A compute-in-memory chip based on resistive random-access memory.

Realizing increasingly complex artificial intelligence (AI) functionalities directly on edge devices calls for unprecedented energy efficiency of edge hardware. Compute-in-memory (CIM) based on resistive random-access memory (RRAM)1 promises to meet such demand by storing AI model weights in dense, analogue and non-volatile RRAM devices, and by performing AI computation directly within RRAM, thus eliminating power-hungry data movement between separate compute and memory2-5. Although recent studies have demonstrated in-memory matrix-vector multiplication on fully integrated RRAM-CIM hardware6-17, it remains a goal for a RRAM-CIM chip to simultaneously deliver high energy efficiency, versatility to support diverse models and software-comparable accuracy. Although efficiency, versatility and accuracy are all indispensable for broad adoption of the technology, the inter-related trade-offs among them cannot be addressed by isolated improvements on any single abstraction level of the design. Here, by co-optimizing across all hierarchies of the design from algorithms and architecture to circuits and devices, we present NeuRRAM-a RRAM-based CIM chip that simultaneously delivers versatility in reconfiguring CIM cores for diverse model architectures, energy efficiency that is two-times better than previous state-of-the-art RRAM-CIM chips across various computational bit-precisions, and inference accuracy comparable to software models quantized to four-bit weights across various AI tasks, including accuracy of 99.0 percent on MNIST18 and 85.7 percent on CIFAR-1019 image classification, 84.7-percent accuracy on Google speech command recognition20, and a 70-percent reduction in image-reconstruction error on a Bayesian image-recovery task.

Predictive factors on postoperative venous thromboembolism after minimally invasive colorectal cancer surgery: a retrospective observational study.

BACKGROUND: Venous thromboembolism (VTE) is a serious and preventable postoperative complication. However, the predictive significance of perioperative biochemical parameters for VTE after minimally invasive colorectal cancer surgery remains unclear. METHODS: A total of 149 patients undergoing minimally invasive colorectal cancer surgery were collected between October 2021 and October 2022. Biochemical parameters related to preoperative and postoperative day 1, day 3, and day 5 were collected, including D-Dimer, mean platelet volume (MPV), and maximum amplitude (MA) of thromboelastography (TEG). Receiver operating characteristic (ROC) curves were used to explore the predictive powers of meaningful biochemical parameters for postoperative VTE, and calibration curves were used to assess predictive accuracy. RESULTS: The overall cumulative incidence of VTE was 8.1% (12/149). The preoperative and postoperative day 3 D-Dimer, postoperative day 3, and day 5 MPV, and postoperative day 1, day 3, and day 5 TEG-MA was significantly higher in the VTE group than in the non-VTE group (P < 0.05). The results of both the ROC curve and the calibration curve indicated that these meaningful D-Dimer, MPV, and TEG-MA had moderate discrimination and consistency for postoperative VTE. CONCLUSIONS: D-Dimer, MPV, and TEG-MA may predict postoperative VTE in patients undergoing minimally invasive surgery for colorectal cancer at specific times in the perioperative period.

Association between Dietary Anthocyanidins and Biliary Cancer Risk in 98,458 Participants: Results from a Prospective Study.

BACKGROUND: Previous studies have suggested anthocyanidins or anthocyanidin-rich foods and extracts exhibit protective effects against various cancers. However, the relationship between dietary anthocyanidins and the risk of biliary cancer remains uncertain. METHODS: This study used data from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial to investigate the relationship between total anthocyanidins intake and biliary cancer incidence. Cox regression analysis was conducted to estimate HRs and corresponding 95% confidence intervals (CI) for the incidence of biliary cancer, with adjustments made for confounding factors. A restricted cubic spline model was employed to examine the dose-response relationship. In addition, subgroup and sensitivity analyses were conducted to evaluate potential interactions and test the model's robustness. RESULTS: During 8.9 years and 872,645.3 person-years of follow-up, 95 cases of biliary cancer were observed. The incidence rate of biliary cancer in this study was 11 cases per 100,000 person-years. Using the fully adjusted Cox regression model, the inverse association was observed between total anthocyanidins intake and the risk of biliary cancer (HR Q4 vs..Q1: 0.52; 95% CI: 0.29-0.91; Ptrend = 0.043). This association remained significant in sensitivity analyses. A linear dose-response relationship (Pnonlinearity = 0.118) and potential interaction with drinking status (Pinteraction = 0.033) were identified. CONCLUSIONS: This study provides evidence of an inverse association between total anthocyanidins intake and biliary cancer incidence. IMPACT: Our study found a total anthocyanidin-rich diet was associated with a reduced risk of biliary cancer in Americans ages 55 to 74 years.

