Predicting microvascular invasion in hepatocellular carcinoma with a CT- and MRI-based multimodal deep learning model.

PURPOSE: To investigate the value of a multimodal deep learning (MDL) model based on computed tomography (CT) and magnetic resonance imaging (MRI) for predicting microvascular invasion (MVI) in hepatocellular carcinoma (HCC). METHODS: A total of 287 patients with HCC from our institution and 58 patients from another individual institution were included. Among these, 119 patients with only CT data and 116 patients with only MRI data were selected for single-modality deep learning model development, after which select parameters were migrated for MDL model development with transfer learning (TL). In addition, 110 patients with simultaneous CT and MRI data were divided into a training cohort (n = 66) and a validation cohort (n = 44). We input the features extracted from DenseNet121 into an extreme learning machine (ELM) classifier to construct a classification model. RESULTS: The area under the curve (AUC) of the MDL model was 0.844, which was superior to that of the single-phase CT (AUC = 0.706-0.776, P < 0.05), single-sequence MRI (AUC = 0.706-0.717, P < 0.05), single-modality DL model (AUCall-phase CT = 0.722, AUCall-sequence MRI = 0.731; P < 0.05), clinical (AUC = 0.648, P < 0.05), but not to that of the delay phase (DP) and in-phase (IP) MRI and portal venous phase (PVP) CT models. The MDL model achieved better performance than models described above (P < 0.05). When combined with clinical features, the AUC of the MDL model increased from 0.844 to 0.871. A nomogram, combining deep learning signatures (DLS) and clinical indicators for MDL models, demonstrated a greater overall net gain than the MDL models (P < 0.05). CONCLUSION: The MDL model is a valuable noninvasive technique for preoperatively predicting MVI in HCC.

Optimizing the Peritumoral Region Size in Radiomics Analysis for Sentinel Lymph Node Status Prediction in Breast Cancer.

RATIONALE AND OBJECTIVES: Peritumoral features have been suggested to be useful in improving the prediction performance of radiomic models. The aim of this study is to systematically investigate the prediction performance improvement for sentinel lymph node (SLN) status in breast cancer from peritumoral features in radiomic analysis by exploring the effect of peritumoral region sizes. MATERIALS AND METHODS: This retrospective study was performed using dynamic contrast-enhanced MRI scans of 162 breast cancer patients. The effect of peritumoral features was evaluated in a radiomics pipeline for predicting SLN metastasis in breast cancer. Peritumoral regions were generated by dilating the tumor regions-of-interest (ROIs) manually annotated by two expert radiologists, with thicknesses of 2 mm, 4 mm, 6 mm, and 8 mm. The prediction models were established in the training set (∼67% of cases) using the radiomics pipeline with and without peritumoral features derived from different peritumoral thicknesses. The prediction performance was tested in an independent validation set (the remaining ∼33%). RESULTS: For this specific application, the accuracy in the validation set when using the two radiologists' ROIs could be both improved from 0.704 to 0.796 by incorporating peritumoral features. The choice of the peritumoral size could affect the level of improvement. CONCLUSION: This study systematically investigates the effect of peritumoral region sizes in radiomic analysis for prediction performance improvement. The choice of the peritumoral size is dependent on the ROI drawing and would affect the final prediction performance of radiomic models, suggesting that peritumoral features should be optimized in future radiomics studies.

Task-based assessment of a convolutional neural network for segmenting breast lesions for radiomic analysis.

PURPOSE: Radiomics allows for powerful data-mining and feature extraction techniques to guide clinical decision making. Image segmentation is a necessary step in such pipelines and different techniques can significantly affect results. We demonstrate that a convolutional neural network (CNN) segmentation method performs comparably to expert manual segmentations in an established radiomics pipeline. METHODS: Using the manual regions of interest (ROIs) of an expert radiologist (R1), a CNN was trained to segment breast lesions from dynamic contrast-enhanced MRI (DCE-MRI). Following network training, we segmented lesions for the testing set of a previously established radiomics pipeline for predicting lymph node metastases using DCE-MRI of breast cancer. Prediction accuracy of CNN segmentations relative to manual segmentations by R1 from the original study, a resident (R2), and another expert radiologist (R3) were determined. We then retrained the CNN and radiomics model using R3's manual segmentations to determine the effects of different expert observers on end-to-end prediction. RESULTS: Using R1's ROIs, the CNN achieved a mean Dice coefficient of 0.71 ± 0.16 in the testing set. When input to our previously published radiomics pipeline, these CNN segmentations achieved comparable prediction performance to R1's manual ROIs, and superior performance to those of the other radiologists. Similar results were seen when training the CNN and radiomics model using R3's ROIs. CONCLUSION: A CNN architecture is able to provide DCE-MRI breast lesion segmentations which are suitable for input to our radiomics model. Moreover, the previously established radiomics model and CNN can be accurately trained end-to-end using ground truth data provided by distinct experts.

The Value of IVIM DWI in Combination with Conventional MRI in Identifying the Residual Tumor After Cone Biopsy for Early Cervical Carcinoma.

