The Dynamics of Pathology Dataset Creation Using Urine Cytology as an Example.

INTRODUCTION: Dataset creation is one of the first tasks required for training AI algorithms but is underestimated in pathology. High-quality data are essential for training algorithms and data should be labelled accurately and include sufficient morphological diversity. The dynamics and challenges of labelling a urine cytology dataset using The Paris System (TPS) criteria are presented. METHODS: 2,454 images were labelled by pathologist consensus via video conferencing over a 14-day period. During the labelling sessions, the dynamics of the labelling process were recorded. Quality assurance images were randomly selected from images labelled in previous sessions within this study and randomly distributed throughout new labelling sessions. To assess the effect of time on the labelling process, the labelled set of images was split into 2 groups according to the median relative label time and the time taken to label images and intersession agreement were assessed. RESULTS: Labelling sessions ranged from 24 m 11 s to 41 m 06 s in length, with a median of 33 m 47 s. The majority of the 2,454 images were labelled as benign urothelial cells, with atypical and malignant urothelial cells more sparsely represented. The time taken to label individual images ranged from 1 s to 42 s with a median of 2.9 s. Labelling times differed significantly among categories, with the median label time for the atypical urothelial category being 7.2 s, followed by the malignant urothelial category at 3.8 s and the benign urothelial category at 2.9 s. The overall intersession agreement for quality assurance images was substantial. The level of agreement differed among classes of urothelial cells - benign and malignant urothelial cell classes showed almost perfect agreement and the atypical urothelial cell class showed moderate agreement. Image labelling times seemed to speed up, and there was no evidence of worsening of intersession agreement with session time. DISCUSSION/CONCLUSION: Important aspects of pathology dataset creation are presented, illustrating the significant resources required for labelling a large dataset. We present evidence that the time taken to categorise urine cytology images varies by diagnosis/class. The known challenges relating to the reproducibility of the AUC (atypical) category in TPS when compared to the NHGUC (benign) or HGUC (malignant) categories is also confirmed.

Challenges Developing Deep Learning Algorithms in Cytology.

BACKGROUND: The incorporation of digital pathology into routine pathology practice is becoming more widespread. Definite advantages exist with respect to the implementation of artificial intelligence (AI) and deep learning in pathology, including cytopathology. However, there are also unique challenges in this regard. SUMMARY: This review discusses cytology-specific challenges, including the need to implement digital cytology prior to AI; the large file sizes and increased acquisition times for whole slide images in cytology; the routine use of multiple stains, such as Papanicolaou and Romanowsky stains; the lack of high-quality annotated datasets on which to train algorithms; and the considerable computer resources required, in terms of both computer infrastructure and skilled personnel, for computing and storage of data. Global concerns regarding AI that are certainly applicable to cytology include the need for model validation and continued quality assurance, ethical issues such as the use of patient data in developing algorithms, the need to develop regulatory frameworks regarding what type of data can be utilized and ensuring cybersecurity during data collection and storage, and algorithm development. Key Messages: While AI will likely play a role in cytology practice in the future, applying this technology to cytology poses a unique set of challenges. A broad understanding of digital pathology and algorithm development is desirable to guide the development of algorithms, as well as the need to be cognizant of potential pitfalls to avoid when incorporating the technology in practice.

Cytological Criteria to Distinguish Phyllodes Tumour of the Breast from Fibroadenoma.

OBJECTIVE: To determine whether there are significant differences between fibroadenomas and phyllodes tumours with regard to selected cytomorphological features. STUDY DESIGN: A 10-year retrospective review was performed of patients who underwent excision of a fibroadenoma or phyllodes tumour and in whom a preoperative fine-needle aspiration was performed. The following cytological criteria were assessed: number of stromal and epithelial fragments, stromal-to-epithelial ratio, stromal cellularity, stromal borders, stromal atypia, and proportion of background wavy spindled cells. Patient age, tumour laterality, and tumour size were recorded. RESULTS: Fifty fibroadenomas and 17 phyllodes tumours were included. Compared to phyllodes tumours, fibroadenomas had a larger number of epithelial fragments, a smaller number of stromal fragments, and a lower stromal-to-epithelial ratio. The stroma tended to be less cellular and less atypical compared to phyllodes tumours and the background cellular population contained fewer spindled cells. CONCLUSION: Fibroadenomas and phyllodes tumours differ with regard to various cytological features, aiding in their distinction on fine-needle aspiration biopsy.

Cytopathology of extracranial ectopic and metastatic meningiomas.

INTRODUCTION: Extracranial meningiomas may infrequently be encountered as ectopic or metastatic tumors. Their rarity and unique cytomorphology often pose significant diagnostic dilemmas. The aim of this study was to report our experience with a series of ectopic and metastatic meningiomas, characterizing their cytomorphology with histological correlation. MATERIALS AND METHODS: A retrospective analysis involving 13 patients with cytological preparations from extracranial meningiomas was performed. Cytology cases were correlated with available surgical resection specimens. Data regarding clinical findings, tumor information, cytomorphology, follow-up histological features and immunohistochemistry were recorded and analyzed. RESULTS: There were 5 cases with metastases and 8 ectopic meningiomas. Metastases occurred in the scalp/skull, lung, paraspinal soft tissue and liver. Primary ectopic meningiomas were located in the paranasal sinuses and ear, orbit and neck. Cytomorphological features characteristic of meningiomas were identified in the majority of samples including tightly cohesive clusters of spindled cells, whorls, intranuclear inclusions, nuclear grooves and psammomatous calcification. Unusual cytomorphological features identified in only a few cases included epithelioid cell predominance, abundant inflammatory cells, small-cell change, papillary structures and pseudoacinar growth. Metastatic tumors exhibited more nuclear atypia and occasionally mitoses or necrosis. Meningiomas were shown to be immunoreactive for epithelial membrane antigen, pancytokeratin and vimentin. CONCLUSION: Although rare, extracranial meningiomas can be encountered in cytologic specimens and should be included in the differential diagnosis when characteristic morphological features of meningiomas are seen. Cytopathologists should be aware that these lesions could be mistaken for other tumors, especially when confounded by atypia and unusual cytomorphological features.

Improving diagnostic capability for HPV disease internationally within the NIH-NIAID Division of AIDS Clinical Trial Networks.

OBJECTIVES: To evaluate an external quality assurance (EQA) program for the laboratory diagnosis of human papillomavirus (HPV) disease that was established to improve international research capability within the Division of AIDS at the National Institute of Allergy and Infectious Disease-supported Adult AIDS Clinical Trials Group network. METHODS: A three-component EQA scheme was devised comprising assessments of diagnostic accuracy of cytotechnologists and pathologists using available EQA panels, review of quality and accuracy of clinical slides from local sites by an outside expert, and HPV DNA detection using a commercially available HPV test kit. RESULTS: Seven laboratories and 17 pathologists in Africa, India, and South America participated. EQA scores were suboptimal for EQA proficiency testing panels in three of seven laboratories. There was good agreement between the local laboratory and the central reader 70% of the time (90% confidence interval, 42%-98%). Performance on the College of American Pathologists' HPV DNA testing panel was successful in all laboratories tested. CONCLUSIONS: The prequalifying EQA round identified correctable issues that will improve the laboratory diagnosis of HPV-related cervical disease at the participating international study sites and will provide a mechanism for ongoing education and continuous quality improvement.