🎯 Volumetric Modulated Arc Therapy (VMAT) Planning Simulator

Arc Geometry Setup

Volumetric Modulated Arc Therapy (VMAT) delivers a full course of intensity-modulated radiotherapy while the linear accelerator gantry sweeps continuously around the patient, instead of stopping at a handful of fixed beam angles like conventional step-and-shoot IMRT. Everything begins with defining the arc geometry: how many arcs, over what angular range, at what collimator and couch angles.

• 358°: Typical single-arc span (near-full rotation, avoids overlap)
• 1–2: Arcs per plan (typical) (complex targets use 2)
• 10–45°: Collimator angle range (reduces interleaf leakage)
• 2008: First clinical concept (Otto, Medical Physics)

From step-and-shoot IMRT to a single continuous arc

Conventional IMRT delivers dose from 5–9 fixed, discrete gantry angles ("step-and-shoot" or sliding-window). The gantry stops, the beam shapes and delivers, then the gantry indexes to the next angle and repeats. VMAT instead treats the entire 360° (or near-360°) of gantry rotation as a continuum of beam angles, each contributing a small, optimized dose increment.

This idea traces to arc therapy concepts from the 1960s–70s (single open-field arcs) and intensity-modulated arc therapy (IMAT) proposed by Yu in 1995, which used multiple overlapping arcs. The breakthrough that made a single optimized arc clinically practical was Karl Otto's 2008 paper "Volumetric modulated arc therapy: IMRT in a single gantry arc" (Medical Physics), which formulated an optimization algorithm capable of simultaneously varying gantry speed, dose rate, and MLC aperture shape at hundreds of control points across one rotation.

Otto's 2008 algorithm is the direct ancestor of every commercial VMAT product in use today — it proved that a single continuously modulated arc could match or exceed fixed-field IMRT plan quality while cutting delivery time roughly five- to ten-fold.

Defining the arc: span, direction, and collimator angle

Arc geometry setup specifies several interlocking parameters before optimization begins:

• Start and stop gantry angle: most single arcs span close to 360° but stop just short of a full circle (e.g., 181° to 179°, a 358° arc) so the beam never re-enters the exact same angle twice, avoiding a dose "seam." • Arc direction: clockwise (CW) or counter-clockwise (CCW). For a two-arc plan, the second arc is almost always rotated in the opposite direction of the first — this halves total delivery time versus resetting the gantry back to the start position between arcs. • Collimator angle: the MLC and jaw assembly is rotated (commonly 10–45°, avoiding exactly 0°/90°) to offset the leaf travel direction relative to the arc, minimizing interleaf leakage and tongue-and-groove effects accumulating along a fixed line. • Couch angle: usually 0° (coplanar) for VMAT; non-coplanar arcs are possible on some platforms for intracranial or spine SBRT. • Collision checking: gantry, couch, and patient/immobilization geometry are virtually checked across the full arc to rule out mechanical collisions before beam-on.

Why one or two arcs, and when more help

A single arc provides one continuous 360°-worth of beam angle diversity, which is often enough to closely surround the PTV with dose and steer around nearby OARs, especially for moderately complex targets like prostate. For anatomically complex or concave-target cases — head and neck with parotid sparing, or targets with two distinct dose levels — a second arc roughly doubles the number of independent control points and beam angle combinations available to the optimizer, generally improving conformity and OAR sparing at the cost of added delivery time.

Dual-arc plans are typically delivered as one CW and one CCW rotation back-to-back, so the gantry never needs to "reset," keeping total delivery time efficient (roughly double a single arc, but still a fraction of step-and-shoot IMRT). Some planning systems also support partial arcs (e.g., avoiding the beam path through a hip prosthesis or one arm) rather than full 360° rotations.

Continuous Gantry Rotation & Dynamic MLC

The defining mechanical feat of VMAT is that the gantry never stops moving during beam-on. As it rotates, the multileaf collimator continuously reshapes the treatment aperture — dozens to hundreds of times per arc — so that at every instant the beam's-eye-view opening conforms to the tumor's projected outline from that specific angle.

