CD, or not CD, that is the question: a digital interobserver agreement study in coeliac disease

OBJECTIVE

Coeliac disease (CD) diagnosis generally depends on histological examination of duodenal biopsies. We present the first study analysing the concordance in examination of duodenal biopsies using digitised whole-slide images (WSIs). We further investigate whether the inclusion of immunoglobulin A tissue transglutaminase (IgA tTG) and haemoglobin (Hb) data improves the interobserver agreement of diagnosis.

DESIGN

We undertook a large study of the concordance in histological examination of duodenal biopsies using digitised WSIs in an entirely virtual reporting setting. Our study was organised in two phases: in phase 1, 13 pathologists independently classified 100 duodenal biopsies (40 normal; 40 CD; 20 indeterminate enteropathy) in the absence of any clinical or laboratory data. In phase 2, the same pathologists examined the (re-anonymised) WSIs with the inclusion of IgA tTG and Hb data.

RESULTS

We found the mean probability of two observers agreeing in the absence of additional data to be 0.73 (±0.08) with a corresponding Cohen's kappa of 0.59 (±0.11). We further showed that the inclusion of additional data increased the concordance to 0.80 (±0.06) with a Cohen's kappa coefficient of 0.67 (±0.09).

CONCLUSION

We showed that the addition of serological data significantly improves the quality of CD diagnosis. However, the limited interobserver agreement in CD diagnosis using digitised WSIs, even after the inclusion of IgA tTG and Hb data, indicates the importance of interpreting duodenal biopsy in the appropriate clinical context. It further highlights the unmet need for an objective means of reproducible duodenal biopsy diagnosis, such as the automated analysis of WSIs using artificial intelligence.

Development and validation of artificial intelligence-based prescreening of large-bowel biopsies taken in the UK and Portugal: a retrospective cohort study

BACKGROUND

Histopathological examination is a crucial step in the diagnosis and treatment of many major diseases. Aiming to facilitate diagnostic decision making and improve the workload of pathologists, we developed an artificial intelligence (AI)-based prescreening tool that analyses whole-slide images (WSIs) of large-bowel biopsies to identify typical, non-neoplastic, and neoplastic biopsies.

METHODS

This retrospective cohort study was conducted with an internal development cohort of slides acquired from a hospital in the UK and three external validation cohorts of WSIs acquired from two hospitals in the UK and one clinical laboratory in Portugal. To learn the differential histological patterns from digitised WSIs of large-bowel biopsy slides, our proposed weakly supervised deep-learning model (Colorectal AI Model for Abnormality Detection [CAIMAN]) used slide-level diagnostic labels and no detailed cell or region-level annotations. The method was developed with an internal development cohort of 5054 biopsy slides from 2080 patients that were labelled with corresponding diagnostic categories assigned by pathologists. The three external validation cohorts, with a total of 1536 slides, were used for independent validation of CAIMAN. Each WSI was classified into one of three classes (ie, typical, atypical non-neoplastic, and atypical neoplastic). Prediction scores of image tiles were aggregated into three prediction scores for the whole slide, one for its likelihood of being typical, one for its likelihood of being non-neoplastic, and one for its likelihood of being neoplastic. The assessment of the external validation cohorts was conducted by the trained and frozen CAIMAN model. To evaluate model performance, we calculated area under the convex hull of the receiver operating characteristic curve (AUROC), area under the precision-recall curve, and specificity compared with our previously published iterative draw and rank sampling (IDaRS) algorithm. We also generated heat maps and saliency maps to analyse and visualise the relationship between the WSI diagnostic labels and spatial features of the tissue microenvironment. The main outcome of this study was the ability of CAIMAN to accurately identify typical and atypical WSIs of colon biopsies, which could potentially facilitate automatic removing of typical biopsies from the diagnostic workload in clinics.

