Deep learning architecture "LightOCT" for diagnostic decision support using optical coherence tomography images of biological samples.

Optical coherence tomography (OCT) is being increasingly adopted as a label-free and non-invasive technique for biomedical applications such as cancer and ocular disease diagnosis. Diagnostic information for these tissues is manifest in textural and geometric features of the OCT images, which are used by human expertise to interpret and triage. However, it suffers delays due to the long process of the conventional diagnostic procedure and shortage of human expertise. Here, a custom deep learning architecture, LightOCT, is proposed for the classification of OCT images into diagnostically relevant classes. LightOCT is a convolutional neural network with only two convolutional layers and a fully connected layer, but it is shown to provide excellent training and test results for diverse OCT image datasets. We show that LightOCT provides 98.9% accuracy in classifying 44 normal and 44 malignant (invasive ductal carcinoma) breast tissue volumetric OCT images. Also, >96% accuracy in classifying public datasets of ocular OCT images as normal, age-related macular degeneration and diabetic macular edema. Additionally, we show ∼96% test accuracy for classifying retinal images as belonging to choroidal neovascularization, diabetic macular edema, drusen, and normal samples on a large public dataset of more than 100,000 images. The performance of the architecture is compared with transfer learning based deep neural networks. Through this, we show that LightOCT can provide significant diagnostic support for a variety of OCT images with sufficient training and minimal hyper-parameter tuning. The trained LightOCT networks for the three-classification problem will be released online to support transfer learning on other datasets.

NOVEL USE OF FLUORESCEIN DYE IN DETECTION OF ORAL DYSPLASIA AND ORAL CANCER.

BACKGROUND: India is now regarded as the country with one of the highest incidence of oral cancer in the world. Considering poor survival in cases with late diagnosis, early detection can reduce morbidity and mortality of cancer patients and may impede malignant transformation in cases of oral potentially malignant disorders (OPMD). Most of the diagnostic aids are expensive and not available for mass screenings in developing countries. There is a need to develop a sensitive and affordable technique for screening of oral cancer, which can be accurate even in hands of health care workers with limited experience. Fluorescein dye has been used for tumour detection in colon, stomach, breast and brain. However, its utility in the diagnosis of oral cancer and OPMD has not yet been explored. METHODS: This is the first study to report the role of fluorescein in the detection of oral cancer and OPMD. The present cross sectional study was conducted at a tertiary care dental centre. It included 100 individuals presenting with 42 OPMDs, 40 oral squamous cell carcinoma (OSCC) and 18 controls. RESULTS: The sensitivity and specificity for the fluorescein detection method for OPMDs and OSCC was found to be 96.6% and 52.4% respectively. The positive predictive value was 73.7% and the negative predictive value was 91.7% for the fluorescein method. The likelihood ratios stood at 2.03 for a positive test and 0.066 for a negative test. CONCLUSION: We conclude that fluorescein staining along with blue light is likely to improve detection of early oral cancers and dysplasia and can play a vital role in mass screening programmes of oral cancer.

Low coherence quantitative phase microscopy with machine learning model and Raman spectroscopy for the study of breast cancer cells and their classification.

Early-stage detection of breast cancer is the primary requirement in modern healthcare as it is the most common cancer among women worldwide. Histopathology is the most widely preferred method for the diagnosis of breast cancer, but it requires long processing time and involves qualitative assessment of cancer by a trained person/doctor. Here, we present an alternate technique based on white light interference microscopy (WLIM) and Raman spectroscopy, which has the capability to differentiate between cancerous and normal breast tissue. WLIM provides quantitative phase information about the biological tissues/cells, whereas Raman spectroscopy can detect changes in their molecular structure and chemical composition during cancer growth. Further, both the techniques can be implemented very quickly without staining the sample. The present technique is employed to perform ex vivo study on a total of 80 normal and cancerous tissue samples collected from 16 different patients. A generalized machine learning model is developed for the classification of normal and cancerous tissues, which is based on texture features obtained from phase maps with an accuracy of 90.6%. The correlation of outcomes from these two techniques can open a new avenue for fast and accurate detection of cancer without any trained personnel.

