SYNCHDMLC geometry check - issue #393 and PR #406

[ojalaj]

There were some problems in SYNCHDMLC QT iso leaf geometry check between zt and zg, as reported in issue #393 and fixed in PR #406, based on Varian's confirmation that zg > zt, as seen in their proprietary figure (they had mixed the numerical values between those parameters in the table and horizontal lines in the figure).

However, I suspect that the check on line 1780 in SYNCHDMLCcm.mortran is still incorrect: IF(ZGROOVE$SYNCHDMLC(5)>ZTONGUE_$SYNCHDMLC(5))[

If zg > zt, shouldn't we check (as originally proposed by @marenaud): IF(ZGROOVE$SYNCHDMLC(5)<ZTONGUE$SYNCHDMLC(5))[ and output an error, if the check fails?

Also the figure in BEAMnrc manual for QT iso is incorrect.

Or do you have some more detailed information on the correct geometry? I apologize, if I have got confused and opened this issue in vain.

[rtownson]

@blakewalters this is still a valid issue, and agrees with your own comments on issue #393 and PR #406.

It gives an error if zg > zt, but the agreement was that zg > zt is correct and the error should occur if zg < zt.

[ftessier]

Was this resolved by merging #455?

[blakewalters]

Yup.

[ojalaj]

@blakewalters , @rtownson, @ftessier Sorry to still bother you with this, but isn't it still (line 1911): IF(ZGROOVE$SYNCHDMLC(5)>ZTONGUE$SYNCHDMLC(5))[ in PR #455,

but it should be IF(ZGROOVE$SYNCHDMLC(5)<ZTONGUE$SYNCHDMLC(5))[

[blakewalters]

Hi @ojalaj:

No worries, we want to make sure this is correct. After discussing this with @marenaud and looking at the TrueBeam Monte Carlo data, though, I came to the conclusion that, for quarter isocenter leaves, ZTONGUE should be > ZGROOVE. Does this not agree with data that you have?

[marenaud]

The latest word I got from Daren at Varian was:

The Monte Carlo data package is correct, if for the quarter leaf isocenter zg is interpreted as the dimension on the left side of the drawing, and zt on the right. Apparently the vertical positions of the horizontal lines are correct, and one needs to extrapolate them to determine what feature they indicate.

[ojalaj]

Hmm, this is bit confusing. If I understand correctly from what is described above, zg would be on the left side and zt on the right side in the TB MC data figure, but why? E.g. for half ISO leaf (and all the other types) zg is on the right side and zt on the left side. Actually I got a confirmation from Varian as well, but it was was vice versa what you have - according to that they had mixed the numerical values between those parameters in the table and horizontal lines in the figure. But maybe we just go with your conclusion for now.

[marenaud]

I'm honestly not too sure how to interpret the statement myself.

The way I interpreted it: "zg is interpreted as the dimension on the left side, and zt on the right" means that the lines in the figure are swapped. There's evidence to that effect, because the line on the left side matches exactly with the MLC dimensions on the right, and vice versa. In the other MLC leaf drawings, the lines indicating the zt/zg dimensions always match with the drawing, except for quarter iso.

The other sentence: "Apparently the vertical positions of the horizontal lines are correct, and one needs to extrapolate them to determine what feature they indicate." kind of supports that interpretation. Because you need to basically swap the lines.

With that interpretation, then the merged PR is correct. It's pretty frustrating that you got basically the opposite answer, also from someone at Varian :(

[blakewalters]

There's also the dimension table on p21 of the TrueBeam MC data handbook in which zg is clearly < zt for the quarter isocenter leaf.

[marenaud]

Makes it all the more confusing, since "the dimension on the left side" on the figure is clearly > "the dimension on the right", which contradicts the table if we interpret "the dimension on the left side" to be zg!

Another way to look at the issue is to ask ourselves whether there are any reasons to believe that quarter iso leaves are different than half-iso, other than having a different thickness. For half iso (and full iso), zg > zt, so having a different pattern for quarter iso is a bit strange? Unless they redesigned the leaf.

[blakewalters]

@marenaud, @ojalaj: I'm just going through old issues and found this unresolved one. The current version of SYNCHDMLC still assumes zg <= zt for QUARTER ISO leaves. Have either of you gained any more insight into this in the interim? If not, can we close this issue? A new issue can always be opened in light of new info. Thanks!

[ojalaj]

Hello @blakewalters and @marenaud ! My understanding is still that zt < zg for QUARTER ISO. If you need further confirmation, I'd would need to show several figures (online).

[marenaud]

Hello! I unfortunately haven't had the chance to delve deeper into the issue since the last time we looked at it. I would defer to @ojalaj.

[ftessier]

Thanks for the feedback, let's leave it open then.