Computer-aided diagnosis of retinopathy in retinal fundus images of preterm infants via quantification of vascular tortuosity

Computer-aided detection of retinopathy of prematurity severity assessment via vessel tortuosity measurement in preterm infants' fundus images

OBJECTIVE

To develop a computer-aided diagnostic system for retinopathy of prematurity (ROP) disease using retinal vessel morphological features.

METHODS

A total of 200 fundus images from 136 preterm infants with stage 1 to 3 ROP were analysed. Two methods were developed to measure vessel tortuosity: the peak-and-valley method and the polynomial curve fitting method. Correlations between temporal artery tortuosity (TAT) and temporal vein tortuosity (TVT) with ROP severity were investigated, and vessel tortuosity relationships with vessel angles (TAA and TVA) and vessel widths (TAW and TVW). A separate dataset from Japan containing 126 images from 97 preterm patients was used for verification.

RESULTS

Both methods identified similar tortuosity in images without ROP and mild ROP cases. However, the polynomial curve fit method demonstrated enhanced tortuosity detection in stages 2 and 3 ROP compared to the peak and valley method. A strong positive correlation was revealed between ROP severity and increased arterial and venous tortuosity (P < 0.0001). A significant negative correlation between TAA and TAT (r = -0.485, P < 0.0001) and TVA and TVT (r = -0.281, P < 0.0001), and a significant positive correlation between TAW and TAT (r = 0.204, P value = 0.0040) were identified. Similar results were found in the test dataset from Japan.

CONCLUSIONS

ROP severity was associated with increased retinal tortuosity and retinal vessel width while displaying a decrease in retinal vascular angle. This quantitative analysis of retinal vessels provides crucial insights for advancing ROP diagnosis and understanding its progression.

A Novel Method to Improve Inter-Clinician Variation in the Diagnosis of Retinopathy of Prematurity Using Machine Learning

PURPOSE

Inter-clinician variation could cause uncertainty in disease management. This is reported to be high in Retinopathy of Prematurity (ROP), a potentially blinding retinal disease affecting premature infants. Machine learning has the potential to quantify the differences in decision-making between ROP specialists and trainees and may improve the accuracy of diagnosis.

METHODS

An anonymized survey of ROP images was administered to the expert(s) and the trainee(s) using a study-designed user interface. The results were analyzed for repeatability as well as to identify the level of agreement in the classification. "Ground truth" was prepared for each individual and a unique classifier was built for each individual using the same. The classifier allowed the identification of the most important features used by each individual.

RESULTS

Correlation and disagreement between the expert and the trainees were visualized using the Dipstick™ diagram. Intra-clinician repeatability and reclassification statistics were assessed for all. The repeatability was 88.4% and 86.2% for two trainees and 92.1% for the expert, respectively. Commonly used features differed for the expert and the trainees and accounted for the variability.

CONCLUSION

This novel, automated algorithm quantifies the differences using machine learning techniques. This will help audit the training process by objectively measuring differences between experts and trainees.

TRANSLATIONAL RELEVANCE

Training for image-based ROP diagnosis can be more objectively performed using this novel, machine learning-based automated image analyzer and classifier.

Retinal vascular geometry detection as a biomarker in diabetes mellitus

OBJECTIVE

To compare the vessel geometry characteristics of color fundus photographs in normal control and diabetes mellitus (DM) patients and to find potential biomarkers for early diabetic retinopathy (DR) based on a neural network vessel segmentation system and automated vascular geometry parameter analysis software.