• ~177: Control points per 360° arc (Varian Eclipse default)
• 2.5–5 mm: Typical MLC leaf width (at isocenter)
• ~2.5 cm/s: Max leaf travel speed (per leaf)
• ~2°: Angular spacing per CP (358°⁄177 control points)

Control points: the discretized backbone of a continuous arc

Although gantry rotation is mechanically continuous, the treatment plan itself is optimized and delivered as a dense sequence of discrete "control points" — commonly 177 per 360° arc on Varian systems (roughly one every 2°), though the number is configurable, often 90–180 per arc depending on planning system and case complexity.

Each control point stores a complete instantaneous machine state: gantry angle, MLC leaf positions (typically 60 leaf pairs across the field), jaw positions, collimator angle, cumulative monitor units, and dose rate. During delivery, the linac interpolates leaf motion and dose delivery smoothly between consecutive control points, so the discrete optimization grid produces apparently fluid, continuous modulation. More control points per arc give the optimizer finer angular resolution to shape dose — visibly smoother, more frequent MLC updates — at the cost of a larger, more complex delivery sequence.

A denser control-point grid does not by itself lengthen delivery time (gantry speed and arc span set that) — it increases the fidelity with which the MLC can track a changing target outline, which is why complex, concave targets benefit most from higher control-point density.

The multileaf collimator as a beam sculptor

A modern MLC consists of 60–160 individually motorized tungsten leaves (commonly 2.5–5 mm wide projected at isocenter) arranged in opposing banks that slide in and out of the beam path. Each leaf can move independently at up to roughly 2–2.5 cm/s, and together the leaf bank silhouette carves an irregular, patient-specific aperture out of the round beam.

In VMAT, the leaf positions at each control point are not static shapes but part of a continuously interpolated trajectory: as the gantry advances from one control point to the next, every leaf glides from its previous position to its next target position while the beam stays on. This produces "sliding-window"-like intensity modulation superimposed on the aperture shaping — leaves can also move faster or slower than their neighbors to sculpt regions of relatively higher or lower fluence within a single aperture, contributing to intensity modulation within the field, not just its outline.

Beam's-eye-view tracking of the target

From the gantry's perspective, the tumor's projected 2D silhouette (the "beam's-eye-view," BEV) changes shape continuously as the gantry rotates — an irregular tumor can look wide and round from one angle and narrow and elongated from another 90° away. The MLC aperture at each control point is optimized to match that instantaneous BEV outline (expanded by a small planning margin), while simultaneously avoiding or partially blocking any OAR that overlaps the field from that angle.

This is fundamentally what separates VMAT from a simple rotating open field: the aperture is not fixed, it is a live, angle-dependent function computed by the treatment planning system's inverse optimization engine, re-evaluated at every one of the ~90–180 control points that make up the arc.

Dose Rate & Gantry Speed Modulation

What makes VMAT optimization genuinely difficult — and genuinely powerful — is that three machine parameters are varied simultaneously and continuously across the arc: MLC aperture shape, gantry rotation speed, and dose rate. Together they give the optimizer far more degrees of freedom per unit time than a fixed-angle IMRT field can offer.

• 100–600: Dose rate range (MU/min, flattened beam)
• up to 2400: FFF beam dose rate (MU/min, flattening-filter-free)
• up to ~6°/s: Gantry rotation speed (≈4.8 rpm max on TrueBeam)
• 3: Modulated parameters (MLC + dose rate + gantry speed)

Why three simultaneous variables, not one

A fixed-gantry IMRT beam only needs to modulate fluence across the 2D aperture (via MLC motion) because the gantry angle is constant during delivery. VMAT instead has to deliver a full angular dose distribution while the gantry keeps moving — so it needs additional handles beyond aperture shape to control how much dose lands at each angle.

Those extra handles are gantry rotation speed and instantaneous dose rate. Slowing the gantry lets more monitor units (MU) accumulate at angles that need heavier weighting (e.g., angles offering the cleanest path to the tumor while avoiding an OAR); speeding it up reduces the MU delivered at less useful angles. Simultaneously varying dose rate (typically 100–600 MU/min for flattened beams, considerably higher for flattening-filter-free, FFF, beams) gives an independent second lever — the optimizer can, for example, use a high dose rate with a fast gantry speed, or a low dose rate with a slow gantry speed, to reach the same MU-per-degree target through different mechanical paths, whichever best respects machine and delivery-time constraints.

Because gantry speed, dose rate, and MLC shape are optimized jointly rather than independently, VMAT reformulates IMRT delivery as a genuinely multi-dimensional optimization problem — Otto's original 2008 formulation described it as progressively refining an initial coarse arc solution through successive, finer control-point sampling.

