Deep learning for digital holography: a review.

Recent years have witnessed the unprecedented progress of deep learning applications in digital holography (DH). Nevertheless, there remain huge potentials in how deep learning can further improve performance and enable new functionalities for DH. Here, we survey recent developments in various DH applications powered by deep learning algorithms. This article starts with a brief introduction to digital holographic imaging, then summarizes the most relevant deep learning techniques for DH, with discussions on their benefits and challenges. We then present case studies covering a wide range of problems and applications in order to highlight research achievements to date. We provide an outlook of several promising directions to widen the use of deep learning in various DH applications.

RedCap: residual encoder-decoder capsule network for holographic image reconstruction.

A capsule network, as an advanced technique in deep learning, is designed to overcome information loss in the pooling operation and internal data representation of a convolutional neural network (CNN). It has shown promising results in several applications, such as digit recognition and image segmentation. In this work, we investigate for the first time the use of capsule network in digital holographic reconstruction. The proposed residual encoder-decoder capsule network, which we call RedCap, uses a novel windowed spatial dynamic routing algorithm and residual capsule block, which extends the idea of a residual block. Compared with the CNN-based neural network, RedCap exhibits much better experimental results in digital holographic reconstruction, while having a dramatic 75% reduction in the number of parameters. It indicates that RedCap is more efficient in the way it processes data and requires a much less memory storage for the learned model, which therefore makes it possible to be applied to some challenging situations with limited computational resources, such as portable devices.

Computational image speckle suppression using block matching and machine learning.

We develop an image despeckling method that combines nonlocal self-similarity filters with machine learning, which makes use of convolutional neural network (CNN) denoisers. It consists of three major steps: block matching, CNN despeckling, and group shrinkage. Through the use of block matching, we can take advantage of the similarity across image patches as a regularizer to augment the performance of data-driven denoising using a pre-trained network. The outputs from the CNN denoiser and the group coordinates from block matching are further used to form 3D groups of similar patches, which are then filtered through a wavelet-domain shrinkage. The experimental results show that the proposed method achieves noticeable improvement compared with state-of-the-art speckle suppression techniques in both visual inspection and objective assessments.

Effect of Hydrogel Stiffness on Chemoresistance of Breast Cancer Cells in 3D Culture.

Chemotherapy is one of the most common strategies for cancer treatment, whereas drug resistance reduces the efficiency of chemotherapy and leads to treatment failure. The mechanism of emerging chemoresistance is complex and the effect of extracellular matrix (ECM) surrounding cells may contribute to drug resistance. Although it is well known that ECM plays an important role in orchestrating cell functions, it remains exclusive how ECM stiffness affects drug resistance. In this study, we prepared agarose hydrogels of different stiffnesses to investigate the effect of hydrogel stiffness on the chemoresistance of breast cancer cells to doxorubicin (DOX). Agarose hydrogels with a stiffness range of 1.5 kPa to 112.3 kPa were prepared and used to encapsulate breast cancer cells for a three-dimensional culture with different concentrations of DOX. The viability of the cells cultured in the hydrogels was dependent on both DOX concentration and hydrogel stiffness. Cell viability decreased with DOX concentration when the cells were cultured in the same stiffness hydrogels. When DOX concentration was the same, breast cancer cells showed higher viability in high-stiffness hydrogels than they did in low-stiffness hydrogels. Furthermore, the expression of P-glycoprotein mRNA in high-stiffness hydrogels was higher than that in low-stiffness hydrogels. The results suggested that hydrogel stiffness could affect the resistance of breast cancer cells to DOX by regulating the expression of chemoresistance-related genes.

Morphological Dependence of Breast Cancer Cell Responses to Doxorubicin on Micropatterned Surfaces.

Cell morphology has been widely investigated for its influence on the functions of normal cells. However, the influence of cell morphology on cancer cell resistance to anti-cancer drugs remains unclear. In this study, micropatterned surfaces were prepared and used to control the spreading area and elongation of human breast cancer cell line. The influences of cell adhesion area and elongation on resistance to doxorubicin were investigated. The percentage of apoptotic breast cancer cells decreased with cell spreading area, while did not change with cell elongation. Large breast cancer cells had higher resistance to doxorubicin, better assembled actin filaments, higher DNA synthesis activity and higher expression of P-glycoprotein than small breast cancer cells. The results suggested that the morphology of breast cancer cells could affect their resistance to doxorubicin. The influence was correlated with cytoskeletal organization, DNA synthesis activity and P-glycoprotein expression.