FINDINGS

A randomly selected subset of all large bowel biopsies was obtained between Jan 1, 2012, and Dec 31, 2017. The AI training, validation, and assessments were done between Jan 1, 2021, and Sept 30, 2022. WSIs with diagnostic labels were collected between Jan 1 and Sept 30, 2022. Our analysis showed no statistically significant differences across prediction scores from CAIMAN for typical and atypical classes based on anatomical sites of the biopsy. At 0·99 sensitivity, CAIMAN (specificity 0·5592) was more accurate than an IDaRS-based weakly supervised WSI-classification pipeline (0·4629) in identifying typical and atypical biopsies on cross-validation in the internal development cohort (p<0·0001). At 0·99 sensitivity, CAIMAN was also more accurate than IDaRS for two external validation cohorts (p<0·0001), but not for a third external validation cohort (p=0·10). CAIMAN provided higher specificity than IDaRS at some high-sensitivity thresholds (0·7763 vs 0·6222 for 0·95 sensitivity, 0·7126 vs 0·5407 for 0·97 sensitivity, and 0·5615 vs 0·3970 for 0·99 sensitivity on one of the external validation cohorts) and showed high classification performance in distinguishing between neoplastic biopsies (AUROC 0·9928, 95% CI 0·9927-0·9929), inflammatory biopsies (0·9658, 0·9655-0·9661), and atypical biopsies (0·9789, 0·9786-0·9792). On the three external validation cohorts, CAIMAN had AUROC values of 0·9431 (95% CI 0·9165-0·9697), 0·9576 (0·9568-0·9584), and 0·9636 (0·9615-0·9657) for the detection of atypical biopsies. Saliency maps supported the representation of disease heterogeneity in model predictions and its association with relevant histological features.

INTERPRETATION

CAIMAN, with its high sensitivity in detecting atypical large-bowel biopsies, might be a promising improvement in clinical workflow efficiency and diagnostic decision making in prescreening of typical colorectal biopsies.

FUNDING

The Pathology Image Data Lake for Analytics, Knowledge and Education Centre of Excellence; the UK Government's Industrial Strategy Challenge Fund; and Innovate UK on behalf of UK Research and Innovation.

Screening of normal endoscopic large bowel biopsies with interpretable graph learning: a retrospective study

OBJECTIVE

To develop an interpretable artificial intelligence algorithm to rule out normal large bowel endoscopic biopsies, saving pathologist resources and helping with early diagnosis.

DESIGN

A graph neural network was developed incorporating pathologist domain knowledge to classify 6591 whole-slides images (WSIs) of endoscopic large bowel biopsies from 3291 patients (approximately 54% female, 46% male) as normal or abnormal (non-neoplastic and neoplastic) using clinically driven interpretable features. One UK National Health Service (NHS) site was used for model training and internal validation. External validation was conducted on data from two other NHS sites and one Portuguese site.

RESULTS

Model training and internal validation were performed on 5054 WSIs of 2080 patients resulting in an area under the curve-receiver operating characteristic (AUC-ROC) of 0.98 (SD=0.004) and AUC-precision-recall (PR) of 0.98 (SD=0.003). The performance of the model, named Interpretable Gland-Graphs using a Neural Aggregator (IGUANA), was consistent in testing over 1537 WSIs of 1211 patients from three independent external datasets with mean AUC-ROC=0.97 (SD=0.007) and AUC-PR=0.97 (SD=0.005). At a high sensitivity threshold of 99%, the proposed model can reduce the number of normal slides to be reviewed by a pathologist by approximately 55%. IGUANA also provides an explainable output highlighting potential abnormalities in a WSI in the form of a heatmap as well as numerical values associating the model prediction with various histological features.

CONCLUSION

The model achieved consistently high accuracy showing its potential in optimising increasingly scarce pathologist resources. Explainable predictions can guide pathologists in their diagnostic decision-making and help boost their confidence in the algorithm, paving the way for its future clinical adoption.

