Explainable AI to improve acceptance of convolutional neural networks for automatic classification of dopamine transporter SPECT in the diagnosis of clinically uncertain parkinsonian syndromes.

PURPOSE: Deep convolutional neural networks (CNN) provide high accuracy for automatic classification of dopamine transporter (DAT) SPECT images. However, CNN are inherently black-box in nature lacking any kind of explanation for their decisions. This limits their acceptance for clinical use. This study tested layer-wise relevance propagation (LRP) to explain CNN-based classification of DAT-SPECT in patients with clinically uncertain parkinsonian syndromes. METHODS: The study retrospectively included 1296 clinical DAT-SPECT with visual binary interpretation as "normal" or "reduced" by two experienced readers as standard-of-truth. A custom-made CNN was trained with 1008 randomly selected DAT-SPECT. The remaining 288 DAT-SPECT were used to assess classification performance of the CNN and to test LRP for explanation of the CNN-based classification. RESULTS: Overall accuracy, sensitivity, and specificity of the CNN were 95.8%, 92.8%, and 98.7%, respectively. LRP provided relevance maps that were easy to interpret in each individual DAT-SPECT. In particular, the putamen in the hemisphere most affected by nigrostriatal degeneration was the most relevant brain region for CNN-based classification in all reduced DAT-SPECT. Some misclassified DAT-SPECT showed an "inconsistent" relevance map more typical for the true class label. CONCLUSION: LRP is useful to provide explanation of CNN-based decisions in individual DAT-SPECT and, therefore, can be recommended to support CNN-based classification of DAT-SPECT in clinical routine. Total computation time of 3 s is compatible with busy clinical workflow. The utility of "inconsistent" relevance maps to identify misclassified cases requires further investigation.

Data-driven identification of diagnostically useful extrastriatal signal in dopamine transporter SPECT using explainable AI.

This study used explainable artificial intelligence for data-driven identification of extrastriatal brain regions that can contribute to the interpretation of dopamine transporter SPECT with 123I-FP-CIT in parkinsonian syndromes. A total of 1306 123I-FP-CIT-SPECT were included retrospectively. Binary classification as 'reduced' or 'normal' striatal 123I-FP-CIT uptake by an experienced reader served as standard-of-truth. A custom-made 3-dimensional convolutional neural network (CNN) was trained for classification of the SPECT images with 1006 randomly selected images in three different settings: "full image", "striatum only" (3-dimensional region covering the striata cropped from the full image), "without striatum" (full image with striatal region removed). The remaining 300 SPECT images were used to test the CNN classification performance. Layer-wise relevance propagation (LRP) was used for voxelwise quantification of the relevance for the CNN-based classification in this test set. Overall accuracy of CNN-based classification was 97.0%, 95.7%, and 69.3% in the "full image", "striatum only", and "without striatum" setting. Prominent contributions in the LRP-based relevance maps beyond the striatal signal were detected in insula, amygdala, ventromedial prefrontal cortex, thalamus, anterior temporal cortex, superior frontal lobe, and pons, suggesting that 123I-FP-CIT uptake in these brain regions provides clinically useful information for the differentiation of neurodegenerative and non-neurodegenerative parkinsonian syndromes.

Exonic SNP in MHC-DMB2 is associated with gene expression and humoral immunity in Japanese quails.

The DMB2 gene is widely expressed at high levels in avian. This gene plays an important role in humoral immunity. The aim of this study was to investigate the effects of 361 G > C Single nucleotide polymorphism (SNP) on DMB2 protein structure and gene expression to determine how the 361 G > C SNP affects humoral immune response in Japanese quails. 0.2 mL of 5% sheep red blood cell (SRBC) was injected into breast muscle of 130 Japanese quails on 28 days. After DNA extraction, PCR was carried out to amplify a 333-base pair DNA fragment from the exon 2 of DMB2 gene. The pattern of all samples was determined through RFLP technique. PCR-RFLP results identified two alleles segregating (C, G) as three genotypes (CC, CG and GG) in Japanese Quails. The antibody response to SRBC with CC genotype was significantly higher than the CG and GG genotypes (P < 0.01). In silico analysis showed that the 361 G > C SNP has no effect on the physicochemical properties and 3D structure. The results of RT-qPCR indicated that the effect of genotype on gene expression is significant, so that the expression of CC genotype is more than CG and GG genotype. It can be inferred that the 361 G > C SNP in the exon 2 of MHC-DMB2 gene is not desirable. This mutation decreases humoral immune response by reducing DMB2 gene expression.

