Denisa Simona Zai

TSRM | Fondazione Monasterio
Comitato Scientifico AITeRTC ETS

[Denisa Simona Zai @dszai93]

Il Noise Power Spectrum è un indicatore avanzato, che analizza il modo in cui il rumore si distribuisce alle diverse frequenze spaziali, definendone una vera e propria "texture" visiva

News completa qui: https://t.co/AAhJKmmAOC

#tsrm #radiology #radiographer

[Denisa Simona Zai @dszai93]

Guida alla caratterizzazione tissutale e differenze di genere nel cuore ipertrofico: il contributo del TSRM https://t.co/X36MB5Mw3B

[Denisa Simona Zai @dszai93]

L'integrazione tra sistemi di automazione e ricostruzioni deep learning possono trasformare il modo con il quale ci approcciamo, quotidianamente, all'ottimizzazione della dose in tomografia computerizzata
News completa: https://t.co/AOHQgY0Q9u

#tsrm #radiographer #radiology

[Denisa Simona Zai @dszai93]

La Fifth Universal Definition of Myocardial Infarction (UDMI 2026) della European Society of Cardiology, pone imaging cardiovascolare e skills del TSRM al centro del percorso diagnostico dell'infarto miocardico
https://t.co/Cr6vKN3ZuH

#ESC #UDMI2026 #escardio #TSRM #PCCT

[Denisa Simona Zai @dszai93]

La precisione nel posizionamento e nel centraggio del paziente può condizionare enormemente la qualità dell’esame in radiodiagnostica.

News completa: https://t.co/wb6BVcaz2N

#tsrm #radiology #radiographer

[Denisa Simona Zai @dszai93]

Possono, le più recenti innovazioni tecnologiche, ridefinire la sicurezza del paziente senza sacrificare l'accuratezza diagnostica?

News completa: https://t.co/BKxzEHSqQx

#tsrm #CardiovascularImaging #PCCT

[Denisa Simona Zai @dszai93]

Per la fisica delle particelle, un miliardesimo di millimetro fa tutta la differenza del mondo. E in radiodiagnostica?
News completa: https://t.co/AEqN3FMa7o

#fermilab #infn #quantumphysics #radiology #radiographer

[Denisa Simona Zai @dszai93]

La Photon Counting CT sta ridefinendo i confini della diagnostica per immagini
News completa: https://t.co/3c8wAw3UkM

#pcct #radiology

[Denisa Simona Zai @dszai93]

Valutare cuore, carotidi e polmoni con un solo protocollo di acquisizione CT, mantenendo una qualità d'immagine elevata senza aumentare la dose al paziente: lo studio ACTA

News completa: https://t.co/3OzpBVTKMP

#tsrm #TomografiaComputerizzata #radiographer #screening #PCCT

[Denisa Simona Zai @dszai93]

Ridurre gli artefatti stair-step: ZeeFree è un algoritmo avanzato, progettato per riallineare il dataset, attraverso l'interpolazione e il ricampionamento virtuale di ogni singolo stack

News completa: https://t.co/UVzYfreFU4

#tsrm #computedtomography #radiographer #radiology

[Denisa Simona Zai @dszai93]

Characterizing a myocardial scar without administering a single drop of gadolinium?

https://t.co/i2YF03OBFY

[JACC Journals @JACCJournals]

In LVH, photon-counting #YesCCT provided #whyCMR level data on LV geometry, function, scar, & ECV. Late iodine enhancement closely matched CMR-LGE & CT-ECV tracked CMR-ECV, while also providing coronary CTA. https://t.co/j6y9PqVgk3

#JACCIMG @mauripieroni72 @MicheleEmdin

[Dr. Filippo Cademartiri @FCademartiri]

# AI in Radiology: If It Saves Time but Loses Money, Is It Really Efficient?

We often discuss AI in radiology as if **improved efficiency automatically means economic value**.