Stepwise photothermal therapy and chemotherapy by composite scaffolds of gold nanoparticles, BP nanosheets and gelatin immobilized with doxorubicin-loaded thermosensitive liposomes.

In recent years, the synergistic effect of photothermal therapy (PTT) and chemotherapy has been recognized as an effective strategy for cancer treatment. Controlling the PTT temperature and drug release profile is desirable for minimizing the unexpected damage to normal cells. In this study, a smart platform of stepwise PTT and chemotherapy has been developed by using composite porous scaffolds of biodegradable black phosphorus (BP) nanosheets, gold nanorods(AuNRs), doxorubicin (Dox)-encapsulated thermosensitive liposomes and biodegradable polymers. Under near-infrared (NIR) laser irradiation, the composite scaffolds could attain high and low local temperatures before and after BP degradation, respectively. Dox release from the composite scaffolds could be controlled by the temperature change. In vitro cell culture and in vivo animal experiments indicated that a strong synergistic effect of PTT and chemotherapy could be achieved at an early stage of treatment before BP degradation, and a mild hyperthermia effect was shown for chemotherapy in the late stage after BP degradation. Moreover, the composite scaffolds after the complete release of Dox could support the proliferation of mesenchymal stem cells. The composite scaffolds showed a synergistic effect of stepwise PTT and chemotherapy for breast cancer elimination and promoted stem cell activities after killing cancer cells.

Autophagy inhibition and microRNA­199a­5p upregulation in paclitaxel­resistant A549/T lung cancer cells.

Multidrug resistance (MDR) is one of the major reasons for the clinical failure of cancer chemotherapy. Autophagy activation serves a crucial role in MDR. However, the specific molecular mechanism linking autophagy with MDR remains unknown. The results of the present study demonstrated that autophagy was inhibited and microRNA (miR)­199a­5p levels were upregulated in MDR model lung cancer cells (A549/T and H1299/T) compared with those in the parental cell lines. Paclitaxel (PTX) treatment increased the expression levels of miR­199a­5p in parental lung cancer cells compared with those in PTX­untreated cells, and these expression levels were negatively correlated with PTX sensitivity of the cells. miR­199a­5p knockdown in A549/T cells induced autophagy and resensitized cells to multiple chemotherapeutic drugs including PTX, taxotere, topotecan, SN38, oxaliplatin and vinorelbine. By contrast, miR­199a­5p overexpression in A549 cells suppressed autophagy and desensitized cells to these chemotherapeutic drugs. Mechanistically, the results of the present study demonstrated that miR­199a­5p blocked autophagy by activating the PI3K/Akt/mTOR signaling pathway and inhibiting the protein expression of autophagy­related 5. Furthermore, p62 protein was identified as a direct target of miR­199a­5p; miR­199a­5p bound to p62 mRNA to decrease its mRNA and protein expression levels. In conclusion, the results of the present study suggested that miR­199a­5p may contribute to MDR development in lung cancer cells by inhibiting autophagy and targeting p62. The regulatory effect of miR­199a­5p on autophagy may provide novel insights for future multidrug­resistant lung cancer chemotherapy.

Folic Acid-Functionalized Composite Scaffolds of Gelatin and Gold Nanoparticles for Photothermal Ablation of Breast Cancer Cells.

Photothermal therapy (PTT) has been developed as a useful therapeutic method for cancer treatment. Localization of PTT agents in cancer sites and targeting capacity are required to further increase therapeutic efficacy. In this study, gold nanoparticles (AuNPs) and gelatin were functionalized with folic acid (FA) and hybridized to prepare FA-functionalized gelatin-AuNPs composite scaffolds. AuNPs with rod and star shapes of three sizes (40, 70, and 110 nm) were used for the hybridization to investigate the influence of AuNPs shape and size. The composite scaffolds showed porous structures with good interconnectivity. Modification with FA increased capture capacity of the composite scaffolds. Hybridization with AuNPs rendered the composite scaffold a good photothermal conversion property under near-infrared (NIR) laser irradiation. Temperature change during laser irradiation increased with the laser power intensity and irradiation time. The shape and size of AuNPs also affected their photothermal conversion property. The composite scaffold of gold nanorods 70 (FA-G/R70) had the highest photothermal conversion capacity. Breast cancer cells cultured in the FA-G/R70 composite scaffold were killed under NIR laser irradiation. Mouse subcutaneous implantation further demonstrated the excellent photothermal ablation capability of FA-G/R70 composite scaffold to breast cancer cells. The FA-functionalized composite scaffolds were demonstrated a high potential for local PPT of breast cancer.