A Machine Learning Model to Predict the Histology of Retroperitoneal Lymph Node Dissection Specimens.

BACKGROUND/AIM: While post-chemotherapy retroperitoneal lymph node dissection (PC-RPLND) benefits patients with teratoma or viable germ cell tumors (GCT), it becomes overtreatment if necrosis is detected in PC-RPLND specimens. Serum microRNA-371a-3p correctly predicts residual viable GCT with 100% sensitivity; however, prediction of residual teratoma in PC-RPLND specimens using current modalities remains difficult. Therefore, we developed a machine learning model using CT imaging and clinical variables to predict the presence of residual teratoma in PC-RPLND specimens. PATIENTS AND METHODS: This study included 58 patients who underwent PC-RPLND between 2005 and 2019 at the University of Tsukuba Hospital. On CT imaging, 155 lymph nodes were identified as regions of interest (ROIs). The ResNet50 algorithm and/or Support Vector Machine (SVM) classification were applied and a nested, 3-fold cross-validation protocol was used to determine classifier accuracy. RESULTS: PC-RPLND specimen analysis revealed 35 patients with necrosis and 23 patients with residual teratoma, while histology of 155 total ROIs showed necrosis in 84 ROIs and teratoma in 71 ROIs. The ResNet50 algorithm, using CT imaging, achieved a diagnostic accuracy of 80.0%, corresponding to a sensitivity of 67.3%, a specificity of 90.5%, and an AUC of 0.84, whereas SVM classification using clinical variables achieved a diagnostic accuracy of 74.8%, corresponding to a sensitivity of 59.0%, a specificity of 88.1%, and an AUC of 0.84. CONCLUSION: Our machine learning models reliably distinguish between necrosis and residual teratoma in clinical PC-RPLND specimens.

Structural Analyses of Designed α-Helix and ß-Sheet Peptide Nanofibers Using Solid-State Nuclear Magnetic Resonance and Cryo-Electron Microscopy and Introduction of Structure-Based Metal-Responsive Properties.

The 21-residue peptide α3, which is artificially designed and consists of three repeats of 7 residues, is known to rapidly assemble into the α-helix nanofiber. However, its molecular structure within the fiber has not yet been fully elucidated. Thus, we conducted a thorough investigation of the fiber's molecular structure using solid-state NMR and other techniques. The molecules were found to be primarily composed of the α-helix structure, with some regions near the C- and N-terminal adopting a 310-helix structure. Furthermore, it was discovered that ß-sheet hydrogen bonds were formed between the molecules at both ends. These intermolecular interactions caused the molecules to assemble parallelly in the same direction, forming helical fibers. In contrast, we designed two molecules, CaRP2 and ßKE, that can form ß-sheet intermolecular hydrogen bonds using the entire molecule instead of just the ends. Cryo-EM and other measurements confirmed that the nanofibers formed in a cross ß structure, albeit at a slow rate, with the formation times ranging from 1 to 42 days. To create peptide nanofibers that instantaneously respond to changes in the external environment, we designed several molecules (HDM1-3) based on α3 by introducing metal-binding sites. One of these molecules was found to be highly responsive to the addition of metal ions, inducing α-helix formation and simultaneously assembling into nanofibers. The nanofibers lost their structure upon removal of the metal ion. The change occurred promptly and was reversible, demonstrating that the intended level of responsiveness was attained.

Prognostic nutritional index of early post-pembrolizumab therapy predicts long-term survival in patients with advanced urothelial carcinoma.

