Energy-efficient high-fidelity image reconstruction with memristor arrays for medical diagnosis.

Medical imaging is an important tool for accurate medical diagnosis, while state-of-the-art image reconstruction algorithms raise critical challenges in massive data processing for high-speed and high-quality imaging. Here, we present a memristive image reconstructor (MIR) to greatly accelerate image reconstruction with discrete Fourier transformation (DFT) by computing-in-memory (CIM) with memristor arrays. A high-accuracy quasi-analogue mapping (QAM) method and generic complex matrix transfer (CMT) scheme was proposed to improve the mapping precision and transfer efficiency, respectively. High-fidelity magnetic resonance imaging (MRI) and computed tomography (CT) image reconstructions were demonstrated, achieving software-equivalent qualities and DICE scores after segmentation with nnU-Net algorithm. Remarkably, our MIR exhibited 153× and 79× improvements in energy efficiency and normalized image reconstruction speed, respectively, compared to graphics processing unit (GPU). This work demonstrates MIR as a promising high-fidelity image reconstruction platform for future medical diagnosis, and also largely extends the application of memristor-based CIM beyond artificial neural networks.

Reduced intensity of early intensification does not increase the risk of relapse in children with standard risk acute lymphoblastic leukemia - a multi-centric clinical study of GD-2008-ALL protocol.

BACKGROUND: The prognosis of childhood acute lymphoblastic leukemia (ALL) is optimistic with a 5-year event-free survival (EFS) rate of 70-85%. However, the major causes of mortality are chemotherapy toxicity, infection and relapse. The Guangdong (GD)-2008-ALL collaborative protocol was carried out to study the effect of reduced intensity on treatment related mortality (TRM) based on Berlin-Frankfurt-Münster (BFM) 2002 backbone treatment. The study was designed to elucidate whether the reduced intensity is effective and safe for children with ALL. METHODS: The clinical data were obtained from February 28, 2008 to June 30, 2016. A total of 1765 childhood ALL cases from 9 medical centers were collected and data were retrospectively analyzed. Patients were stratified into 3 groups according to bone marrow morphology, prednisone response, age, genotype, and karyotype information: standard risk (SR), intermediate risk (IR) and high risk (HR). For SR group, daunorubicin was decreased in induction IA while duration was reduced in Induction Ib (2 weeks in place of 4 weeks). Doses for CAM were same in all risk groups - SR patients received one CAM, others got two CAMs. RESULTS: The 5-year and 8-year overall survival (OS), event-free survival (EFS) and cumulative incidence of relapse (CIR) were 83.5±0.9% and 83.1±1.0%, 71.9±1.1% and 70.9±1.2%, and 19.5±1.0% and 20.5±1.1%, respectively. The 2-year treatment-related mortality (TRM) was 5.2±0.5%. The 5-year and 8-year OS were 90.7±1.4% and 89.6±1.6% in the SR group, while the 5-year and 8-year EFS were 81.5±1.8% and 80.0±2.0%. In the SR group, 74 (15.2%) patients measured minimal residual disease (MRD) on Day 15 and Day 33 of induction therapy. Among them, 7 patients (9.46%) were MRD positive (≥ 0.01%) on Day 33. The incidence of relapse in the MRD Day 33 positive group (n=7) was 28.6%, while in the MRD Day 33 negative group (n=67) was 7.5% (p=0.129). CONCLUSIONS: The results of GD-2008-ALL protocol are outstanding for reducing TRM in childhood ALL in China with excellent long term EFS. This protocol provided the evidence for further reducing intensity of induction therapy in the SR group according to the risk stratification. MRD levels on Day 15 and Day 33 are appropriate indexes for stratification.

Multichannel parallel processing of neural signals in memristor arrays.

Fully implantable neural interfaces with massive recording channels bring the gospel to patients with motor or speech function loss. As the number of recording channels rapidly increases, conventional complementary metal-oxide semiconductor (CMOS) chips for neural signal processing face severe challenges on parallelism scalability, computational cost, and power consumption. In this work, we propose a previously unexplored approach for parallel processing of multichannel neural signals in memristor arrays, taking advantage of their rich dynamic characteristics. The critical information of neural signal waveform is extracted and encoded in the memristor conductance modulation. A signal segmentation scheme is developed to adapt to device variations. To verify the fidelity of the processed results, seizure prediction is further demonstrated, with high accuracy above 95% and also more than 1000× improvement in power efficiency compared with CMOS counterparts. This work suggests that memristor arrays could be a promising multichannel signal processing module for future implantable neural interfaces.

