Two frameworks answer two questions. SIMAC Digital Twin answers the third.
TG-40/142 and TG-100 have carried radiotherapy QA for thirty years, and neither is going anywhere. But between them they only answer two questions — and the third is the one physicists are actually asked in the morning huddle.
Measurement against a fixed baseline. Parameter by parameter. A snapshot in time.
Risk prioritised in advance, by expert judgement, before any measurement is taken.
Answering this requires a model of the machine itself, built from physics. That's SIMAC Digital Twin.
A model of the machine — not a fit to its data.
Most systems that call themselves a “digital twin” are an empirical curve fit to commissioning data — accurate near the conditions used to build them, untested everywhere else. SIMAC Digital Twin is built the opposite way.
Models the accelerator, not a curve
Takes into account all of the underlying accelerator physics — high-voltage supply, RF chain and waveguide, target, flattening filter, jaws and MLC, gantry and isocenter, dose calculation — rather than relying on empirical curve fitting.
Current by construction
Uses routine quality assurance measurements to automatically update and maintain beam accuracy throughout the machine's lifecycle — no new hardware, no extra beam time.
Full characterization, few measurements
Predicts complete beam characteristics from a limited set of measurements, reducing or eliminating the need for expensive physical water tank commissioning. Patent pending.
Closing the QA loop.
Plan, do, check, act — the field has always been strong on the first three. Act is where QA still runs on individual experience, and where the direct connection between QA data and machine state has been missing.
Measure
Routine daily, monthly and annual QA — output, symmetry, flatness, energy, MLC, isocenter — on the schedule you already run.
Organise & triage
Results normalised, attributed to machine and energy, trended, and screened for signal before they leave the database.
Model
The physics model re-fits its subsystem parameters to explain the current data — HV supply, RF chain, target, filter, jaws, MLC, geometry.
Act
A subsystem-level cause, a dose consequence in Gy, and a defensible answer to treat or delay — instead of a judgement call.
The loop then repeats. Every measurement makes the model more current; a more current model makes the next measurement more meaningful.
I've spent twenty-five years in radiotherapy QA. We got very good at answering whether a parameter moved. We never had a good answer to whether it mattered. That's not a process gap — it's a modeling gap. SIMAC Digital Twin is what closes it.
The same physics model, three more places it pays off.
Patient-specific QA
Gamma analysis passes the vast majority of plans and still misses clinically meaningful dose errors — TG-218 says so directly. SIMAC Dose Calculation recalculates the plan on today's machine state instead of comparing it to a phantom measurement.
§6.A in the white paperLife-cycle management
Return-to-service decisions currently have no quantitative backing. For manufacturers: which component tolerances actually matter to clinical dose. For clinics: a documented history that outlives staff turnover.
§6.B in the white paperWorkforce training
Physicists and service engineers have no common language for the machine. SIMAC Training gives both a model of your specific linac, safe to interrogate without clinical risk.
§6.C in the white paper
This white paper — and the pilot behind it — was co-authored with Image Owl, whose TotalQA® platform is the QA data pipeline SIMAC Digital Twin has been validated against in early deployments. SIMAC Digital Twin itself is data-source agnostic; TotalQA is simply where we started.
Toward Digital Twin-Enabled Management of Medical Linear Accelerators
Marco Carlone (Linax Technologies) and Matt Whitaker (Image Owl) lay out what TG-142 and TG-100 leave unanswered, how a physics-based twin closes the gap, and the governance questions the field still needs to settle.
Toward Digital Twin-Enabled Management of Medical Linear Accelerators
Linax Technologies · Image Owl
- §What TG-142 and TG-100 each deliver, and where each stops
- §The physics model: subsystems, updating, and the Virtual Water Tank
- §Cumulative dose impact as a new QA metric
- §Continuous commissioning and root cause intelligence
- §Patient-specific QA: why gamma is a weak proxy for dose accuracy
- §Life-cycle management, workforce training, and open governance questions
This is a white paper and an active pilot — not a finished product.
We'd rather say that plainly than dress up a roadmap as a shipping feature. Here's where things actually stand.
SIMAC Digital Twin's subsystem models are validated against commissioning data and already used for training and root-cause simulation, without touching a live accelerator.
Connecting SIMAC Digital Twin to a continuous QA data stream — currently via Image Owl's TotalQA® — for live continuous commissioning and root-cause detection.
Model versioning, audit trails, and the regulatory boundary between an independent QA model and a licensed treatment planning system. The paper says so too.
Every LINAC deserves a digital twin.