Cytologic scoring of equine exercise-induced pulmonary hemorrhage: Performance of human experts and a deep learning-based algorithm.

Exercise-induced pulmonary hemorrhage (EIPH) is a relevant respiratory disease in sport horses, which can be diagnosed by examination of bronchoalveolar lavage fluid (BALF) cells using the total hemosiderin score (THS). The aim of this study was to evaluate the diagnostic accuracy and reproducibility of annotators and to validate a deep learning-based algorithm for the THS. Digitized cytological specimens stained for iron were prepared from 52 equine BALF samples. Ten annotators produced a THS for each slide according to published methods. The reference methods for comparing annotator's and algorithmic performance included a ground truth dataset, the mean annotators' THSs, and chemical iron measurements. Results of the study showed that annotators had marked interobserver variability of the THS, which was mostly due to a systematic error between annotators in grading the intracytoplasmatic hemosiderin content of individual macrophages. Regarding overall measurement error between the annotators, 87.7% of the variance could be reduced by using standardized grades based on the ground truth. The algorithm was highly consistent with the ground truth in assigning hemosiderin grades. Compared with the ground truth THS, annotators had an accuracy of diagnosing EIPH (THS of < or ≥ 75) of 75.7%, whereas, the algorithm had an accuracy of 92.3% with no relevant differences in correlation with chemical iron measurements. The results show that deep learning-based algorithms are useful for improving reproducibility and routine applicability of the THS. For THS by experts, a diagnostic uncertainty interval of 40 to 110 is proposed. THSs within this interval have insufficient reproducibility regarding the EIPH diagnosis.

Deep Learning-Based Quantification of Pulmonary Hemosiderophages in Cytology Slides.

Exercise-induced pulmonary hemorrhage (EIPH) is a common condition in sport horses with negative impact on performance. Cytology of bronchoalveolar lavage fluid by use of a scoring system is considered the most sensitive diagnostic method. Macrophages are classified depending on the degree of cytoplasmic hemosiderin content. The current gold standard is manual grading, which is however monotonous and time-consuming. We evaluated state-of-the-art deep learning-based methods for single cell macrophage classification and compared them against the performance of nine cytology experts and evaluated inter- and intra-observer variability. Additionally, we evaluated object detection methods on a novel data set of 17 completely annotated cytology whole slide images (WSI) containing 78,047 hemosiderophages. Our deep learning-based approach reached a concordance of 0.85, partially exceeding human expert concordance (0.68 to 0.86, mean of 0.73, SD of 0.04). Intra-observer variability was high (0.68 to 0.88) and inter-observer concordance was moderate (Fleiss' kappa = 0.67). Our object detection approach has a mean average precision of 0.66 over the five classes from the whole slide gigapixel image and a computation time of below two minutes. To mitigate the high inter- and intra-rater variability, we propose our automated object detection pipeline, enabling accurate, reproducible and quick EIPH scoring in WSI.

Effects of on-slide fixation on the cell quality of cytocentrifuged equine bronchioalveolar lavage fluid.

BACKGROUND: During bronchoalveolar lavage fluid (BALF) sample preparation in horses, several technical aspects can affect sample variability. To date, the effects of different fixatives on prepared equine BALF films have been insufficiently investigated. OBJECTIVES: This study aimed to determine the effect of various on-slide fixation methods on cell quality, including spray fixation of wet films, and acetone and methanol fixation of air-dried samples in comparison with unfixed, air-dried films. METHODS: Cytocentrifuged BALF samples from 5 horses were fixed in a wet state using a commercially available fixation spray. They were also fixed with acetone or methanol after air-drying using standard protocols or were air-dried with no fixation. After different postfixation storage durations and temperatures, the samples were stained with May-Grünwald Giemsa or immunocytochemistry stains. Subsequently, differential cell counts (DCCs) were performed, cell areas were measured, and cell morphologies and immunocytochemical staining intensities were assessed semiquantitatively. RESULTS: Optimal cell morphology results were achieved with the wet-spray fixation method. Acetone and methanol fixation, especially when performed at -20°C, caused reduced cell morphology quality, thereby significantly altering DCCs. For storage of unstained samples for 1 week at room temperature, no significant changes in cell morphology were observed for either fixation method. Wet-spray fixation resulted in enhanced preservation of macrophage, granulocyte, and mast cell sizes compared with air-drying techniques. Immunocytochemical staining of unfixed and acetone-fixed samples was the most intense. CONCLUSIONS: Wet-spray fixation resulted in the best preservation of cellular morphology and less cell shrinkage compared with unfixed specimens and is, therefore, recommended for BALF cytology.

Influence of bronchoalveolar lavage on thoracic radiography in the horse.

Bronchoalveolar lavage (BAL) and thoracic radiography are routinely performed diagnostic procedures. We hypothesized that BAL increases the interstitial opacity of caudoventral and caudodorsal thoracic radiographs. Fifty-three horses, including 8 clinic owned and 45 from a referral hospital population, were classified as healthy controls (n = 12), severe equine asthma (recurrent airway obstruction, n = 12) or mild-to-moderate equine asthma (inflammatory airway disease, n = 21) based on the results of a clinical scoring system. Eight were excluded due to different diagnoses and poor image quality. Four randomized thoracic radiographs of each horse were scored by two blinded observers, who were also asked to identify the image as obtained before or after a BAL procedure. In severe equine asthma, the chance (adjusted odds) of misinterpretation of the correct imaging time was approximately 5 times higher than in controls (odds ratio [OR] = 5.373, p = 0.028). The chance of misinterpretation was approximately 4 times lower in caudodorsal images than in caudoventral projections (OR = 0.241, p = 0.004). Identification of the correct imaging time was highly correlated with an increase in interstitial opacity (OR = 9.976, p ＜ 0.0001). In conclusion, we recommend performing BAL after thoracic radiography to avoid possible misinterpretation.