Dosimetric Analysis Of A Microselectron-HDR Ir-192 Source For Cervical Cancer Brachytherapy Based On AAPM TG-43 Using MCNP6.2

Cervical cancer remains a major public-health challenge worldwide, with incidence and mortality that continue to demand effective, organ-sparing treatment strategies.¹ Brachytherapy is a cornerstone of curative treatment for locally advanced cervical cancer because it permits delivery of very high doses directly to the tumor while rapidly reducing dose with distance to spare surrounding normal tissues.² High-dose-rate (HDR) brachytherapy using Ir-192 afterloading sources is widely adopted in clinical practice due to its favorable physical and radiobiological properties, including short half-life, high specific activity, and a gamma-ray energy spectrum that is well suited to intracavitary and interstitial applications.^3,4 Accurate dosimetry in HDR brachytherapy is therefore essential both to ensure tumor control and to minimize the dose to organs at risk (OAR) such as bladder and rectum, which lie in close proximity to the cervical target volume.

The American Association of Physicists in Medicine Task Group 43 (AAPM TG-43) formalised a practical and widely used framework for brachytherapy dosimetry that defines a set of key parameters geometry factor, radial dose function, anisotropy function, air-kerma strength and dose-rate constant  enabling consistent intercomparison of sources and validation of treatment planning systems.5,6 TG-43-based parameterisation remains the clinical standard for routine brachytherapy dose calculations, yet the protocol is defined for water-equivalent media and does not explicitly account for patient-specific heterogeneities or complex clinical geometries. Consequently, TG-43 values obtained in homogeneous water phantoms may fail to capture local perturbations in dose that arise from realistic tissue composition, organ contours, and interfaces encountered in pelvic brachytherapy. The literature therefore emphasises the need for independent verification methods and patient-based evaluations when precise dosimetric knowledge is required for clinical decision making.6

Monte Carlo particle-transport codes have become an established tool for high-accuracy brachytherapy dosimetry because they can model the full physics of photon interactions and arbitrary three-dimensional geometries, including heterogeneities and detailed source design.7 Previous Monte Carlo investigations of Ir-192 sources8,9 using a variety of codes (e.g., GATE, MCNP5) have demonstrated strong agreement with TG-43 reference data for standard, homogeneous phantoms while also enabling investigation of effects not captured by TG-43, such as encapsulation-induced anisotropy and perturbations near material interfaces.8,9 Nevertheless, many prior studies modelled sources in simple, water-equivalent geometries and relied on simulated geometries divorced from actual clinical source-to-organ distances derived from patient treatment plans; such simplifications limit the direct clinical translatability of the results.

A further consideration in cervical brachytherapy is the specific anatomy and organ composition of the treated population. Regional anatomical phantoms and reference compositions that better reflect the body habitus and organ densities of Asian populations have been published and can materially affect Monte Carlo dose estimates when used in lieu of generic, homogeneous water phantoms.10 Incorporating patient-relevant organ densities and clinically measured source-to-OAR distances into Monte Carlo models therefore holds the potential to reveal clinically meaningful deviations from TG-43 predictions particularly at short distances and near interfaces where attenuation and scatter differ from water assumptions. This need motivated the present patient-based inhomogeneous modeling approach adopted in this study, which employs Taiwanese reference phantom organ densities together with measured distances from a clinical treatment-planning case at the hospital (anonymized).

Against this background, the present work models a microSelectron-HDR Ir-192 source and quantifies the TG-43 dosimetric parameters (geometry factor G(r,θ), radial dose function g(r), and anisotropy function F(r,θ)) using the MCNP6.2 Monte Carlo code within a patient-based inhomogeneous pelvic phantom. The specific aims are: (1) to compute TG-43 parameters for a geometrically realistic Ir-192 source and to compare the results with TG-43 reference data, and (2) to evaluate the influence of patient-relevant tissue inhomogeneities and clinically measured source-to-OAR separations (extracted from treatment-planning system images) on local dose distribution around the cervix and to the bladder and rectum. By combining source-level modeling, Monte Carlo transport and clinically observed geometry, the study seeks to bridge the gap between idealised phantom dosimetry and the practical dose uncertainties encountered in routine clinical brachytherapy.

The outcomes reported in this study therefore address two interlinked needs of modern brachytherapy physics: first, provision of high-fidelity, independently computed dosimetric parameters for a commonly used HDR Ir-192 source that can serve as an external verification of TPS calculations; and second, an assessment of whether and when patient-specific heterogeneity corrections materially alter dose metrics to OARs in cervical treatments. These results aim to inform both clinical physicists tasked with quality assurance and researchers developing next-generation planning algorithms that incorporate heterogeneity corrections where clinically warranted.

