PDF
Citation
Print
Introduction
Advanced radiotherapy techniques, such as intensity-modulated radiotherapy (IMRT) and volumetric modulated arc therapy (VMAT), employ highly modulated multi-leaf collimator (MLC) motions to improve target conformity. This, in turn, increases the plan complexity and demands higher beam delivery accuracy [1-6]. In such cases, the delivered beam may deviate from the planned values, leading to errors in patient dose predictions. To prevent such issues, patient-specific quality assurance (PSQA) is performed before treatment by delivering the plan to a QA phantom and comparing the measured dose distribution with the calculated distribution using gamma analysis. Although its clinical relevance has been questioned, it remains widely adopted in clinical practice because no practical alternative has yet been established [7,8]. Despite acceptable agreements between the treatment plan and PSQA results, beam delivery errors can still occur during actual treatments. Real-time beam monitoring verifies beam delivery during treatment, allowing timely error detection. One such method uses a detector mounted on the gantry head to measure the transmitted beam signal and analyze it segment-by-segment (SBS).
The Integral Quality Monitor (IQM; iRT Systems) is a large-area, transmission-type ionization chamber mounted on the treatment head [9]. By comparing the measured signal with the planned signal in real time, the system can be used for both real-time delivery monitoring and PSQA. Unlike conventional QA systems that require additional phantom setups and detector alignment, the gantry-mounted IQM enables rapid and reproducible measurements, improving workflow efficiency and reducing operator dependency.
We commissioned a newly installed IQM system on a Versa HD linear accelerator (LINAC) (Elekta AB) to characterize its response for four photon beams—6 MV, 10 MV, 6 MV flattening filter free (FFF), and 10 MV-FFF—and to confirm that the overall beam characteristics were acceptable. This study summarizes the commissioning process and evaluates the suitability of the IQM system.
Materials and Methods
1. IQM system and workflow
Fig. 1 shows a schematic of the IQM alongside a photograph of the IQM mounted on the gantry head of a LINAC. The IQM comprises a large-area parallel-plate ionization chamber with a gradient-sensitive volume with electrodes and insulator, as well as an electrometer, an inclinometer, and associated control electronics. As the photon beam passes through the chamber, the air between the aluminum electrodes becomes ionized, and the resulting collected charge is recorded in real time. Since the thickness of the sensitive volume varies laterally, the measured signal depends on the beam’s incident position along the lateral direction. The accumulated signal indicates the fluence delivered through the detector for the given beam aperture.
Fig. 2 illustrates the IQM-based PSQA workflow. The system first calculates the expected SBS signal using the treatment plan MLC sequence contained in the DICOM file. Here, a “segmenter,” as defined by the manufacturer, refers to the IQM sampling unit and may include one or more control points (or segments), depending on the assigned monitor unit (MU). The user sends the treatment plan DICOM file to the IQM server, which computes the expected signals, while the IQM measures the actual signal for each segmenter during delivery. The segmenter-by-segmenter cumulative signals are then compared with the calculated signals.
2. Beam model generation for the IQM
Since the IQM calculates the expected segmenter-by-segmenter cumulative signal for each beam, the beam model must accurately reproduce the LINAC’s beam characteristics. To construct this model, inline and crossline profiles from the Versa HD were acquired using a BluePhantom2 scanning system (IBA Dosimetry GmbH). A CC13 ionization chamber (IBA Dosimetry GmbH) was used as the field detector, and a Stealth chamber (IBA Dosimetry GmbH) served as the reference detector, both operated under a common control unit [10]. The measurement data were acquired and processed using the MyQA Accept software (version 9.0.17; IBA Dosimetry GmbH).
For 6 MV and 10 MV beams, profiles were measured in water at a source-to-surface distance (SSD) of 100 cm, at the depth of maximum dose (dmax), using a 40×40 cm2 field. For 6 MV-FFF and 10 MV-FFF beams, profiles were measured in the air, following the manufacturer’s recommendations, at a source-to-detector distance of 100 cm, using the same 40×40 cm2 field and the same detectors. Acrylic buildup caps (30 mm for 6 MV-FFF and 40 mm for 10 MV-FFF) were used to ensure charged-particle equilibrium. All profiles were smoothed using the least-squares method, interpolated at 0.2-mm intervals, and normalized to the central axis. The measured profiles were sent to the manufacturer to generate the beam model.