Automated and robust organ segmentation for 3D-based internal dose calculation.

PURPOSE: In this work, we address image segmentation in the scope of dosimetry using deep learning and make three main contributions: (a) to extend and optimize the architecture of an existing convolutional neural network (CNN) in order to obtain a fast, robust and accurate computed tomography (CT)-based organ segmentation method for kidneys and livers; (b) to train the CNN with an inhomogeneous set of CT scans and validate the CNN for daily dosimetry; and (c) to evaluate dosimetry results obtained using automated organ segmentation in comparison with manual segmentation done by two independent experts. METHODS: We adapted a performant deep learning approach using CT-images to delineate organ boundaries with sufficiently high accuracy and adequate processing time. The segmented organs were consequently used as binary masks for further convolution with a point spread function to retrieve the activity values from quantitatively reconstructed SPECT images for "volumetric"/3D dosimetry. The resulting activities were used to perform dosimetry calculations with the kidneys as source organs. RESULTS: The computational expense of the algorithm was sufficient for clinical daily routine, required minimum pre-processing and performed with acceptable accuracy a Dice coefficient of [Formula: see text] for liver segmentation and of [Formula: see text] for kidney segmentation, respectively. In addition, kidney self-absorbed doses calculated using automated segmentation differed by [Formula: see text] from dosimetry performed by two medical physicists in 8 patients. CONCLUSION: The proposed approach may accelerate volumetric dosimetry of kidneys in molecular radiotherapy with 177Lu-labelled radiopharmaceuticals such as 177Lu-DOTATOC. However, even though a fully automated segmentation methodology based on CT images accelerates organ segmentation and performs with high accuracy, it does not remove the need for supervision and corrections by experts, mostly due to misalignments in the co-registration between SPECT and CT images. Trial registration EudraCT, 2016-001897-13. Registered 26.04.2016, www.clinicaltrialsregister.eu/ctr-search/search?query=2016-001897-13 .

Aptamer-Based Sandwich Assay for Measurement of Thymidine Kinase 1 in Serum of Cancerous Patients.

Thymidine kinase 1 (TK1) is traditionally a serum biomarker that is elevated in the early stages of malignancies. The diagnostic and prognostic role of TK1 for screening and monitoring human malignancies has recently been investigated. Anti-human TK1 aptamers were selected through 12 iterative rounds of systematic evolution of ligands by exponential enrichment from a DNA library. The aptamer pool of round 12 was amplified, and the polymerase chain reaction product was cloned on the TA vector. Of the 85 colonies obtained, 52 were identified as positive clones. These aptamers were screened for TK1 with surface plasmon resonance, where apta37 and apta69 showed the highest affinity for TK1. The TK1_apta37 and TK1_apta69 aptamers were used in a sandwich assay platform and successfully detected TK1 in the concentration range of 54-3500 pg mL-1. Clinical samples from 60 cancerous patients were also tested with this assay system and compared using the conventional antibody-based enzyme-linked immunosorbent assay kit. The aptamer sandwich assay demonstrated a dynamic range for TK1 at clinically relevant serum levels, covering subpicogram per milliliter concentrations. The new approach offers a simple and robust method for detecting serum biomarkers that have low and moderate abundance. The results of this study demonstrate the screening capability of the aptamer sandwich assay platform and its potential applicability to the point-of-care testing system.

Myostatin knockout using zinc-finger nucleases promotes proliferation of ovine primary satellite cells in vitro.

Myostatin (MSTN) has previously been shown to negatively regulate the proliferation and differentiation of skeletal muscle cells. Satellite cells are quiescent muscle stem cells that promote muscle growth and repair. Because the mechanism of MSTN in the biology of satellite cells is not well understood, this study was conducted to generate MSTN mono-allelic knockout satellite cells using the zinc-finger nuclease mRNA (MSTN-KO ZFN mRNA) and also to investigate the effect of this disruption on the proliferation and differentiation of sheep primary satellite cells (PSCs). Nineteen biallelic and four mono-allelic knockout cell clones were obtained after sequence analysis. The homologous mono-allelic knockout cells with 5-bp deletion were used to further evaluations. The results demonstrated that mono-allelic knockout of MSTN gene leads to translation inhibition. Real-time quantitative PCR results indicated that knockout of MSTN contributed to an increase in CDK2 and follistatin and a decrease in p21 at the transcript level in proliferation conditions. Moreover, MSTN knockout significantly increased the proliferation of mutant clones (P < 0.01). Consistent with the observed increase in CDK2 and decrease in p21 in cells lacking MSTN, cell cycle analysis showed that MSTN negatively regulated the G1 to S progression. In addition, knockout of myostatin resulted in a remarkable increase in MyoD and MyoG expression under differentiating conditions but had no effect on Myf5 expression. These results expanded our understanding of the regulation mechanism of MSTN. Furthermore, the MSTN-KO ZFN mRNA system in PSCs could be used to generate transgenic sheep in the future.

