There will be differences in symmetry because of individual variation as you suggested and because of differences based on how they are manually drawn. This particular atlas contributed by Eickhoff and Zilles was transformed from MNI space in their Anatomy toolbox for SPM. Whether this is an artifact of how they have drawn the region is hard to say. I've attached an image of the superior temporal gyri for each hemisphere. I was able to verify the same ratio on the original and the Talairach-transformed version distributed with AFNI, and they duplicated your calculation of a ratio of about 0.72:

[~/spm5/toolbox/Anatomy_v15] dglen% 3dcopy ~/spm5/toolbox/Anatomy_v15/MacroLabels.hdr STG2+tlrc
++ 3dcopy: AFNI version=AFNI_2008_07_18_1710 (Sep 4 2009) [64-bit]
*+ WARNING: Assuming ANALYZE orientaion is LPI.
++ To change orientation or silence this message,
++ Set AFNI_ANALYZE_ORIENT to the proper orientation
++ in your .afnirc file.
++ e.g.: AFNI_ANALYZE_ORIENT = LPI
*+ WARNING: Assuming view is +orig.
++ To change view or silence this message,
++ Set AFNI_ANALYZE_VIEW to the proper view
++ in your .afnirc file.
++ e.g.: AFNI_ANALYZE_VIEW = orig
*+ WARNING: ANALYZE origin ignored in file /Users/dglen/spm5/toolbox/Anatomy_v15/MacroLabels.hdr
++ If datasets are out of alignment,
++ Set AFNI_ANALYZE_ORIGINATOR = YES
++ in your .afnirc file.
[~/spm5/toolbox/Anatomy_v15] dglen% ls STG2
STG2+tlrc.BRIK STG2+tlrc.HEAD
[~/spm5/toolbox/Anatomy_v15] dglen% ls STG2+tlrc.
[~/spm5/toolbox/Anatomy_v15] dglen% 3dclust 2 10 STG2+orig'<81..82>'
++ 3dclust: AFNI version=AFNI_2008_07_18_1710 (Sep 4 2009) [64-bit]
++ Authored by: RW Cox et al
** Warning: skipping dataset STG2+orig<81..82>
[~/spm5/toolbox/Anatomy_v15] dglen% 3dclust 2 10 STG2+tlrc'<81..82>'
++ 3dclust: AFNI version=AFNI_2008_07_18_1710 (Sep 4 2009) [64-bit]
++ Authored by: RW Cox et al
#
#Cluster report for file STG2+tlrc<81..82>
#[Connectivity radius = 2.00 mm Volume threshold = 10.00 ]
#[Single voxel volume = 1.0 (microliters) ]
#[Voxel datum type = byte ]
#[Voxel dimensions = 1.000 mm X 1.000 mm X 1.000 mm ]
#[Coordinates Order = RAI ]
#Mean and SEM based on Absolute Value of voxel intensities:
#
#Volume CM RL CM AP CM IS minRL maxRL minAP maxAP minIS maxIS Mean SEM Max Int MI RL MI AP MI IS
#------ ----- ----- ----- ----- ----- ----- ----- ----- ----- ------- ------- ------- ----- ----- -----
25258 -133.6 -88.7 74.3 -148.5 -110.5 -116.5 -50.5 51.5 91.5 82 0 82 -121.5 -112.5 51.5
18307 -22.3 -89.8 74.6 -40.5 -6.5 -118.5 -57.5 52.5 91.5 81 0 81 -30.5 -100.5 52.5
#------ ----- ----- ----- ----- ----- ----- ----- ----- ----- ------- ------- ------- ----- ----- -----
# 43565 -87.2 -89.2 74.4 81.58 0.0024 [~/spm5/toolbox/Anatomy_v15] dglen% 3dclust 2 10 CA_N27_ML::Superior_Temporal_Gyrus
++ 3dclust: AFNI version=AFNI_2008_07_18_1710 (Sep 4 2009) [64-bit]
++ Authored by: RW Cox et al
#
#Cluster report for file CA_N27_ML::Superior_Temporal_Gyrus
#[Connectivity radius = 2.00 mm Volume threshold = 10.00 ]
#[Single voxel volume = 1.0 (microliters) ]
#[Voxel datum type = byte ]
#[Voxel dimensions = 1.000 mm X 1.000 mm X 1.000 mm ]
#[Coordinates Order = RAI ]
#Mean and SEM based on Absolute Value of voxel intensities:
#
#Volume CM RL CM AP CM IS minRL maxRL minAP maxAP minIS maxIS Mean SEM Max Int MI RL MI AP MI IS
#------ ----- ----- ----- ----- ----- ----- ----- ----- ----- ------- ------- ------- ----- ----- -----
18262 -54.9 23.2 8.2 -69.0 -33.0 -3.0 58.0 -11.0 22.0 82 0 82 -44.0 0.0 -11.0
13218 49.4 22.1 8.4 33.0 64.0 -4.0 52.0 -11.0 22.0 81 0 81 40.0 9.0 -11.0
#------ ----- ----- ----- ----- ----- ----- ----- ----- ----- ------- ------- ------- ----- ----- -----
# 31480 -11.4 22.7 8.3 81.58 0.0028
[~/spm5/toolbox/Anatomy_v15] dglen% ccalc '13218/18262'
0.723798
[~/spm5/toolbox/Anatomy_v15] dglen% ccalc '18307/25258'
0.724800

