ColoRectalCADx: Expeditious Recognition of Colorectal Cancer with Integrated Convolutional Neural Networks and Visual Explanations Using Mixed Dataset Evidence.

Colorectal cancer typically affects the gastrointestinal tract within the human body. Colonoscopy is one of the most accurate methods of detecting cancer. The current system facilitates the identification of cancer by computer-assisted diagnosis (CADx) systems with a limited number of deep learning methods. It does not imply the depiction of mixed datasets for the functioning of the system. The proposed system, called ColoRectalCADx, is supported by deep learning (DL) models suitable for cancer research. The CADx system comprises five stages: convolutional neural networks (CNN), support vector machine (SVM), long short-term memory (LSTM), visual explanation such as gradient-weighted class activation mapping (Grad-CAM), and semantic segmentation phases. Here, the key components of the CADx system are equipped with 9 individual and 12 integrated CNNs, implying that the system consists mainly of investigational experiments with a total of 21 CNNs. In the subsequent phase, the CADx has a combination of CNNs of concatenated transfer learning functions associated with the machine SVM classification. Additional classification is applied to ensure effective transfer of results from CNN to LSTM. The system is mainly made up of a combination of CVC Clinic DB, Kvasir2, and Hyper Kvasir input as a mixed dataset. After CNN and LSTM, in advanced stage, malignancies are detected by using a better polyp recognition technique with Grad-CAM and semantic segmentation using U-Net. CADx results have been stored on Google Cloud for record retention. In these experiments, among all the CNNs, the individual CNN DenseNet-201 (87.1% training and 84.7% testing accuracies) and the integrated CNN ADaDR-22 (84.61% training and 82.17% testing accuracies) were the most efficient for cancer detection with the CNN+LSTM model. ColoRectalCADx accurately identifies cancer through individual CNN DesnseNet-201 and integrated CNN ADaDR-22. In Grad-CAM's visual explanations, CNN DenseNet-201 displays precise visualization of polyps, and CNN U-Net provides precise malignant polyps.

Automated system for classification of COVID-19 infection from lung CT images based on machine learning and deep learning techniques.

The objectives of our proposed study were as follows: First objective is to segment the CT images using a k-means clustering algorithm for extracting the region of interest and to extract textural features using gray level co-occurrence matrix (GLCM). Second objective is to implement machine learning classifiers such as Naïve bayes, bagging and Reptree to classify the images into two image classes namely COVID and non-COVID and to compare the performance of the three pre-trained CNN models such as AlexNet, ResNet50 and SqueezeNet with that of the proposed machine learning classifiers. Our dataset consists of 100 COVID and non-COVID images which are pre-processed and segmented with our proposed algorithm. Following the feature extraction process, three machine learning classifiers (Naive Bayes, Bagging, and REPTree) were used to classify the normal and covid patients. We had implemented the three pre-trained CNN models such as AlexNet, ResNet50 and SqueezeNet for comparing their performance with machine learning classifiers. In machine learning, the Naive Bayes classifier achieved the highest accuracy of 97%, whereas the ResNet50 CNN model attained the highest accuracy of 99%. Hence the deep learning networks outperformed well compared to the machine learning techniques in the classification of Covid-19 images.

Non-invasive technique to detect diabetic retinopathy based on Electrooculography signal using machine learning classifiers.

Single-channel Electrooculogram (EOG) is proposed for detecting diabetic retinopathy. The Corneal-retinal potential of the eyes plays a vital role in the acquisition of Electrooculography. Diabetes is the most prevalent disease and for one out of three people with diabetes above 40 years, diabetic retinopathy occurs. It is necessary for the early detection of diabetic retinopathy as it is one of the primary reasons for blindness in the population. The potential difference between cornea and retina leads to the acquisition of EOG signal. The proposed study aims to design a low-cost miniaturized hardware circuit to obtain EOG signal using second order filters without compromising in accuracy of the outcome signal and to classify the signal into normal and diabetic retinopathy subjects by extracting the statistical features like kurtosis, mean, median absolute deviation, standard deviation, and range from software filtered EOG signal. Among the classifiers used, Support vector machine (SVM) shows a higher accuracy of 93.33%. The sensitivity, specificity and Area Under Curve (AUC) values of SVM are 96.43%, 90.625%, 0.93% is considered as more favorable outcome for the proposed method and it supports the developed prototype and processing methodology. The novelty of the research is based on proposing and exploring a non-invasive methodology for Diabetic retinopathy diagnosis based on EOG signal. Thus, the designed hardware is simple in operation and cost effective, provides an affordable and non-invasive diagnostic tool for diabetic retinopathy patients.

