- Purpose: Model-based dose calculation algorithms (MBDCAs) have become
clinically available within the past several years in commercially available
brachytherapy planning systems. This work investigates the application of a 510(k)
approved MBDCA Acuros™ BV and establishes a reproducible method of
validation. The research is broken into three primary research goals. The first
objective was to focus on helping medical physicists reproducibly validate the
implementation of MBDCA in the clinic by using direct measurements and
comparison to Monte Carlo (MC). The second objective was to commission a very
low energy spectrum using Attila™, the parent software to Acuros™ BV, by
comparison to MC. The third area retrospectively examines the application of
Acuros™ BV to a large group of previously treated patients by accelerated partial
breast irradiation brachytherapy.
Methods: A Quality Assurance phantom (QAP) was fabricated by taking a solid
(30x30x5 cm3) piece of water and drilling a 3 cm diameter hole to accommodate a
variety of pre-manufactured density plugs. The phantom was used to validate the
clinical implementation of the commercially available MBDC algorithm AcurosTM
BV. Once the QAP was imported into BrachyVision, the AcurosTM BV 192Ir dose
distributions were calculated. The QAP was modeled using Monte Carlo N-Particle
version 5 (MCNP5) and dose predictions were compared to AcurosTM BV at 0.2 cm
increments out to 4.5 cm. Additionally, physical measurements taken using Diode and
a MOSFET were also compared to AcurosTM BV. The second part of the
investigation used a similar QAP simulated three dimensionally in Attila™. This
investigation replaced the 192Ir source with a 50 kVp source spectrum provided by
Xoft, Inc. Attila™ and MCNP5 were used to simulate this low energy spectrum as the
source in the QAP at multiple points for comparison. The final investigation
examined two hundred accelerated partial breast irradiated (APBI) brachytherapy
patients that had been treated from 2008 to 2012 by the original TG43 formalism and
recalculated using AcurosTM BV. Clinically relevant endpoints such as V100, V90,
erythema and moist desquamation were compared between TG43 and Acuros™ BV.
Results: Over twenty points extending through different heterogeneities were
compared using an 192Ir source and calculated for both AcurosTM BV and MCNP.
Physical measurements taken on the QAP using solid-state detectors compared within
1% of Acuros™ BV calculations. The absolute dose of AcurosTM BV compared with
MC results varied from -3.0% up to 7.2%. When investigating the impact of
heterogeneities from a low kV spectrum, the absolute difference from Attila™ and
MC ranged from -15.1% to 9.0%. The final retrospective study showed noticeable
changes in recalculated dosimetry, with the most significant changes observed with
balloon devices. The difference in dose when AcurosTM BV was subtracted from
TG43 calculations showed balloons have reduced skin doses that ranged from 0.4%
to 10.2%. The differences for SAVI™ showed skin dose changes from -9.5% to
7.8%. For balloon devices, the average dose differences for V100, V95, V90 were
5.6%, 4.9% and 3.8%, respectively. The SAVI™ dose differences for V100, V95,
V90 were 2.2%, 2.1%, and 1.8%, respectively.
Conclusions: A phantom was fabricated to serve as a tool for clinical medical
physicists to efficiently and reproducibly validate the commissioning of MBDCAs.
This QAP aids in a fully inclusive system check from CT acquisition, to AcurosTM
BV calculations and finally diode or MOSFET measurements. Commissioning for
very low energy spectrum, such as eBx, was accomplished by importing this phantom
into Attila™ and comparing to MCNP calculations when encountering
heterogeneities. The final retrospective study suggest further investigation is needed
to assign uniform densities to heterogeneities encountered in the patient's CT images.
The recalculated patient dose using Acuros™ BV did predict a reduction in dose at
key points, such as tissue-air interface, V100 and V90 coverage, however, acute and
chronic skin dose thresholds showed no trending in data and were not predictable.
- description.provenance : Approved for entry into archive by Julie Kurtz(firstname.lastname@example.org) on 2017-06-04T00:32:40Z (GMT) No. of bitstreams: 2license_rdf: 1370 bytes, checksum: cd1af5ab51bcc7a5280cf305303530e9 (MD5)RogersKevinL2017.pdf: 3387414 bytes, checksum: 3a127afd74d0d5347c435f39cebaf5de (MD5)
- description.provenance : Submitted by Kevin Rogers (email@example.com) on 2017-05-31T04:17:44ZNo. of bitstreams: 2license_rdf: 1370 bytes, checksum: cd1af5ab51bcc7a5280cf305303530e9 (MD5)RogersKevinL2017.pdf: 3387414 bytes, checksum: 3a127afd74d0d5347c435f39cebaf5de (MD5)
- description.provenance : Made available in DSpace on 2017-06-12T20:54:42Z (GMT). No. of bitstreams: 2license_rdf: 1370 bytes, checksum: cd1af5ab51bcc7a5280cf305303530e9 (MD5)RogersKevinL2017.pdf: 3387414 bytes, checksum: 3a127afd74d0d5347c435f39cebaf5de (MD5) Previous issue date: 2017-05-16
- description.provenance : Approved for entry into archive by Steven Van Tuyl(firstname.lastname@example.org) on 2017-06-12T20:54:42Z (GMT) No. of bitstreams: 2license_rdf: 1370 bytes, checksum: cd1af5ab51bcc7a5280cf305303530e9 (MD5)RogersKevinL2017.pdf: 3387414 bytes, checksum: 3a127afd74d0d5347c435f39cebaf5de (MD5)