I came across a repeatable issue. The sensitivity appears inverted along the x-axis when reconstructing SPECT data. As a consequence attenuation correction can be wrong.
This happens in both master and in #445 but with inverted axis.