Dr Danuta Sampson
Dr Danuta Sampson is a Research Fellow at the University of Surrey. She holds a joint appointment between Centre for Vision, Speech and Signal Processing within Faculty of Engineering and Physical Sciences and the Section of Clinical Medicine and Ageing within Faculty of Health & Medical Sciences and . Between 2014-2018 she held a research position at the Lions Eye Institute and the University of Western Australia. Her expertise is in optics, optical microscopy, and image processing. She acquired this expertise during her PhD at the Nicolaus Copernicus University in Torun, Poland, in the Optical Biomedical Imaging Group. Her research focuses on developing hardware and software methodologies for visualization, assessment and correlation of subtle changes in the structure and function of the human retina. Beyond research, she has been involved in many outreach programs aimed at raising general public interest in optics and science. Recently, she has been selected to serve as The Optical Society (OSA) Ambassador for 2017. This lifetime distinction recognizes OSA young professionals for their continuing efforts to serve the wider optics and photonics community.
Areas of specialism
To develop a fully automatic method, based on deep learning algorithms, for determining the locations of cone photoreceptors within adaptive optics scanning laser ophthalmoscope images and evaluate its performance against a dataset of manually segmented images.
A fully convolutional network (FCN) based on U-Net architecture was used to generate prediction probability maps and then used a localization algorithm to reduce the prediction map to a collection of points. The proposed method was trained and tested on two publicly available datasets of different imaging modalities, with Dice overlap, false discovery rate, and true positive reported to assess performance.
The proposed method achieves a Dice coefficient of 0.989, true positive rate of 0.987, and false discovery rate of 0.009 on the first confocal dataset; and a Dice coefficient of 0.926, true positive rate of 0.909, and false discovery rate of 0.051 on the second split detector dataset. Results compare favorably with a previously proposed method, but this method provides quicker (25 times faster) evaluation performance.
The proposed FCN-based method demonstrates that deep learning algorithms can achieve accurate cone localizations, almost comparable to a human expert, while labeling the images.
Manual cone photoreceptor identification is a timeconsuming task due to the large number of cones present within a single image; using the proposed FCN-based method could support the image analysis task, drastically reducing the need for manual assessment of the photoreceptor mosaic.
All automated image quality indicators for en face optical coherence tomography angiography (OCTA) images require gold standard validation for determining optimum thresholds.
A manual grading system (gold standard) for OCTA images was validated and compared to two automated image quality indicators: signal strength index (SSI) and scan quality index (SQI) generated by different software versions of the Optovue OCTA device.
Retrospective cross?sectional study.
A total of 52 eyes of 52 healthy individual and 77 eyes of 51 patients with retinal vascular diseases.
A total of 129 OCTA images of the superficial vascular plexus were graded manually by three independent examiners. Each image was assigned grades 1 to 4 (1?2, unacceptable; 3?4, acceptable) masked to the software?generated quality indicators.
Main Outcome Measures
Inter?grader agreement and comparison of the utility of SSI and SQI in discriminating between acceptable and unacceptable OCTA images.
There was a substantial agreement between the three graders (º = 0.63). Mean SSI and SQI was significantly different between acceptable and unacceptable images (P Â .001). SQI outperformed SSI in separating acceptable from unacceptable images (areas under the receiver operating characteristic curve: 0.87 vs 0.80) and the optimum cut?off was e7 for SQI and e70 for SSI for acceptable images. Up to 30% of images with quality indicators reaching the optimum SQI and SSI cut?off thresholds still had unacceptable quality on manual grading. Unacceptable images were found in 33% and 66% of healthy and diseased eyes, respectively.
Conclusions and Relevance
SQI is closely related to manual grading but we caution reliance on the optimized threshold to determine image quality. SQI is superior to SSI in discriminating between acceptable and unacceptable images.
Significance: Pulsatility is a vital characteristic of the cardiovascular system. Characterization of the pulsatility pattern locally in the peripheral microvasculature is currently not readily available and would provide an additional source of information which may prove important in understanding the pathophysiology of arterial stiffening, vascular ageing, and their linkage with cardiovascular disease development.
Aim: We aim to confirm the suitability of speckle decorrelation optical coherence tomography angiography (OCTA) under various non-contact/contact scanning protocols for the visualization of pulsatility patterns in vessel-free tissue and in the microvasculature of peripheral human skin.
Results: Results from 5 healthy subjects show distinct pulsatile patterns both in vessel-free tissue with either non-contact or contact imaging and in individual microvessels with contact imaging; respectively, likely caused by the pulsatile pressure and pulsatile blood flow. The pulse rates show good agreement with those from pulse oximetry, confirming that the pulsatile signatures reflect pulsatile hemodynamics.
Conclusions: This study demonstrates the potential of speckle decorrelation OCTA for measuring localized peripheral cutaneous pulsatility and defines scanning protocols necessary to undertake such measurements. Non-contact imaging should be used for the study of pulsatility in vessel-free tissue and contact imaging with strong mechanical coupling in individual microvessels. Further studies of microcirculation based upon this method and protocols are warranted.