Pembrolizumab has been widely used to treat advanced urothelial carcinoma that has progressed after first-line platinum-based chemotherapy. Because its clinical benefits are limited, biomarkers that can predict a good response to pembrolizumab are required. The prognostic nutritional index (PNI), calculated using the serum albumin level and peripheral lymphocyte count, has been evaluated as a predictive biomarker in cancer immunotherapy. The present study investigated the application of PNI as a predictive biomarker for pembrolizumab response in patients with advanced urothelial cancer. A retrospective study was conducted on 34 patients treated with pembrolizumab at Shiga University of Medical Science Hospital between January 2018 and July 2022. The posttreatment PNI (post-PNI) was calculated within 2 months of starting pembrolizumab. The present study investigated the association between post-PNI and objective response, overall survival (OS) and progression-free survival (PFS). The patient cohort was stratified into two categories, high and low post-PNI groups, with a cutoff value of post-PNI at 40. The higher post-PNI group demonstrated a better disease control rate than the lower post-PNI group (complete response + partial response + stable disease, 75 vs. 21%, P=0.004). Regarding median OS, the higher post-PNI group exhibited a significantly longer survival time than the lower post-PNI group (23.1 vs. 2.9 months, P<0.001). Similarly, the higher post-PNI group exhibited a significantly longer PFS than the lower post-PNI group (10.2 vs.1.9 months, P<0.001). Multivariate analysis showed that a higher post-PNI value was an independent predictor for OS (hazard ratio, 0.04; 95% confidence interval, 0.01-0.14; P<0.001) and PFS (hazard ratio, 0.12; 95% confidence interval, 0.04-0.35; P<0.001). The present study indicated that the post-PNI was a predictor of favorable clinical outcomes in patients treated with pembrolizumab for advanced urothelial carcinoma.

Urine cell image recognition using a deep-learning model for an automated slide evaluation system.

OBJECTIVES: To develop a classification system for urine cytology with artificial intelligence (AI) using a convolutional neural network algorithm that classifies urine cell images as negative (benign) or positive (atypical or malignant). PATIENTS AND METHODS: We collected 195 urine cytology slides from consecutive patients with a histologically confirmed diagnosis of urothelial cancer (between January 2016 and December 2017). Two certified cytotechnologists independently evaluated and labelled each slide; 4637 cell images with concordant diagnoses were selected, including 3128 benign cells (negative), 398 atypical cells, and 1111 cells that were malignant or suspicious for malignancy (positive). This pathologically confirmed labelled dataset was used to represent the ground truth for AI training/validation/testing. Customized CutMix (CircleCut) and Refined Data Augmentation were used for image processing. The model architecture included EfficientNet B6 and Arcface. We used 80% of the data for training and validation (4:1 ratio) and 20% for testing. Model performance was evaluated with fivefold cross-validation. A receiver-operating characteristic (ROC) analysis was used to evaluate the binary classification model. Bayesian posterior probabilities for the AI performance measure (Y) and cytotechnologist performance measure (X) were compared. RESULTS: The area under the ROC curve was 0.99 (95% confidence interval [CI] 0.98-0.99), the highest accuracy was 95% (95% CI 94-97), sensitivity was 97% (95% CI 95-99), and specificity was 95% (95% CI 93-97). The accuracy of AI surpassed the highest level of cytotechnologists for the binary classification [Pr(Y > X) = 0.95]. AI achieved >90% accuracy for all cell subtypes. In the subgroup analysis based on the clinicopathological characteristics of patients who provided the test cells, the accuracy of AI ranged between 89% and 97%. CONCLUSION: Our novel AI classification system for urine cytology successfully classified all cell subtypes with an accuracy of higher than 90%, and achieved diagnostic accuracy of malignancy superior to the highest level achieved by cytotechnologists.

Development of in-house fully residual deep convolutional neural network-based segmentation software for the male pelvic CT.

