Validation of algorithms to identify colorectal cancer patients from administrative claims data of a Japanese hospital.

BACKGROUND: Administrative claims data are a valuable source for clinical studies; however, the use of validated algorithms to identify patients is essential to minimize bias. We evaluated the validity of diagnostic coding algorithms for identifying patients with colorectal cancer from a hospital's administrative claims data. METHODS: This validation study used administrative claims data from a Japanese university hospital between April 2017 and March 2019. We developed diagnostic coding algorithms, basically based on the International Classification of Disease (ICD) 10th codes of C18-20 and Japanese disease codes, to identify patients with colorectal cancer. For random samples of patients identified using our algorithms, case ascertainment was performed using chart review as the gold standard. The positive predictive value (PPV) was calculated to evaluate the accuracy of the algorithms. RESULTS: Of 249 random samples of patients identified as having colorectal cancer by our coding algorithms, 215 were confirmed cases, yielding a PPV of 86.3% (95% confidence interval [CI], 81.5-90.1%). When the diagnostic codes were restricted to site-specific (right colon, left colon, transverse colon, or rectum) cancer codes, 94 of the 100 random samples were true cases of colorectal cancer. Consequently, the PPV increased to 94.0% (95% CI, 87.2-97.4%). CONCLUSION: Our diagnostic coding algorithms based on ICD-10 codes and Japanese disease codes were highly accurate in detecting patients with colorectal cancer from this hospital's claims data. The exclusive use of site-specific cancer codes further improved the PPV from 86.3 to 94.0%, suggesting their desirability in identifying these patients more precisely.

Mu-net: Multi-scale U-net for two-photon microscopy image denoising and restoration.

Advances in two two-photon microscopy (2PM) have made three-dimensional (3D) neural imaging of deep cortical regions possible. However, 2PM often suffers from poor image quality because of various noise factors, including blur, white noise, and photo bleaching. In addition, the effectiveness of the existing image processing methods is limited because of the special features of 2PM images such as deeper tissue penetration but higher image noises owing to rapid laser scanning. To address the denoising problems in 2PM 3D images, we present a new algorithm based on deep convolutional neural networks (CNNs). The proposed model consists of multiple U-nets in which an individual U-net removes noises at different scales and then yields a performance improvement based on a coarse-to-fine strategy. Moreover, the constituent CNNs employ fully 3D convolution operations. Such an architecture enables the proposed model to facilitate end-to-end learning without any pre/post processing. Based on the experiments on 2PM image denoising, we observed that our new algorithm demonstrates substantial performance improvements over other baseline methods.

Calcineurin knockout mice show a selective loss of small spines.

Calcineurin is required for long-term depression and activity-dependent spine shrinkage, and calcineurin mutations have been identified in patients with schizophrenia. Moreover, mice with conditional knockout of calcineurin B (CNB-KO) exhibit behavioral abnormalities suggestive of schizophrenia. Changes in the dendritic spines of these mice, however, have not been investigated. We therefore examined the dendritic spines of CNB-KO mice, and observed a significant reduction in small spines and an increase in large spines in the prefrontal and visual cortices. The effect of CNB-KO on the spine sizes was relatively moderate, possibly due to the presence of spontaneous fluctuations (dynamics) in the dendritic spines themselves. Thus, CNB-KO mice showed a spine phenotype similar to those recently reported in patients with schizophrenia.

Specific transformation of assembly with actin filaments and molecular motors in a cell-sized self-emerged liposome.

Eukaryotes, by the same combination of cytoskeleton and molecular motor, for example actin filament and myosin, can generate a variety of movements. For this diversity, the organization of biological machineries caused by the confinement and/or crowding effects of internal living cells, may play very important roles.

Transformation of actoHMM assembly confined in cell-sized liposome.