Semantic annotation for computational pathology: multidisciplinary experience and best practice recommendations

Recent advances in whole‐slide imaging (WSI) technology have led to the development of a myriad of computer vision and artificial intelligence‐based diagnostic, prognostic, and predictive algorithms. Computational Pathology (CPath) offers an integrated solution to utilise information embedded in pathology WSIs beyond what can be obtained through visual assessment. For automated analysis of WSIs and validation of machine learning (ML) models, annotations at the slide, tissue, and cellular levels are required. The annotation of important visual constructs in pathology images is an important component of CPath projects. Improper annotations can result in algorithms that are hard to interpret and can potentially produce inaccurate and inconsistent results. Despite the crucial role of annotations in CPath projects, there are no well‐defined guidelines or best practices on how annotations should be carried out. In this paper, we address this shortcoming by presenting the experience and best practices acquired during the execution of a large‐scale annotation exercise involving a multidisciplinary team of pathologists, ML experts, and researchers as part of the Pathology image data Lake for Analytics, Knowledge and Education (PathLAKE) consortium. We present a real‐world case study along with examples of different types of annotations, diagnostic algorithm, annotation data dictionary, and annotation constructs. The analyses reported in this work highlight best practice recommendations that can be used as annotation guidelines over the lifecycle of a CPath project.

Deep learning based digital cell profiles for risk stratification of urine cytology images

Urine cytology is a test for the detection of high-grade bladder cancer. In clinical practice, the pathologist would manually scan the sample under the microscope to locate atypical and malignant cells. They would assess the morphology of these cells to make a diagnosis. Accurate identification of atypical and malignant cells in urine cytology is a challenging task and is an essential part of identifying different diagnosis with low-risk and high-risk malignancy. Computer-assisted identification of malignancy in urine cytology can be complementary to the clinicians for treatment management and in providing advice for carrying out further tests. In this study, we presented a method for identifying atypical and malignant cells followed by their profiling to predict the risk of diagnosis automatically. For cell detection and classification, we employed two different deep learning-based approaches. Based on the best performing network predictions at the cell level, we identified low-risk and high-risk cases using the count of atypical cells and the total count of atypical and malignant cells. The area under the receiver operating characteristic (ROC) curve shows that a total count of atypical and malignant cells is comparably better at diagnosis as compared to the count of malignant cells only. We obtained area under the ROC curve with the count of malignant cells and the total count of atypical and malignant cells as 0.81 and 0.83, respectively. Our experiments also demonstrate that the digital risk could be a better predictor of the final histopathology-based diagnosis. We also analyzed the variability in annotations at both cell and whole slide image level and also explored the possible inherent rationales behind this variability.

Hover-Net: Simultaneous segmentation and classification of nuclei in multi-tissue histology images

Nuclear segmentation and classification within Haematoxylin & Eosin stained histology images is a fundamental prerequisite in the digital pathology work-flow. The development of automated methods for nuclear segmentation and classification enables the quantitative analysis of tens of thousands of nuclei within a whole-slide pathology image, opening up possibilities of further analysis of large-scale nuclear morphometry. However, automated nuclear segmentation and classification is faced with a major challenge in that there are several different types of nuclei, some of them exhibiting large intra-class variability such as the nuclei of tumour cells. Additionally, some of the nuclei are often clustered together. To address these challenges, we present a novel convolutional neural network for simultaneous nuclear segmentation and classification that leverages the instance-rich information encoded within the vertical and horizontal distances of nuclear pixels to their centres of mass. These distances are then utilised to separate clustered nuclei, resulting in an accurate segmentation, particularly in areas with overlapping instances. Then, for each segmented instance the network predicts the type of nucleus via a devoted up-sampling branch. We demonstrate state-of-the-art performance compared to other methods on multiple independent multi-tissue histology image datasets. As part of this work, we introduce a new dataset of Haematoxylin & Eosin stained colorectal adenocarcinoma image tiles, containing 24,319 exhaustively annotated nuclei with associated class labels.

