eMC - The Homogeneous Set

Standard Applicator Inserts

The first task is simple but time-consuming: calculating all combinations of energy and applicator with the default inserts that come with the applicators (not home-made inserts) using the standard SSD of 100cm. The goal is to check relative dose distributions and absolute dose (MU), with the calculation settings given above.

The planning is easy because the field sizes are exact (except for the corners, which have 8mm triangles filled out). This can be seen in the following calculation(blue) and measurement (red) of isodoses parallel to the surface in 20mm depth, for a 20MeV field and field size 4x4cm:

It is interesting to see that for 20MeV, the low dose region outside the field (still under the cerrobend cutout material) is a little underestimated by eMC: the measurement (red) is higher. The reason could be in the measurement (higher sensitivity of the electron diode for low-energy scatter) or the algorithm (IPS model, see eMC1.htm). All our home-made cutouts (see below) are molded a little thicker, so the verification of these could answer the question.

Anyway, taking into account that this is the smallest field size we allow and the highest energy, the agreement here is still very good:

The following table shows the number of particle histories that had to be processed until 1% accuracy was reached:

Particles	E06	E09	E12	E16	E20
4x4orig	6910000	12010000	22690000	41720000	63180000
6x6orig	7180000	9780000	16240000	27040000	38820000
10x6orig	11090000	14430000	25010000	40970000	58940000
10x10orig	15920000	22260000	37470000	61560000	85770000
15x15orig	37990000	48910000	78560000	122860000	272610000
20x20orig	64580000	86480000	138720000	213380000	295290000
25x25orig	100780000	138010000	209790000	316440000	447070000

(Values in bold were recalculated with Eclipse 8.1 using the Distributed Calculation Framework with a parallelization factor between 6 and 12).

The calculation times given in the next table are taken from the "Calculation Log". They are realistic in the sense that the calculations were done mid-day, under normal clinical conditions, between all other photon calculations. Since the workload on the calculation servers is not constant, there may be large variations in calculation time, for the same field.

Conditions for calculation were: 3D-image with 300 slices (slice thickness 1mm), grid 2.5mm, 1% accuracy, low 3D-gaussian smoothing.

Calc.Time [s]	E06	E09	E12	E16	E20
4x4orig	87	138	194	274	433
6x6orig	301	436	1175	1069	1827
10x6orig	942	834	3686	4901	3468
10x10orig	1266	1048	2027	5585	7140
15x15orig	2590	3589	3882	7263	1687
20x20orig	5917	6198	24585	14661	16368
25x25orig	4909	7844	20059	17513	25739

Dividing the two above tables gives the number of particle histories per second (note the increase in performance for the DCF-calculated fields in bold):

Histories/second	E06	E09	E12	E16	E20
4x4orig	79425	87029	116959	152263	145912
6x6orig	23854	22431	13821	25295	21248
10x6orig	11773	17302	6785	8360	16995
10x10orig	12575	21240	18485	11022	12013
15x15orig	14668	13628	20237	16916	161595
20x20orig	10914	13953	5642	14554	18041
25x25orig	20530	17594	10459	18069	17369

Home-Made Applicator Inserts

For the home-made inserts, which are usually rectangular or round, having all kinds of field sizes and which are sometimes asymmetric, care has to be taken when planning the field in Eclipse. The actual field sizes in isocenter somtimes differ a little (up to 3mm) from the nominal field sizes. This is due to the molding process. To avoid misinterpretation of the results, it is therefore a good idea first to do the measuement, determine the actual field size, and then calculate the field with the field size determined in the water phantom.

The following images show a typical comparison of a calculated (blue) and a measured (red) crossplane distribution along the depth axis, for a 4x8cm asymmetric applicator insert and 12MeV. The Gantry is at 0°:

Rotating the Gantry to 25° gives oblique incidence on the water surface (9MeV, SSD100, 6x6orig insert):

The Gamma Index is used here to evaluate the agreement. The profiles cut vertically and horizontally through the "failed" region (blue is the calculation):

The gamma criterium used was 2mm/2%, but with a suppression of doses below 5%:

back to Beam Verification Set