To construct a simple model of a cellular system equipped with motor proteins, cell-sized giant liposomes encapsulating various amounts of actoHMM, the complexes of actin filaments (F-actin) and heavy meromyosin (HMM, an actin-related molecular motor), with a depletion reagent to mimic the crowding effect of inside of living cell, were prepared. We adapted the methodology of the spontaneous transfer of water-in-oil (W/O) droplets through a phospholipid monolayer into the bulk aqueous phase and successfully prepared stable giant liposomes encapsulating the solution with a physiological salt concentration containing the desired concentrations of actoHMM, which had been almost impossible to obtain using currently adapted methodologies such as natural swelling and electro-formation on an electrode. We then examined the effect of ATP on the cytoskeleton components confined in those cell-sized liposomes, because ATP is known to drive the sliding motion for actoHMM. We added α-hemolysin, a bacterial membrane pore-forming toxin, to the bathing solution and obtained liposomes with the protein pores embedded on the bilayer membrane to allow the transfer of ATP inside the liposomes. We show that, by the ATP supply, the actoHMM bundles inside the liposomes exhibit specific changes in spatial distribution, caused by the active sliding between F-actin and HMM. Interestingly, all F-actins localized around the inner periphery of liposomes smaller than a critical size, whereas in the bulk solution and also in larger liposomes, the actin bundles formed aster-like structures under the same conditions.

Phase behavior of crowded like-charged mixed polyelectrolytes in a cell-sized sphere.

We studied the phase behavior of a mixture of two semiflexible negatively charged polyelectrolytes, giant DNA and alginate, under crowded condition in a cell-sized sphere (5-40 µm in diameter), where the persistence length of giant DNA is 50 nm and that of alginate is 5 nm. Through microscopic observation, we found that the polymer mixture exhibits a unique phase behavior, which depends on the size of the sphere, whereas the mixture remains homogeneous and isotropic in bulk solution. When the sphere is small, DNA is completely depleted on the surface. When the sphere is medium sized, a portion of the DNA is depleted on the surface, and the remainder stays within the sphere. By introducing a curvature-dependent term for the interaction between DNA and the surface into the Flory-Huggins model, we interpret the observed characteristics of the phase behavior in terms of the relative importance of the surface-to-volume effect.

Cooperation between giant DNA molecules and actin filaments in a microsphere.

We report the appearance of a spatially segregated state in a microsphere as a result of cooperation between actin filaments and giant DNA molecules. When the coexisting actin concentration is high enough, DNA molecules are excluded toward the surface by forming an assembly with actin filaments. With a decrease in the actin concentration, actin filaments tend to dissolve within the sphere whereas DNA molecules remain to be excluded onto the surface. When the actin concentration becomes still lower, DNA molecules dissolve within the sphere by avoiding surface attachment. We interpret these experimental trends by introducing the concept of an "exclusion zone" for actin filaments within a microsphere.

Mismatch of bulk viscosity reduces interfacial diffusivity at an aqueous-oil system.

A significant decrease in the diffusivity of phospholipid molecules at an aqueous-oil monolayer was observed experimentally when the bathing solutions had different viscosities eta(1) not equal eta(2). Near the point of isoviscosity eta(1) approximately eta(2), the diffusion constant at the monolayer can be drastically affected by a slight change in the viscosity mismatch. We interpreted this phenomenon by considering the effect of vertical fluctuations which affect the long-term dynamics of two-dimensional lateral diffusion at the monolayer.

Protein synthesis in giant liposomes using the in vitro translation system of Thermococcus kodakaraensis.

An in vitro translation system, based on cell components of the hyperthermophilic archaeon, Thermococcus kodakaraensis, has previously been developed. The system has been optimized and applied for protein production at high temperatures (60-65 degrees C). In this paper, we have examined the possibilities to utilize this system at a lower temperature range using green fluorescence protein (GFP) as the reporter protein. By optimizing the composition of the reaction mixture, and adding chaperonins from the mesophilic Escherichia coli, the yield of protein production at 40 degrees C was increased by fivefold. For liposome encapsulation of the optimized system, water-in-oil cell-sized emulsions were prepared by adding the translation system/GFP mRNA mixture to mineral oil supplemented with 1,2-dioleoyl-sn -glycero-3-phosphatidylcholine (DOPC). Giant liposomes were formed when these emulsions passed across a water/oil interface occupied with DOPC. The liposomes were incubated at 40 degrees C for 90 min, and fluorescence was examined by laser confocal microscopy. A significant increase in average fluorescence intensity was observed in liposomes with GFP mRNA, but not in those without mRNA. Our results indicate that the T. kodakaraensis in vitro translation system is applicable for protein production within giant liposomes, and these artificial cell models should provide the methodology to reconstitute various cell functions from a constitutional biology approach.