Edge learning using a fully integrated neuro-inspired memristor chip.

Learning is highly important for edge intelligence devices to adapt to different application scenes and owners. Current technologies for training neural networks require moving massive amounts of data between computing and memory units, which hinders the implementation of learning on edge devices. We developed a fully integrated memristor chip with the improvement learning ability and low energy cost. The schemes in the STELLAR architecture, including its learning algorithm, hardware realization, and parallel conductance tuning scheme, are general approaches that facilitate on-chip learning by using a memristor crossbar array, regardless of the type of memristor device. Tasks executed in this study included motion control, image classification, and speech recognition.

Effect of nHA-Coated Femoral Stem Prosthesis Combined with Platelet-Rich Plasma in Hemi Hip Replacement of Femoral Neck Fracture in the Elderly.

Purpose: This study investigated the efficacy of nanohydroxyapatite- (nHA-) coated biological prosthesis combined with platelet-rich plasma (PRP) in hemi hip replacement of femoral neck fracture (FNF) in the elderly. Methods: From September 2018 to September 2021, 102 elderly patients with FNF treated in our hospital were chosen and divided into two groups according to different intervention methods. Fifty-one patients in the bone cement group were treated with bone cement prosthesis, and the rest 51 patients in the observation group were treated with nHA biological prosthesis combined with PRP in hemi hip replacement. In order to explore the osteogenic effect of nHA and PRP, osteoblasts were cultured. Results: It was found that nHA and PRP could both effectively promote the proliferation of osteoblasts and improve their mineralization ability, especially when used in combination. In the course of clinical therapy, we found that the use of biological prosthesis combined with PRP could effectively reduce the level of serum procollagen type I carboxy terminal peptide (PICP) and better improve the levels of bone alkaline phosphatase (BALP) and bone Gla protein (BGP), so as to reduce the bone conversion rate and promote the formation of new bone around the prosthesis. In addition, no significant difference was found in intraoperative bleeding, operation time, hospital stay, 48 h drainage volume, partial weight-bearing time, and complete weight-bearing time between two groups. Otherwise, the use of biological prosthesis could effectively avoid the occurrence of adverse reactions such as bone cement crisis and fracture around femoral prosthesis, so as to better restore the hip function and improve patients' life quality. Conclusions: Therefore, in hemi hip replacement of FNF in the elderly, nHA biological prosthesis combined with PRP can effectively promote the formation of new bone around the prosthesis stem, so as to obtain good initial stability, enable patients to carry out early weight-bearing exercise, and effectively avoid adverse reactions caused by bone cement prosthesis, thus improving patients' hip function and life quality.

Comprehensive analysis of the immune implication of FABP4 in colon adenocarcinoma.

BACKGROUND: Fatty acid-binding protein 4 (FABP4) has been reported to be associated with tumor progress and poor prognosis in various cancers. However, the relationship between FABP4 expression and tumor immunity in colon adenocarcinoma (COAD) is still poorly understood. METHODS: FABP4 mRNA expression was analyzed using The Cancer Genome Atlas (TCGA)-COAD data. FABP4 protein staining was performed by immunohistochemistry (IHC) staining in our 10 paired COAD samples and corresponding adjacent noncancerous tissues. The association between FABP4 and immune cell infiltration was evaluated by Tumor Immune Estimation Resource (TIMER) database. FABP4 coexpressed genes were identified based on Cancer Cell Line Encyclopedia (CCLE) database, which were employed for further enrichment analysis. FABP4 related immunomodulators was identified by Tumor and Immune System Interaction Database (TISIDB) database, and a prognostic risk signature was constructed based on FABP4-related immunomodulators using stepwise Cox regression analysis. A nomogram consists of FABP4 related immunomodulators signature and clinical parameters was developed to predict the overall survival (OS). RESULTS: In TCGA data, we found that the decreased FABP4 mRNA expression in COAD samples compared with normal samples, and low FABP4 mRNA expression was associated with B cells, CD4+ T cells, CD8+ T cells, myeloid dendritic cells, macrophages, and neutrophils. In our 10 paired samples, the protein levels of COAD were lower in all COAD tissues than in their adjacent noncancerous tissues. Functional enrichment analysis revealed that FABP4 coexpressed genes were mostly enriched in immune-related pathways. Based on 54 FABP4-related immunomodulators, a 2-gene FABP4-related prognostic risk signature was developed, and the signature stratified the patients into the high-risk and low-risk groups with statistically different survival outcomes. The Nomogram consists of the prognostic signature and clinical parameters had a certain predictability for prognosis of COAD patients. CONCLUSION: These findings suggest that FABP4 is associated with 2-gene immune signature which also correlate with the prognosis of COAD patients.

