METHODS: In this paper, we propose a novel approach to distinguish colonic polyps by integrating several techniques, including a modified deep residual network, principal component analysis and AdaBoost ensemble learning. A powerful deep residual network architecture, ResNet-50, was investigated to reduce the computational time by altering its architecture. To keep the interference to a minimum, median filter, image thresholding, contrast enhancement, and normalisation techniques were exploited on the endoscopic images to train the classification model. Three publicly available datasets, i.e., Kvasir, ETIS-LaribPolypDB, and CVC-ClinicDB, were merged to train the model, which included images with and without polyps.
RESULTS: The proposed approach trained with a combination of three datasets achieved Matthews Correlation Coefficient (MCC) of 0.9819 with accuracy, sensitivity, precision, and specificity of 99.10%, 98.82%, 99.37%, and 99.38%, respectively.
CONCLUSIONS: These results show that our method could repeatedly classify endoscopic images automatically and could be used to effectively develop computer-aided diagnostic tools for early CRC detection.
CASE REPORT: We described a case of extragonadal vaginal YST in a one year and seven months old girl who presented with vaginal discharge and bleeding, and discuss its differential diagnosis and potential pitfalls in immunohistochemistry. She was found to have a suprapubic mass on examination. The serum alpha fetoprotein was 11919.4 ng/mL. Computed tomography of the pelvis revealed a large 6.4 cm heterogenous pelvic mass. Colposcopic examination of the pelvis showed a fungating vaginal mass that was subsequently confirmed as a yolk sac tumour. Immunohistochemically, the malignant cells were positive toward CKAE1/AE3, AFP and glypican-3, as well as CD117.
DISCUSSION: Solid pattern extragonadal vaginal YST may morphologically resemble dysgerminoma that is also CD117 positive, while the glandular pattern YST may have clear cytoplasm and is positive for cytokeratin; hence, may resemble clear cell carcinoma. Being mindful of these potential diagnostic caveats is necessary to prevent misdiagnosis.
METHODS: A cross-sectional diagnostic study was performed on patients who had undergone nasal endoscopy and allergy testing. Allergy status was determined by positive serology or epicutaneous testing. Endoscopy was reviewed by blinded assessors for middle turbinate head edema. Appearance was graded as either normal, focal, multifocal, diffuse, or polypoid edema. Receiver-operator (ROC) analysis, likelihood ratio (LR), sensitivity, specificity, and positive predictive value (PPV) were determined.
RESULTS: One hundred eighty-seven patients representing 304 nasal cavities were assessed (42% female, age 39.74 ± 14.7 years, 57% allergic). Diffuse edema (PPV 91.7%/LR = 8) and polypoid edema (PPV 88.9%/LR = 6.2) demonstrated the strongest association with inhalant allergy. Multifocal edema was used as a cut-off to represent inhalant allergy from ROC analysis, which demonstrated 94.7% specificity and 23.4% sensitivity. The PPV for multifocal was 85.1% and LR = 4.4.
CONCLUSION: Middle turbinate edema is a useful nasal endoscopic feature to predict presence of inhalant allergy and, although not sensitive, has excellent PPV.
BACKGROUND: Metastatic tumours are more frequent in older individuals and can be the only sign/symptom of an undiagnosed primary malignancy.
CASE REPORT: A 69-year-old lady presented with an enlarging mandibular swelling. Incisional biopsy and imaging studies were suggestive of metastatic follicular thyroid carcinoma.
CONCLUSION: Metastatic tumours should be considered in the differential diagnosis of mandibular swellings especially in older individuals.
METHODS: This is a retrospective observational study utilising 187 archived data sets of women presenting with lower urinary tract symptoms and/or pelvic organ prolapse between August 2012 and November 2013. Evaluation included a standardised interview, ICS-POPQ, rectal examination and 4D translabial ultrasound. The main outcome measure was the diagnosis of rectocele by digital rectal palpation on Valsalva manoeuvre. This diagnosis correlated with the sonographic diagnosis of rectocele to determine agreement between digital examination and ultrasound findings.
RESULTS: Complete data sets were available for 180 participants. On imaging, the mean position of the rectal ampulla was 11.07 (-36.3 to 44.3) mm below the symphysis pubis; 42.8% (77) had a rectocele of a depth of ≥10 mm. On palpation, a rectocele was detected in 60 women (33%). Agreement between palpation and imaging was observed in 77%; the kappa was 0.52 (CI 0.39-0.65). On receiver operator characteristic analysis, the area under the curve was 0.854 for the relationship between rectocele pocket depth and the detection of rectocele on palpation.
CONCLUSION: Moderate agreement was found between digital rectal examination for rectocele and translabial ultrasound findings of a "true rectocele". Digital rectal examination may be used to identify these defects in clinical practice. Extending the clinical examination of prolapse to include rectal examination to palpate defects in the rectovaginal septum may reduce the need for defecatory proctograms for the assessment of obstructive defecation and may help triage patients in the management of posterior compartment prolapse.