AAV2-mediated follistatin overexpression induces ovine primary myoblasts proliferation.

BACKGROUND: Follistatin (FST) has been shown to bind to some TGF-ß family members and can function as a potent myostatin (MSTN) antagonist. Recent studies have revealed that over-expression of FST by adeno-associated viruses increases muscle growth in mice, humans and nonhuman primates. In the present study, to determine the effect of FST on ovine primary myoblast (OPM) proliferation, FST was over-expressed using an adeno-associated virus serotype 2 (AAV 2) vector. RESULTS: Western blot results showed that AAV induced the expression of FST protein in transduced OPM cells. Real-time quantitative PCR results indicated that over-expression of FST resulted in a dramatic increase in Akt I and CDK2 expression and a decrease in p21 expression. Moreover, cell cycle analysis confirmed that FST down-regulated p21, a CDK inhibitor, and increased the level of CDK2 expression in OPM cells. Hence, follistatin positively regulated the G1 to S progression. Our results showed that FST induced proliferation through a down-regulation of p21, as only the p21 expression level was down-regulated as a result of FST over-expression in myoblasts, whereas no change was observed in the level of p57 expression. CONCLUSIONS: These results expanded our understanding of the regulation mechanism of FST in ovine primary myoblasts. Our results provide the first evidence that the AAV viral system can be used for gene transfer in ovine myoblast cells. Moreover, the results showed that an AAV vector can successfully induce the expression of FST in OPM cells in vitro. These findings demonstrated that FST over-expression induces proliferation through a down-regulation of the p21 gene under proliferating conditions.

CXC chemokines CXCL1, CXCL9, CXCL10 and CXCL12 are variably expressed in patients with sickle cell disease and carriers: are they predictive tools for disease complications?

BACKGROUND: Sickle cell hemoglobinopathies are amongst a group of genetic disorders resulting from a single base-pair DNA mutation at the beta chain of hemoglobin. Chemokines and cytokines play a part in the pathogenesis of inflammatory and infectious diseases. They are also involved in balancing angiogenesis/angiostasis processes to form new vascular networks. We aimed the present study to measure the circulating CXC chemokines CXCL1, CXCL9, CXCL10, and CXCL12 in the plasma of sickle cell patients (SCD). METHODS: This cross-sectional study was conducted at the Kerman Special Disease Center and Rafsanjan Molecular Medicine Research Center during 2010 to 2011. Peripheral blood specimens were collected from 77 children with SCD and 70 controls. Serum samples were isolated and CXCL1, CXCL9, CXCL10, and CXCL12 were measured using ELISA. RESULTS: The findings of this study demonstrated that serum concentrations of CXCL1 and CXCL12 were elevated in SCD patients when compared with controls. Results also showed that the circulating levels of CXCL9 and CXCL10 were decreased in SCD patients in comparison to control subjects. However, we found increased levels of CXC chemokines in SCD patients suffering from pain crisis but the difference was not significant. CONCLUSIONS: According to the results of this study it can probably be concluded that the balance between angiogenesis/angiostasis CXC chemokines is an important predictive factor for initiation of complications in SCD patients. The elevated level of pro-inflammatory CXC chemokines may also be related to inflammatory responses associated with SCD complication.

The adapter proteins of TLRs, TRIF and MYD88, are upregulated in depressed individuals.

BACKGROUND AND AIMS: TRIF and MYD88 are intracellular adaptor proteins for TLR signaling, and altered expression of these molecules can lead to defective or unregulated immune responses. Furthermore, previous studies revealed that depression may alter immune responses, but its mechanisms of action are unclear yet. There is a possibility that immunity and depression are linked through molecules such as TRIF and MYD88, thus, the aim of this study was to evaluate the mRNA levels of TRIF and MYD88 in the PBMCs isolated from depressed medical students. MATERIAL AND METHODS: The current study examined 38 depressed medical students studying in Iran and 43 healthy students from the same cohort as a control group. The mRNA levels of TRIF and MYD88 were examined in parallel with a housekeeping gene using real-time PCR. RESULTS: Our results demonstrated that expression of TRIF and MYD88 were significantly elevated in PBMCs isolated from depressed patients when compared to healthy subjects. CONCLUSIONS: Based on the current results, it seems that chronic inflammation in depressed patients correlates to the over expression of TRIF and MYD88 genes. Our results show a possible link between the reported increases of chronic inflammation in depressed individuals with unbalanced expression of genes that regulate immunity.

