Validation of a Deep Learning Model for Detecting Chest Pathologies from Digital Chest Radiographs.

Purpose: Manual interpretation of chest radiographs is a challenging task and is prone to errors. An automated system capable of categorizing chest radiographs based on the pathologies identified could aid in the timely and efficient diagnosis of chest pathologies. Method: For this retrospective study, 4476 chest radiographs were collected between January and April 2021 from two tertiary care hospitals. Three expert radiologists established the ground truth, and all radiographs were analyzed using a deep-learning AI model to detect suspicious ROIs in the lungs, pleura, and cardiac regions. Three test readers (different from the radiologists who established the ground truth) independently reviewed all radiographs in two sessions (unaided and AI-aided mode) with a washout period of one month. Results: The model demonstrated an aggregate AUROC of 91.2% and a sensitivity of 88.4% in detecting suspicious ROIs in the lungs, pleura, and cardiac regions. These results outperform unaided human readers, who achieved an aggregate AUROC of 84.2% and sensitivity of 74.5% for the same task. When using AI, the aided readers obtained an aggregate AUROC of 87.9% and a sensitivity of 85.1%. The average time taken by the test readers to read a chest radiograph decreased by 21% (p < 0.01) when using AI. Conclusion: The model outperformed all three human readers and demonstrated high AUROC and sensitivity across two independent datasets. When compared to unaided interpretations, AI-aided interpretations were associated with significant improvements in reader performance and chest radiograph interpretation time.

Levels of Autonomous Radiology.

Radiology, being one of the younger disciplines of medicine with a history of just over a century, has witnessed tremendous technological advancements and has revolutionized the way we practice medicine today. In the last few decades, medical imaging modalities have generated seismic amounts of medical data. The development and adoption of artificial intelligence applications using this data will lead to the next phase of evolution in radiology. It will include automating laborious manual tasks such as annotations, report generation, etc, along with the initial radiological assessment of patients and imaging features to aid radiologists in their diagnostic and treatment planning workflow. We propose a level-wise classification for the progression of automation in radiology, explaining artificial intelligence assistance at each level with the corresponding challenges and solutions. We hope that such discussions can help us address challenges in a structured way and take the necessary steps to ensure the smooth adoption of new technologies in radiology.

A deep learning approach for automated diagnosis of pulmonary embolism on computed tomographic pulmonary angiography.

BACKGROUND: Computed tomographic pulmonary angiography (CTPA) is the diagnostic standard for confirming pulmonary embolism (PE). Since PE is a life-threatening condition, early diagnosis and treatment are critical to avoid PE-associated morbidity and mortality. However, PE remains subject to misdiagnosis. METHODS: We retrospectively identified 251 CTPAs performed at a tertiary care hospital between January 2018 to January 2021. The scans were classified as positive (n = 55) and negative (n = 196) for PE based on the annotations made by board-certified radiologists. A fully anonymized CT slice served as input for the detection of PE by the 2D segmentation model comprising U-Net architecture with Xception encoder. The diagnostic performance of the model was calculated at both the scan and the slice levels. RESULTS: The model correctly identified 44 out of 55 scans as positive for PE and 146 out of 196 scans as negative for PE with a sensitivity of 0.80 [95% CI 0.68, 0.89], a specificity of 0.74 [95% CI 0.68, 0.80], and an accuracy of 0.76 [95% CI 0.70, 0.81]. On slice level, 4817 out of 5183 slices were marked as positive for the presence of emboli with a specificity of 0.89 [95% CI 0.88, 0.89], a sensitivity of 0.93 [95% CI 0.92, 0.94], and an accuracy of 0.89 [95% CI 0.887, 0.890]. The model also achieved an AUROC of 0.85 [0.78, 0.90] and 0.94 [0.936, 0.941] at scan level and slice level, respectively for the detection of PE. CONCLUSION: The development of an AI model and its use for the identification of pulmonary embolism will support healthcare workers by reducing the rate of missed findings and minimizing the time required to screen the scans.

Observer performance evaluation of the feasibility of a deep learning model to detect cardiomegaly on chest radiographs.