METHODS

A total of 102 consecutive patients with type 2 DM (T2DM) and 132 healthy controls were recruited. All participants underwent general ophthalmic examinations, and retinal fundus photographs were taken with a digital fundus camera without mydriasis. Color fundus photographs were input into a dense-block generative adversarial network (D-GAN)-assisted retinal vascular segmentation system (http://www.gdcerc.cn:8081/#/login) to obtain binary images. These images were then analyzed by customized software (ocular microvascular analysis system V2.9.1) for automatic processing of vessel geometry parameters, including the monofractal dimension (D_box), multifractal dimension (D₀), vessel area ratio (R), max vessel diameter (d_max), average vessel diameter (d_ave), arc-chord ratio (A/C), and tortuosity (τ_n). Geometric differences between the healthy subjects and DM patients were analyzed. Then, regression analysis and receiver operating characteristic (ROC) curve analysis were performed to evaluate the diagnostic efficiency of the vascular geometry parameters.

RESULTS

No significant differences were observed between the baseline characteristics of each group. DM patients had lower D_box and D₀ values (1.330 ± 0.041; 1.347 ± 0.038) than healthy subjects (1.343 ± 0.048, p < 0.05; 1.362 ± 0.042, p < 0.05) and showed increasing values of d_max, d_ave, A/C, and τ_n compared with normal controls, although only the differences in d_ave and τ_n between the groups were statistically significant. In the regression analysis, d_ave and τ_n showed a good correlation with diabetes (d_ave, OR 1.765, 95% CI 1.319-2.362, p < 0.001; τ_n, OR 9.323, 95% CI 1.492-58.262, p < 0.05).

CONCLUSIONS

We demonstrated the relationship between retinal vascular geometry and the process in DM patients, showing that D_box, D₀, d_ave, and τ_n may be indicators of morphological changes in retinal vessels in DM patients and can be early biomarkers of DR.

Improving Blood Vessel Tortuosity Measurements via Highly Sampled Numerical Integration of the Frenet-Serret Equations

Measures of vascular tortuosity-how curved and twisted a vessel is-are associated with a variety of vascular diseases. Consequently, measurements of vessel tortuosity that are accurate and comparable across modality, resolution, and size are greatly needed. Yet in practice, precise and consistent measurements are problematic-mismeasurements, inability to calculate, or contradictory and inconsistent measurements occur within and across studies. Here, we present a new method of measuring vessel tortuosity that ensures improved accuracy. Our method relies on numerical integration of the Frenet-Serret equations. By reconstructing the three-dimensional vessel coordinates from tortuosity measurements, we explain how to identify and use a minimally-sufficient sampling rate based on vessel radius while avoiding errors associated with oversampling and overfitting. Our work identifies a key failing in current practices of filtering asymptotic measurements and highlights inconsistencies and redundancies between existing tortuosity metrics. We demonstrate our method by applying it to manually constructed vessel phantoms with known measures of tortuousity, and 9,000 vessels from medical image data spanning human cerebral, coronary, and pulmonary vascular trees, and the carotid, abdominal, renal, and iliac arteries.

A computer-aided diagnosis system for plus disease in retinopathy of prematurity with structure adaptive segmentation and vessel based features

Retinopathy of Prematurity (ROP) is a blinding disease affecting the retina of low birth-weight preterm infants. Accurate diagnosis of ROP is essential to identify treatment-requiring ROP, which would help to prevent childhood blindness. Plus disease, which characterizes abnormal twisting, widening and branching of the blood vessels, is a significant symptom of treatment requiring ROP. In this paper, we have developed and evaluated a computer-based analysis system for objective assessment of plus disease in ROP, which best mimics the clinical method of disease diagnosis by identifying unique vessel based features. The proposed system consists of an initial segmentation stage, which will efficiently extract blood vessels of varying width and length by utilizing structure adaptive filtering, connectivity analysis and image fusion. The paper proposes the usage of additional retinal features namely leaf node count and vessel density, to portray the abnormal growth and branching of the blood vessels and to complement the commonly used features namely tortuosity and width. The test results show a better classification of plus disease in terms of sensitivity (95%) and specificity (93%), emphasizing the superiority of the proposed segmentation algorithm and vessel-based features. An additional advantage of the proposed system is that the process of selection of relevant vessels for feature extraction is fully automated, which makes the system highly useful to the non-physician graders, owing to the unavailability of a sufficient number of ROP specialists.