
Sanjay Juneja: AI Moves from Theory to the Clinic Floor – And Why It’s Key to Safer, Saster, More Equitable Cancer Care
Sanjay Juneja, Vice President of Clinical AI Operations at Tempus AI, shared on LinkedIn:
“With oncology, AI has evolved from fascinating to functional to, now? Foundational. And knowing the key differences in LLM architecture is crucial if considering investing in AI. Let me explain.
‘Does your AI solution use function calling, or tooling?’
Function Calling is like a helpful assistant: great at narrow, rule-based tasks–filling out a prior auth for a breast MRI? Requires an AI to simply know the rules, the CPT codes, and what to pull from the EMR. Queue: populate form.
Tool Use, on the other hand, is where AI goes from ‘autocomplete on steroids’… to an adaptive agent.
Tool-using AI can browse the web, perform vector search, retrieve guidelines, integrate with diagnostic tools, and reason across UNcertain inputs. And then? ‘reason’ or ‘decide’ what to do next. Retry when gotten to an ‘error’.
Example: A patient with a new lung cancer.
Instead of one input, we’re juggling:
- Imaging data (CT, PET)
- Pathology results
- Clinical history
- Labs
- Staging criteria
A tool-using agent can synthesize this multi-modal data and help clinicians make complex decisions: What lymph nodes look sketchy–N1 or N2? What does that mean in terms of staging? Okay, now staged–is this resectable? Or does data show pre-op therapy is better?
It’s not just smart – it’s context-aware, and real-time adaptive.
Why It Matters:
- it learns by DOING (like medical trainees)
- is both modular and scalable (don’t you want to ‘plug in’ dig path slide analysis when you can, rather than just read reports?)
- Works in ambiguity (aka all of medicine)
Most bold envisionment?
- Bring specialist-level triage to underserved settings and deserts!
This is how AI moves from theory to the clinic floor – and why it’s key to safer, faster, more equitable cancer care. And needless to say, its every bit what we’re pursuing at Tempus.
Poll time: Were you considering this evolution in LLM architecture with your AI ‘shopping’?”
More posts featuring Sanjay Juneja.
-
Challenging the Status Quo in Colorectal Cancer 2024
December 6-8, 2024
-
ESMO 2024 Congress
September 13-17, 2024
-
ASCO Annual Meeting
May 30 - June 4, 2024
-
Yvonne Award 2024
May 31, 2024
-
OncoThon 2024, Online
Feb. 15, 2024
-
Global Summit on War & Cancer 2023, Online
Dec. 14-16, 2023