Search

Deep learning (DL) has been proposed to automate image segmentation and provide accuracy, consistency, and efficiency. Accurate segmentation of lipomatous tumors (LTs) is critical for correct tumor radiomics analysis and localization. The major challenge of this task is data heterogeneity, including tumor morphological characteristics and multicenter scanning protocols. To mitigate the issue, we aimed to develop a DL-based Super Learner (SL) ensemble framework with different data correction and normalization methods. Pathologically proven LTs on pre-operative T1-weighted/proton-density MR images of 185 patients were manually segmented. The LTs were categorized by tumor locations as distal upper limb (DUL), distal lower limb (DLL), proximal upper limb (PUL), proximal lower limb (PLL), or Trunk (T) and grouped by 80%/9%/11% for training, validation and testing. Six configurations of correction/normalization were applied to data for fivefold-cross-validation trainings, resulting in 30 base learners (BLs). A SL was obtained from the BLs by optimizing SL weights. The performance was evaluated by dice-similarity-coefficient (DSC), sensitivity, specificity, and Hausdorff distance (HD95). For predictions of the BLs, the average DSC, sensitivity, and specificity from the testing data were 0.72 [Formula: see text] 0.16, 0.73 [Formula: see text] 0.168, and 0.99 [Formula: see text] 0.012, respectively, while for SL predictions were 0.80 [Formula: see text] 0.184, 0.78 [Formula: see text] 0.193, and 1.00 [Formula: see text] 0.010. The average HD95 of the BLs were 11.5 (DUL), 23.2 (DLL), 25.9 (PUL), 32.1 (PLL), and 47.9 (T) mm, whereas of SL were 1.7, 8.4, 15.9, 2.2, and 36.6 mm, respectively. The proposed method could improve the segmentation accuracy and mitigate the performance instability and data heterogeneity aiding the differential diagnosis of LTs in real clinical situations.

The blue color of glass and ceramic glazes produced in Apulia and Basilicata (Southern Italy) between the 13th and 14th centuries and connected to the Norman-Swabian Emperor Frederick II, has been, for a long time, under archaeometric investigation. On the one hand, it has usually been associated with lapis lazuli, due to the finding of the polysulphide blue chromophores typical of lazurite. Moreover, the observation that the mineral haüyne, which belongs to the sodalite group as well as lazurite, can be blue and/or can gain a blue color after heating, due to the same chromophores, has caused this automatic attribution to be questioned, and also considering that the mineral is characteristic of the rock haüynophyre of Melfi (Potenza, Southern Italy), a location of interest for glass and pottery findings. In this paper, for the first time, several haüyne crystals were found in the blue glaze of a ceramic dish found at Melfi Castle, leading to the hypothesis that, in this case, the local haüyne-bearing source could have been used as the coloring raw material. The discovery was possible thanks to SEM-EDS and Raman analyses that, respectively, highlighted the typical numerous presence of very fine sulphur-based inclusions in the crystals and the characteristic Raman signal of blue haüyne. This study was also focused on the composition of the crystals inclusions, aided by SEM-EDS and Raman maps, since the original very fine pyrrhotite was transformed into Cu and Pb phases (copper sulphates, copper sulphides, and lead oxide) due to reactions with cations that had mobilized from the glaze, while the migration of Si from the glass allowed the transformation of the rim of the haüyne, a silica-undersaturated mineral, into a corona of small euhedral and neomorphic Pb-rich feldspars, a silica-saturated phase.