![\"Electron](\"https:\/\/rtmedical.com.br\/wp-content\/uploads\/2026\/06\/electron-transport.jpg\")
This article describes the physical basis of each family, the main commercial implementations, known failure situations, and practical commissioning and validation criteria. References include the regulatory report AAPM TG-105<\/a> and clinical validation studies indexed in PubMed.<\/p>\n Photons deposit dose primarily through secondary electrons released in photoelectric interactions, Compton, and pair production. Dose deposition occurs over relatively long distances from the point of primary interaction, and the influence of heterogeneities is mediated mainly by variations in the linear attenuation coefficient. The lateral scattering of a photonic beam is small in relation to the treatment depth \u2014 which supports algorithms such as AAA and Acuros XB.<\/p>\n Electrons interact in fundamentally different ways. When passing through matter, they lose energy through collision (ionization and excitation) and radiation (bremsstrahlung), and undergo successive elastic scattering events with atomic nuclei and electrons. The result is a tortuous trajectory, not straight, which gives rise to three unparalleled physical characteristics in photons:<\/p>\n Finite range<\/strong>: 6 MeV electrons travel, in water, approximately 3 cm before being absorbed. The depth-dose curve shows a peak near the surface followed by an abrupt drop. This range depends on the energy, the electron density of the medium and the statistical fluctuations of deposited energy (straggling<\/em>).<\/p>\n Multiple Coulomb Scattering (MCS)<\/strong>: the lateral broadening of the beam increases with depth in a non-linear fashion. In low atomic number materials (soft tissues), scattering is moderate and relatively predictable. At interfaces with dense (bone, metals) or low-density (lung, air) materials, local scattering changes dramatically, creating regions of increased or reduced dose that algorithms based on homogeneous media cannot represent.<\/p>\n Extreme sensitivity to<\/strong>configuration geometry: the air between the applicator and the skin surface, the irregular anatomical contour, the presence of bolus or cutouts \u2014 all of these factors modify the fluency reaching the patient in ways that photon algorithms rarely need to consider at this level of detail. The difference in effective atomic numbers between air, water, cortical bone and lung influences both the stopping power (stopping power<\/em>) and the average angular scattering, making density scaling \u2014 valid for photons in the first approximation \u2014 insufficient for electrons in complex geometries.<\/p>\n The Pencil Beam model for electrons was formally formulated by Hogstrom, Mills and Almond based on the Fermi-Eyges multiple scattering theory. The central idea is to decompose the beam into a set of infinitely thin pencil beams. Each pencil propagates along its central axis and accumulates lateral scatter described by a Gaussian distribution. The total dose at a point results from the superposition of all Gaussians weighted by the local fluence.<\/p>\n The Fermi-Eyges theory is exact for a homogeneous and unlimited medium. The extension to heterogeneous media is done by effective density projection: instead of calculating the actual electron transport through the heterogeneous medium, the algorithm “flattens” the density variations into an effective density profile along the axis of each pencil. This works well for inhomogeneities parallel to the beam axis (e.g., lung behind chest wall in direct incidence), but fails when the inhomogeneities are lateral to the axis or when the interface creates asymmetries in the scattering that the Gaussian cannot represent.<\/p>\n Awareness of these limitations motivated the development of stochastic transport-based algorithms for clinical use, culminating in eMC.<\/p>\n The Electron Monte Carlo (eMC) available from TPS Eclipse (Varian Medical Systems) is based on the Monte Carlo Macro method developed by Neuenschwander and Born. Rather than simulating each individual collision event \u2014 as the EGSnrc, GEANT4, and PENELOPE reference codes do \u2014 eMC groups trajectory segments into \u201cmacro steps,\u201d each described by a pre-calculated distribution of energy and angle. This reduces the computational cost while maintaining a statistical representation of the actual transport.<\/p>\n From a physical point of view, eMC solves problems that PB cannot:<\/p>\n The clinical validation of the eMC was the subject of a specific publication for small fields \u2014 a particularly challenging situation for PB \u2014 demonstrating that the eMC reproduces experimental measurements with lower error in configurations with reduced fields (PMID 19692969)<\/a>. For heterogeneities, independent validation studies confirmed that eMC outperforms PB in lung and bone scenarios, although the magnitude of the deviations depends on the energy and thickness of the heterogeneous material (PMID 22088999)<\/a>.