[hnRNP K Contributes to Adriamycin Resistance in Acute Myeloid Leukemia through Regulating Autophagy].

OBJECTIVE: To explore the role of heterogeneous nuclear ribonucleoprotein(hnRNP) K regulating autophagy in the drug resistance of acute myeloid leukemia, so as to provide a new molecular marker for treatment of leukemia. METHODS: The relationship between the expression level of hnRNP K and the drug resistance of myeloid leukemia was verified by fluorescence quantitative PCR; the expression of autophagy related protein LC3I/ II was detected by Western blot after the hnRNP K was modulated by RNA interference technology; the sensitivity of leukemia cells to doxorubicin was analyzed before and after the expression of hnRNP K were modulatd. RESULTS: The expression of hnRNP K and LC3I/II significantly increased in bone marrow nonremission patients and in drug resistant cell line, however, the expression of LC3I/ II decreased when the expression of hnRNP K were reduced, while the sensitivity of cells to adriamycin could be recovered. CONCLUSION: hnRNP K may be involved in the formation of adriamycin resistance in acute myeloid leukemia by regulating autophagy.

HnRNP K contributes to drug resistance in acute myeloid leukemia through the regulation of autophagy.

The goal of this study was to explore the role of heterogeneous nuclear ribonucleoprotein K (hnRNP K) in drug resistance through the regulation of autophagy in acute myeloid leukemia (AML). First, we used fluorescence quantitative polymerase chain reaction (PCR) to verify the connection between the expression level of hnRNP K and the level of drug resistance in AML. We then used Western blotting to determine the expression level of the autophagy-related proteins microtubule-associated protein light chain 3 I and II (LC3 I/II) after the modulation of hnRNP K by ribonucleic acid (RNA) interference. Finally, an analysis of adriamycin drug sensitivity was conducted before and after the modulation of hnRNP K expression. hnRNP K and LC3 I/II were significantly overexpressed in the bone marrow of nonremission patients and in drug-resistant cell lines; however, the expression of LC3 I/II was decreased when the expression of hnRNP K was reduced and drug sensitivity to adriamycin could be restored. hnRNP K may be involved in the development of adriamycin resistance in AML through the regulation of autophagy.

Versatile Molding Process for Tough Cellulose Hydrogel Materials.

Shape-persistent and tough cellulose hydrogels were fabricated by a stepwise solvent exchange from a homogeneous ionic liquid solution of cellulose exposure to methanol vapor. The cellulose hydrogels maintain their shapes under changing temperature, pH, and solvents. The micrometer-scale patterns on the mold were precisely transferred onto the surface of cellulose hydrogels. We also succeeded in the spinning of cellulose hydrogel fibers through a dry jet-wet spinning process. The mechanical property of regenerated cellulose fibers improved by the drawing of cellulose hydrogel fibers during the spinning process. This approach for the fabrication of tough cellulose hydrogels is a major advance in the fabrication of cellulose-based structures with defined shapes.

[Bioinformatic analysis of chronic myeloid leukemia progression and preliminary experimental verification].

This study was aimed to explore the progression mechanism of chronic myeloid leukemia, so as to provide the new molecular markers for evaluation of CML clinical outcome and selection of treatment. The microarray data of genes related with progression from different phase of chronic myeloid leukemia (CML) were collected from public data depository GEO (Gene expression datasets). SAM analysis, fold change filtering, cross comparison were used to analyze the data and identify different genes. Moreover, MeV and pSTIING sofewares were used to analyze the key differential genes and signal pathways. At last, Q-PCR were used to confirm the predicted key gene. The results indicated that after comparison, 9 genes were differentially expressed from AP to BC, and the integrin-mediated cell adhesion , focal adhesion, regulation of actin cytoskeleton were the principal pathways during CML progression. Network construction analysis found that AP-related genes or pathways may be the original signals; and MLLT4, WDR35 and EPHB4 were the key genes for CML progression. EPHB4 was confirmed by Q-PCR in CML BC patients and CP patients. It is concluded that MLLT4, WDR35, EPHB4, integrin-mediated cell adhesion, focal adhesion and regulation of actin cytoskeleton are the principal genes and pathways during CML progression.