Non-invasive thyroid detection based on electroglottogram signal using machine learning classifiers.

Thyroid is a butterfly shaped gland located in the neck region. Hormones are secreted by the thyroid gland that is responsible for various functions that maintain metabolism of the body. The variance in secretion of the hormones causes disorders such as Hyperthyroidism or Hypothyroidism. Electroglottography signal is a bio signal which represents the impedance that exist between the glottis regions. The study aims at design and development of an hardware circuit for the acquisition of Electroglottogram signal from normal and thyroid subjects is proposed followed by feature extraction from the acquired bio signal is performed. Further, machine learning classifiers were used to classify the normal and thyroid individuals. This modality of acquisition is non-invasive. Performance evaluation is done by testing various classifiers to study the accuracy. The classifiers tested were Random Forest, Random Tree, Bayes Net, Multilayer Perceptron, Simple Logistic classifier, and One-R classifier. Classifiers such as Random Forest, Random Tree, and Multilayer Perceptron showed high accuracy. The accuracy estimated by these classifiers was tested and its ROC curves with AUC scores were derived. The highest accuracy was reported for Simple Logistic classifier which was about 95.1%. Random Forest and Random Tree reported 93.5% and 91.9% respectively. Similarly, Multilayer Perceptron and Bayes Net gave 93.5% and 91.9%. The One-R classifier algorithm reported the lowest accuracy of 90.3% among the studied classifier algorithms. The ROC-AUC score for the classifiers were also reported to be more than 0.9 which is considered more promising and supports the acquisition and processing methodology. Hence the proposed technique can be efficiently used to diagnose thyroid non-invasively.

Study of a retinal layer model to generate a spike waveform for a color deficient and strabismus individual.

The physiological modeling of retinal layers to provide an insight into how the incoming image is converted into its equivalent spike train that can be decoded by the human brain is a key issue. Most of the retinal layer models concentrate mainly on image compression, edge detection and image reconstruction. A retinal layer model to generate spike waveform corresponding to the visual information is not covered much in the literature. The aim of this study was to develop a mathematical model of retinal layers that has complex neural structures, that can detect the incoming signal and transform the signal into the equivalent spike train. The proposed retinal layer model includes a photoreceptor, an outer plexiform (OPL), an inner plexiform (IPL) and ganglion cell layers exhibiting the properties of compression, luminance and spatial temporal filtering in the processing of visual information. The photoreceptor layer enhances the contrast visibility in the dark region and maintains the same in the bright regions. The OPL is modeled to enhance the contour of the image. The finer detail of the image is extracted by mathematically modeling the IPL. The ganglion cell layer is modeled using the Hodgkin-Huxley model to generate the spike train for the incoming information. The spike train was generated for color deficient individuals namely protanopia, deuteranopia, tritanopia and for individuals suffering from night blindness. Simulation results showed a spike train was generated only for a certain threshold stimulus value. The differences in spike pattern for a normal and visually impaired individual were studied. This may lead to a methodology for earlier diagnosis.

HATRICK-Synchronous Triple Primary Tumors of Thyroid.

Multiple primary cancer (MPC) has an incidence of 1.8% and is defined as having two or more cancers in a single patient. Synchronous tumors are defined as ≥ 2 primary tumors occurring within 6 months of diagnosis of the first primary tumor. We present a case of a 27-year-old female patient who presented with a painless, gradually progressive right-sided neck swelling for the last 1 year with no systemic complaints. Examination revealed a 4 × 3-cm, firm, smooth surfaced swelling on right lobe of thyroid. USG neck showed a hypoechoic solid nodule on the right lobe and the left lobe was normal. FNAC showed features of adenomatous colloid nodule, Bethesda II. Right hemithyroidectomy specimen revealed evidence of triple tumors-not otherwise-specified (NOS) tumor, papillary carcinoma thyroid, and medullary carcinoma thyroid, which was confirmed with positivity on IHC with synaptophysin, CEA, and chromogranin. Concurrent appearance of NOS, PTC, and medullary carcinoma thyroid in the very same patient is extremely rare and has not been previously reported in the English literature.

