Academic Courses > DOS 522 > Monitor Unit Discrepancies
DOS 522 - Week 5 Discussion
Writing Prompt
What is an acceptable MU/dose discrepancy (i.e. treatment planning system vs. hand calculation)? Does it vary on treatment modality, technique, or body site?
Initial Post: Computerized MU Checks vs Hand Calculations
Monitor unit (MU) calculations performed by either electronic second-check software or by hand calculations are a useful tool for judging whether a radiation therapy plan is potentially problematic. At Loyola University Health System, the guidelines for MU discrepancies between RadCalc second check calculations and Eclipse plan calculations are not interpreted as hard rules, but as guidelines, so that room is left for professional judgement (Mike Mysz, oral communication, February 18, 2015). The reason that room for professional judgement is required is that differences in computation algorithms, different anatomic sites, and other perturbing factors can all throw off calculations significantly in some scenarios. We generally have high expectations and prefer to see MU differences in the 1% or less range, but we will not be concerned until differences are greater than 3%. At that point, we will look into the plan in more detail to determine if the anomaly is acceptable or if corrective action needs to be taken.a
Mike explained that two common cases where values can vary significantly are electron plans and breast tangents. Electrons behave very differently from photons when interacting with matter, and simplistic models in second check systems can not usually take into account the subtleties of these interactions. In cases such as breast tangents, RadCalc's simple tissue inhomogeneity calculation model, which relies on pre-computed effective depth, can not take into account side-scatter differences where air or lung tissue may not be contributing to the dose at the calculation point to the extent that RadCalc expects. Small field-in-field strips in these plans can also be problematic, because in some cases, the calculation point may actually be blocked by multileaf collimator (MLC) leaves, requiring the creation of a secondary calc point to get a reasonable value. The majority of a small field-in-field strip's volume may be in close proximity to lung tissue, leading to higher than expected dose because of the aforementioned scatter contribution problems.
For VMAT second checks, RadCalc has an option to use actual CT data for tissue inhomogeneity correction, and this is used in our checks. This provided much better accuracy for these types of plans, but it is not considered to be necessary for checks of simpler plans like conformal plans.
Judgement can also come into play when determining whether a discrepancy is worth addressing. If a small field-in-field only has 5 MU assigned to it, even a 10% discrepancy may have minimal impact on a plan. If it is determined that the source of the discrepancy is well understood, even discrepancies in this range may be determined to be acceptable.
Mike explained that two common cases where values can vary significantly are electron plans and breast tangents. Electrons behave very differently from photons when interacting with matter, and simplistic models in second check systems can not usually take into account the subtleties of these interactions. In cases such as breast tangents, RadCalc's simple tissue inhomogeneity calculation model, which relies on pre-computed effective depth, can not take into account side-scatter differences where air or lung tissue may not be contributing to the dose at the calculation point to the extent that RadCalc expects. Small field-in-field strips in these plans can also be problematic, because in some cases, the calculation point may actually be blocked by multileaf collimator (MLC) leaves, requiring the creation of a secondary calc point to get a reasonable value. The majority of a small field-in-field strip's volume may be in close proximity to lung tissue, leading to higher than expected dose because of the aforementioned scatter contribution problems.
For VMAT second checks, RadCalc has an option to use actual CT data for tissue inhomogeneity correction, and this is used in our checks. This provided much better accuracy for these types of plans, but it is not considered to be necessary for checks of simpler plans like conformal plans.
Judgement can also come into play when determining whether a discrepancy is worth addressing. If a small field-in-field only has 5 MU assigned to it, even a 10% discrepancy may have minimal impact on a plan. If it is determined that the source of the discrepancy is well understood, even discrepancies in this range may be determined to be acceptable.
|
Academic Courses > DOS 522 > Monitor Unit Discrepancies
|
Written February 18, 2015
Second Semester, 2 Months into Internship |