Machine constraints the optimizer must respect

The optimization is not free to pick any combination of speed and dose rate — real linear accelerators impose hard mechanical and dosimetric limits that the plan must satisfy at every control point:

• Maximum gantry rotation speed: roughly 4.8 rpm (about 6°/second) on modern platforms such as Varian TrueBeam or Elekta Versa HD/Agility-equipped machines. • Maximum and minimum dose rate: typically 600 MU/min ceiling for flattened 6–15 MV beams (much higher, up to ~2400 MU/min, for FFF beams used in SBRT); dose rate cannot be commanded to zero while the beam is nominally on. • Maximum leaf travel speed: leaves must be able to reach their next control point's target position within the time the gantry takes to travel between control points, or the plan is mechanically undeliverable. • MU-per-degree limits: extremely rapid dose-rate changes or very slow/fast gantry segments both have practical bounds tied to accelerator response time.

A deliverable VMAT plan is one where every control-point transition respects all of these simultaneously — which is why VMAT optimization is computationally heavier than fixed-field IMRT optimization.

Clinical payoff: monitor unit and time efficiency

Despite the added optimization complexity, VMAT plans are typically more MU-efficient than step-and-shoot IMRT plans for the same case, because continuous delivery avoids the dead time and duplicated segments inherent to discrete-angle delivery, and because the optimizer can spread the same total dose more smoothly across many more effective beam angles.

Lower total MU for an equivalent dose distribution translates into less integral (whole-body) dose from head leakage and scatter, shorter beam-on time, and — because the linac is doing useful modulation throughout, rather than idling between segments — dramatically shorter overall treatment slots. This is the practical foundation for VMAT's roughly 2-minute single-arc delivery versus 10–20 minutes for equivalent step-and-shoot IMRT.

Dose Accumulation Over the Arc

No single control point delivers a clinically meaningful dose by itself — each contributes only a thin angular wedge. The conformal, steep-gradient dose distribution that defines a good VMAT plan only emerges once dose from every control point across the full arc (or arcs) is summed inside the patient.

• ~180: Control points summed (per 360° arc, typical)
• 100/90/70/50%: Isodose lines reported (standard reporting levels)
• ~1.0–1.2: Conformity index target (prescription isodose / PTV volume)
• 1–3 mm: Dose calc grid resolution (typical TPS voxel size)

From angular wedges to a conformal cloud

The treatment planning system's dose calculation engine (typically a convolution-superposition or Monte Carlo algorithm) computes the dose deposited in the patient by every individual control point's beam segment — a narrow angular "wedge" of dose shaped by that control point's MLC aperture, weighted by its MU. Summed over ~90–180 control points per arc, spaced roughly 2–4° apart, these overlapping wedges reinforce each other most strongly inside the PTV (where nearly every angle contributes) and cancel out or stay thin everywhere else — exactly the geometric principle that has underpinned rotational therapy since single-arc treatments in the 1950s–60s, now made intensity-modulated.

Because the PTV receives contributions from the full angular sweep while any single OAR typically only intersects the beam path from a limited angular range, the accumulated dose distribution naturally concentrates on the target and falls off sharply just outside it — producing the steep dose gradients VMAT is known for.

Isodose lines and conformity

The accumulated 3D dose distribution is visualized clinically as isodose lines/surfaces — contours connecting points receiving the same percentage of the prescription dose. Standard reporting levels include the 100% (prescription) line, which should tightly wrap the PTV, and lower levels such as 90%, 70%, and 50%, which describe how quickly dose falls off moving away from the target.

Plan quality is often summarized with a conformity index (CI), comparing the volume enclosed by the prescription isodose to the PTV volume itself (ideal CI ≈ 1.0), and a gradient index, describing how sharply dose falls off outside that volume. Because VMAT integrates dose from a near-continuous range of angles rather than 5–9 discrete ones, it typically achieves tighter conformity and steeper gradients around concave or irregularly shaped targets than fixed-field IMRT, while often using fewer total monitor units.

A conformity index near 1.0 means the prescription isodose volume matches the PTV volume almost exactly — CI values noticeably above 1 indicate the high-dose region "spills" beyond the target into surrounding normal tissue.