Computer-based automated analysis of X-ray and thermal imaging of knee region in evaluation of rheumatoid arthritis.

The aim and objectives of the study are as follows: (1) to perform automated segmentation of knee X-ray images using fast greedy snake algorithm and feature extraction using gray level co-occurrence matrix method, (2) to implement automated segmentation of knee thermal image using RGB segmentation method and (3) to compare the features extracted from the segmented knee region of X-ray and thermal images in rheumatoid arthritis patients using a biochemical method as standard. In all, 30 rheumatoid arthritis patients and 30 age- and sex-matched healthy volunteers were included in the study. X-ray and thermography images of knee regions were acquired, and biochemical tests were carried out subsequently. The X-ray images were segmented using fast greedy snake algorithm, and feature extractions were performed using gray level co-occurrence matrix method. The thermal image was segmented using RGB-based segmentation method and statistical features were extracted. Statistical features extracted after segmentation from X-ray and thermal imaging of knee region were correlated with the standard biochemical parameters. The erythrocyte sedimentation rate shows statistically significant correlations (p < 0.01) with the X-ray parameters such as joint space width and % combined cortical thickness. The skin surface temperature measured from knee region of thermal imaging was highly correlated with erythrocyte sedimentation rate. Among all the extracted features namely mean, variance, energy, homogeneity and difference entropy depict statistically significant percentage differences between the rheumatoid arthritis and healthy subjects. From this study, it was observed that thermal infrared imaging technique serves as a potential tool in the evaluation of rheumatoid arthritis at an earlier stage compared to radiography. Hence, it was predicted that thermal imaging method has a competency in the diagnosis of rheumatoid arthritis by automated segmentation methods.

Physiological modeling for detecting degree of perception of a color-deficient person.

Physiological modeling of retina plays a vital role in the development of high-performance image processing methods to produce better visual perception. People with normal vision have an ability to discern different colors. The situation is different in the case of people with color blindness. The aim of this work is to develop a human visual system model for detecting the level of perception of people with red, green and blue deficiency by considering properties like luminance, spatial and temporal frequencies. Simulation results show that in the photoreceptor, outer plexiform and inner plexiform layers, the energy and intensity level of the red, green and blue component for a normal person is proved to be significantly higher than for dichromats. The proposed method explains with appropriate results that red and blue color blindness people could not perceive red and blue color completely.

DNA adducts-chemical addons.

DNA adduct is a piece of DNA covalently bond to a chemical (safrole, benzopyrenediol epoxide, acetaldehyde). This process could be the start of a cancerous cell. When a chemical binds to DNA, it gets damaged resulting in abnormal replication. This could be the start of a mutation and without proper DNA repair, this can lead to cancer. It is this chemical that binds with the DNA is our prime area of concern. Instead of performing the whole body analysis for diagnosing cancer, this test could be carried out for early detection of cancer. When scanning tunneling microscope is used, the DNA results can be obtained earlier. DNA adducts in scientific experiments are used as biomarkers.

Cutaneous amyloidosis and insulin with coexistence of acanthosis nigricans.

Skin is one of the important organs affected by amyloidosis which is characterized by extracellular deposition of fibrillary proteins having homogenous, eosinophilic on routine staining with distinct tinctorial properties. Nodular cutaneous amyloidosis is rare and may affect dermis, subcutis and also vascular walls. Nodular amyloid deposits in the deeper dermis occurring at the site of insulin injection are a rare observation, which is described here. This description indicates that cutaneous amyloidosis may be associated with local subcutaneous injections of insulin and may clinically mimic a neoplasm or lipodystrophic lesion.

A study of histopathologic spectrum of nodular hidradenoma.