Background: Cardiothoracic ratio (CTR) is the ratio of the diameter of the heart to the diameter of the thorax. An abnormal CTR (>0.55) is often an indicator of an underlying pathological condition. The accurate prediction of an abnormal CTR chest X-rays (CXRs) aids in the early diagnosis of clinical conditions. Purpose: We propose a deep learning (DL)-based model for automatic CTR calculation to assist radiologists with rapid diagnosis of cardiomegaly and thus optimise the radiology flow. Material and Methods: The study population included 1012 posteroanterior CXRs from a single institution. The Attention U-Net DL architecture was used for the automatic calculation of CTR. An observer performance test was conducted to assess the radiologist's performance in diagnosing cardiomegaly with and without artificial intelligence assistance. Results: U-Net model exhibited a sensitivity of 0.80 [95% CI: 0.75, 0.85], specificity >99%, precision of 0.99 [95% CI: 0.98, 1], and a F1 score of 0.88 [95% CI: 0.85, 0.91]. Furthermore, the sensitivity of the reviewing radiologist in identifying cardiomegaly increased from 40.50% to 88.4% when aided by the AI-generated CTR. Conclusion: Our segmentation-based AI model demonstrated high specificity (>99%) and sensitivity (80%) for CTR calculation. The performance of the radiologist on the observer performance test improved significantly with provision of AI assistance. A DL-based segmentation model for rapid quantification of CTR can therefore have significant potential to be used in clinical workflows by reducing radiologists' burden and alerting to an abnormal enlarged heart early on.

RING finger protein TOPORS modulates the expression of tumor suppressor SMAR1 in colorectal cancer via the TLR4-TRIF pathway.

TOP1-binding arginine/serine-rich protein (TOPORS), a really interesting new gene finger protein, has the ability to bind to a palindromic consensus DNA sequence that enables it to function as a potential transcriptional regulator. However, its role in regulating the transcription of cancer-associated genes is yet to be explored. As Toll-like receptor 4 (TLR4) agonists are known to regress solid tumors, we observed that lipopolysaccharide (LPS) induces TOPORS via a TLR4-TIR domain-containing adapter-inducing interferon-ß-dependent pathway, which in turn modulates the transcription of tumor suppressor scaffold/matrix attachment region-binding protein 1 (SMAR1, also known as BANP). ChIP analysis showed that TOPORS binds on the SMAR1 promoter and its occupancy increases upon LPS treatment. A previous study from our laboratory revealed that SMAR1 acts as a repressor of signal transducer and activator of transcription 3 (STAT3) transcription. Tumor growth, as well as tumor-associated macrophage polarization, depends on the status of the STAT1:STAT3 ratio. LPS-induced SMAR1 expression decreases STAT3 expression and also skews the macrophage polarization toward M1 phenotype. In contrast, LPS failed to polarize tumor-associated macrophages to M1 phenotype in a SMAR1-silenced condition, which shows the involvement of SMAR1 in dictating the fate of colorectal cancer progression. Identification of the molecular mechanism behind LPS-mediated tumor regression would be crucial for designing cancer treatment strategies involving bacterial components.

Chromatin remodeling protein SMAR1 regulates adipogenesis by modulating the expression of PPARγ.

Adipogenesis is described as the process of conversion of pre-adipocytes into differentiated lipid-laden adipocytes. Adipogenesis is known to be regulated by a myriad of transcription factors and co-regulators. However, there is a dearth of information regarding the mechanisms that regulate these transcription factors and hence control adipogenesis. PPARγ is the master transcriptional regulator of adipogenesis and its expression is essential for adipocyte differentiation. Herein, we identified that scaffold/matrix attachment region-binding protein 1 (SMAR1) negatively regulates adipogenesis. We observed that SMAR1 gets downregulated during adipocyte differentiation and knockdown of SMAR1 promotes lipid accumulation and adipocyte differentiation. Mechanistically, we have shown that SMAR1 suppresses PPARγ through recruitment of the HDAC1/mSin3a repressor complex to the PPARγ promoter. We further identified cell division cycle 20 (cdc20) mediated proteasomal degradation of SMAR1 during adipogenesis. Moreover, knockdown of cdc20 resulted in stabilization of SMAR1 and a reduction in adipocyte differentiation. Taken together, our observations suggest that SMAR1 functions as a negative regulator of adipogenesis by inhibiting PPARγ expression in differentiating adipocytes.

Tumor suppressor SMAR1 regulates PKM alternative splicing by HDAC6-mediated deacetylation of PTBP1.

