Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj) header
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj) profile picture

Abdo/Cancer Imaging & Deputy Lead for Faculty Development @Imagingtoronto, Vice Chair EDI @UofTMedIm & Asst. Prof @UofT, 🇮🇳President 🏅.Proud husband.🐇Pa

Toronto, Ontario
Joined July 2009
Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)Photo by Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ (@ankushjaj)
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
So @nyulangone built @OpenEvidence into its Epic record systemwide, and now clinicians can ask clinical questions in plain language from inside a patient's chart. Northwestern Medicine did something similar earlier. Clinical-knowledge AI is moving from a separate app into systems clinicians are already using. Could this mean that the crappy 2006-UI EHRs I'm constantly forced to use become less crappy? Let's hope so. buff.ly/bFMzKnt #ClinicalAI
10
19
2
161
33,974
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
They're calling it the greatest opening paragraph of any recent oncology article
33
273
16
2,291
95,672
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Fun fact: @AbdulElSayed will have the highest h-index of any US Senator ever: 37. (Pictured: AJR article that was one of several collaborations with @WBrinjikji, then radiology resident.)
10
45
1
380
24,175
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Proud to announce my latest addition to the RADS Classification Scheme, standardising our approach to colleagues, reducing ambiguity, informing referring clinicians COCK-RADS™ Whilst lacking true empirical data, I feel it would happily sit within the @RadiologyACR’s scheme
14
19
4
129
18,851
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Epidural lipomatosis Top: before GLP-1 agonist Bottom: after GLP-1 agonist ajronline.org/doi/epdf/10.22…
17
38
3
280
20,233
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Cancer therapy has genotoxicity, as reflected by mutational signatures. We need to know more about the "genomic afterlife" of people treated with cancer. "For long-term cancer survivors, the biological consequences of treatment do not end after the original cancer is cured." @jclinicalinvest jci.org/articles/view/213544
28
62
13
233
27,874
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
In medicine, there are two schools of thought. "I had it bad in my training. Now it's your turn to suffer.” Or “I had it bad in my training. I want things to be better and will advocate for you.” Don't become part of the problem. Take the path less traveled.
24
54
10
332
17,729
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Thanks to Sam Ajmera , another Indo- Canadian making big donation . Sam is originally from Bombay and founded FGF brands . Successful immigrants create jobs for Canadians and support 🇨🇦 philanthropically . Thank an immigrant .
6
51
2
286
11,672
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Open your eyes. If you have trouble recognizing typical AI writing, look at these 10 patterns (from the detector GPTZero).
🤖 Made with AI
4
6
1
28
5,483
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
In this article, radiologists from different countries share compassionate data-driven strategies and action plans to address individual reading inefficiency and maintain equity in a large practice. ajronline.org/doi/10.2214/AJ…
3
16
2,199
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
My review... Overall it's the most impressive general CT model I've seen so far. BUT! I would say it's still not very good. Maybe at the level of a first year radiology resident. I tested it on a few easy cases and in each case the model has hallucinated findings (e.g. lung nodule that doesn't exist, radiation fibrosis that doesn't exist, foley catheter in the urinary bladder on a CT chest that did not include the urinary bladder) or missed findings. See example image below.
70,111 CT scans trained this model. one NVIDIA GPU turned a full chest CT into a radiology report in under 30 seconds 🔥 top finding: a spiculated nodule, right upper lobe. the expert's tumor marking it never saw? right lung. open weights from @nvidia on @huggingface 🤗
2
6
1
85
14,559
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
The invisible health risks of AI, and how to address them nature.com/articles/s44360-0…
19
117
3
367
41,797
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
I'm giving away 25 copies of #VIBE. To enter, repost and comment by October 13. adamgrant.net/vibe
607
816
128
1,542
150,854
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
We are incredibly grateful for the generosity of the anonymous donor that has contributed the single largest gift for cancer care and research in Canadian history. A truly inspirational launch for #CarryTheFire, our campaign to create a world free from the fear of cancer.
Every major breakthrough at UHN's @pmcancercentre has been powered by philanthropy. Now, in this new "golden age" of cancer research, care and education, the Princess Margaret Cancer Foundation is launching Canada's largest cancer fundraising campaign: $2.5B to help create a world free from the fear of cancer. thestar.com/news/gta/princes…
6
1
36
1,732
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Proud to host PM @MarkJCarney, Diana Fox Carney and @EvanLSolomon. UHN is excited to partner in shaping the future of AI-enabled health care .
5
20
4
63
6,753
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Just remember every time you hear an ad for "whole-body MRI" it's based on a pack of lies and against the clear recommendations of Canada's credible radiologists. car.ca/wp-content/uploads/20…
4
41
5
105
9,182
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
At @JAMA_current today, 2 radiologists publish what should be the consent form for a total body MRI in healthy people Note: "no major medical society recommends whole-body MRI screening in the general population because it is unproven, and the harms likely outweigh the benefits." jamanetwork.com/journals/jam…
48
253
31
939
159,234
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
Terrific @MontefioreRad Grand Rounds on the controversial topic of “For-Profit Whole Body MRI” by Dr. Matthew Davenport @UMichRadiology @MontefioreNYC @EinsteinMed
6
1
34
3,381
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
AI cuts MRI scan time from 23 to 9 minutes at Amsterdam cancer center. “We are already seeing that we can perform 18 more examinations in a week than before[...] We can also take more time for each patient, and staff can take breaks more often.”
10
27
1
143
9,801
Ankush Jajodiaᅠᅠᅠᅠᅠᅠᅠᅠ retweeted
This week, the "AI replacing doctors" debate is back. The CEO of America's largest public hospital system says he's ready to replace radiologists with AI. The Stanford-Harvard NOHARM study shows top models outperforming generalists. The discourse is moving fast. I run AI at @UHN, the largest hospital in Canada. Here's what I actually see. We've developed AI models across imaging, pathology, and clinical decision support. In controlled conditions, the accuracy numbers are real. In some narrow tasks, models genuinely outperform. That's not hype. But the operational reality of running these systems inside a large hospital teaches you things benchmarks never will. The errors that hurt patients aren't the confident wrong answers. They're the quiet omissions, i.e., the thing the model didn't flag because it wasn't in the training distribution. NOHARM found 76.6% of AI errors were omissions. We see this too. And in a hospital, a missed finding doesn't just affect one case. It propagates: the downstream physician trusts the AI read, the patient waits, the window closes. The accountability structure also doesn't exist yet. When an AI-assisted diagnosis leads to harm, who is responsible: the physician, the hospital, the vendor? In Canada, we don't have a clear answer. No hospital system deploying AI at scale does. That's not a regulatory delay. That's a fundamental gap in the infrastructure for AI-in-medicine. What I'm genuinely optimistic about: AI is already changing how our radiologists work. Not replacing them, but changing the shape of the job. Routine reads get faster. Their time shifts toward complex cases, clinical correlation, cases where the AI flags uncertainty. That's the right direction. But "ready to replace radiologists" skips 10 hard years of work on deployment infrastructure, liability frameworks, clinician training, and failure mode monitoring that nobody wants to talk about because it's less exciting than accuracy benchmarks. The capability question is nearly answered. The deployment question has barely been asked. CEO story: beckershospitalreview.com/ra… NOHARM paper: arxiv.org/abs/2512.01241
33
82
17
306
64,349