RATIONALE AND OBJECTIVES: To investigate the value of intravoxel incoherent motion (IVIM) diffusion-weighted imaging (DWI) in combination with conventional MRI in identifying the residual tumor after biopsy for early cervical carcinoma. MATERIALS AND METHODS: Eighty patients with histologically proven early cervical carcinoma were enrolled into this study. MRI sequences included two sets of MRI sequences including conventional MRI (T1WI, T2WI, and dynamic contrast-enhanced MRI) and IVIM DWI/conventional MRI combinations. The patients were classified into residual tumor and nonresidual tumor group after biopsy. IVIM parameters were quantitatively analyzed and compared between two groups. The diagnostic ability of two sets of MRI sequences were calculated and compared. RESULTS: The mean D and f values were significantly lower in residual tumor group than in nonresidual tumor group (p < 0.05). The areas under receiver operating characteristic curves of D and f for discriminating between residual tumor and nonresidual tumor group were 0.848 and 0.767, respectively. The sensitivity and accuracy of conventional MRI/IVIM DWI combinations for the detection of residual tumor were 82.7% and 83.8%, respectively, while the sensitivity and accuracy of conventional MRI were 52.4% and 53.8%, respectively. CONCLUSION: The addition of IVIM DWI to conventional MRI considerably improves the sensitivity and accuracy of the detection of residual tumor after biopsy for early cervical carcinoma.

Oil-Repellent Antifogging Films with Water-Enabled Functional and Structural Healing Ability.

Healable oil-repellent antifogging films are fabricated by layer-by-layer assembly of hyaluronic acid (HA) and branched poly(ethylenimine) (bPEI), followed by immersion in the aqueous solutions of perfluorooctanesulfonic acid potassium salt (PFOS). The loading of PFOS endows the HA/bPEI films with oil repellency while maintaining its original hydrophilicity. The resulting films have an excellent antifogging ability, and various organic liquids can easily slide down the slightly tilted films (<10°) without any residue. Through water-assisted migration of PFOS and polyelectrolytes, oil-repellent antifogging films are able to repetitively and autonomously recover their damaged oil repellency and transparency caused by plasma etching, cutting, or scratching, prolonging their life span. The as-developed healable oil-repellent antifogging films have potential application as antifingerprint coatings for touch screens, antigraffiti coatings for signs and shop windows, and antifogging coatings for lenses, mirrors, and windshields.

Intumescent flame-retardant and self-healing superhydrophobic coatings on cotton fabric.

Flame-retardant and self-healing superhydrophobic coatings are fabricated on cotton fabric by a convenient solution-dipping method, which involves the sequential deposition of a trilayer of branched poly(ethylenimine) (bPEI), ammonium polyphosphate (APP), and fluorinated-decyl polyhedral oligomeric silsesquioxane (F-POSS). When directly exposed to flame, such a trilayer coating generates a porous char layer because of its intumescent effect, successfully giving the coated fabric a self-extinguishing property. Furthermore, the F-POSS embedded in cotton fabric and APP/bPEI coating produces a superhydrophobic surface with a self-healing function. The coating can repetitively and autonomically restore the superhydrophobicity when the superhydrophobicity is damaged. The resulting cotton fabric, which is flame-resistant, waterproof, and self-cleaning, can be easily cleaned by simple water rinsing. Thus, the integration of self-healing superhydrophobicity with flame retardancy provides a practical way to resolve the problem of washing durability of the flame-retardant coatings. The flame-retardant and superhydrophobic fabric can endure more than 1000 cycles of abrasion under a pressure of 44.8 kPa without losing its flame retardancy and self-healing superhydrophobicity, showing potential applications as multifunctional advanced textiles.

Robust and flexible free-standing films for unidirectional drug delivery.

Robust and flexible free-standing polymer films for unidirectional drug delivery are fabricated by sandwiching drug-containing polyelectrolyte multilayer films between poly(lactic-co-glycolic acid) (PLGA) barrier and capping layers. The drug-containing films are fabricated by layer-by-layer (LbL) assembly of chemically cross-linked poly(allylamine hydrochloride)-dextran (PAH-D) microgel and hyaluronic acid (HA), which can load negatively charged cancer-inhibiting drug, methotrexate (MTX). Because the PLGA barrier layer effectively blocks MTX release, MTX can be predominantly released from the PLGA capping layer of the free-standing film. This increases the efficacy of released MTX to cancer cells while minimizing its side effects on the normal tissues. We believe that the unidirectional drug delivery free-standing films can open a new avenue to design of highly efficient drug delivery systems for biomedical application.

Layer-by-layer assembled microgel films with high loading capacity: reversible loading and release of dyes and nanoparticles.

Multilayer films containing microgels of chemically cross-linked poly(allylamine hydrochloride) (PAH) and dextran (named PAH-D) were fabricated by layer-by-layer deposition of PAH-D and poly(styrene sulfonate) (PSS). The successful fabrication of PAH-D/PSS multilayer films was verified by quartz crystal microbalance measurements and cross-sectional scanning electron microscopy. The as-prepared PAH-D/PSS multilayer films can reversibly load and release negatively charged dyes such as methyl orange (MO) and fluorescein sodium and mercaptoacetic acid-stabilized CdTe nanoparticles. The loading capacity of the film for MO can be as large as approximately 3.0 microg/cm2 per bilayer, which corresponds to a MO density of 0.75 g/cm3 in the film. The high loading capacity of the PAH-D/PSS films originates from the cross-linked film structure with sufficient binding groups of protonated amine groups, as well as their high swelling capability by solvent. The loaded material can be released slowly when immersing the films in 0.9% normal saline. Meanwhile, the PAH-D/PSS multilayer films could deposit directly on either hydrophilic or hydrophobic substrates such as quartz, polytetrafluoroethylene, polystyrene, poly(ethylene terephthalate), and polypropylene. The microgel films of PAH-D/PSS are expected to be widely useful as matrixes for loading functional guest materials and even for controlled release.