BACKGROUND: This study aimed to (1) develop a fully residual deep convolutional neural network (CNN)-based segmentation software for computed tomography image segmentation of the male pelvic region and (2) demonstrate its efficiency in the male pelvic region. METHODS: A total of 470 prostate cancer patients who had undergone intensity-modulated radiotherapy or volumetric-modulated arc therapy were enrolled. Our model was based on FusionNet, a fully residual deep CNN developed to semantically segment biological images. To develop the CNN-based segmentation software, 450 patients were randomly selected and separated into the training, validation and testing groups (270, 90, and 90 patients, respectively). In Experiment 1, to determine the optimal model, we first assessed the segmentation accuracy according to the size of the training dataset (90, 180, and 270 patients). In Experiment 2, the effect of varying the number of training labels on segmentation accuracy was evaluated. After determining the optimal model, in Experiment 3, the developed software was used on the remaining 20 datasets to assess the segmentation accuracy. The volumetric dice similarity coefficient (DSC) and the 95th-percentile Hausdorff distance (95%HD) were calculated to evaluate the segmentation accuracy for each organ in Experiment 3. RESULTS: In Experiment 1, the median DSC for the prostate were 0.61 for dataset 1 (90 patients), 0.86 for dataset 2 (180 patients), and 0.86 for dataset 3 (270 patients), respectively. The median DSCs for all the organs increased significantly when the number of training cases increased from 90 to 180 but did not improve upon further increase from 180 to 270. The number of labels applied during training had a little effect on the DSCs in Experiment 2. The optimal model was built by 270 patients and four organs. In Experiment 3, the median of the DSC and the 95%HD values were 0.82 and 3.23 mm for prostate; 0.71 and 3.82 mm for seminal vesicles; 0.89 and 2.65 mm for the rectum; 0.95 and 4.18 mm for the bladder, respectively. CONCLUSIONS: We have developed a CNN-based segmentation software for the male pelvic region and demonstrated that the CNN-based segmentation software is efficient for the male pelvic region.

Grouping of chemicals based on the potential mechanisms of hepatotoxicity of naphthalene and structurally similar chemicals using in vitro testing for read-across and its validation.

Integrated Approaches to Testing and Assessment provides a framework to improve the reliability of read-across for chemical risk assessment of systemic toxicity without animal testing. However, the availability of only a few case studies hinders the use of this concept for regulatory purposes. Thus, we compared the biological similarity of structurally similar chemicals using in vitro testing to demonstrate the validity of this concept for grouping chemicals and to extract key considerations in read-across. We analyzed the hepatotoxicity of naphthalene and three chemicals structurally similar to naphthalene (2,7-naphthalenediol, 1,5-naphthalenediol, and 1-naphthol) for which 90-day repeated dose toxicity data are available. To elucidate and compare their potential mechanisms, we conducted in vitro microarray analysis using rat primary hepatocytes and validated the results using a biomarker and metabolic activation analysis. We observed that 2,7-naphthalenediol, 1,5-naphthalenediol, and 1-naphthol had similar potential mechanisms, namely, induction of oxidative stress by their metabolic activation. Conversely, naphthalene did not show a similar toxicity effect. The existing in vivo data confirmed our grouping of chemicals based on this potential mechanism. Thus, our findings suggest that in vitro toxicogenomics and related biochemical assays are useful for comparing biological similarities and grouping chemicals based on their toxicodynamics for read-across.

Syndrome of inappropriate antidiuretic hormone secretion as a side effect of chemotherapy for testicular cancer: A case report.

INTRODUCTION: Inappropriate antidiuretic hormone secretion syndrome can be a serious adverse event of cisplatin-based chemotherapy. Cisplatin had to be changed to other drugs or chemotherapy completely discontinued in earlier reported cases. CASE PRESENTATION: Three cycles of bleomycin, etoposide, and cisplatin chemotherapy were planned for a 40-year-old man with a diagnosis of lymph node recurrence of testicular cancer. On day 9, he suffered from vomiting and mental disturbance. Severe hyponatremia (110 mEq/L) with low plasma osmolality led to a diagnosis of a syndrome of inappropriate antidiuretic hormone secretion, and infusions of hypertonic saline and salt intake were prescribed. Second and third courses of bleomycin, etoposide, and cisplatin chemotherapy could then be given with careful electrolyte management. CONCLUSION: Continuation of cisplatin administration with precise electrolyte adjustment can be a treatment option in regimens where cisplatin is essential for achieving optimal antitumor efficacy.

Inhibitory effect of donepezil on bradykinin-induced increase in the intracellular calcium concentration in cultured cortical astrocytes.

