Mapping malignant infiltration into lymph nodes can have a critical impact on patient decision making. We predicted that multigradient echo (MGE) experiments could reflect cellularity and cell-size changes due to underlying susceptibility distributions. Lymph nodes extracted from rectal cancer patients exhibited the predicted non-monotonic and non-mono-exponential MGE signal decay, which provided insights into the underlying microstructure. A simple model distinguished benign from malignant nodal tissue and the differences were at least partially explained by differences in cellularity and cell size. These results can impact lymph node staging accuracy, as already corroborated by our pilot results in-vivo, upon rectal cancer staging, at 1.5T.