RESUMEN
Background: Predicting whether T790M emerges early is crucial to the adjustment of targeted drugs for non-small cell lung cancer (NSCLC) patients. This study aimed to evaluate the risk of T790M resistance in progressive new brain metastases (BMs) based on multisequence magnetic resonance imaging (MRI) radiomics. Methods: This retrospective study included 405 consecutive patients (training cohort: 294 patients; testing cohort: 111 patients) with proven NSCLC with disease progression of new BM. The radiomics features were separately extracted from T2-weighted imaging (T2WI), T2 fluid-attenuated inversion recovery (T2-FLAIR), diffusion-weighted imaging (DWI), and contrast-enhanced T1-weighted imaging (T1-CE) sequence of baseline MRI. Then, we calculated radiomics scores (rad-score) of the 4 sequences respectively and established predictive models (lesion- or patient-level) to evaluate T790M resistance within up to 14 months using random forest classifier. Receiver operating characteristic (ROC) curves and F1 scores were used to validate the performance of two models in both the training and testing cohort. Results: There were significant differences in rad-scores of the four sequences between T790M-positive and negative groups whether in the training or testing cohort (P<0.05). The lesion-level model consisting of rad-scores showed excellent discrimination, with an area under the curve (AUC) and F1-score of 0.879 and 0.798 in the training cohort, and 0.834 and 0.742 in the testing cohort, respectively. The patient-level model also showed a favorable discriminatory ability with an AUC and F1 score of 0.851 and 0.837, which was confirmed with an AUC and F1 score of 0.734 and 0.716 in the testing cohort. Conclusions: The MRI-based radiomics signatures may be new markers to identify patients at high risk of developing resistance in the early period.
RESUMEN
Known as the third pole of the world, the Qinghai-Tibet Plateau has been experiencing rapid permafrost warming and thawing over the last few decades. However, the impact of permafrost distribution and hydrogeology on river hydrochemistry in alpine areas remains unclear. This study conducted four sampling campaigns to reveal the temporal and spatial variations in and factors driving river hydrochemistry in the upper reaches of the Heihe River, the northeastern Qinghai-Tibet Plateau. We found that the concentrations of major ions and total dissolved solids (TDS) in river water showed substantial seasonal variations; the concentrations were generally lower during the initial thawing and thawed periods than during the initial freezing period. However, solute fluxes during the thawed period were much higher than those during the frozen period. The concentrations of major ions and TDS gradually decreased to a minimum from the permafrost meander (PM) section to the seasonal frost meander (SFM) section and then increased the seasonal frost canyon section. Using the revised forward model, we found that river solutes were contributed by carbonate weathering (mean 38.9%) > sulfide oxidation (22.9%) > evaporite dissolution (20.2%) > atmospheric precipitation (8.7%) > silicate weathering (5.0%) > glacial meltwater (4.3%). The higher TDS, Na+, Cl-, Ca2+, Mg2+, and SO42- concentrations in the PM section reflected the influence of freeze-out fractionation. The concentrations of major ions and TDS were lowest in the SFM section, indicating that the riparian porous aquifer was essential in regulating river hydrochemistry, thus reducing its spatiotemporal variations in the alpine area. In the mountain glacier-hillslope-riparian porous aquifer-river system, the river was mainly recharged by groundwater with insufficient water-rock interactions due to the rapid flow owing to the high elevation difference and high permeability of the riparian quaternary porous aquifers. Our findings provide insights into the construction of hydrogeochemical models in alpine areas and are practically important for the scientific management of water resources in the Qinghai-Tibet Plateau.