Automated reporting of cervical biopsies using artificial intelligence.

When detected at an early stage, the 5-year survival rate for people with invasive cervical cancer is 92%. Being aware of signs and symptoms of cervical cancer and early detection greatly improve the chances of successful treatment. We have developed an Artificial Intelligence (AI) algorithm, trained and evaluated on cervical biopsies for automated reporting of digital diagnostics. The aim is to increase overall efficiency of pathological diagnosis and to have the performance tuned to high sensitivity for malignant cases. Having a tool for triage/identifying cancer and high grade lesions may potentially reduce reporting time by identifying areas of interest in a slide for the pathologist and therefore improving efficiency. We trained and validated our algorithm on 1738 cervical WSIs with one WSI per patient. On the independent test set of 811 WSIs, we achieved 93.4% malignant sensitivity for classifying slides. Recognising a WSI, with our algorithm, takes approximately 1.5 minutes on the NVIDIA Tesla V100 GPU. Whole slide images of different formats (TIFF, iSyntax, and CZI) can be processed using this code, and it is easily extendable to other formats.

Detection of malignancy in whole slide images of endometrial cancer biopsies using artificial intelligence.

In this study we use artificial intelligence (AI) to categorise endometrial biopsy whole slide images (WSI) from digital pathology as either "malignant", "other or benign" or "insufficient". An endometrial biopsy is a key step in diagnosis of endometrial cancer, biopsies are viewed and diagnosed by pathologists. Pathology is increasingly digitised, with slides viewed as images on screens rather than through the lens of a microscope. The availability of these images is driving automation via the application of AI. A model that classifies slides in the manner proposed would allow prioritisation of these slides for pathologist review and hence reduce time to diagnosis for patients with cancer. Previous studies using AI on endometrial biopsies have examined slightly different tasks, for example using images alongside genomic data to differentiate between cancer subtypes. We took 2909 slides with "malignant" and "other or benign" areas annotated by pathologists. A fully supervised convolutional neural network (CNN) model was trained to calculate the probability of a patch from the slide being "malignant" or "other or benign". Heatmaps of all the patches on each slide were then produced to show malignant areas. These heatmaps were used to train a slide classification model to give the final slide categorisation as either "malignant", "other or benign" or "insufficient". The final model was able to accurately classify 90% of all slides correctly and 97% of slides in the malignant class; this accuracy is good enough to allow prioritisation of pathologists' workload.

Weakly supervised learning and interpretability for endometrial whole slide image diagnosis.

Fully supervised learning for whole slide image-based diagnostic tasks in histopathology is problematic due to the requirement for costly and time-consuming manual annotation by experts. Weakly supervised learning that utilizes only slide-level labels during training is becoming more widespread as it relieves this burden, but has not yet been applied to endometrial whole slide images, in iSyntax format. In this work, we apply a weakly supervised learning algorithm to a real-world dataset of this type for the first time, with over 85% validation accuracy and over 87% test accuracy. We then employ interpretability methods including attention heatmapping, feature visualization, and a novel end-to-end saliency-mapping approach to identify distinct morphologies learned by the model and build an understanding of its behavior. These interpretability methods, alongside consultation with expert pathologists, allow us to make comparisons between machine-learned knowledge and consensus in the field. This work contributes to the state of the art by demonstrating a robust practical application of weakly supervised learning on a real-world digital pathology dataset and shows the importance of fine-grained interpretability to support understanding and evaluation of model performance in this high-stakes use case.

British Gynaecological Cancer Society (BGCS) cervical cancer guidelines: Recommendations for practice.

Cervix cancer in many countries is declining and screening programmes and immunisation will reduce the incidence in the next few decades. This guideline attempts to cover management of invasive disease reflecting diagnosis and imaging including new imaging and sentinel lymph node biopsies. Smaller volume disease is usually managed surgically whereas advanced disease is treated with (chemo)- radiation. It also includes discussion of fertility sparing procedures. Practices are changing frequently for all aspects of care usually in attempts to reduce complications and improve quality of life. The management of advanced disease is treated by chemotherapy and the use of newer agents is also discussed. Other sections discuss specialist situations such as cancer in pregnancy, rare cervical tumours, late effects and supportive measures and fertility preserving approaches.

New Directions in Vulvar Cancer Pathology.

PURPOSE OF REVIEW: The aim of this article is to provide clinicians and pathologists with an understanding of the aetiopathology, pathogenesis and classification of vulval neoplasia and their molecular correlates. RECENT FINDINGS: There is an increased understanding of subcellular changes in vulvar malignancies. These provide the direction for further research and aid personalised treatment for patients. The article explores concepts of the aetiology of vulvar cancer and updates the reader with the equivalence of terminology of preneoplastic vulval disease. The differential diagnosis of squamous neoplasia and their clinicopathological correlation is detailed. The salient findings from recent literature into the understanding of the disease of squamous cell neoplasia and rare vulvar malignancies are summarised.