BACKGROUND: Hidradenomas are rare benign adnexal neoplasms that encompass a morphological gamut with a range of differentiation. As a consequence, there is a great likelihood of being mistaken for other primary and metastatic tumors. Though conventionally regarded as eccrine, they have been reclassified into eccrine and apocrine types. OBJECTIVE: This study aims to document the histological spectrum of nodular hidradenomas, with particular reference to categorizing them into eccrine or apocrine tumors. RESULTS: A total of 15 cases with features of nodular hidradenoma with their age ranging from 18 years to 73 years were studied. Most of the cases were solitary, circumscribed, solid and cystic, dermal, symmetrical, lobulated tumors with a sheet-like and papillary architecture. The cells were chiefly eosinophilic with a regular oval grooved nucleus and a small inconspicuous nucleolus. Clear cells were also seen. Squamous differentiation was an important feature, with most showing a infundibular type of keratinization. Sebaceous differentiation is also common. The stroma varied from hyaline to myxoid. Only 1 case showed poroid differentiation. CONCLUSIONS: This study describes the assortment of histologic characteristics in hidradenomas. Apocrine hidradenomas are more common, contrary to earlier belief that favored an eccrine origin.

Actinomycosis in urachal remnants: A rare cause of pseudotumor.

Actinomycosis is a chronic inflammatory condition caused by Actinomyces israeli, a gram positive anaerobic bacterium. It can have a variety of clinical manifestations and can mimic a malignancy. We present one such case of urachal actinomycosis that mimicked a tumor. A 28-year-old man presented with abdominal pain of 20 days duration. Per abdominal palpation revealed a firm mass with ill-defined borders in the suprapubic region. Computed tomography and magnetic resonance imaging scans of the pelvis showed an irregular lesion in the urinary bladder extending to the umbilicus, giving the impression of urachal remnants with inflammation. Peroperatively, an irregular, hard mass measuring 6 × 5 cm, involving the anterior and posterior bladder walls, the appendix, the terminal ileum and sigmoid colon, was seen, which was suspicious for a malignancy. Frozen sections from the mass showed extensive inflammation and a florid fibroblastic proliferation, giving the impression of an inflammatory pseudotumor. The tissue was extensively sampled for paraffin sections and only one of them revealed a colony of Gram, PAS and GMS- positive organisms, conclusive for Actinomycosis. It is important to be aware of this uncommon, yet significant, presentation of a common infectious disease in order to avoid misdiagnosis and over-treatment as a malignancy.

Histopathologic profile of alopecia areata in Indian patients.

BACKGROUND: Alopecia Areata (AA) is a "non-scarring" alopecia that has an autoimmune basis. Though clinically distinctive, problems arise in diagnosis depending on the temporal stage of the disease at presentation; some of them progress to scarring alopecia and predicting its prognosis is difficult. Histological changes depend on the disease stage and site of the biopsy. OBJECTIVES: To describe the spectrum of histologic features in AA. MATERIALS AND METHODS: A prospective and retrospective study of H and E sections of all biopsies signed out as AA between 2001 and 2009 (20 cases) was undertaken. RESULTS: The diagnosis was made on vertical sections in all cases. The total number of hair follicles ranged from 1 to 24 with an average of 7 and comprised mainly terminal follicles. Vellus follicles were scanty. Anagen to non-anagen ratio was 1:1.62. Miniaturization of follicles was noted in five (25%) cases. Peribulbar inflammation was seen in all the cases with a dominance of lymphocytes. Perifollicular fibrosis was noted in 12 (60%) and pigment casts in 5 (25%) cases. Scarring was seen in two cases. In these cases, a diagnosis of AA was rendered on the basis of even spacing of the fibrotic units and remnants of the catagenic basement membrane within the scars. The epidermis and interfollicular dermis were normal in all the cases. CONCLUSION: The most consistent features of AA are an increase in non-anagen terminal follicles and peribulbar lymphocytic infiltrate. The etiology can be determined even in cases that have progressed to scarring.

Pilomatricoma: a tumor with hidden depths.

