It depends on what is to be pasteurized. If one aims for pasteurizing the surface only, then the shape is more or less unimportant. If however one wants to pasteurize the core, then the shape will affect the times.
To be on the safe size, measure the thickness where the meat is thickest.
Myhrwold writes in http://forums.egullet.org/index.php?/topic/116617-sous-vide-recipes-techniques-equipment-2004-2010/page_st_120_p_982720#entry982720 about cooking times (not pasteurization times):
This is true for any kind of cooking actually. Most of what we are
taught about cooking is actually wrong. For example, anytime somebody
tells you it is “10 minutes per pound” they are saying something that
cannot possibly be accurate, because this would imply that cooking
time is proportional to the weight. If you take something like a whole
bird and scale it up you will find that cooking time is actually
proportional to something like (weight)^(2/3) – weight to the 2/3
power. This is because increasing the weight scales up the thickness
by the cube root. Since most people are not accustomed to taking
things to fractional powers, people substitute a linear relationship.
That might work out OK in practice over a small range, but it can’t be
accurate over a large range. For example, if you double the weight,
the linear relationship would tell you that you double the cooking
time. The 2/3 power would tell you to increase it by 59% - that is a
pretty big difference.
As a general rule of thumb heat diffusion times go as the square of
the thickness increase. So, doubling the thickness results in FOUR
TIMES the cooking time. That is a rough general rule of thumb, which
is not perfect but it illustrates the nonlinearity of the system. I
don’t know why this isn’t taught to chefs more often because it is a
fairly easy thing to grasp and use.
A side note:
If you have an iPad or iPhone then I can recommend SousVideDash. One must enter meat type, shape and size, initial and final temparature. Then graphs are drawn that shows pasteurization times for both surface and core (or more precisely graphs of the reduction of salmonella and listeria).
See the website:
http://www.sousvidedash.com/