The beam model was verified by evaluating signal difference and profile symmetry. IQM signals were measured for MLC fields at various positions with a field size of 4×4 cm2, as illustrated in Fig. 3. Signals from a total of 60 fields were evaluated, with each field delivered using 50 MU. Signals at different field positions were normalized to that of the central field. For each position, calculated and measured signals were compared along the gradient and non-gradient directions, with tolerances of ±5% for the gradient direction and ±10% for the non-gradient direction. Symmetry was verified to be within 3%.
3. Commissioning of the IQM
1) Area integrated output factor
The area integrated output factor (AOF) is a correction factor that accounts for field-size-dependent output variations. The factor is defined as the relative changes between the beam outputs of various rectangular fields and the output of the reference field size, which was 10×10 cm2. AOFs were measured for all energies with field sizes from 1×1 cm2 to 40×40 cm2 (50 MU per field) and compared with the values calculated by the IQM. The maximum allowable difference between measured and calculated AOFs was ±1% for all field sizes, except for the 1×1cm2 field, which had a tolerance of 2%.
2) Dose output correction factor
The dose output correction factors (DOCF) was defined based on the absolute dose calibration under the reference condition. All beams were calibrated to 1 cGy/MU at dmax in water according to the AAPM TG-51 protocol [11]. A Farmer-type ionization chamber (TW30010; PTW GmbH) and a Unidos Tango electrometer (PTW GmbH) were used in a 10×10 cm2 field at 100 cm SSD, and the reading at 10 cm depth was corrected to dmax using the percent depth dose at 10 cm.
The IQM reference signal was then measured using a 20×20 cm2 field and 200 MU. Three repeated measurements were required to demonstrate a standard deviation ≤0.5%. The DOCF was determined from the reference signal and was required to be within ±3% of unity.
3) Dosimetric properties
The following dosimetric properties were evaluated using a 6 MV beam and a 10×10 cm² field.
• Reproducibility: 100 MU was delivered 10 consecutive times; the percent standard deviation of the IQM signal was required to be ≤0.5%.
• MU linearity: 5, 10, 20, 50, 100, 200, and 500 MU were delivered; the signal was required to demonstrate linearity with R2>0.999.
• Dose-rate dependency: 50 MU was delivered at 100, 200, 300, 400, and 600 MU/min; the signal variation with dose rate was required to be ≤0.5%.
4) Miscellaneous characteristics
Temperature and pressure readings from the IQM were compared with calibrated instruments (XP100 and XA1000; Lufft) and were required to agree within ±1°C and ± 2 mmHg. The inclinometer calibration was verified by rotating the gantry from −150° to 180° in 30° increments and the collimator angle from −135° to 180° in 45° increments. The correlations between the set angles and the IQM readings (roll, pitch, yaw) were analyzed, and the chi-square value was required to be less than 1.
4. Clinical verification of the IQM
The IQM evaluates the beam delivery in two ways: cumulative and SBS. The cumulative evaluation compares the running sum of the measured signals with the calculated cumulative signal for the entire field, thereby reflecting the overall delivery accuracy (Fig. 4a). The SBS evaluation compares the measured and calculated signals for each segmenter, providing local information on delivery accuracy (Fig. 4b).
Clinical verification was performed for all energies, and Table 1 summarizes the number of plans by energy and delivery technique (dynamic MLC [DMLC], VMAT, stereotactic radiosurgery [SRS]/stereotactic body radiation treatment [SBRT]). DMLC refers to fixed-gantry sliding-window IMRT; VMAT refers to single- or multi-arc rotational delivery with variable modulation; and SRS/SBRT refers to single- or multi-arc plans using small fields, typically with one or more separate apertures averaging ≤3 cm in dimension and an average MU per degree of gantry rotation >20 MU/°.
For the cumulative evaluation, all measurements were required to fall within the watch level (±3.0%), and the overall distribution of measurements was expected to remain within the wider action level (−7.1% to +5.5%) specified by the manufacturer [12,13]. The action level was derived from the correlation between IQM signal deviations and dose-volume histogram changes in plans with intentional MLC or MU errors, as reported in previous clinical studies.
For the SBS evaluation, only the manufacturer-driven watch level was applied, set at ±20%, indicating that at least 95% of the segmenters had to fall within ±20% of the expected signal [13].
Results
1. Generated beam model of the IQM
The inline and crossline profiles for each beam energy are illustrated in Fig. 5. All profiles exhibited acceptable symmetry, within 1%. Additionally, no observable differences were found between inline and crossline profiles, with local maximum differences within 2%. Table 2 presents the results of signal differences and symmetry between measured and calculated values.
