Deep learning segmentation of endothelial cell images using an active learning paradigm with guided label corrections.

Purpose: To create Guided Correction Software for informed manual editing of automatically generated corneal endothelial cell (EC) segmentations and apply it to an active learning paradigm to analyze a diverse set of post-keratoplasty EC images. Approach: An original U-Net model trained on 130 manually labeled post-Descemet stripping automated endothelial keratoplasty (EK) images was applied to 841 post-Descemet membrane EK images generating "uncorrected" cell border segmentations. Segmentations were then manually edited using the Guided Correction Software to create corrected labels. This dataset was split into 741 training and 100 testing EC images. U-Net and DeepLabV3+ were trained on the EC images and the corresponding uncorrected and corrected labels. Model performance was evaluated in a cell-by-cell analysis. Evaluation metrics included the number of over-segmentations, under-segmentations, correctly identified new cells, and endothelial cell density (ECD). Results: Utilizing corrected segmentations for training U-Net and DeepLabV3+ improved their performance. The average number of over- and under-segmentations per image was reduced from 23 to 11 with the corrected training set. Predicted ECD values generated by networks trained on the corrected labels were not significantly different than the ground truth counterparts (p=0.02, paired t-test). These models also correctly segmented a larger percentage of newly identified cells. The proposed Guided Correction Software and semi-automated approach reduced the time to accurately segment EC images from 15 to 30 to 5 min, an ∼80% decrease compared to manual editing. Conclusions: Guided Correction Software can efficiently label new training data for improved deep learning performance and generalization between EC datasets.

Machine Learning Analysis of Postkeratoplasty Endothelial Cell Images for the Prediction of Future Graft Rejection.

Purpose: This study developed machine learning (ML) classifiers of postoperative corneal endothelial cell images to identify postkeratoplasty patients at risk for allograft rejection within 1 to 24 months of treatment. Methods: Central corneal endothelium specular microscopic images were obtained from 44 patients after Descemet membrane endothelial keratoplasty (DMEK), half of whom had experienced graft rejection. After deep learning segmentation of images from all patients' last and second-to-last imaging, time points prior to rejection were analyzed (175 and 168, respectively), and 432 quantitative features were extracted assessing cellular spatial arrangements and cell intensity values. Random forest (RF) and logistic regression (LR) models were trained on novel-to-this-application features from single time points, delta-radiomics, and traditional morphometrics (endothelial cell density, coefficient of variation, hexagonality) via 10 iterations of threefold cross-validation. Final assessments were evaluated on a held-out test set. Results: ML classifiers trained on novel-to-this-application features outperformed those trained on traditional morphometrics for predicting future graft rejection. RF and LR models predicted post-DMEK patients' allograft rejection in the held-out test set with >0.80 accuracy. RF models trained on novel features from second-to-last time points and delta-radiomics predicted post-DMEK patients' rejection with >0.70 accuracy. Cell-graph spatial arrangement, intensity, and shape features were most indicative of graft rejection. Conclusions: ML classifiers successfully predicted future graft rejections 1 to 24 months prior to clinically apparent rejection. This technology could aid clinicians to identify patients at risk for graft rejection and guide treatment plans accordingly. Translational Relevance: Our software applies ML techniques to clinical images and enhances patient care by detecting preclinical keratoplasty rejection.

Quantitative and qualitative evaluation of deep learning automatic segmentations of corneal endothelial cell images of reduced image quality obtained following cornea transplant.

We are developing automated analysis of corneal-endothelial-cell-layer, specular microscopic images so as to determine quantitative biomarkers indicative of corneal health following corneal transplantation. Especially on these images of varying quality, commercial automated image analysis systems can give inaccurate results, and manual methods are very labor intensive. We have developed a method to automatically segment endothelial cells with a process that included image flattening, U-Net deep learning, and postprocessing to create individual cell segmentations. We used 130 corneal endothelial cell images following one type of corneal transplantation (Descemet stripping automated endothelial keratoplasty) with expert-reader annotated cell borders. We obtained very good pixelwise segmentation performance (e.g., Dice coefficient = 0.87 ± 0.17 , Jaccard index = 0.80 ± 0.18 , across 10 folds). The automated method segmented cells left unmarked by analysts and sometimes segmented cells differently than analysts (e.g., one cell was split or two cells were merged). A clinically informative visual analysis of the held-out test set showed that 92% of cells within manually labeled regions were acceptably segmented and that, as compared to manual segmentation, automation added 21% more correctly segmented cells. We speculate that automation could reduce 15 to 30 min of manual segmentation to 3 to 5 min of manual review and editing.

Machine learning for segmenting cells in corneal endothelium images.

