Colorectal cancer is one of the most common cancers worldwide, accounting for an annual estimated 1.8 million incident cases. With the increasing number of colonoscopies being performed, colorectal biopsies make up a large proportion of any histopathology laboratory workload. We trained and validated a unique artificial intelligence (AI) deep learning model as an assistive tool to screen for colonic malignancies in colorectal specimens, in order to improve cancer detection and classification; enabling busy pathologists to focus on higher order decision-making tasks. The study cohort consists of Whole Slide Images (WSI) obtained from 294 colorectal specimens. Qritive’s unique composite algorithm comprises both a deep learning model based on a Faster Region Based Convolutional Neural Network (Faster-RCNN) architecture for instance segmentation with a ResNet-101 feature extraction backbone that provides glandular segmentation, and a classical machine learning classifier. The initial training used pathologists’ annotations on a cohort of 66,191 image tiles extracted from 39 WSIs. A subsequent application of a classical machine learning-based slide classifier sorted the WSIs into ‘low risk’ (benign, inflammation) and ‘high risk’ (dysplasia, malignancy) categories. We further trained the composite AI-model’s performance on a larger cohort of 105 resections WSIs and then validated our findings on a cohort of 150 biopsies WSIs against the classifications of two independently blinded pathologists. We evaluated the area under the receiver-operator characteristic curve (AUC) and other performance metrics. The AI model achieved an AUC of 0.917 in the validation cohort, with excellent sensitivity (97.4%) in detection of high risk features of dysplasia and malignancy. We demonstrate an unique composite AI-model incorporating both a glandular segmentation deep learning model and a classical machine learning classifier, with excellent sensitivity in picking up high risk colorectal features. As such, AI plays a role as a potential screening tool in assisting busy pathologists by outlining the dysplastic and malignant glands.

No abstract

BackgroundHeterotopic pancreas most commonly occurs in the upper gastrointestinal tract of adults, usually as an incidental finding. It seldom occurs at the umbilicus, and even rarely in the pediatric age group.Case presentationHere we present a case of heterotopic pancreatic tissue in the omphalomesenteric duct remnant of a 9-month-old baby girl. She presented with redness at the base of the umbilicus associated with occasional mild wetness. A urachal fistula was suspected by ultrasound. Histology from subsequent resection revealed fibrous tissue with heterotopic pancreatic tissue and accompanying small bowel mucosa. The patient’s umbilical redness resolved after the surgery.ConclusionsUpon literature search, we found only 17 other cases of heterotopic pancreas reported in the umbilicus. They described a high male to female ratio, frequent association with omphalomesenteric duct remnant and presentation of umbilical discharge. The Heinrich system is frequently used to classify heterotopic pancreas into 3 types, based on the presence of acini, islets and ducts. Several mechanisms have been proposed on the pathogenesis of heterotopic pancreas, including misplacement, metaplasia and totipotent cell theories. Heterotopic pancreas can manifest clinically with diseases of the pancreas, including malignant transformation, reported as high as 12.7% in a series. Awareness of this finding in the biopsy aids the suitable treatment decisions for the patient.

No abstract

Primary cardiac lymphoma (PCL) is a rare neoplasm involving the heart, pericardium or both. Patients with PCL have a median survival of approximately 7 months based on literature. We report a unique case of a 63-year-old woman presenting with worsening exertional dyspnoea and symptoms of superior vena cava obstruction. Investigations revealed a large right atrial (RA) tumour with obstructive features. In the past, PCL was frequently diagnosed at postmortem but with modern imaging and biopsy techniques, this allows early diagnosis and treatment of this rapidly fatal disease. Our patient had a trans-jugular venous biopsy of the RA mass under trans-oesophageal echocardiography guidance. This confirmed the diagnosis of PCL. She subsequently completed seven cycles of chemotherapy and is currently in remission 20 months after diagnosis.

Background/aimsThe programmed cell death receptor 1 (PD-1) checkpoint inhibitor, nivolumab, has been approved for the treatment of metastatic renal cell carcinoma (RCC). However, the understanding of the expression and distribution of PD ligand 1 (PD-L1) in the tumour immune microenvironment and its prognostic role in an Asian cohort is limited. Our group investigated PD-L1 protein expression in a cohort of Asian patients with RCC of mixed ethnicity, using two commercially available antibody clones.MethodsE1L3N and SP263 anti-PD-L1 clones were used to categorise RCCs of various histological subtypes, diagnosed at our institution between 1995 and 2008, into PD-L1-positive or PD-L1-negative groups, based on a 1% Tumour Proportion Score (TPS) cut-off.ResultsIn total, 267 (83%) clear cell (cc)RCC and 55 (17%) non-ccRCC cases were studied. Overall PD-L1 protein expression rates for the entire cohort were 13% and 8% for the E1L3N and SP263 clones, respectively. Patients bearing PD-L1-positive tumours experienced significantly decreased disease-free survival (DFS; E1L3N: p=0.01; SP263: p=0.03) but not overall survival, compared with those with PD-L1-negative tumours. Multivariate survival analysis further confirmed the results of the E1L3N clone (HR 1.85, 95% CI 1.10 to 3.13, p=0.02), but not SP263, after adjusting for pathological stage, histological subtype and grade. The addition of PD-L1 (E1L3N) TPS to clinicopathological features significantly increased the prognostic value for DFS (∆LRχ2=5.25; p=0.022), compared with clinicopathological features alone.ConclusionsPD-L1 protein expression was associated with an unfavourable prognosis in our study cohort. PD-L1 (E1L3N) expression was an independent prognostic indicator of clinical outcome in all RCCs when using a 1% cut-off.

IgG4-related disease of the thyroid gland is a recently recognized subtype of thyroiditis, often with rapid progression requiring surgical treatment. It is considered as a spectrum of disease varying from early IgG4-related Hashimoto's thyroiditis (HT) pattern to late fibrosing HT or Riedel's thyroiditis patterns. Here, we report a 47-year-old Malay woman presenting with progressively painless neck swelling over 3 years and subclinical hypothyroidism. Computed tomography (CT) scan revealed diffuse thyroid enlargement (up to 13 cm) with retrosternal extension and without regional lymphadenopathy. Fine needle aspiration of the thyroid showed a limited number of follicular epithelial cell groups with widespread Hurthle cell change and scanty background colloid, but no evidence of lymphocytosis. The cytologic features fell into the category of 'atypia of undetermined significance'. Subsequently, the patient developed hypercapnic respiratory failure secondary to extrinsic upper airway compression by the thyroid mass and underwent emergent total thyroidectomy. Histology of the thyroid showed diffuse dense lymphoplasmacytic infiltrate and fibrosis. Follicular cells exhibited reactive nuclear features and some Hurthle cell change. IgG4+ plasma cells were over 40/high power field while overall IgG4/IgG ratio was above 50%. The overall features suggest the diagnosis of IgG4-related disease of the thyroid gland in the form of IgG4-related HT. Post-surgery, the patient was found to have markedly elevated serum IgG4 concentration but PET/CT did not show significant increased fludeoxyglucose uptake in other areas. Her recovery was complicated by a ventilator-associated pneumonia with empyema, limiting early use of corticosteroids for treatment of IgG4-related disease.

No abstract