Donepezil is a potent and selective acetylcholinesterase inhibitor developed for the treatment of Alzheimer's disease. In the present study, we investigated the responses of astrocytes to bradykinin, an inflammatory mediator, and the effect of donepezil on these responses using cultured cortical astrocytes. Bradykinin induced a transient increase of intracellular calcium concentration ([Ca2+]i) in cultured astrocytes. Bradykinin-induced [Ca2+]i increase was inhibited by the exposure to thapsigargin, which depletes Ca2+ stores on endoplasmic reticulum, but not by the exclusion of extracellular Ca2+. Twenty four hours pretreatment of donepezil reduced the bradykinin-induced [Ca2+]i increase. This reduction was inhibited not only by mecamylamine, a nAChR antagonist, but also by PI3K and Akt inhibitors. In addition, donepezil inhibited bradykinin-induced increase of the intracellular reactive oxygen species level in astrocytes. These results suggest that donepezil inhibits the inflammatory response induced by bradykinin via nAChR and PI3K-Akt pathway in astrocytes.

DJ-1 activates SIRT1 through its direct binding to SIRT1.

The DJ-1 gene is a ras-dependent oncogene and also a causative gene for a familial form of Parkinson's disease park7. DJ-1 is a multi-functional protein and plays roles in regulation of cell growth, cells death, metabolism and mitochondrial homeostasis against oxidative stress. To explore various functions, DJ-1 associates with a number of proteins localized in the nucleus, cytoplasm and mitochondria. The oxidative status of a cysteine residue at an amino acid number 106 (C106) of DJ-1 determines the active level of DJ-1. Precise molecular mechanism of exploration of DJ-1 function is, however, not resolved. In this study, we identified Sirtuin family proteins (SIRT1, 2, and 4-6) as DJ-1-binding proteins, and DJ-1 associated with SIRT1 in cells. Sirtuins like DJ-1 also regulates growth, death and metabolism of cells and mitochondrial homeostasis. We found that DJ-1 stimulated deacetylase activity of SIRT1 and that SIRT1-suppressed transcriptional activity of SIRT1-target p53 was further decreased by DJ-1. Furthermore, SIRT1 activity was reduced in DJ-1-knockout cells, and this reduced activity was restored by re-introduction of wild-type DJ-1 but not of C106-mutant DJ-1 into DJ-1-knockout cells. It is first report showing direct connection of DJ-1 with SIRT1.

Polycomb-dependent epigenetic landscape in adult T-cell leukemia.

Adult T-cell leukemia-lymphoma (ATL) shows global gene expression alterations that confer cellular characteristics and unfavorable prognosis. However, molecular mechanisms of the sustained expression changes are largely unknown, because there is no study addressing the relationship between landscapes of the gene expression and epigenetic modifications. Here, we analyzed ATL epigenome and integrated it with transcriptome from primary ATL cells and those from corresponding normal CD4(+)T cells to decipher ATL-specific "epigenetic code" that was critical for cell identity. We found that polycomb-repressive complex 2 (PRC2)-mediated trimethylation at histone H3Lys27 (H3K27me3) was significantly and frequently reprogrammed at half of genes in ATL cells. A large proportion of the abnormal gene downregulation was detected at the early stage of disease progression and was explained by H3K27me3 accumulation. The global H3K27me3 alterations involved ATL-specific gene expression changes that included several tumor suppressors, transcription factors, epigenetic modifiers, miRNAs, and developmental genes, suggesting diverse outcomes by the PRC2-dependent hierarchical regulation. Interestingly, a key enzyme, EZH2, was sensitive to promiscuous signaling network including the NF-κB pathway and was functionally affected by human T-cell leukemia virus type I (HTLV-1) Tax. The Tax-dependent immortalized cells showed H3K27me3 reprogramming that was significantly similar to that of ATL cells. Of note, a majority of the epigenetic silencing has occurred in leukemic cells from indolent ATL and also in HTLV-1-infected T cells from asymptomatic HTLV-1 carriers. Because pharmacologic inhibition of EZH2 reversed epigenetic disruption and selectively eliminated leukemic and HTLV-1-infected cells, targeting the epigenetic elements will hold great promise in treatment and prevention of the onset of ATL and HTLV-1-related diseases.