<\/p>\n eMC is a stochastic method. The calculated dose contains inherent noise, the magnitude of which depends on the number of simulated particle histories. The clinician faces a direct trade-off: more stories reduce noise and uncertainty in the volumes of interest, but increase calculation time. AAPM TG-105<\/a> dedicates a specific section to adequate statistical resolution for plan approval, advising that local stochastic uncertainty in treatment volume be known and documented before clinical approval.<\/p>\n The quantity reported by a Monte Carlo engine depends on the implementation. It is not safe to assume that every eMC delivers D_m nor that every PB delivers D_w. Comparison between plans must record the dose convention and clinical version settings, especially in bone and high-density materials.<\/p>\n The configuration geometry of an electron field has a much greater impact on the dose distribution than in photons. Each element needs to be modeled appropriately in TPS.<\/p>\n Applicators<\/strong>: each combination of accelerator, energy and applicator size defines a specific depth-dose curve and side profile. Commissioning requires measurement for each combination; Unvalidated interpolation is a source of error. The output factor (output factor<\/em>) varies with the field size \u2014 especially below 6\u00d76 cm\u00b2 \u2014 and must be measured directly.<\/p>\n Cutouts<\/strong>: molded in lead or cerrobend, the cutouts define the field in an irregular shape. The output factor of a clipping is not simply interpolated from the open field data; it must be measured or calculated with the already validated model. PB often fails to correctly predict the output factor of small indentations because the lateral particle balance is disrupted.<\/p>\n Bolus<\/strong>: tissue-equivalent material positioned on the skin to modify dose distribution \u2014 raising the maximum dose region to the surface or compensating for oblique surfaces. TPS needs to represent the exact bolus geometry. Small differences in thickness between the modeled bolus and the actual bolus can cause clinically relevant deviations in surface dose distribution. Boluses manufactured by 3D printing with calibrated density and post-manufacturing CT in planning reduce this source of error.<\/p>\n Irregular surfaces<\/strong>: nose, ear, armpit, scrotum \u2014 regions with recesses or air cavities close to the field. PB tends to incorrectly calculate dose in regions where the surface creates steep lateral creep gradients. eMC treats these geometries with greater fidelity, but statistical noise at the field edges can mask real dose gradients if the number of stories is insufficient.<\/p>\n Electrons pass through the lung with less lateral scattering and greater effective range than in soft tissues. Range extension can be estimated by density scaling, but lateral beam broadening is underrepresented by PB. The practical result is that PB may underestimate the dose to critical structures positioned lateral to the field when there is lung between the applicator and the target \u2014 exactly the scenario of the post-mastectomy chest wall with regional lymph node chain.<\/p>\n Thick bone increases lateral spread and reduces effective range. At bone-soft tissue interfaces, backscattered electrons from the high-density interface deposit additional dose in the adjacent tissue, a phenomenon that PB does not adequately represent. eMC captures this behavior through stochastic simulation, but the magnitude of the effect depends on bone thickness and beam energy.<\/p>\n The presence of titanium, stainless steel, or amalgam within or adjacent to the treatment field creates local dose perturbations that simplified algorithms do not represent with sufficient accuracy. Monte Carlo reference is the only method capable of modeling electronic transport in high atomic number materials with adequate reliability for these cases. The usual clinical recommendation is to avoid direct incidence of electrons on metallic implants or to explicitly document dosimetric uncertainty in the patient’s medical record.<\/p>\n Lateral particle-to-electron equilibrium requires that the field be wide enough so that scattered contributions from the edges do not disturb the dose at the center. Below approximately 3\u00d73 cm\u00b2, this balance is incomplete, and the dose in the center of the field falls in relation to that predicted by models based on large fields. PB underestimates this drop. The eMC, with specific commissioning for small fields (PMID 19692969)<\/a>, provides more faithful representation \u2014 as long as statistical uncertainty is managed appropriately and commissioning measurements include these fields.<\/p>\nIn this Article<\/h2>\n
\n
Why electrons are a different problem than photons<\/h2>\n
Pencil Beam and the Fermi-Eyges approximation<\/h2>\n
Strengths of Pencil Beam<\/h3>\n
\n
Known limitations of Pencil Beam<\/h3>\n
\n
What changes with the Electron Monte Carlo from Eclipse<\/h2>\n
\n
The price: statistical noise and calculation time<\/h3>\n
Dose to medium versus dose to water<\/h3>\n
Applicators, clippings, boluses and irregular surfaces<\/h2>\n
Heterogeneities, interfaces and small fields<\/h2>\n
Lung and low-density tissues<\/h3>\n
Cortical bone and high-density materials<\/h3>\n
Metallic implants<\/h3>\n
Small fields<\/h3>\n