Volumetric analysis of breast cancer tissues using machine learning and swept-source optical coherence tomography.

In breast cancer, 20%-30% of cases require a second surgery because of incomplete excision of malignant tissues. Therefore, to avoid the risk of recurrence, accurate detection of the cancer margin by the clinician or surgeons needs some assistance. In this paper, an automated volumetric analysis of normal and breast cancer tissue is done by a machine learning algorithm to separate them into two classes. The proposed method is based on a support-vector-machine-based classifier by dissociating 10 features extracted from the A-line, texture, and phase map by the swept-source optical coherence tomographic intensity and phase images. A set of 88 freshly excised breast tissue [44 normal and 44 cancers (invasive ductal carcinoma tissues)] samples from 22 patients was used in our study. The algorithm successfully classifies the cancerous tissue with sensitivity, specificity, and accuracy of 91.56%, 93.86%, and 92.71% respectively. The present computational technique is fast, simple, and sensitive, and extracts features from the whole volume of the tissue, which does not require any special tissue preparation nor an expert to analyze the breast cancer as required in histopathology. Diagnosis of breast cancer by extracting quantitative features from optical coherence tomographic images could be a potentially powerful method for cancer detection and would be a valuable tool for a fine-needle-guided biopsy.

Physics for Surgeons Part 3: Why Cyst Is Spherical in Shape?

Physical examination of any swelling is the first step in making a diagnosis. Many a times we see a patient with a spherical swelling, which is usually a cyst. The interpretation of physical signs should be based on sound principles of physics. In the present paper, we explain physical characteristics of a swelling (cyst) using principles of fluid mechanics.

What Is a False Negative Sentinel Node Biopsy: Definition, Reasons and Ways to Minimize It?

Sentinel node biopsy helps in assessing the involvement of axillary lymph node without the morbidity of full axillary lymph node dissection, namely arm and shoulder pain, paraesthesia and lymphoedema. The various methods described in the literature identify the sentinel lymph nodes in approximately 96 % of cases and associated with a false negativity rate of 5 to 10 %. A false negative sentinel node is defined as the proportion of cases in whom sentinel node biopsy is reported as negative, but the rest of axillary lymph node(s) harbours cancer cells. The possible causes of a false negative sentinel lymph node may be because of blocked lymphatics either by cancer cells or following fibrosis of previous surgery/radiotherapy, and an alternative pathway opens draining the blue dye or isotope to another uninvolved node. The other reasons may be two lymphatic pathways for a tumour area, the one opening to a superficial node and the other in deep nodes. Sometimes, lymphatics do not relay into a node but traverse it going to a higher node. In some patients, the microscopic focus of metastasis inside a lymph node is so small-micrometastasis (i.e. between 0.2 and 2 mm) or isolated tumour cells (i.e. less than 0.2 mm) that is missed by the pathologist. The purpose of this review is to clear some fears lurking in the mind of most surgeons about the false negative sentinel lymph node (FNSLN).

Time resolved spectroscopy and gain studies of Fullerenes C60 and C70.

The fluorescence decay time of Fullerenes C60 and C70 in pure form as well as in mixture with Coumarin C440 and Quinizarine dyes are studied. Results indicate that the decay of pure fullerenes is constant throughout the solute concentration and it is also independent of excitation wavelength, whereas in the case of mixture with dyes different behavior is noticed. We have also calculated the Stern-Volmer quenching constant and optical gain of both the fullerenes from which it is found that the optical gain is positive for Fullerene C70 only in a very narrow range of concentration.

Förster's resonance energy transfer between Fullerene C60 and Coumarin C440.

The interaction between Coumarin C440 with Fullerene C60 has been studied by fluorescence and time resolved spectroscopic techniques. The Coumarin C440-Fullerene C60 pair shows Forster's resonance energy transfer (FRET) from Coumarin C440 (donor) to Fullerenes C60 (acceptor). The FRET efficiency of this pair increases with the increase of the acceptor concentration. The critical energy transfer distance (R0) at which transfer efficiency is 50% is found to be 34Ǻ. Stern-Volmer plot indicates static as well as dynamic quenching. However, the FRET studies show highest efficiency at the critical stage of dimer formation.