On the other hand, the ratio of the superior temporal gyri from the TT_daemon is very close to 1.

[~/spm5/toolbox/Anatomy_v15] dglen% 3dclust 2 10 TT_Daemon::Superior_Temporal_Gyrus
++ 3dclust: AFNI version=AFNI_2008_07_18_1710 (Sep 4 2009) [64-bit]
++ Authored by: RW Cox et al
#
#Cluster report for file TT_Daemon::Superior_Temporal_Gyrus
#[Connectivity radius = 2.00 mm Volume threshold = 10.00 ]
#[Single voxel volume = 1.0 (microliters) ]
#[Voxel datum type = byte ]
#[Voxel dimensions = 1.000 mm X 1.000 mm X 1.000 mm ]
#[Coordinates Order = RAI ]
#Mean and SEM based on Absolute Value of voxel intensities:
#
#Volume CM RL CM AP CM IS minRL maxRL minAP maxAP minIS maxIS Mean SEM Max Int MI RL MI AP MI IS
#------ ----- ----- ----- ----- ----- ----- ----- ----- ----- ------- ------- ------- ----- ----- -----
34972 50.1 16.9 0.3 20.0 69.0 -22.0 65.0 -42.0 30.0 36 0 36 29.0 -9.0 -42.0
34635 -51.0 16.8 0.2 -69.0 -21.0 -22.0 65.0 -42.0 30.0 36 0 36 -34.0 -9.0 -42.0
#------ ----- ----- ----- ----- ----- ----- ----- ----- ----- ------- ------- ------- ----- ----- -----
# 69607 -0.2 16.8 0.2 36 0
[~/spm5/toolbox/Anatomy_v15] dglen% ccalc '34635/34972'
0.990364


Interestingly, several months ago, just looking at the overlap of symmetry for a single subject dataset we have here, I found the greatest difference in overlap - not by number of voxels but in a voxel-wise mirroring approach at the superior temporal gyrus. I used a script similar to the one I had posted here:

[afni.nimh.nih.gov]

[afni.nimh.nih.gov]

With the addition of a line to look at the ratio of the two hemispheres,

3dcalc -a lefthemi+tlrc. -b righthemi_flip_al2left+tlrc. -expr 'abs(a/b)' -prefix flipratio

The attached image shows the result for an axial view where differences were largest. In the overlay, red is at a level of 2 or more, green is less than one, and yellow is about one. Looking up that region with whereami showed the crosshairs to be the left superior temporal gyrus (and throughout the temporal gyrus really). A literature search shows that region has been examined for asymmetry recently too. This recent study for instance looked at this region as a predictor of verbal ability in children.

[dx.doi.org]

And this other study looked at the decrease in asymmetry in Williams Syndrome patients, citing asymmetry as the norm for superior temporal gyrus.

[ovidsp.ovid.com]

The planum temporale region within the superior temporal lobe is described here as an important location of asymmetry in this review article about asymmetry by Toga and Thompson,
[www.nature.com]