BACKGROUND: Highly proliferating cancer cells exhibit the Warburg effect by regulation of PKM alternative splicing and promoting the expression of PKM2. Majority of the alternative splicing events are known to occur in the nuclear matrix where various MARBPs actively participate in the alternative splicing events. SMAR1, being a MARBP and an important tumor suppressor, is known to regulate the splicing of various cancer-associated genes. This study focuses on the regulation of PKM alternative splicing and inhibition of the Warburg effect by SMAR1. METHODS: Immunohistochemistry was performed in breast cancer patient samples to establish the correlation between SMAR1 and PKM isoform expression. Further, expression of PKM isoforms upon modulation in SMAR1 expression in breast cancer cell lines was quantified by qRT-PCR and western blot. The acetylation status of PTBP1 was estimated by immunoprecipitation along with its enrichment on PKM pre-mRNA by CLIP in SMAR1 knockdown conditions. The role of SMAR1 in tumor metabolism and tumorigenesis was explored by in vitro enzymatic assays and functional assays upon SMAR1 knockdown. Besides, in vivo tumor formation by injecting adeno-SMAR1-transduced MDA-MB-231 cells in NOD/SCID mice was performed. RESULTS: The expression profile of SMAR1 and PKM isoforms in breast cancer patients revealed that SMAR1 has an inverse correlation with PKM2 and a positive correlation with PKM1. Further quantitative PKM isoform expression upon modulation in SMAR1 expression also reflects that SMAR1 promotes the expression of PKM1 over tumorigenic isoform PKM2. SMAR1 deacetylates PTBP1 via recruitment of HDAC6 resulting in reduced enrichment of PTBP1 on PKM pre-mRNA. SMAR1 inhibits the Warburg effect, tumorigenic potential of cancer cells, and in vivo tumor generation in a PKM2-dependent manner. CONCLUSIONS: SMAR1 regulates PKM alternative splicing by causing HDAC6-dependent deacetylation of PTBP1, resulting in reduced enrichment of PTBP1 on PKM pre-mRNA. Additionally, SMAR1 suppresses glucose utilization and lactate production via repression of PKM2 expression. This suggests that tumor suppressor SMAR1 inhibits tumor cell metabolism and tumorigenic properties of cancer cells via regulation of PKM alternative splicing.

Epigenetic Regulation of Adipogenesis in Development of Metabolic Syndrome.

Obesity is one of the biggest public health concerns identified by an increase in adipose tissue mass as a result of adipocyte hypertrophy and hyperplasia. Pertaining to the importance of adipose tissue in various biological processes, any alteration in its function results in impaired metabolic health. In this review, we discuss how adipose tissue maintains the metabolic health through secretion of various adipokines and inflammatory mediators and how its dysfunction leads to the development of severe metabolic disorders and influences cancer progression. Impairment in the adipocyte function occurs due to individuals' genetics and/or environmental factor(s) that largely affect the epigenetic profile leading to altered gene expression and onset of obesity in adults. Moreover, several crucial aspects of adipose biology, including the regulation of different transcription factors, are controlled by epigenetic events. Therefore, understanding the intricacies of adipogenesis is crucial for recognizing its relevance in underlying disease conditions and identifying the therapeutic interventions for obesity and metabolic syndrome.

SMAR1 favors immunosurveillance of cancer cells by modulating calnexin and MHC I expression.

Down-regulation or loss of MHC class I expression is a major mechanism used by cancer cells to evade immunosurveillance and increase their oncogenic potential. MHC I mediated antigen presentation is a complex regulatory process, controlled by antigen processing machinery (APM) dictating immune response. Transcriptional regulation of the APM that can modulate gene expression profile and their correlation to MHC I mediated antigen presentation in cancer cells remain enigmatic. Here, we reveal that Scaffold/Matrix-Associated Region 1- binding protein (SMAR1), positively regulates MHC I surface expression by down-regulating calnexin, an important component of antigen processing machinery (APM) in cancer cells. SMAR1, a bonafide MAR binding protein acts as a transcriptional repressor of several oncogenes. It is down-regulated in higher grades of cancers either through proteasomal degradation or through loss of heterozygosity (LOH) at the Chr.16q24.3 locus where the human homolog of SMAR1 (BANP) has been mapped. It binds to a short MAR region of the calnexin promoter forming a repressor complex in association with GATA2 and HDAC1. A reverse correlation between SMAR1 and calnexin was thus observed in SMAR1-LOH cells and also in tissues from breast cancer patients. To further extrapolate our findings, influenza A (H1N1) virus infection assay was performed. Upon viral infection, the levels of SMAR1 significantly increased resulting in reduced calnexin expression and increased MHC I presentation. Taken together, our observations establish that increased expression of SMAR1 in cancers can positively regulate MHC I surface expression thereby leading to higher chances of tumor regression and elimination of cancer cells.

Wnt Signaling and Its Significance Within the Tumor Microenvironment: Novel Therapeutic Insights.

Wnt signaling is one of the central mechanisms regulating tissue morphogenesis during embryogenesis and repair. The pivot of this signaling cascade is the Wnt ligand, which binds to receptors belonging to the Frizzled family or the ROR1/ROR2 and RYK family. This interaction governs the downstream signaling cascade (canonical/non-canonical), ultimately extending its effect on the cellular cytoskeleton, transcriptional control of proliferation and differentiation, and organelle dynamics. Anomalous Wnt signaling has been associated with several cancers, the most prominent ones being colorectal, breast, lung, oral, cervical, and hematopoietic malignancies. It extends its effect on tumorigenesis by modulating the tumor microenvironment via fine crosstalk between transformed cells and infiltrating immune cells, such as leukocytes. This review is an attempt to highlight the latest developments in the understanding of Wnt signaling in the context of tumors and their microenvironment. A dynamic process known as immunoediting governs the fate of tumor progression based on the correlation of various signaling pathways in the tumor microenvironment and immune cells. Cancer cells also undergo a series of mutations in the tumor suppressor gene, which favors tumorigenesis. Wnt signaling, and its crosstalk with various immune cells, has both negative as well as positive effects on tumor progression. On one hand, it helps in the maintenance and renewal of the leucocytes. On the other hand, it promotes immune tolerance, limiting the antitumor response. Wnt signaling also plays a role in epithelial-mesenchymal transition (EMT), thereby promoting the maintenance of Cancer Stem Cells (CSCs). Furthermore, we have summarized the ongoing strategies used to target aberrant Wnt signaling as a novel therapeutic intervention to combat various cancers and their limitations.