MATERIALS AND METHODS

A computational, patient-based Monte Carlo approach was adopted to determine AAPM TG-43 dosimetric parameters for a microSelectron-HDR Ir-192 source. All modeling, simulation and post-processing steps described below follow the procedures documented in this study. The overall workflow combined detailed physical modeling of the Ir-192 source and its encapsulation, construction of an inhomogeneous pelvic region based on reference organ compositions, extraction of clinically measured source-to-organ distances from an institutional treatment-planning case, Monte Carlo transport using MCNP6.2, and post-processing to obtain G(r,θ), g(r) and F(r,θ) according to the TG-43 formalism.

This study used de-identified TPS data from a single clinical case. The protocol was reviewed and approved by the local Institutional Review Board / Medical Ethics Committee (approval number: KE/FK/1048/EC/2025). All patient data were anonymized and handled in accordance with local regulations.

1. Source geometry and material composition

The Ir-192 source was modelled to represent a microSelectron-HDR afterloader seed following the actual device dimensions used at the hospital (anonymized). The active core was represented as a cylindrical rod with active length ≈ 3.5 mm and diameter 0.6 mm and material density 22.42 g·cm⁻³. The encapsulation was modelled as a concentric cylindrical shell of stainless steel AISI-316L with a total capsule length of 4.5 mm, capsule wall length 0.9 mm (capsule geometry as implemented in the MCNP input), and density ~8.02 g·cm⁻³. The capsule composition used the elemental fractions (Fe ≈ 68%, Cr ≈ 17%, Ni ≈ 12%, Mn ≈ 2%, Si ≈ 1%).¹¹ The emission spectrum for 192Ir was represented by the principal gamma lines (0.205, 0.296, 0.308, 0.316, 0.468, 0.484, 0.589, 0.604 and 0.612 MeV)¹² and implemented in the MCNP source definition to reproduce the physical photon emission.

2. Phantom construction and clinical geometry

A patient-based inhomogeneous pelvic region was constructed using organ compositions and densities drawn from the Taiwanese reference.¹⁰ The phantom explicitly included the cervix/tumor region and the two principal organs at risk for intracavitary cervical brachytherapy: bladder and rectum. Clinically measured source-to-organ distances were extracted from a single treatment-planning system (TPS) case at the hospital (anonymized) using the TPS measurement tools; the nearest recorded distances used for analysis included rectum points at ~3.89 cm and bladder at ~4.07 cm from the source (TPS), and these clinical distances were used to position scoring volumes relative to the source in the inhomogeneous phantom. The phantom was discretised for dose scoring by placing small water spheres at the measurement points described below.

3. Monte Carlo transport and scoring parameters

All particle transport simulations were performed with MCNP6.2. Photon histories were simulated from the defined Ir-192 source with the full energy spectrum described above. Dose scoring used the F6 energy-deposition tally (MeV·g⁻¹ per source particle) recorded in small water volumes (spherical detectors) positioned at radial distances from the source center. Tally spheres of radius 0.05 cm were placed along radial lines at r = 0.5–9.0 cm with an increment of 0.5 cm for the radial dose function measurements; radial directions included 90° and 270° for g(r) determination and a full polar angle set (0°, 15°, 30°, …, 165°, 180°) for anisotropy F(r,θ). The simulations used a total number of histories sufficient to produce low statistical uncertainties (the documented results used large sample sizes e.g., up to 1×10⁹ histories for key runs where stated) and that standard MCNP variance estimates were used to report statistical error. The MCNP input also contained material cards for the organ compositions and the stainless-steel capsule as noted above.

4. Activity normalization and unit conversion

Simulated tallies (MeV·g⁻¹ per history) were converted to clinically interpretable dose rates (Gy·s⁻¹) by applying the energy-to-dose conversion factor (1 MeV·g⁻¹ = 1.602×10⁻¹⁰ Gy per photon) and scaling with an assumed source activity. For the work presented in this study an example activity of 10 Ci (3.7×10¹¹ decays·s⁻¹) was used when converting tally results to absolute dose rate; this yielded a multiplicative factor FM ≈ 59.274 Gy·s⁻¹ for the stated activity when combined with the tally units, as documented in the raw calculation workflow. Where normalization to TG-43 parameter definitions was required (for example normalizing g(r) at r₀ = 1 cm, θ₀ = 90°), the computed dose rates were divided by the dose rate at the reference point to obtain dimensionless radial and anisotropy functions.

5. TG-43 parameter calculation and geometric functions

The geometry factor G(r,θ) was calculated analytically for a line-source geometry consistent with TG-43 definitions using the modeled active length L of the source. The radial dose function g(r) was obtained by dividing the MCNP-derived dose rate along the transverse axis (θ = 90°/270°) by the geometric factor and then normalising to the value at r = 1 cm. The anisotropy function F(r,θ) was computed as the ratio of the MCNP dose rate, corrected for the geometry factor, at each polar angle θ to the corresponding transverse-axis (θ₀ = 90°) value at the same r, again following TG-43 prescriptions.⁶ Data processing and tabulation of G(r,θ), g(r) and F(r,θ) were performed using Microsoft Excel; plots and tabulated comparisons versus AAPM reference values used the processed numeric arrays produced in this step.