Images of the endothelial cell layer of the cornea can be used to evaluate corneal health. Quantitative biomarkers extracted from these images such as cell density, coefficient of variation of cell area, and cell hexagonality are commonly used to evaluate the status of the endothelium. Currently, fully-automated endothelial image analysis systems in use often give inaccurate results, while semi-automated methods, requiring trained image analysis readers to identify cells manually, are both challenging and time-consuming. We are investigating two deep learning methods to automatically segment cells in such images. We compare the performance of two deep neural networks, namely U-Net and SegNet. To train and test the classifiers, a dataset of 130 images was collected, with expert reader annotated cell borders in each image. We applied standard training and testing techniques to evaluate pixel-wise segmentation performance, and report corresponding metrics such as the Dice and Jaccard coefficients. Visual evaluation of results showed that most pixel-wise errors in the U-Net were rather non-consequential. Results from the U-Net approach are being applied to create endothelial cell segmentations and quantify important morphological measurements for evaluating cornea health.

Synthesis and characterization of CaSr-Metal Organic Frameworks for biodegradable orthopedic applications.

Metal-organic frameworks (MOFs) formed from metals and organic ligands, are crystalline materials that are degradable in aqueous medium, and capable of releasing Ca and Sr ions. In this manuscript, the ability of MOFs to degrade and release osteogenic Ca and Sr ions was investigated. MOFs were generated by choosing osteoinductive Ca and Sr metals, and an organic ligand 1,3,5 tricarboxylicbenzene (H3BTC) as a linker. These MOFs were able to induce in vitro biomineralization from pre-osteoblastic MC3T3 cells and human mesenchymal stem cells (hMSCs). Moreover, these MOFs (when loaded with dimethyloxalylglycine (DMOG)) induced vascular endothelial production from hMSCs. qRT-PCR analysis performed on hMSCs (isolated from femoral heads of patients undergoing joint arthroplasty) treated with MOFs crystals suggested that the CaSr-MOFs by themselves can upregulate osteogenic genes in hMSCs, which is the first time to our knowledge that this has been observed from MOFs.

Ethnic differences in the epidemiological characteristics of severe trauma due to falls from heights among children in northern Israel.

BACKGROUND: In the western world, trauma is the leading cause of disability and mortality in the 1-39 years age group. Road accidents constitute the most frequent cause of mortality among children older than 1 year and falls from heights are the most frequent cause of injuries requiring hospitalization. OBJECTIVES: To analyze the epidemiology and characteristics of severe pediatric trauma due to falls from a height in northern Israel. This analysis should aid in planning an effective intervention plan. METHODS: This observational study included all patients aged 0-14 who died or were admitted to an intensive care unit in northern Israel following a fall from a height. Demographic and clinical data were collected retrospectively for 3 years and prospectively for 1 year. RESULTS: A total of 188 children were severely injured or died following such a fall, with an annual rate of 11.4 per 100,000 children. Over 85% of severe injuries due to falls occurred among non-Jewish children, with an incidence rate 6.36 times higher than among Jewish children (20.17 and 3.17 per 100,000 children, respectively). In the non-Jewish sector 93.7% of the falls occurred at or around the child's home, mainly from staircases, balconies and roofs. CONCLUSIONS: A very high incidence of severe trauma due to domestic falls from a height was found among non-Jewish children in northern Israel. Domestic falls represent an important epidemiological problem in the non-Jewish pediatric sector, and an effective prevention plan should include measures to modify parents' attitudes towards safety issues and the creation of a safe domestic environment.

Clinical outcomes of breast cancer in carriers of BRCA1 and BRCA2 mutations.

BACKGROUND: Some features of breast cancer in women with a BRCA1 mutation suggest that hereditary breast cancer has a poor outcome. We conducted a national population-based study of Israeli women to determine the influence, if any, of a BRCA1 or a BRCA2 mutation on the prognosis in breast cancer. METHODS: We obtained data on all incident cases of invasive breast cancer that were diagnosed from January 1, 1987, to December 31, 1988, and recorded in the Israel National Cancer Registry. We requested a paraffin-embedded tumor block or an unstained slide and the corresponding pathological and clinical records for all such cases. DNA extracted from the tumor specimens was analyzed for the three founder mutations in BRCA1 and BRCA2. For each subject, available pathological and oncologic records were reviewed. RESULTS: We were able to retrieve a pathological sample from 1794 of 2514 subjects (71%). Among those women, we obtained medical records for 1545 (86%). A BRCA1 or BRCA2 mutation was identified in 10% of the women who were of Ashkenazi Jewish ancestry. The adjusted hazard ratios for death from breast cancer were not significantly different among mutation carriers and noncarriers (hazard ratio among BRCA1 carriers, 0.76; 95% confidence interval [CI], 0.45 to 1.30; P=0.31; hazard ratio among BRCA2 carriers, 1.31; 95% CI, 0.80 to 2.15; P=0.28). Among women who were treated with chemotherapy, the hazard ratio for death among BRCA1 carriers was 0.48 (95% CI, 0.19 to 1.21; P=0.12). CONCLUSIONS: Breast cancer-specific rates of death among Israeli women are similar for carriers of a BRCA founder mutation and noncarriers.