Verifying the accumulated plan before treatment

Before a VMAT plan is ever delivered to a patient, the summed dose distribution is checked against the prescription using dose-volume histograms (DVHs), which plot the fraction of a structure's volume receiving at least a given dose — allowing planners to confirm PTV coverage (e.g., "95% of the PTV receives ≥100% of prescription dose") and OAR constraint compliance (e.g., "rectum V70Gy < 20%") in a single quantitative summary.

Because VMAT plans are dosimetrically complex — hundreds of control points with jointly varying MLC, dose rate, and gantry speed — every patient-specific plan additionally undergoes machine-specific quality assurance: the plan is delivered to a phantom with an embedded 2D/3D detector array, and the measured dose distribution is compared to the calculated one (commonly using gamma analysis, e.g., 3%/2mm criteria) before the plan is approved for clinical use.

Treatment Delivery & Verification

On the treatment day, the entire optimized arc — every gantry angle, MLC shape, dose rate, and speed value computed during planning — is delivered as one uninterrupted rotation. Imaging performed immediately before and often during that rotation verifies that the patient is correctly positioned and that the beam matches the plan.

• ~2 min: Single-arc delivery time (vs. 10–20 min step-and-shoot IMRT)
• ~3–4 min: Dual-arc delivery time (complex targets)
• RapidArc: Varian trade name (introduced 2008)
• VMAT (Elekta): Elekta trade name (Agility/Beam Modulator MLC)

A single uninterrupted rotation vs. step-and-shoot

The clearest practical distinction between VMAT and conventional IMRT is delivery time. Step-and-shoot IMRT stops the gantry at each of 5–9 fixed angles, delivers a shaped segment, then indexes to the next angle — repeated dead time, leaf repositioning, and beam interruptions typically stretch total beam-on and setup time to 10–20 minutes per fraction. A single VMAT arc, by contrast, keeps the beam on essentially continuously through the full ~358° rotation, completing in roughly 90 seconds to 2 minutes; a two-arc plan for a more complex target takes roughly 3–4 minutes.

Shorter delivery has real clinical value beyond throughput: less time on the table reduces the chance of intrafraction patient motion, improves comfort (especially relevant for elderly, pediatric, or claustrophobic patients), and lets departments treat substantially more patients per linac per day without compromising plan quality.

Because VMAT typically halves-to-tenths the beam-on time of step-and-shoot IMRT for comparable or better plan quality, it has become the default delivery technique at most modern radiotherapy centers for prostate, head and neck, lung, and many SBRT indications.

Commercial VMAT platforms

VMAT was commercialized under several vendor-specific trade names built on the same underlying continuous-arc, triple-modulation principle:

• RapidArc (Varian Medical Systems, now Siemens Healthineers) — the first commercial VMAT product, launched in 2008 following Otto's algorithm, delivered on Varian Trilogy/TrueBeam linacs with the Millennium or High-Definition (HD120) MLC. • VMAT (Elekta) — delivered on Elekta Versa HD/Synergy platforms using the Agility (160-leaf) or Beam Modulator MLC; Elekta also historically offered "Ergo++" arc planning tools. • SmartArc (Philips Pinnacle³ treatment planning system) — a planning-system-level VMAT optimization module usable across compatible linac vendors.

While branding differs, all rely on the same core physics: hundreds of jointly optimized control points across one or more continuous gantry rotations, with simultaneous MLC, dose-rate, and gantry-speed modulation.

Image guidance and clinical applications

Modern VMAT delivery is almost always paired with image-guided radiotherapy (IGRT): cone-beam CT (CBCT) or orthogonal kV/MV imaging is acquired immediately before treatment to verify patient positioning against the planning CT, with couch shifts applied to correct any offset before the beam turns on. The electronic portal imaging device (EPID) mounted opposite the gantry can also capture transit images during or after the arc, which are compared against the predicted portal dose as an additional per-fraction verification step.

Clinically, VMAT is now standard of care across a wide range of indications: prostate cancer (single or dual arc, sparing rectum and bladder), head and neck cancer (typically dual arc, sparing parotids, spinal cord, and brainstem while covering complex concave PTVs), lung cancer (including hybrid VMAT-SBRT techniques combining steep dose gradients with hypofractionated regimens), and intracranial, spine, and other SBRT/SRS sites where fast, highly conformal delivery limits normal-tissue dose. Comparative studies generally show VMAT achieves plan quality equal to or better than fixed-field IMRT, with substantially fewer monitor units and severalfold shorter delivery time.