A Varian linear accelerator with HDTSe mode was utilized to deliver 6 MeV electron beams for the treatment. No scatters or degraders were placed in front of the patient. To achieve a larger field for a single beam, the source-to-surface distance (SSD) was extended from the 100 cm SSD isocenter by 232 cm, resulting in a 332 cm SSD and a field size of 133 cm by 133 cm. A combined dual electron field with gantry angles of 70˚ and 110˚ relative to the horizontal (90˚) were used to cover the vertical length of 242 cm. The schematic view of the treatment geometry is shown in Figure 1, with the phantom anterior surface located 232 cm from the 100 cm isocenter.

A custom-made rotational platform was used, rotating at a speed of 1.11 revolutions per minute (rpm). The speed of rotation was critical and affected the rotational dose rate and the dose absorbed by the superficial layer. A speed of 1.11 rpm appeared comfortable and tolerated by patients. Patients stand on the rotational platform at an extended SSD of 332 cm. The gantry angles of 70˚ and 110˚ were chosen to ensure that the beams could point above the patient’s head and below the patient’s feet, respectively. Daily treatments with a machine rate

of 600 MU/min typically required approximately 30 minutes, with 15 minutes allocated for patient set-up and 15 minutes for the beam-on time to deliver the prescribed dose.

During treatment, the patient’s eyes were shielded internally, and nails and toes were shielded externally. Commercial tungsten eye shields (3 mm tungsten coated with 1 mm aluminum) were used for eye shielding [31] . Lead sheets with a thickness of 3 mm were used to shield the nails and toes after delivering a dose of 12.5 Gy [32] .

Dosimetry measurements included 1) measuring the static dose rate at the extended SSD, 2) rotational dose rate for a rotating phantom, and 3) a correction to individual field profile for SSD variation over the skin surface.

1) Static Measurements: Static measurements were performed at an SSD of 332 cm using a NACP parallel plate ionization chamber in a polystyrene phantom. The measurements were carried out in a 10 × 10 cm² field at an isocenter SSD of 100 cm for a single horizontal 6 MeV beam. The dosimetry standard was used to calibrate the measurements.

2) Rotational Calibration: The rotational dose rate was measured by comparing the rotational dose to the static dose for one revolution at the extended SSD and phantom. The rotational correction factor was determined as the ratio of the rotational dose to the static dose.

3) Individual Field Profile: To determine the combined dose distribution of the dual field, the individual field profile was measured along the vertical direction using a parallel plate chamber. The measurements were performed for the beam with a gantry angle of 90˚ at the extended SSD. The composite dose profile was obtained by combining the individual beam profiles, matching the 50% dose line.

In order to ensure safe and accurate radiation delivery to patients during RTSET treatment, in-vivo dose measurements were conducted using film, MOSFET, and optically stimulated luminescent dosimeter (OSLD).

1) Film Dosimetry: Kodak XV film (Kodak, Rochester, NY) was used for in-vivo dosimetry measurements. Film characteristics were obtained by exposing a series of films in a 10 × 10 cm² field at an SSD of 100 cm with doses ranging from 0 - 150 cGy. The films were analyzed using a densitometer to obtain dose-versus-reading data. The irradiated films were calibrated using a curve between densitometer reading and delivered dose. The film strips were securely attached to the patient’s skin to ensure no air gaps.

2) MOSFET Dosimetry (Best Medical Canada, formerly Thomson and Nielsen Electronics Ltd., Ottawa, Ontario Canada): High-sensitivity MOSFET dosimeters, composed of two MOSFET devices mounted together, were used for in-vivo surface dose measurements [33] [34] [35] . The dosimeters were calibrated with and without buildup for a 10 × 10 cm² static field at an SSD of 100 cm for 6 MeV electron beams.

3) OSLD Dosimetry: OSLD detectors, consisting of aluminum oxide doped with carbon, were used for in-vivo dose measurements (MicroStar Reader, Landauer, Inc., Glenwood, IL) [35] [36] [37] . Calibration curves were created by irradiating the OSLDs with a 6 MeV electron beam in a 10 × 10 cm² field at d_max (=1.4 cm) and an SSD of 100 cm with doses ranging from 0 - 150 cGy. Four nanoDot detectors were used for each calibration point, and dose points were monitored in a solid water phantom using a Farmer ionization chamber. The reader software processed the conversion directly to dose in cGy by taking into account all the calibration factors involved. In the in-vivo dosimetry measurements, each detector was read three times, and the average reading was used for each measurement point.