An Assessment of a Public Health Initiative of Homeopathy for Primary Teething.

BACKGROUND: During primary teething, children suffer from running nose, mild fever, diarrhoea and other mild irritations and inflammations. A public health programme, 'Homoeopathy for the Healthy Child', was undertaken on a pilot basis focusing on promotion of healthy teething by provision of home-based care through six pre-identified homeopathic medicines for complaints commonly observed during primary teething. This article assesses the feasibility of this programme and reports the impact of this initiative on teething profile in children and episodes of diarrhoea and upper respiratory tract infection (URTI). MATERIALS AND METHODS: Accredited Social Health Activists (ASHAs) were trained in child care and usage of a kit comprising six medicines, namely Calcarea phosphoricum 6X (CP), Ferrum phosphoricum 3X, Magnesium phosphoricum 6X, Belladonna 30C, Chamomilla 30C and Podophyllum 30C. Calcarea phosphoricum was given regularly to each participating child from 6 months to 1 year of age. Home-based care for diarrhoea, URTI and mild fever was provided by ASHAs using the other five medicines in the kit. Dentition pattern and diarrhoea/URTI episodes were recorded over a period of the next 12 months. RESULTS: Eleven thousand four-hundred and twenty-six children were followed up regularly. Amongst those who enrolled at 6-7 months, a larger proportion of children were approaching expected teething in successive months as compared with children enrolled at 12 months, thus indicating that teething delays, if any, were overcome during this period. Incidence of diarrhoea and URTI showed decrease in the months after enrolment. Children responded favourably to the medicines given by ASHAs at the time of diarrhoea/URTI episodes, and ASHAs expressed satisfaction with the programme. CONCLUSION: An approach with regular use of CP and home-based care with homeopathy through health workers for common problems in teething children is acceptable to the community and enhances outreach of services to the public at large. Observations in terms of the healthy teething period may be further validated through studies of homeopathy with suitable comparator group.

IL-4 promotes asymmetric dimethylarginine accumulation, oxo-nitrative stress, and hypoxic response-induced mitochondrial loss in airway epithelial cells.

BACKGROUND: Obesity is known to increase asthma risk and severity. Increased levels of asymmetric dimethylarginine (ADMA), an endogenous nitric oxide synthase inhibitor, are associated with mitochondrial toxicity, asthma, and metabolic syndrome. IL-4 upregulates the expression of protein arginine methyltransferases, which are essential for ADMA formation. Importantly, cross-talk between IL-4, ADMA, and mitochondrial dysfunction could explain how obesity and IL-4 can synergize to exacerbate allergic inflammation. OBJECTIVE: We sought to investigate how IL-4, a key asthma-associated cytokine, can influence ADMA-related effects on lungs. METHODS: BEAS2B (bronchial epithelial) cells were treated with IL-4 followed by ADMA and investigated for oxo-nitrative stress and resultant mitochondrial toxicity after 48 hours by using flow cytometry, confocal imaging, immunoblotting, and fluorimetric assays. RESULTS: IL-4-induced mitotoxicity in BEAS2B cells was significantly higher in the presence of exogenous ADMA. IL-4 treatment led to proteolytic degradation of dimethylarginine dimethylaminohydrolase 2, which catabolizes ADMA. IL-4 pretreatment was associated with increased intracellular ADMA accumulation and increased ADMA-induced mitotoxicity. Airway epithelial cells treated with IL-4 followed by ADMA showed exaggerated oxo-nitrative stress and potent induction of the cellular hypoxic response, despite normoxic conditions. The hypoxic response was associated with reduced mitochondrial function but was reversible by overexpression of the mitochondrial biogenesis factor, mitochondrial transcription factor A. CONCLUSION: We conclude that IL-4 promotes intracellular ADMA accumulation, leading to mitochondrial loss through oxo-nitrative stress and hypoxic response. This provides a novel understanding of how obesity, with high ADMA levels, and asthma, with high IL-4 levels, might potentiate each other and highlights the potential of mitochondrial-targeted therapeutics in obese subjects with asthma.