Artificial intelligence (AI) is expected to support clinical judgement in medicine. We constructed a new predictive model for diabetic kidney diseases (DKD) using AI, processing natural language and longitudinal data with big data machine learning, based on the electronic medical records (EMR) of 64,059 diabetes patients. AI extracted raw features from the previous 6 months as the reference period and selected 24 factors to find time series patterns relating to 6-month DKD aggravation, using a convolutional autoencoder. AI constructed the predictive model with 3,073 features, including time series data using logistic regression analysis. AI could predict DKD aggravation with 71% accuracy. Furthermore, the group with DKD aggravation had a significantly higher incidence of hemodialysis than the non-aggravation group, over 10 years (N = 2,900). The new predictive model by AI could detect progression of DKD and may contribute to more effective and accurate intervention to reduce hemodialysis.

Although several murine mAbs that have been humanized became useful therapeutic agents against a few malignancies, therapeutic Abs are not yet available for the majority of the human cancers because of our lack of knowledge of which antigens (Ags) can become useful targets. In the present study we established a procedure for comprehensive identification of such Ags through the extensive isolation of human mAbs that may become therapeutic. Using the phage-display Ab library we isolated a large number of human mAbs that bind to the surface of tumor cells. They were individually screened by immunostaining, and clones that preferentially and strongly stained the malignant cells were chosen. The Ags recognized by those clones were isolated by immunoprecipitation and identified by MS. We isolated 2,114 mAbs with unique sequences and identified 21 distinct Ags highly expressed on several carcinomas. Of those 2,114 mAbs 356 bound specifically to one of the 21 Ags. After preparing complete IgG1 Abs the in vitro assay for Ab-dependent cell-mediated cytotoxicity (ADCC) and the in vivo assay in cancer-bearing athymic mice were performed to examine antitumor activity. The mAbs converted to IgG1 revealed effective ADCC as well as antitumor activity in vivo. Because half of the 21 Ags showed distinct tumor-specific expression pattern and the mAbs isolated showed various characteristics with strong affinity to the Ag, it is likely that some of the Ags detected will become useful targets for the corresponding carcinoma therapy and that several mAbs will become therapeutic agents.phage Ab library ͉ therapeutic Ab ͉ tumor-associated antigen S ince the discovery of a method to produce mAbs numerous scientists have been trying to identify and produce mAbs that could be used for immunotherapy against various malignancies. The success for example of alemtuzumab against CD52, trastuzumab against HER2, and rituximab against CD20 for treatment of chronic lymphocytic leukemia, breast cancer, and nonHodgkins lymphoma, respectively (1-3), suggests that mAbs are likely to become very important therapeutic agents also against a wider range of cancers. However, for the majority of the human cancers useful therapeutic Abs are not yet available because of our lack of knowledge of which antigens (Ags) are likely to become useful targets (4). Therefore, several groups of investigators have been trying to identify other potential Ags as targets for immunotherapy using microarray technology (5, 6). Although many differences in transcripts have been revealed between malignant cells and the normal counterpart cells, it will take more time and laborious work to examine which Ags could be targets and to prepare therapeutic Abs against them. Furthermore, the presence of a large amount of transcripts does not always indicate expression of a large amount of the proteins.Our experimental approach was designed in the opposite way to the strategy with the microarray technology mentioned above and was based on the phage-display technology (7). First we isolate...

No abstract

Because of a worldwide shortage of renal grafts, kidneys procured from donors after cardiac death (DCD) have recently become an important source of renal transplants. However, DCD kidneys often have complications with delayed graft function (DGF) and recipients require hemodialysis (HD) in the early period after kidney transplantation (KTx). This study evaluated serum NGAL as a potential specific parameter to predict early functional recovery of transplanted DCD kidneys. The average serum neutrophil gelatinase-associated lipocalin (NGAL) level in normal samples was 53 +/- 30 ng/ml, while that in patients with chronic renal failure requiring HD was markedly raised at 963 +/- 33 ng/ml. In patients undergoing a living-related KTx from a living donor (n=11), serum NGAL level decreased rapidly after KTx, and only in two cases, with serum NGAL levels over 400 ng/ml on postoperative day 1 (POD1), was HD required due to DGF. In contrast, all patients undergoing a KTx from a DCD (n=5) required HD due to DGF. Even in these cases, serum NGAL levels decreased rapidly several days after a KTx prior to the recovery of urine output and preceding the decrease in serum creatinine level. The pattern of decline in serum NGAL was biphasic, the decrease after the second peak indicating a functional recovery within the next several days. These data suggest that monitoring of serum NGAL levels may allow us to predict graft recovery and the need for HD after a KTx from a DCD.

The expression pattern of Notch family receptors during mouse spermatogenesis was examined by immunohistochemistry. The entire cytoplasm of spermatogonia, spermatocytes and spermatids showed staining with antibodies against extracellular domains of Notch1, 2 and 4. In contrast, the nuclei of spermatogonia showed staining with an antibody against the intracellular domain of Notch3, and the nuclei of spermatocytes and spermatids showed staining with antibodies against the intracellular domains of Notch1 and 4. During regeneration of spermatogonia in busulfan-treated mice, the nuclei of all proliferating cells showed staining for the intracellular domain of Notch3. Western blot analysis showed that the molecular weights of the intracellular domains of Notch1 and 3 localizing in the nuclear fraction were smaller than those in the cytoplasmic fraction. This was consistent with the theory that the intracellular domain of Notch was cleaved in the cytoplasm and translocated to the nucleus. These results suggest that different Notch signals are sequentially activated during mouse spermatogenesis and control the proliferation and differentiation of spermatogenic stem cells.