Cell culture, sex determination and single cell cloning of ovine transgenic satellite cells in vitro.

BACKGROUND: This study was performed to describe the basic methods to isolate and culture of primary satellite cells (PSCs) obtained from 50 to 60-day-old sheep fetuses, single cell cloning of transfected PSCs and sexing of ovine PSCs based on the ZFY/ZFX, amelogenin and high-motility-group (HMG) box sequences. RESULTS: Three-step enzymatic digestion method increased PSCs isolation from tissue and reduced the damage of cells during long time incubation with enzymes. The results of cloning showed that the 103 and 81 clones (from a total of 184 clones) were derived from feeder and bFGF treatment, respectively. The overall sexing efficiency in the present study was 100%. Southern blot results of sex determination were in complete agreement with PCR-amplified bands which confirmed that the HMG box of SRY gene amplified from the ovine genome and that was specific for male. CONCLUSIONS: We successfully isolated and cultured sheep primary satellite cells via mechanical and enzymatic disaggregation. Our finding demonstrated that use of feeder and addition of bFGF to the culture medium improved cloning efficiency. The results of sex detection demonstrated that these methods can be applied to detect the sex of primary satellite cells and to determine the sex of sheep embryo prior to produce sheep embryos by somatic cell nuclear transfer technique in vitro. Nevertheless, our findings suggested that sex determination of satellite cells base on amelogenin sequence can be accurate, relatively simple, rapid, and inexpensive.

Expression of CC chemokines CCL2, CCL5, and CCL11 is associated with duration of disease and complications in type-1 diabetes: a study on Iranian diabetic patients.

BACKGROUND: Type-1 diabetes (T1D) is defined as a heterogeneous autoimmune disease. Immune system related factors are important in the pathogenesis of T1D. Chemokines are important factors in the pathogenesis of several autoimmune diseases, including T1D. They are potent chemotactic cytokines with various functions such as maturation, trafficking of leukocytes, angiogenesis, and homing of stem cells. Therefore, the current study was aimed to examine whether expression of CC chemokines CCL2, CCL5, and CXCL11 is associated with disease duration and complications in Iranian T1D patients. METHODS: In this experimental study, blood samples were collected from 108 T1D patients and 189 healthy controls in EDTA pre-coated tubes. The serum levels of CC chemokines were measured by ELISA. Demographic data were also collected along with experimental examinations in a questionnaire which was designed specifically for this study. RESULTS: Results of the present study demonstrated that the expression of CCL2 was decreased while CCL5 and CCL11 were increased in T1D patients in comparison to controls. These results demonstrated that CCL2, CCL5, and CCL1 were elevated in T1D patients with duration of disease. Again, our findings demonstrated that CCL2, CCL5, and CCL11 were elevated in T1D patients with age. But there was not a significant difference between circulating level of CC chemokines studied in T1D patients regarding their gender and they have followed a similar pattern of expression in both genders. Our findings also showed that all three CC chemokines were elevated in TID patients suffering from diabetes complications. CONCLUSIONS: According to the results of our study, elevated levels of CCL5 and CCL11 are in parallel with decreased level of CCL2 and are useful tools in the differential diagnosis of T1D from other types of metabolic disorders. Elevated levels of these CC chemokines probably could be implicated as predictive factors for occurrence of T1D complications. These results may also re-emphasize the prominent therapeutic role(s) of these CC chemokines in control of either T1D or its associated complications.

Defective NF-kB transcription factor as the mediator of inflammatory responses: a study on depressed Iranian medical students.

BACKGROUND: NF-kB is a transcription factor that is a downstream target of several cell signaling systems including TLRs. Defective expression of the molecule can lead to inappropriate immune responses. Previous studies revealed that depression can affect immune responses, but its molecular mechanisms are yet to be fully understood. Thus, the main aim of this study was to identify if mRNA levels of NF-kB are changed in the PBMCs isolated from Iranian depressed medical students. METHODS: This cross-sectional study was done on 38 Iranian depressed medical students and 43 healthy students as a control group. The mRNA levels of NF-kB were assessed in parallel with beta-actin (as the housekeeping gene) using Real-Time PCR technique. RESULTS: Our results showed that mRNA levels of NF-kB were significantly decreased in isolated PBMCs from depressed patients compared to healthy controls. CONCLUSIONS: According to the results obtained in the present study, it seems that depressed patients are unable to appropriately express NF-kB at mRNA levels which may in turn lead to defective molecule expression.