The shielding for patient’s eyes, nails, and toes during treatment was crucial to ensuring patient safety. An eye shields was commercially available to protect a patient’s lens during radiation therapy. A tungsten eye shield was made from a millimeter thick anodized aluminum cap to reduce the electron backscatter to the eyelids. The effectiveness of 3 mm lead shielding to fingernails and toes was sufficient to provide adequate (95%) shielding for normal organ at 6 MeV electron beam. A soft bandage sheet on top of the lead sheet provided effective shielding reducing potential backscatter toward treated skin.

X-ray bremsstrahlung contamination on the central axis of stationary electron beams, with nominal energies of 6 - 20 MeV, typically contributed between 1% to 7% of the maximum dose [2] [3] . Without precautions, this can result in a significant absorbed dose from x-ray contamination throughout the patient’s body. The DFR-TSET technique utilized the dual angled electron fields to reduce x-ray contamination by directing the central axis away from the patient as shown in Figure 1 [3] [4] .

The beam profiles [29] [38] were measured using a parallel plate ionization chamber in the phantom. The dose distribution of a single field along the vertical direction was measured in 10 cm increments, and the data points are shown as diamond symbols in Figure 2(a). To obtain a combined profile, the 50% dose line was matched at a position of z = 75 cm from the horizontal position, as shown in Figure 2(b). The combined profile demonstrated a deviation uniformity of less than ±6%.

Regarding dosimetry characteristics, the static dose rate at a source-to-surface distance (SSD) of 332 cm was determined to be 0.0751 cGy/MU, based on comparison with the dose at SSD of 100 cm and a 10 × 10 cm² field. The rotational dose rate, which is the dose ratio of the static phantom at its d_max over the rotational phantom at its own d_max, was found to be 0.39 for one complete revolution at a machine rate of 600 MU/min.

Due to variations in the patient’s abdomen thickness, the SSD correction was necessary. As electron beams do not follow an inverse square law, a power law correction was applied to account for SSD variations. In this setup, an empirically determined factor of 2.25 was used to describe this power law. Therefore, the dose to be delivered per revolution (DR) can be calculated using Equation (1):

DR = DRs × DRr × ( SSD o / SSD ) p × DRM × Pr (1)

where DRs is the static dose rate, DRr is the rotational rate ratio, SSD_o is the reference SSD (332 cm in this case), SSD is the patient-specific SSD, p is the power law factor, DRM is the machine rate, and Pr is the platform rotational rate.

Based on the measurements and setup described above, the following values were obtained:

DRs = 0.0751 cGy/MU;

DRr = 0.39;

p = 2.25;

Pr = 1/0.9 = 1.11 rpm.

For example, if a patient has an SSD of 325 cm and the machine rate is 600 MU/min, the dose per revolution can be calculated as follows:

DR = 0.0751 cGy/MU × (332/325)^2.25 × 0.39 × 600 (MU/min)/(1.11 rpm)

=16.6 cGy per revolution or 16.6 cGy per 540 MU (=32.53 MU/cGy).

If the prescription is 125 cGy per fraction, it would require approximately 7.5 revolutions, which is equivalent to 4066 MU. Since 7.5 revolutions is not an integer number, 7 and 8 revolutions were alternated every other day, resulting in delivered doses of 117 cGy and 133 cGy to the patient, respectively.

Charged Particle Equilibrium (CPE) status [39] may not be well established due to the use of extended SSDs for an electron beam. During static dose rate measurements at SSD of 332, a non-CPE feature was observed with the maximum depth dose moved to the surface in the percent depth dose (PDD) curve. We can use virtual SSD or effective SSD to describe electron beams. However, virtual SSDs might not provide accurate inverse square law corrections to the output of extended SSDs. Also, effective SSDs would vary with energy and field size, especially for extremely large field sizes [38] . Therefore, a modified power law was used to describe extended SSD differences between an individual patient and a calibrated phantom on the rotational platform. This power law deviated from the inverse square behavior by a factor of 2.25.

In-vivo dosimetry measurements were performed during patient treatments to verify the delivered dose as prescribed and assess the uniformity of dose distribution [31] . In vivo dosimetry was a direct method of measuring radiation doses to patients receiving radiation treatment. It was a suitable method both to monitor treatment delivery, and to detect or prevent various errors early in course of treatment. Monitoring the dose delivery constituted a safety measure, being a part of the patient radiation protection. Moreover, the in vivo measurement provided a treatment record confirming that the dose was delivered correctly within the expected tolerance. Three different measurement methods were employed: XV film, MOSFET, and OSLD.