MILD-Net: Minimal information loss dilated network for gland instance segmentation in colon histology images

The analysis of glandular morphology within colon histopathology images is an important step in determining the grade of colon cancer. Despite the importance of this task, manual segmentation is laborious, time-consuming and can suffer from subjectivity among pathologists. The rise of computational pathology has led to the development of automated methods for gland segmentation that aim to overcome the challenges of manual segmentation. However, this task is non-trivial due to the large variability in glandular appearance and the difficulty in differentiating between certain glandular and non-glandular histological structures. Furthermore, a measure of uncertainty is essential for diagnostic decision making. To address these challenges, we propose a fully convolutional neural network that counters the loss of information caused by max-pooling by re-introducing the original image at multiple points within the network. We also use atrous spatial pyramid pooling with varying dilation rates for preserving the resolution and multi-level aggregation. To incorporate uncertainty, we introduce random transformations during test time for an enhanced segmentation result that simultaneously generates an uncertainty map, highlighting areas of ambiguity. We show that this map can be used to define a metric for disregarding predictions with high uncertainty. The proposed network achieves state-of-the-art performance on the GlaS challenge dataset and on a second independent colorectal adenocarcinoma dataset. In addition, we perform gland instance segmentation on whole-slide images from two further datasets to highlight the generalisability of our method. As an extension, we introduce MILD-Net⁺ for simultaneous gland and lumen segmentation, to increase the diagnostic power of the network.

Novel digital signatures of tissue phenotypes for predicting distant metastasis in colorectal cancer

Distant metastasis is the major cause of death in colorectal cancer (CRC). Patients at high risk of developing distant metastasis could benefit from appropriate adjuvant and follow-up treatments if stratified accurately at an early stage of the disease. Studies have increasingly recognized the role of diverse cellular components within the tumor microenvironment in the development and progression of CRC tumors. In this paper, we show that automated analysis of digitized images from locally advanced colorectal cancer tissue slides can provide estimate of risk of distant metastasis on the basis of novel tissue phenotypic signatures of the tumor microenvironment. Specifically, we determine what cell types are found in the vicinity of other cell types, and in what numbers, rather than concentrating exclusively on the cancerous cells. We then extract novel tissue phenotypic signatures using statistical measurements about tissue composition. Such signatures can underpin clinical decisions about the advisability of various types of adjuvant therapy.

Validation of an NGS mutation detection panel for melanoma

Background

Knowledge of the genotype of melanoma is important to guide patient management. Identification of mutations in BRAF and c-KIT lead directly to targeted treatment, but it is also helpful to know if there are driver oncogene mutations in NRAS, GNAQ or GNA11 as these patients may benefit from alternative strategies such as immunotherapy.

Methods

While polymerase chain reaction (PCR) methods are often used to detect BRAF mutations, next generation sequencing (NGS) is able to determine all of the necessary information on several genes at once, with potential advantages in turnaround time. We describe here an Ampliseq hotspot panel for melanoma for use with the IonTorrent Personal Genome Machine (PGM) which covers the mutations currently of most clinical interest.

Results

We have validated this in 151 cases of skin and uveal melanoma from our files, and correlated the data with PCR based assessment of BRAF status. There was excellent agreement, with few discrepancies, though NGS does have greater coverage and picks up some mutations that would be missed by PCR. However, these are often rare and of unknown significance for treatment.

Conclusions

PCR methods are rapid, less time-consuming and less expensive than NGS, and could be used as triage for patients requiring more extensive diagnostic workup. The NGS panel described here is suitable for clinical use with formalin-fixed paraffin-embedded (FFPE) samples.

Electronic supplementary material

The online version of this article (doi:10.1186/s12885-017-3149-0) contains supplementary material, which is available to authorized users.

