A paper$^1$ I'm reading says fluence measures the incoming radiance from all directions and that fluence is similar to irradiance. It's defined by $\phi(x) = \int_{4 \pi} L(x, \vec{\omega'}) d\omega' $ (The tick mark seems to emphasize $\vec{\omega'}$ varies over the integration, rather than being a fixed direction.)
Intuitively, radiance is the amount of light given off in a direction but attenuated by how far off the surface is from being perpendicular to the light direction. $L=\frac{d^2\Phi}{d\omega dA \cos \theta}$
Irradiance is the area light density of a surface. $E=\frac{d\Phi}{dA}=\int_{2 \pi } L_i(\vec{\omega}) \cos \theta d\omega$
Intensity is the amount of light hitting a surface from a particular direction. $I(\vec{\omega})=\frac{d\Phi}{d\omega}$
It also looks like you can write the radiance as irradiance times intensity. $L=E\cdot I(\vec{\omega})$
Looking at the formulas, it seems like irradiance sums radiance over the hemisphere over a surface point and attenuated by how much the surface orients toward the light source, while fluence sums up radiance over the whole sphere surrounding a point (to measure inscattering).
Wikipedia says, "radiant fluence is the radiant energy received by a surface per unit area, or equivalently the irradiance of a surface integrated over time of irradiation." But this definition doesn't seem to match the one given above since there's no time involved.
http://omlc.org/education/ece532/class1/irradiance.html
- Real-Time Rendering of Translucent Materials with Directional Subsurface Scattering, Alessandro Dal Corso, 2014