Chapter 3 - Construction of cell-sized liposomes encapsulating actin and actin-cross-linking proteins.

To shed light on the mechanism underlying the active morphogenesis of living cells in relation to the organization of internal cytoskeletal networks, the development of new methodologies to construct artificial cell models is crucial. Here, we describe the successful construction of cell-sized liposomes entrapping cytoskeletal proteins. We discuss experimental protocols to prepare giant liposomes encapsulating desired amounts of actin and cross-linking proteins including molecular motor proteins, such as fascin, alpha-actinin, filamin, myosin-I isolated from brush border (BBMI), and heavy meromyosin (HMM). Subfragment 1 (S-1) is also studied in comparison to HMM, where S-1 and HMM are single-headed and double-headed derivatives of conventional myosin (myosin-II), respectively. In the absence of cross-linking proteins, actin filaments (F-actin) are distributed homogeneously without any order within the liposomes. In contrast, when actin is encapsulated together with an actin-cross-linking protein, mesh structures emerge that are similar to those in living motile cells. Optical microscopic observations on the active morphological changes of the liposomes are reported.

Emergence of superstructures from a homogeneous lipid sphere.

The spontaneous generation of a periodic hexagonal superstructure on a giant phospholipid sphere (GPS) with a diameter of 20-200 microm was studied. The GPS was composed of ternary phospholipids consisting of dioleoylphosphatidylethanolamine (DOPE), dioleoylphosphatidylcholine (DOPC), and dioleoylphosphatidylinositol-bisphosphate (DOPIP(2)). GPSs were prepared by natural swelling of a lipid film formed on a glass substrate. A GPS with a homogeneous lipid mixture tends to form a two-layered structure between the surface and inner parts; the surface layer is attributed to a DOPIP(2) rich region (we call this layer SL), and the interior is rich in DOPE and DOPC (we call this layer IL). A hexagonal superstructure develops in the SL, and the topology then changes to form multiple-doughnut structures. Finally, myelin-like tubes are generated through symmetry breaking of the doughnutlike structures. The time-dependent change in the surface-area expansion of a GPS is shown to obey the logistic growth model, and this is attributed to the kinetic process of phase segregation between the surface and bulk phase of the GPS.

Entrapping desired amounts of actin filaments and molecular motor proteins in giant liposomes.

We have successfully prepared cell-sized giant liposomes encapsulating desired amounts of actoHMM, a mixture of actin filament (F-actin) and heavy meromyosin (HMM, an actin-related molecular motor), in the presence of 5 mM MgCl 2 and 50 mM KCl. We employed a spontaneous transfer method to prepare those liposomes. In the absence of HMM, F-actin was distributed homogeneously inside the liposomes. In contrast, when F-actin was encapsulated in liposomes together with HMM, network structures were generated. Such network structures are attributable to the cross-linking of F-actin by HMM.

How does the mobility of phospholipid molecules at a water/oil interface reflect the viscosity of the surrounding oil?

The mobility of phospholipid molecules at a water/oil interface on cell-sized phospholipid-coated microdroplets was investigated through the measurement of diffusion constants by fluorescence recovery after photobleaching. It is found that the diffusion constant of phospholipids showed the relation D approximately (eta water + eta oil) -0.85, where D is the diffusion constant, eta water is the viscosity of water, and eta oil is the viscosity of oil. This observation indicates that the viscosity of the surrounding oil is the primary factor that determines the diffusibility of phospholipids at a water/oil interface.