2. Results of the IQM commissioning
Tables 3 and 4 summarize the results of the IQM commissioning. DOCF and AOF for all energies were within the predefined tolerance ranges, including the measurements for the 1×1 cm2 field. The variability of repeated measurements was minimal, and the MU–signal relationship was highly linear across the tested range. The IQM response showed no clinically significant dependence on dose rate, and the built-in temperature/pressure sensors agreed with the reference instruments within the manufacturer’s specified limits. The angle calibration test for gantry and collimator rotation also demonstrated accurate geometric mapping.
3. Clinical verification of the IQM to be used in patient-specific quality assurance
All beam energies and treatment techniques achieved acceptable cumulative and SBS pass rates, as summarized in Table 5. Fig. 6 shows the distribution of IQM signal deviations for each segment, along with the watch and action levels; the majority of segments remained within the watch level.
The lowest cumulative and SBS pass rates were observed for the 10 MV DMLC cases, whereas 6 MV beams exhibited the highest pass rates regardless of delivery technique. In most cases, SBS pass rates were slightly lower than the corresponding cumulative pass rates, except for the 10 MV-FFF cases. FFF beams tended to yield lower pass rates than their flattened counterparts at the same nominal energy, consistent with the wider spread of IQM signals shown in Fig. 6.
Discussion
The IQM system was installed and commissioned on the Versa HD LINAC, and the beam model demonstrated acceptable agreement between calculated and measured values within the manufacturer’s predefined tolerance. Profile differences in both gradient and non-gradient directions, as well as their symmetries, were within the predefined tolerances, confirming that the modeled transmission through the detector is clinically acceptable. The AOF for a 1×1 cm2 field was evaluated separately due to higher measurement uncertainty in smaller fields, but it remained within the acceptable tolerance. The observation that the largest AOF differences were negative (measurement<calculation) is consistent with small-field conditions, where backscatters from the jaws and MLC to the monitoring chamber can slightly overestimate the delivered MU, resulting in lower IQM signals for those fields [14,15]. The dosimetric property tests also met the acceptance criteria, and the built-in temperature/pressure and angular sensors performed within the manufacturer’s limits, supporting the routine use of the device for PSQA.
Clinical verification using representative DMLC, VMAT, and SRS cases demonstrated that both cumulative and SBS evaluations met the tolerance thresholds. More scattered signal differences were observed in the initial segments, along with larger tolerance levels. Random errors due to limited signal statistics were reduced as the IQM signals accumulated during beam delivery [12,13]. Here, the watch and action levels for cumulative evaluation were defined as the values converged after gradual reduction during the initial segments [13]. As expected, SBS pass rates were lower than cumulative pass rates because individual segments deliver much smaller signals and are therefore more affected by random variations. FFF beams showed a higher fraction of segments near or outside the watch/action levels, which is consistent with the higher modulation and smaller apertures typically used in these plans, rather than indicating any deficiency of the IQM itself.
Unlike conventional PSQA based on 2D/3D gamma analysis, the IQM provides an online comparison of measured transmitted fluence against a pre-calculated reference, using cumulative and SBS thresholds instead of spatial dose agreement. This approach enables real-time detection of delivery deviations while requiring site-specific establishment of pass levels. Several studies have demonstrated a relationship between IQM signals and gamma results, as well as target/organ at risk dose−volume parameters [12,13,16-18]. There was moderate statistical correlation between the two values (R2<0.7) [16]. Meanwhile, the target mean dose and organ dmax showed a strong correlation with the IQM signal (R2>0.9) [12]. Although these studies support the clinical relevance of the IQM, larger multi-institutional datasets are still needed to establish broadly applicable tolerances for diverse techniques and beam qualities.
This system also has practical limitations. If the detector is to remain mounted during actual patient treatments for real-time monitoring, a separate “IQM-on” beam model must be created in addition to the conventional “IQM-off” model, because the presence of the transmission chamber alters the beam characteristics [19]. With both models commissioned and the detector kept in place, the system can be used not only for PSQA but also for real-time treatment monitoring, reducing operator dependence and improving delivery traceability. Although the present work focused on commissioning and clinical verification, the results support future efforts to expand continuous, real-time beam monitoring in the clinic.
Conclusions
The IQM was successfully commissioned on the Versa HD LINAC and demonstrated adequate accuracy and stability for use in PSQA. The calculation model and detector properties agreed within predefined tolerances, and clinical verification yielded reliable pass rates across various photon treatment techniques. These findings provide practical background for installing and commissioning the IQM for routine PSQA and indicate its potential extension to real-time treatment monitoring when an IQM-on-beam model is implemented.
XML Download