, I got the distinct impression the pixel was smirking at me. Did it just flip me off? I must be overdue for a break. It was an ungodly hour, after all, and I’d lost count of how many cases I’d read thus far. Come off it, I told myself; it’s probably just an artifact. Blow it off and move on.
“Go ahead,” the pixel sneered. “I dare you.”
Yep…definitely overdue for some coffee. I went and got myself a cup of mud. When I returned, though, the pixel hadn’t miraculously vanished. Nor had the impish persona with which my beleaguered mind had imbued it: “Well? Are you going to deal with me now, or are we going to look at each other some more?”
“I’m not afraid of you,” I said. What the hey, I work from home. The only individual who might notice me acting crazy, talking to a computer screen, is my cat, and even she was asleep.
“Oh, suuuuure you’re not,” the pixel retorted. “Confident, assertive radiologist you are. Let’s go back and count how many weasel-words and hedges you’ve put in your reports so far tonight. What’s a good over/under for that-100? 200?”
“Imaging doesn’t always give a conclusive answer,” I half-pleaded as if the pixel would see my side of things and gracefully go away. “I’m not doing patients any favors if I pretend otherwise.”
“Oh, so you think I might be ‘real’ pathology then? Go on, say so in your report. Better yet, call up the referring doc and tell him he’s got a surgical emergency on his hands because of me.”
I glared at the pixel. “I’m not doing that. You know as well as I do that you’re a diagnostic fake out.”
“Put your dictaphone where your mouth is, then,” it taunted. “Don’t even mention me in your report. I’m sure you won’t be hearing about it from the surgeon tomorrow. Or the peer-review committee. Or some guy handing you malpractice-claim papers in a couple of years.”
I reached for the phone. The pixel was astonished. “You-you’re really going to call me in as a positive case? Wow, even I didn’t expect that!” I ignored it as I got on the line with the technologist and asked for some additional images to be sent my way.
Call completed, I turned back to my antagonist. It was not amused. “Think that’s going to resolve the issue? How do you know I won’t be on the next set of images, too? Or you’ll see one of my buddies, giving you a whole new problem?”
But the pixel’s spell over me was broken, at least for the time being, and it was banished from my screen with a click of the “next case” button.
Burnout in Radiology: Key Risk Factors and Promising Solutions
June 9th 2025Recognizing the daunting combination of increasing imaging volume and workforce shortages, these authors discuss key risk factors contributing to burnout and moral injury in radiology, and potential solutions to help preserve well-being among radiologists.
Mammography AI Platform for Five-Year Breast Cancer Risk Prediction Gets FDA De Novo Authorization
June 2nd 2025Through AI recognition of subtle patterns in breast tissue on screening mammograms, the Clairity Breast software reportedly provides validated risk scoring for predicting one’s five-year risk of breast cancer.
The Reading Room: Artificial Intelligence: What RSNA 2020 Offered, and What 2021 Could Bring
December 5th 2020Nina Kottler, M.D., chief medical officer of AI at Radiology Partners, discusses, during RSNA 2020, what new developments the annual meeting provided about these technologies, sessions to access, and what to expect in the coming year.
Can Emerging AI Software Offer Detection of CAD on CCTA on Par with Radiologists?
May 14th 2025In a study involving over 1,000 patients who had coronary computed tomography angiography (CCTA) exams, AI software demonstrated a 90 percent AUC for assessments of cases > CAD-RADS 3 and 4A and had a 98 percent NPV for obstructive coronary artery disease.
Could AI-Powered Abbreviated MRI Reinvent Detection for Structural Abnormalities of the Knee?
April 24th 2025Employing deep learning image reconstruction, parallel imaging and multi-slice acceleration in a sub-five-minute 3T knee MRI, researchers noted 100 percent sensitivity and 99 percent specificity for anterior cruciate ligament (ACL) tears.