Breast cancer related lymphoedema of the arm (BCRL) is a peripheral swelling which results from impaired lymphatic drainage after breast cancer treatment [1]. The pathophysiology of BCRL is not well understood, one of the main limiting factors being a lack of anatomical and functional information on both superficial (epi-fascial) and deep (sub-fascial) lymphatic collecting vessels. We have recently proposed a novel contrast-enhanced magnetic resonance lymphangiography (CE-MRL) protocol which minimizes venous enhancement and allows the identification of upper limb lymphatics via the use of associated contrast uptake curves [2]. In this work we quantify differences between affected and unaffected arms in a cohort of patients with unilateral BCRL.Both arms of 10 BCRL patients (table in Figure 1) were imaged at 1.5 T (MAGNETOM Aera, Siemens AG, Erlangen, Germany) on separate visits. A high resolution 3D T1W fast-spoiled gradient-echo pulse sequence was employed (TE/TR=2.77/6.14 ms, flip angle=12°, voxel size=1x1x1 mm, spectrally selective fat suppression, FOV=300 mm in the superior/inferior direction, acquisition time=66 s). Contrast was administered with a 1 ml total volume intradermal injection to each of the four inter-digital spaces (0.02 ml of gadoteridol, 0.1 ml of 1% lidocaine and 0.88 ml of saline, with resulting [Gd]=0.01 M) [2]. The forearm was imaged with 15 dynamic acquisitions, in order to follow the passage of the contrast. In this analysis we evaluate the presence of lymphatic collecting vessels in the final image volume, subtracted from the first reference frame. The forearm (wrist to elbow) of each patient was divided into 3 sections of equal length (distal, medial, proximal) and the number of vessel segments in each section was counted, in order to characterize vessel density and contrast progression along the arm. The total number of vessels within the forearm, the number of long vessels crossing more than one section, and the presence of deep lymphatic vessels were also recorded. Paired sets of vessel counts were compared (affected versus unaffected arms) using a two-tailed Wilcoxon signed-rank test (p<0.05). Vessel size was estimated by processing maximum intensity projection (MIP) images to obtain the Frangi’s vesselness measure [3]. The scale (σ) of the Frangi’s filter, which provides an estimate of the vessel radius, was extracted for all visible vessels along the entire length of the vessel. Cumulative distributions of vessel size measurements from all images were compared (affected versus unaffected arms, two-tailed Mann-Whitney test, p<0.05).Vessels exhibiting persistent contrast uptake were identified as lymphatics, by inspection of MIP images which retained for every voxel the correspondent uptake curve (Figure 2). In 9 patients the images revealed the presence of lymphatic collecting vessels in both arms, whilst in one patient the unaffected arm showed no contrast uptake. Lymphatic vessels were visualized at different depths in the arm, but primarily within the epi-fascial compartment. In total, a similar number of vessels were detected in the affected and unaffected arms (43 and 46, respectively) and a comparable number of long vessels were present (22 and 18, respectively). Vessel size histograms indicate that lymphatics in the affected arm have larger diameter: the associated density function (Figure 3) peaks at 1.8 mm, while it is skewed towards 1 mm for the unaffected arms. Furthermore, the two vessel diameter distributions were significantly different (p<0.0001). The median number of vessel segments in each forearm was higher in the unaffected arms within the distal and medial sections, and in total (Figure 4). However, differences in vessel counts between affected and unaffected arms did not reach statistical significance according to the Wilcoxon signed-rank test. Besides enhancing vessels, additional patterns seen in the affected arm included: a) diffuse enhancement associated with contrast leakage, and b) characteristic dermal backflow (Figure 5).A higher number of lymphatics vessels enhancing or further progression of contrast along the arm could indicate improved ability to transport lymphatic fluid. However, our analysis did not detect significant differences in vessel counts between affected and unaffected sides within different sections of the forearm. Nevertheless, there was a statistically significant difference between the two distributions of vessel diameters. Lymphatic vessels within the affected arms presented with a larger diameter. This might be an effect of increased permeability of the vessel walls, which leads to leakage of contrast around the vessel. Alternatively, vessels dilate if intraluminal pressure rises, as a consequence of either downstream obstruction or increased lymph flow. In future work we will quantify changes of vessel diameter with time and evaluate flow through analysis of the uptake curves [2], in order to clarify the mechanisms involved.