For XV film measurements, film strips were attached laterally across specific regions of the patient’s body, including the head, left arm, left anterior abdomen, left posterior thigh, left internal leg, and right calf. The prescribed dose for one patient per fraction was 116 cGy as shown in Figure 3(a), and the measured doses for these locations were as follows: 116 cGy, 104 cGy, 119 cGy, 108 cGy, 107 cGy, and 108 cGy, respectively. The maximum dose was observed on the abdomen (119 cGy), while the minimum dose was on the left arm (104 cGy), which was 94% of the maximum dose. The uniformity of the measured dose was within 6%, and the average measured dose was 110 cGy, approximately 5% lower than the prescribed dose.

In the case of MOSFET measurements, probes were attached to specific locations on the patient’s body, including the neck, upper anterior abdomen, low anterior abdomen, and left posterior thigh. The prescribed dose to this patient was 125 cGy as shown in Figure 3(b), and the measured doses were 124 cGy, 124 cGy, 128 cGy, and 127 cGy, respectively. The measured doses ranged from 99% to 102% of the prescribed dose.

Both the film and MOSFET measurements indicated good dose uniformity and agreement between the delivered and prescribed doses in the dual field RTSET approach. These results supported the adequacy of the dosimetry calibration and treatment protocols. However, among the various detector types, OSLDs were found to have advantages over films and MOSFET in terms of better handling, quick reading, and exporting of measurement results. OSLDs also allowed for the use of multiple detectors due to their small size, enabling more precise dose distribution information during in-vivo dosimetry measurements.

OSLD detectors were attached across the patient’s whole body as shown in Figure 4, and four patients were selected for the in-vivo measurements. The measured doses for different locations were summarized in Table 1. The prescription dose varied for each patient based on the physician’s protocol.

Remark: PA (Posterior-Anterior) is the reference point for dose calibration, indicating the patient’s position in relation to the linear accelerator. PA2 is located 15 cm vertically above PA in the superior direction. AP (Anterior-Posterior) represents the calibration level on the anterior skin. NA (Not Available) indicates that data for this specific position is not provided.

In vivo dosimetry measurements using OSLDs were helpful in identifying underdosed regions that were blocked during treatment on the patient’s skin [5] [37] [40] . Boost fields to local areas may be necessary based on clinical judgment

from the underdosed regions and their geometric sizes identified by the in-vivo OSLD results. OSLDs played an important role in providing the delivered dose distribution to avoid serious overdosing or underdosing when delivering the boost field. Boost local fields were delivered at conventional SSDs of 100-110 cm using standard electron cones and customized cutouts.

The specific regions requiring boost fields were not clearly defined and varied from one patient to another. OSLD in-vivo measurements assisted in identifying these regions. Generally, the boosted organ areas included the soles of the feet, the palms of the hands, the top of the head, and the upper parts of the arms in the outer region blocked by the head in DFR-TSET. These areas had already received certain doses from the RTSET treatment. The boost prescription dose would compensate for the partially delivered dose based on OSLD measurements during DFR-TSET.

A MOSFET dosimeter was being compared to OSLD in terms of their angular dependence for in-vivo skin dose measurements. A MOSFET dosimeter was based on a semiconductor material and exhibits an angular dependence in its response. The measured dose can vary depending on the angle at which the radiation beam interacts with the dosimeter. For in-vivo skin dose measurements, a MOSFET dosimeter showed a relatively high angular dependence of around 5% - 7% [34] or even higher, [35] indicating that the measured dose can significantly change with varying angles of incidence.

OSLDs also showed a small angular dependence. The reported angular dependence for OSLDs in the MeV energy range was about 3% - 4% [41] . This suggested that the measured dose with a OSLD detector was less affected by changes in the angle of beam incidence.

In the specific treatment scenario, DFR-TSET, where the patient was positioned in a rotational platform with a radiation beam on, the dose angular dependence appeared to be reduced. The rotational positioning would average out the effects of different incident angles, leading to an averaged effect in terms of surface dose absorption. This implied that the angular dependence of the dosimeter become less effective because the dose received on the skin was averaged over the patient rotation.

Overall, in-vivo skin dose measurements can exhibit certain directional dependencies. However, in the DFR-TSET technique, the angular dependence can be further minimized, resulting in an average effect. One potential reason for this was the rotational positioning of the patient during the delivery of the radiation beam, which tended to mitigate a specific angular effect.