Simultaneous automatic scoring and co-registration of hormone receptors in tumor areas in whole slide images of breast cancer tissue slides

Automation of downstream analysis may offer many potential benefits to routine histopathology. One area of interest for automation is in the scoring of multiple immunohistochemical markers to predict the patient's response to targeted therapies. Automated serial slide analysis of this kind requires robust registration to identify common tissue regions across sections. We present an automated method for co-localized scoring of Estrogen Receptor and Progesterone Receptor (ER/PR) in breast cancer core biopsies using whole slide images. Regions of tumor in a series of fifty consecutive breast core biopsies were identified by annotation on H&E whole slide images. Sequentially cut immunohistochemical stained sections were scored manually, before being digitally scanned and then exported into JPEG 2000 format. A two-stage registration process was performed to identify the annotated regions of interest in the immunohistochemistry sections, which were then scored using the Allred system. Overall correlation between manual and automated scoring for ER and PR was 0.944 and 0.883, respectively, with 90% of ER and 80% of PR scores within in one point or less of agreement. This proof of principle study indicates slide registration can be used as a basis for automation of the downstream analysis for clinically relevant biomarkers in the majority of cases. The approach is likely to be improved by implantation of safeguarding analysis steps post registration. © 2016 International Society for Advancement of Cytometry.

Breastfeeding practice, oral contraceptive use and risk of rheumatoid arthritis among Chinese women: the Guangzhou Biobank Cohort Study

OBJECTIVE

Hormonal and reproductive factors are implicated in the aetiology of RA, but results of previous studies have been mixed. The aim of this cross-sectional study was to assess the relationships between RA, use of oral contraceptives (OCs) and history of breastfeeding in a population of older women from South China.

METHODS

We used baseline data from 7349 women ≥ 50 years of age in the Guangzhou Biobank Cohort. Questionnaires were used to obtain socio-demographic, lifestyle and obstetric history data, including parity, OC use and breastfeeding practices. The main outcome was RA. Women were asked about history of RA and were examined to assess joint swelling. RF levels were measured. The presence of RA was defined in two ways: (i) as reporting physician-diagnosed RA or pain and swelling in at least three joints (including the wrist), and (ii) also having at least one of the following: positive RF, morning stiffness or objective swelling of the small joints of the hands.

RESULTS

Compared with those who had never breastfed, breastfeeding was associated with half the risk of RA. The risk was lower with increasing duration of breastfeeding [adjusted odds ratio (OR) 0.54 (95% CI 0.29, 1.01) for breastfeeding at least 36 months; P for trend = 0.04]. OC use had no relationship with RA.

CONCLUSION

Breastfeeding (especially longer duration) but not OC use is associated with a lower risk of RA. This has potentially important implications for future RA disease burden, given the declining rates of breastfeeding and the one-child policy in China. Further research is needed to explain the biological mechanism.

Uncertainties in coupled thermal-hydrological processes associated with the Drift Scale Test at Yucca Mountain, Nevada

Understanding thermally driven coupled hydrological, mechanical, and chemical processes in unsaturated fractured tuff is essential for evaluating the performance of the potential radioactive waste repository at Yucca Mountain, Nevada. The Drift Scale Test (DST), intended for acquiring such an understanding of these processes, has generated a huge volume of temperature and moisture redistribution data. Sophisticated thermal-hydrological (TH) conceptual models have yielded a good fit between simulation results and those measured data. However, some uncertainties in understanding the TH processes associated with the DST still exist. This paper evaluates these uncertainties and provides quantitative estimates of the range of these uncertainties. Of particular interest for the DST are the uncertainties resulting from the unmonitored loss of vapor through an open bulkhead of the test. There was concern that the outcome from the test might have been significantly altered by these losses. Using alternative conceptual models, we illustrate that predicted mean temperatures from the DST are within 1 degrees C of the measured mean temperatures through the first 2 years of heating. The simulated spatial and temporal evolution of drying and condensation fronts is found to be qualitatively consistent with measured saturation data. Energy and mass balance computation shows that no more than 13% of the input energy is lost because of vapor leaving the test domain through the bulkhead. The change in average saturation in fractures is also relatively small. For a hypothetical situation in which no vapor is allowed to exit through the bulkhead, the simulated average fracture saturation is not qualitatively different enough to be discerned by measured moisture redistribution data. This leads us to conclude that the DST, despite the uncertainties associated with open field testing, has provided an excellent understanding of the TH processes.