This pragmatic analysis makes an important point: **it doesn't.**

The authors modeled three real-world AI applications—intracranial hemorrhage triage, pulmonary embolism triage, and breast cancer detection—including implementation costs, radiologist time, reimbursement, and downstream revenues. Efficiency and Financial Gains From Artificial Intelligence Algorithm Implementation.pdf

The results are surprisingly different.

For intracranial hemorrhage, AI saved **1.15 min/case** and produced a positive **63.3% ROIC**.

For pulmonary embolism, AI saved **0.83 min/case**—but still produced a **−76.6% ROIC**. It would need to save ~3.55 minutes per case, or cost only ~$47,000/year instead of $200,000, to become financially favorable.

And mammography tells an even more interesting story.

AI actually **reduced interpretation profitability**, with a −194% ROIC based on reading alone. But once the downstream value of **35 additional cancers detected** was included, ROIC became **+274.8%**.

## The Critical Point
This exposes a major weakness in how we currently evaluate radiology AI.
**Not every algorithm creates value in the same way.**

Saving seconds is not necessarily valuable enough to justify another expensive, single-purpose AI tool.

And detecting more disease may generate enormous value—but only if the healthcare system actually captures that downstream value.

The authors also acknowledge that their model does **not** measure diagnostic accuracy, patient outcomes, radiologist satisfaction, or burnout. Efficiency and Financial Gains From Artificial Intelligence Algorithm Implementation.pdf

## My Take
This is exactly why the current model of dozens of separate AI applications, each solving one small task and each carrying its own licensing and integration costs, may ultimately be economically unsustainable.

The future cannot simply be:
**more algorithms = better radiology.**

It must be **integrated intelligence that creates measurable clinical value across the entire imaging pathway.**

And this becomes even more relevant with **Photon-Counting CT**.

PCCT will generate increasingly complex anatomical, spectral, quantitative, and biological information. Adding one separate algorithm for every possible output is probably not the answer.

The real opportunity is an integrated AI layer capable of turning that enormous PCCT dataset into **better diagnosis, better workflow, and better patient management—all within the same ecosystem.**

Because the question is no longer:
**“Does the AI work?”**

It is:
**“Does it create enough clinical value to justify what it costs?”**

#ArtificialIntelligence #Radiology #HealthcareAI #PhotonCountingCT #PCCT #MedicalImaging #HealthcareEconomics #DigitalHealth

[Denisa Simona Zai @dszai93]

It was a huge honor to participate as speaker in the first event in the #CardiacCT webinar series organized by Associazione Italiana TSRM in
Radiodiagnostica e TC. An event packed with incredibly interesting content
#tsrm #education #CT #radiology

[Denisa Simona Zai @dszai93]

Dose beyond the reconstructed volume: unseen yet measurable: https://t.co/GYO8hJkBdM

#xray #radiology #computedtomography #radtech

[Denisa Simona Zai @dszai93]

Set your priorities between radiation protection, emerging technologies, technical optimization, and professional liability.
Complete the survey ➡️ https://t.co/iDF9mqaXF4

#radiographers #TSRM
#radiologytechnologist #RadiationProtection #EmergingTechnologies

[Denisa Simona Zai @dszai93]

Dalla materia al pixel: Tecnica, AI e Qualità per l’Imaging

https://t.co/lkbdLlbxWC

#ECM #TSRM #AI #imaging #Tech

[Denisa Simona Zai @dszai93]

CTDI shows how the scanner works; on the other hand, the scan range is representative of technical - diagnostic appropriateness. Optimizing it reflects the radiographer’s skills and operational accuracy

https://t.co/42zng8Of7U

#Radiology #RadiationProtection #TSRM #CT

[Denisa Simona Zai @dszai93]

#PhotonCountingCT #memes #radiology #TSRM

[Denisa Simona Zai @dszai93]

📚Formazione AlTeRTC FAD asincrona creata dal 2° Congresso Nazionale 2025 🎯27 ECM
🔗 https://t.co/nrK4wK1cV7

#ECM #TSRM #radiology #FormazioneContinua #AITeRTC