6. Clinical comparators, uncertainty assessment and limitations

Simulated TG-43 parameters and dose rates were compared with AAPM TG-43 reference data and with the TPS-derived clinical distances/dose points to assess the influence of tissue inhomogeneities. Uncertainty analysis considered MC statistical uncertainties reported by MCNP, geometric modeling assumptions (e.g., source and capsule dimensions), and the limitation inherent to using a single clinical TPS case for clinical positioning. This study explicitly notes these limitations and recommends larger clinical sample sizes and further parameter studies (e.g., air-kerma strength, dose-rate constant) for follow-up work. All computational runs and processing steps were logged and archived per the documented study workflow to enable traceability and reproducibility.

RESULTS

The MCNP6.2 simulations produced detailed dosimetric data for a microSelectron-HDR Ir-192 source modelled with an active length of ≈3.5 mm and diameter ≈0.6 mm, encapsulated in stainless steel (AISI-316L). Photon emission was represented by the principal Ir-192 gamma lines and simulations were performed for radial points r = 0.5–9.0 cm (Δr = 0.5 cm) and polar angles θ covering 0°–180°. Tallying used small water spheres (radius 0.05 cm) and the MCNP F6 energy-deposition tally; results were converted to dose rate (Gy·s⁻¹) using the conversion factor 1 MeV/g = 1.602×10⁻¹⁰ Gy and an example activity of 10 Ci (3.7×10¹¹ s⁻¹), giving a multiplicative factor FM ≈ 59.274 Gy·s⁻¹ for absolute scaling. The simulations used large numbers of histories to reduce statistical uncertainty (documented runs with up to 1×10⁹ histories for key cases).

Geometry-factor G(r,θ) calculations followed the TG-43 line-source formulation and were used to normalise the raw MCNP dose-rate data for g(r) and F(r,θ). The computed G(r,θ) behaved as expected for the chosen active length and was applied consistently in subsequent normalisations (see Table 1 for representative values).

Table 1. geometric function value of Ir-192 radiation source.

Using the transverse axis (θ = 90°/270°) dose rates, the radial dose function g(r) was determined by normalising the geometry-corrected dose to the reference point r0 = 1.0 cm, θ0 = 90°. The resulting g(r) curve exhibits the characteristic monotonic decrease with distance and closely follows the AAPM TG-43 reference behavior. Quantitatively, g(r = 1 cm) matched the reference value and the mean percentage difference of g(r) relative to the AAPM TG-43 dataset across the tested range was reported to be < 1%, indicating excellent numerical agreement between the Monte Carlo model and TG-43 benchmarks. Representative numerical results and the comparison plot are provided in the Table 2 and Figure 1

Table 2. radial dose function value of Ir-192 radiation source.

r	g(r)	90°		270°
	AAPM TG 43	g(r)	Diff.	g(r)	Diff.
0,5	0,997	1,0036	0,0066	1,002	0,005
1	1	1	0	1	0
1,5	1,003	1,0023	0,0007	1,0026	0,0004
2	1,005	1,002	0,003	1,0028	0,0022
2,5	N/A	1,0027	N/A	1,0031	N/A
3	1,008	1,0022	0,0058	1,0037	0,0043
3,5	N/A	1,0009	N/A	1,0048	N/A
4	1,007	1,0007	0,0063	1,0013	0,0057
4,5	N/A	1,0025	N/A	0,9964	N/A
5	1,003	0,9965	0,0065	0,9926	0,0104
5,5	N/A	0,9997	N/A	0,9923	N/A
6	0,996	0,9962	0,0002	0,9873	0,0087
6,5	N/A	0,9985	N/A	0,9743	N/A
7	N/A	0,9888	N/A	0,9651	N/A
7,5	N/A	0,9826	N/A	0,9637	N/A
8	0,972	0,9658	0,0062	0,9671	0,0049
9	N/A	0,9669	N/A	0,9476	N/A
Mean			0,003922		0,004622

Figure 1. Comparison of Radial Dose Function Graphs of Simulated Angles of 90° (orange) and 270° (grey) against the AAPM TG-43 reference (blue).

Anisotropy function F(r,θ) results show the expected angular dependence imposed by the finite active length and encapsulation: maximal values near the transverse plane (θ ≈ 90°) and pronounced reductions toward θ = 0° and 180°, consistent with increased self-shielding and pathlength through the capsule at polar extremes. This study reports that F(r,θ) curves from MCNP6.2 follow TG-43 trends and the simulated anisotropy curves reproduce the qualitative shape of reference anisotropy plots; departures from the reference occur predominantly at short distances and in angular sectors near material interfaces, where encapsulation and heterogeneity effects are most influential (see Figure 2 and Table 3).

Table 3. anisotropy function of Ir-192 radiation source in the direction 0° to 180°.

Figure 2. Anisotropy function of Ir-192 radiation source.