Clinical profiles of human immunodeficiency virus-associated lymphoma in Hong Kong

OBJECTIVE

To identify the clinical and prognostic features of human immunodeficiency virus-associated lymphoma in the local population with a view to designing more effective treatment strategies.

DESIGN

Retrospective review.

SETTING

Referral hospital, Hong Kong.

SUBJECTS AND METHODS

All patients (n=10) with human immunodeficiency virus-associated lymphoma managed at Queen Elizabeth Hospital from January 1995 to December 2001.

RESULTS

All patients were men with a median age of 39 years. The median CD4 cell count at the time of diagnosis of lymphoma was 0.056 x 10(9)/L. All tumours were diffuse large B-cell lymphomas, with the exception of one systemic Burkittlike lymphoma. Systemic lymphoma was diagnosed in seven patients and three had primary central nervous system lymphoma. Combined antiretroviral therapy was continued or given to five of the six patients who received some form of chemotherapy or radiotherapy treatment. Of the two patients with primary central nervous system lymphoma who received whole brain irradiation therapy, one patient survived 41 months in clinical remission after diagnosis and the other patient died of sepsis while in partial remission 19 months after diagnosis. The four patients with systemic lymphoma who received standard- or reduced-dose chemotherapy with cyclophosphamide, doxorubicin, vincristine, and prednisone had a median survival of 3 months.

CONCLUSION

The clinical profiles of these patients were similar to those of patients with human immunodeficiency virus-associated lymphoma in western countries. The overall survival of patients was poor with conventional chemoradiotherapy. Other innovative treatment approaches should be investigated to prolong the survival of this patient group.

Effect of heterogeneity in fracture permeability on the potential for liquid seepage into a heated emplacement drift of the potential repository

A numerical model was used to investigate the effect of spatial variability in fracture permeability on liquid seepage and moisture distribution in the vicinity of a waste emplacement drift in the unsaturated zone (UZ) of Yucca Mountain. The model is based on a two-dimensional, cross-sectional, dual-permeability model of the unsaturated zone at Yucca Mountain and uses a stochastic approach to investigate the effect of small-scale heterogeneous features. The studies were conducted using one uniform fracture permeability case, three realizations of stochastically generated fracture permeability, one discrete permeability feature case, and one increased ambient liquid flux case. In all cases, the models predict that completely dry drift conditions will develop above and below the drift in 10-100 years and remain dry for 1000-2000 years. During this period, the models predict no seepage into drifts, although liquid flux above the drifts and within the drift pillars may increase by up to two orders of magnitude above ambient flux. This is because the heat released by the emplaced waste is sufficient to vaporize liquid flux of one to two orders of magnitude higher than present-day ambient flux for over 1000 years. The results also show that unsaturated zone thermal-hydrological (TH) models with uniform layer permeability can adequately predict the evolution of seepage and moisture distribution in the rock mass surrounding the repository drifts. The models further show that although variability in fracture permeability may focus and enhance liquid flow in regions of enhanced liquid saturation (due to condensation above the drifts), vaporization and vapor diffusion can maintain a dry environment within the drifts for thousands of years.

Phase II trial of systemic recombinant interleukin-2 in the treatment of refractory nasopharyngeal carcinoma

BACKGROUND

Interleukin-2 (IL-2) is a cytokine produced by activated T cells, which has shown powerful immunostimulatory and antineoplastic properties. Nasopharyngeal carcinoma (NPC) is an Epstein-Barr virus-associated cancer with abundant lymphocyte infiltration histologically. The activity of IL-2 in the treatment of NPC patients is currently unknown. A phase II study was, therefore, initiated to evaluate the efficacy, toxicity and immunological consequences of intravenous bolus IL-2 in patients with recurrent/metastatic NPC.