The regulatory impacts of Morus Alba leaf extract on some enzymes involved in glucose metabolism pathways in diabetic rat liver.

BACKGROUND: Many traditional therapies have been proposed as alternative regimens for treatment of diabetes mellitus. The Morus Alba (MA) leaf is a natural therapeutic compound which is shown to have antidiabetic properties. The aim of the present study was to determine whether MA leaf extract is capable of regulating liver enzymes that are involved in glucose metabolism pathways in normal and Streptozotocin (STZ)-induced diabetic rats. METHODS: Forty healthy adult male Wistar rats (eight weeks old) weighing about 250 +/- 10 g were taken for this experiment. The rats were divided into 4 groups with 10 rats in each group and treated through a gavage tube for a period of two months as follows: group I: non diabetic control rats with distilled water; group II: non diabetic rats with 1.0 g/kg per day; group III: diabetic control rats with distilled water and group IV: diabetic rats with MA 1.0 g/kg per day. At the end of the 8th week, serum glucose, insulin and hepatic glucokinase activity were measured using standard methods and compared between diabetic and healthy rats. We also assessed the expression of phosphofructokinase-1 enzyme at the level of mRNA, using a Real Time-PCR method. RESULTS: Findings of the present study demonstrated that MA leaf extract can significantly increase liver glucokinase activity and serum insulin levels in diabetic rats (p < 0.05). It also significantly attenuated the serum glucose level in rats compared to the control groups (p < 0.05). Also, the body weight of diabetic rats was significantly (p < 0.05) decreased as compared to their initial weight. However, the body weights of diabetic rats treated with MA increased in the same way as normal control rats. CONCLUSIONS: The present findings suggest that the antihyperglycemic action of MA is mediated by increasing liver glucokinase activity and serum insulin level. These results are additional, definite evidence supporting MA as traditional medicine for diabetic patients.

Increased circulating levels of CXC chemokines is correlated with duration and complications of the disease in type-1 diabetes: a study on Iranian diabetic patients.

BACKGROUND: Type-1 diabetes (T1D) is characterized as a heterogenous autoimmune disease. Immune system factors are important in the pathogenesis of T1D. Chemokines as crucial members of the immune system are key factors in the pathogenesis of several autoimmune diseases, including T1D. They are potent chemotactic cytokines with various functions varied from maturation, trafficking of leukocytes, to angiogenesis, angiostasis, and homing of stem cells. Therefore, the current study was aimed to examine if the expression of pro-angiogenic CXC chemokines like CXCL1 and anti-angiogenic chemochines such as CXCL9 are associated with duration and complications of T1D in Iranian diabetic patients. METHODS: In this experimental study, blood samples were collected from 209 T1D patients and 189 healthy controls. The serum levels of CXCL1 and CXCL9 were measured by ELISA. Demographic data were also collected on a questionnaire which was designed specifically for this study. RESULTS: Increased plasma levels of chemokines studied (CXCL1 and CXCL9) were observed in T1D patients compared to controls. Current findings also demonstrated that there was a close association between chemokines and complications of T1D and chemokines were elevated in T1D patients suffering complications. CONCLUSIONS: Our results probably suggest that the serum levels of CXCL1 and CXCL9 play important roles in T1D pathogenesis. It is also worth noting that these factors are useful prognostic and/or diagnostic biological markers in T1D patients.

Is the IL-10 promoter polymorphism at position -592 associated with immune system-related diseases?

Immune responses are the main causes of immune system-related diseases such as hypersensitivities and autoimmunity. It has also been established that cytokines play key roles in the regulation of immune responses which have been shown to be important in the pathogenesis of the diseases. IL-10, the main anti-inflammatory cytokine, is produced by several immune cells such as T regulatory and Th2 lymphocytes, activated macrophages, B regulatory lymphocytes as well as other cell types. It plays a key role in the regulation of immune responses after microbe elimination (homeostasis) and against self-antigens to prevent hypersensitivity and autoimmune diseases, respectively. Studies showed that a single nucleotide polymorphism (SNP) at the -592 position of IL-10 is associated with its regulation of expression. This review addresses the recent information regarding the association of the polymorphism at position -592 of IL-10 with immune-related diseases including type 2 diabetes with and without nephropathy, multiple sclerosis, and asthma with an emphasize on Iranian patients.