Case report: A case report and literature review of extrapancreatic solid pseudopapillary neoplasm.

Background: Solid pseudopapillary neoplasm (SPN) is a rare tumor with low malignant potential, which typically occurs in the pancreas. Extrapancreatic SPN is also extremely rare worldwide. Case presentation: We report a case of a 70-year-old woman hospitalized with abdominal pain and bloating. The patient did not have any underlying diseases, such as diabetes, coronary heart disease, or hypertension. More than 30 years ago, the patient underwent surgery for "ectopic pregnancy". The patient had no family history of hereditary disease, nor did any immediate family members have a history of cancer. Laboratory tests showed that her hemoglobin and albumin levels were low and she had a high level of cancer antigen 125 (CA125). Enhanced computed tomography (CT) showed a large tumor in the abdomen and pelvis. The patient subsequently underwent surgery, and it was found that the tumor was attached to the terminal ileum. Pathological findings suggested that the tumor was an extrapancreatic SPN, with an ectopic pancreas found in the tumor tissue. The patient did not receive chemotherapy or radiotherapy after surgery. After 13 months of follow-up, the patient was admitted again with abdominal pain. CT showed tumor recurrence with extensive systemic metastases. The patient and her family refused reoperation and biopsy, and the patient was discharged after the abdominal pain and anemia resolved. Conclusion: We report a rare case of extrapancreatic SPN of ileal origin, which could be the first report worldwide. It had aggressive biological features, with recurrence and metastasis 13 months after surgery. For extrapancreatic SPN, the risk of recurrence should be assessed, and for tumors suspected of malignant behavior, a longer follow-up after discharge may be needed. Although SPN generally has a good prognosis after surgery, there is no consensus on whether postoperative chemotherapy and other treatments are needed for patients with high recurrence risk.

Comprehensive analysis of the correlation between GSTM1 and tumor immunity in colon cancer.

Background: Glutathione S-transferase mu 1 (GSTM1) is one of the major glutathione conjugation enzymes. Its expression and activity have been suggested to correlate with the occurrence of colon cancer; however, the role of GSTM1 in tumor immunity remains unclear. Methods: Relevant data downloaded from The Cancer Genome Atlas (TCGA), Clinical Proteomic Tumor Analysis Consortium (CPTAC), and Human Protein Atlas (HPA) was used to perform a multi-dimensional expression analysis of GSTM1 in colon adenocarcinoma (COAD). The correlation between GSTM1 and tumor immunity was analyzed with multiple online tools. Then protein-protein interaction (PPI) network and functional enrichment analyses of GSTM1-associated immunomodulators were performed. Further, we developed the Cox regression model based on the GSTM1-related immunomodulators. Finally, a GSTM1-based clinical nomogram and a calibration curve was established to predict the probability and accuracy of long-term survival. Result: GSTM1 was significantly downregulated in COAD versus normal tissues. Infiltration levels of B cells, CD8+ T cells, and dendritic cells were closely correlated to GSTM1 gene copy number deletion, and GSTM1 expression levels in COAD positively correlated with dendritic cell, B cell, neutrophil, and macrophage infiltration. Functional enrichment analysis indicated 36 GSTM1-related immunomodulators are involved in immune-related pathways of regulating T cell activation and lymphocytic activation. A 2-gene prognostic risk signature based on the 36 GSTM1-related immunomodulators was built using the Cox regression model, and the risk signature in combination with stage had an area under the curve (AUC) value of 0.747 by the receiver operating characteristic method. patients with higher risk scores-calculated based on 2 gene prognostic risk characteristics and further identified as an independent prognostic factor-were associated with worse survival using the Kaplan-Meier analysis. Together, the clinical nomogram and calibration curve based on GSTM1 suggested a good prediction accuracy for long-term survival probability. Conclusions: Our study provided evidence supporting the significant role of GSTM1 in COAD immunity and suggests GSTM1 as a potential novel target for COAD immunotherapy.