BACKGROUND: Pilomatricoma is a benign tumor of hair matrix differentiation and has been classically described as comprising of basaloid and shadow cells admixed with multinucleated giant cells and areas of calcification. However, there are a diverse range of histologic features this tumor displays that are often unrecognized. AIMS: This study was undertaken to record the histopathologic features of pilomatricoma with an emphasis on the occurrence of other forms of differentiation. METHODS: The study included all skin biopsy specimens over a 13-year period from 1995 to 2007 that had a histologic diagnosis of pilomatricoma. Hematoxylin and eosin-stained slides were reviewed. RESULTS: This study included 21 cases of pilomatricoma. Supramatrical differentiation was seen in all cases and three-quarters of the cases showed matrical differentiation. Also observed in some of the cases were clear cell differentiation toward the outer root sheath, infundibular differentiation, calcification, ossification and secondary inflammation with a foreign body giant cell reaction. Epidermal induction in the form of a downward plate-like growth of the epidermis was seen in a few cases. CONCLUSION: Pilomatricoma, although considered a tumor of hair matrix differentiation, can show cellular evolution toward the other parts of the hair follicle, such as the outer and inner root sheaths, sebaceous and infundibular components and, therefore, can be considered a panfollicular neoplasm.

Follicular mycosis fungoides - A report of four Indian cases.

BACKGROUND: Follicular Mycosis Fungoides (FMF) is an under-recognized disease in India. Its clinical mimics include Hansen's disease and Sarcoidosis. AIMS: To describe the clinical and pathological features of FMF. MATERIALS AND METHODS: All cases of FMF between January and December 2007 were retrieved. Cases of conventional epidermotropic MF with a minor follicular component were excluded. Slides were reviewed by two observers. The following criteria were assessed: degree and density of folliculotropism of lymphocytes, location of folliculotropism (infundibular / isthmic / bulbar), follicular mucin, eosinophils, granulomas, and conventional epidermotropism. Each feature was assigned a semi-quantitative grade. RESULTS: There were four cases of FMF, with an equal gender distribution and a mean age of 17.5 years. All lesions were on the face. They presented as: hypopigmented patches (2) and erythematous plaques (2). Alopecia was seen in two cases. The clinical diagnosis was Hansen's disease in all four, with a differential of Alopecia mucinosa / Sarcoidosis in two cases.The histological features seen were: disproportionate folliculotropism, lymphocyte tagging with haloes, follicular mucin, and nucleomegaly / convolution in all four cases, prominent eosinophils (2), epithelioid granulomas (1), eccrine infiltration (4), parakeratosis at the follicular ostia (2), and sebaceotropism (1). The infiltrate was bulbar (4) and isthmic (2). The rest of the epidermis showed no hint of conventional MF. CONCLUSION: The preferential features for FMF were involvement of face, dominant folliculotropism, nuclear atypia and convolution, and follicular mucin. Presence of granulomas and eosinophils necessitated exclusion of infectious causes. The absence of findings of MF in the rest of the epidermis should not deter pathologists from rendering this diagnosis.

Early mycosis fungoides vs. inflammatory mimics: how reliable is histology?

BACKGROUND: The histologic diagnosis of early mycosis fungoides (MF) and its distinction from inflammatory dermatoses is challenging, owing to the overlap of several features. AIMS: 1) To assess the efficacy of histologic criteria to diagnose early MF, 2) to study their utility in differentiating inflammatory mimics of MF. METHODS: We retrospectively reviewed slides from 50 cases clinically/histologically suspicious for MF. The diagnoses were established based on response to treatment and follow-up. The slides were analyzed double-blinded by two observers independently. Twenty-eight histologic criteria were assessed and each criterion was graded. Univariate analysis was performed on the results. RESULTS: There were 17 cases of MF and 33 of inflammatory dermatoses. Of the 28 criteria, the following 15 achieved significance on univariate analysis: disproportionate epidermotropism, tagging of lymphocytes along the basal layer, haloed lymphocytes, convoluted lymphocytes, Pautrier's abscesses, larger epidermal lymphocytes, wiry dermal collagen, absence of edema, eccrine infiltration, folliculotropism, follicular mucin, involvement of papillary and reticular dermis, monomorphous infiltrates, and atypia of dermal lymphocytes. The criteria that were 100% specific for MF included convoluted lymphocytes, eccrine infiltration, and follicular mucin. Absence of edema was 100% sensitive and specific in distinguishing MF from its inflammatory mimics. CONCLUSIONS: A combination of histologic patterns and cytology of lymphocytes is reliable in distinguishing MF from inflammatory dermatoses. No single criterion is effective in achieving this. Rather than merely recording the presence or absence of a criterion, grading each of them adds objectivity to the diagnosis.