Sinonasal glomangiopericytoma: Is anything new?

More than 100 cases of sinonasal hemangiopericytoma have been reported in the literature, but only a handful of cases of nasal glomangiopericytoma. In this article, we report a case of a nasal glomangiopericytoma that was treated with endonasal surgical excision. We also attempt to clarify the confusion that attends to the nomenclature surrounding the terms glomangiopericytoma and hemangiopericytoma, which are often used interchangeably. Although glomangiopericytomas are histologically similar to sinonasal hemangiopericytomas, they sometimes behave in a different clinical manner. To further enhance our understanding of nasal glomangiopericytomas, more cases need to be reported. This may improve our ability to establish specific treatment modalities for this type of neoplasm and to predict clinical outcomes.

Multifocal FIGO Stage IA1 Squamous Carcinoma of the Cervix: Criteria for Identification, Staging, and its Good Clinical Outcome.

Multifocal squamous cervical carcinomas account for up to 25% of IA1 tumors identified on excisional biopsy, yet there are no uniformly accepted histopathologic criteria for defining and staging these lesions. Here, we use a strict case definition and meticulous specimen processing from colposcopist to pathologist to identify and follow-up 25 cases of multifocal IA1 cervical squamous carcinomas identified in excisional biopsies. We stage these tumors using the dimensions of the largest focus and a minimum of 2 mm between each foci to define multifocality. The cases are followed up for a median of 7 yr with no episodes of tumor recurrence or metastasis. We also show that the prevalence of residual preinvasive (20%) and invasive disease (5%) on repeat excision/surgery are comparable to data available for unifocal IA1 cases. Our study supports the hypothesis that multifocal lesions should be staged according to largest individual focus of invasion and we emphasize the importance of meticulous specimen handling to appropriately identify multifocal tumors. In addition, our analysis suggests that outcomes are comparable to unifocal lesions and supports the hypothesis that they may be managed in a similar manner.

Adequacy of cervical sampling in hysterectomy specimens for endometrial cancer.

AIM: To determine whether sampling one section, which includes the anterior and posterior cervical lips in hysterectomy specimens, provides the best prospect for detecting stromal invasion in hysterectomy specimens for endometrial cancers. METHODS: To assess the most likely section in which cervical stromal invasion was identified in 29 cases. RESULTS: Cervical stromal invasion was detected in first section in 75.8% cases, whereas 24.1% detected in random sections other than first section. CONCLUSION: Cervical stromal invasion is most likely to be identified in one section taken in the midline, which includes anterior and posterior lips. However, in a significant number of cases, this one section missed invasion, which was detected in other random cervical sections. Therefore, sampling of one block (anterior and posterior) from the cervix is insufficient for detecting cervical stromal invasion for endometrial cancer.

Paraganglioma of the skull base presenting as nasal polyps.

The authors report a case of paraganglioma of the skull base presenting as nasal polyps. A 29-year-old patient presented with epistaxis and was found to have nasal polyps. The patient underwent a nasal polypectomy. After pathology showed an unusual appearance, the mass was subsequently excised endoscopically using radiofrequency coblation, and it was found to be originating from the skull base. The diagnosis was made using a combination of clinical findings, radiology, and histopathology examination. It is important to consider paraganglioma in the differential diagnosis of unusual tumors of the nose and skull base.

Gamma tubulin: a promising indicator of recurrence in squamous cell carcinoma of the larynx.

OBJECTIVE: Centrosome amplification as detected by gamma tubulin (GT) immunostaining is associated with genetic instability and tumor aggressiveness. We assessed GT for its ability to predict recurrence of squamous cell carcinoma of the larynx (SCCL). STUDY DESIGN: Case series with chart review. MATERIALS AND METHODS: Five micron sections of 35 archival SCCL samples were subjected to antigen retrieval and immunostaining with antibody to GT. The keratin antibody CK5 served as a positive control for antigen retrieval, and tonsillar tissue was used as a negative control. RESULTS: Of the 35 tumors analyzed, 22 were associated with recurrence(R) and 13 were not (NR). Fourteen of the 22 R tumors, but 0 of 13 of the NR tumours had a GT staining score of 2+ or 3+ (P < 0.0002). GT was also related to recurrence in node-negative tumors (P < 0.006) but was unrelated to T stage (P = 0.726). CONCLUSIONS: GT staining appears to be a better predictor of tumor recurrence than T stage and also predicts recurrence in N0 tumors.