METHODS

Between November 1996 and April 1997, 14 patients with recurrent/metastatic NPC were entered into the study. Recombinant IL-2 (Proleukin, Chiron) was injected by intravenous bolus every 8 h at 72,000 IU/kg for a maximum of 15 doses. After 7 days, patients were retreated with a second identical cycle of therapy. Those patients who were stable or responding to treatment 5-6 weeks later went on to receive another course (two cycles) of therapy. All patients received prophylactic antibiotics and antipyretic medicine. Response and toxicities were evaluated. Serial plasma level of TNF-alpha, IL-6, soluble IL-2 receptor, IL-10 and soluble CD8 were determined.

RESULTS

Fourteen patients received a total of 34 cycles of therapy. No response was observed. Fifty percent had stable disease, 50% had progressive disease after a median of two cycles of therapy. There was one treatment-related death from acute myocardial infarction. Body weight increase (>5%) occurred in 80% of cycles, and hypotension (BP <80 mm Hg systolic) occurred in 53%. Serum creatinine increase (>2 mg%) occurred in 24% of cycles, and SGOT/SGPT increase (>3x) in 10% of cycles. Symptoms of somnolence, general malaise, nausea and vomiting, pruritus, xerostomia, desquamation were generally mild to moderate but rapidly reversible.

CONCLUSION

The single modality of intravenous bolus IL-2 at the dose level of 72,000 IU/kg is clinically ineffective in NPC patients. Potential mechanisms of the ineffectiveness of IL-2 therapy on NPC patients are discussed.

Brain abscess formation in radiation necrosis of the temporal lobe following radiation therapy for nasopharyngeal carcinoma

BACKGROUND

Radiation necrosis is a known complication following radiation therapy for extracranial as well as intracranial tumours. However, brain abscess formation in radiation necrosis has not been reported in the literature. We report the clinical data of 6 patients suffering from this condition.

METHOD

Twenty-eight patients with radiation necrosis of the temporal lobe following radiotherapy for nasopharyngeal carcinoma were treated surgically at the Department of Neurosurgery, Queen Elizabeth Hospital, Hong Kong between January 1992 and July 1999. Of these, 6 cases were complicated by brain abscess formation. The clinical data of these 6 patients are retrospectively reviewed.

FINDINGS

The patients were 5 males and 1 female, ranging in age from 41 to 67 years. Three patients had previous treatment with steroids for the symptomatic radiation necrosis. A history of nasal infection or otitis media was recognised in all 6 patients. All patients were treated surgically by temporal lobectomy and excision of the necrotic tissue together with the abscess cavity. Intra-operatively, a bony defect was observed between the middle cranial fossa and the sphenoid sinus in 3 patients and the bony defect was repaired with a temporalis muscle flap. The species of organisms could only be identified in 3 patients. In 3 patients, the pus smear was positive but the culture was negative. Subsequently, 4 patients recovered and 2 patients died.

INTERPRETATION

Cerebral radiation necrosis is a predisposing cause of brain abscess formation. Surgical excision is recommended as the treatment of choice in this group of patients.

Polymorphous low grade adenocarcinoma of the palate in a child

We report a case of polymorphous low grade adenocarcinoma involving the palate of a 12-year-old girl, the first example of this tumour occurring in the paediatric age group. The tumour displayed infiltrative growth, neural invasion, variegated histological patterns, and minimal cytological atypia. The patient remained disease-free four years after wide local excision of the tumour. The distinction of polymorphous low grade adenocarcinoma from pleomorphic adenoma and adenoid cystic carcinoma is also discussed.

Primary mucoid adenocarcinoma of the larynx

Mucous glands carcinomas of the larynx are rare, with most being represented by 'non-specific' adenocarcinoma and adenoid cystic carcinoma. Here we report a unique case of mucoid adenocarcinoma of the larynx occurring in a 46-year-old woman. Despite the presence of regional lymph node metastasis, she remained well four years after surgery.