A deep learning-based stripe self-correction method for stitched microscopic images.

Stitched fluorescence microscope images inevitably exist in various types of stripes or artifacts caused by uncertain factors such as optical devices or specimens, which severely affects the image quality and downstream quantitative analysis. Here, we present a deep learning-based Stripe Self-Correction method, so-called SSCOR. Specifically, we propose a proximity sampling scheme and adversarial reciprocal self-training paradigm that enable SSCOR to utilize stripe-free patches sampled from the stitched microscope image itself to correct their adjacent stripe patches. Comparing to off-the-shelf approaches, SSCOR can not only adaptively correct non-uniform, oblique, and grid stripes, but also remove scanning, bubble, and out-of-focus artifacts, achieving the state-of-the-art performance across different imaging conditions and modalities. Moreover, SSCOR does not require any physical parameter estimation, patch-wise manual annotation, or raw stitched information in the correction process. This provides an intelligent prior-free image restoration solution for microscopists or even microscope companies, thus ensuring more precise biomedical applications for researchers.

Resection-inspired histopathological diagnosis of cerebral cavernous malformations using quantitative multiphoton microscopy.

Rationale: Cerebral cavernous malformation (CCM) is prone to recurring microhemorrhage, which can lead to drug-resistant epilepsy. Surgical resection is the first choice to control seizures for CCM-associated epilepsy. At present, removal of resection-related tissue only depends on cautious visual identification of CCM lesions and perilesional yellowish hemosiderin rim by the neurosurgeon. Inspired by the resection requirements, we proposed quantitative multiphoton microscopy (qMPM) for a histopathology-level diagnostic paradigm to assist clinicians in precisely complete resection. Methods: A total of 35 sections specimens collected from 12 patients with the CCM-related epilepsy were included in this study. First, qMPM utilized a label-free multi-channel selective detection to image the histopathological features based on the spectral characteristics of CCM tissues. Then, qMPM developed three customized algorithms to provide quantitative information, a vascular patterns classifier based on linear support vector machine, visualization of microhemorrhage regions based on hemosiderin-related parameters, and the CCM-oriented virtual staining generative adversarial network (CCM-stainGAN) was constructed to generate two types of virtual staining. Results: Focused on CCM lesion and perilesional regions, qMPM imaged malformed vascular patterns and micron-scale hemosiderin-related products. Four vascular patterns were automatically identified by the classifier with an area under the receiver operating characteristic curve of 0.97. Moreover, qMPM mapped different degrees of hemorrhage regions onto fresh tissue while providing three quantitative hemosiderin-related indicators. Besides, qMPM realized virtual staining by the CCM-stainGAN with 98.8% diagnostic accuracy of CCM histopathological features in blind analysis. Finally, we provided pathologists and surgeons with the qMPM-based CCM histopathological diagnostic guidelines for a more definitive intraoperative or postoperative diagnosis. Conclusions: qMPM can provide decision-making supports for histopathological diagnosis, and resection guidance of CCM from the perspectives of high-resolution precision detection and automated quantitative assessment. Our work will promote the development of MPM diagnostic instruments and enable more optical diagnostic applications for epilepsy.

Macular ganglion cell complex injury in different stages of anterior ischemic optic neuropathy.

BACKGROUND: Anterior ischemic optic neuropathy (AION) is a group of ophthalmic diseases in which the optic nerve is injured causing blindness. However, the pathogenesis, clinical manifestations, and clinical treatments of AION are yet elusive. Only a few related experimental or clinical reports are available on the disease. In this study, spectral domain optical coherence tomography (SD-OCT) was used to examine the morphology of thickness swelling and atrophic changes of macular ganglion cell complex (mGCC) in the different stages of AION that were then compared with the visual fields. Thus, the clinical value of mGCC examination was alleged to be similar to that of the visual field. AIM: To explore the mGCC injury at different stages in AION and the clinical significance. METHODS: Cases with AION were analyzed in a retrospective study. SD-OCT was used to analyze the correlation between mGCC and peripapillary retinal nerve fiber layer thicknesses at different stages of AION and the changes in the corresponding stages of visual fields. RESULTS: A total of 21 cases (28 eyes) presented AION. The onset time of AION was defined as early stage (within 3 wk of onset), middle stage (from 3 wk to 2 mo), and late stage (disease span > 2 mo). In the early stage, the mGCC thickness of SD-OCT was within the normal high limit, and the perioptic nerve fibers thickness was more than the normal. The changes in the visual field in early stage were not consistent with the swelling changes in mGCC and peri-disc nerve fibers. In addition, atrophy and thinning appeared in mGCC, and the perioptic nerve fibers were swollen. However, the thickness was lower in the middle period than that in the early stage. The change in visual field was consistent with that of mGCC in this period. In the late stage, mGCC shrank and thinned, and the thickness of the nerve fibers around the optic disc in the corresponding region shrank and thinned. CONCLUSION: The changes in mGCC thickness in patients with AION showed early, middle, and late stages of development by SD-OCT. Although the early stage visual field changes of AION were not consistent with the swelling changes of mGCC, the horizontal delimitation or annular atrophy of mGCC was consistent with that in the middle and late stage of the disease. The atrophy of peripheral nerve fibers was later than that of the mGCC atrophy.

Patterns of Retinal Ganglion Cell Damage in Nonarteritic Anterior Ischemic Optic Neuropathy Assessed by Swept-Source Optical Coherence Tomography.

OBJECTIVE: To evaluate the ability of macular ganglion cell and inner plexiform layer (mGCIPL) and retinal nerve fiber layer (RNFL) thickness measurements by long-wavelength swept-source optical coherence tomography (SS-OCT) to assess retinal ganglion cell (RGC) damage in nonarteritic anterior ischemic optic neuropathy (NAION). METHODS: A retrospective study of 20 patients with unilateral NAION was performed. SS-OCT scanning of the macular and peripapillary areas was performed to measure the total and six-sector thicknesses of macular RNFL (mRNFL) and mGCIPL, as well as peripapillary RNFL (pRNFL) thicknesses in global and 12 clock-hour sectors. Further comparison of these thicknesses between NAION involved eyes and uninvolved counterparts was performed in 12 of the 20 patients at 4 visits. The thickness map and en face images generated by volume data of the posterior pole over a 12 × 9-mm area were used for RNFL analysis. RESULTS: Median time intervals between the visual symptom onset and first thinning occurrences of mGCIPL, mRNFL, and pRNFL were 17 days (95% Confidence Interval [CI] 14-18 days), 43 days (95% CI 32-48 days), and 70 days (95% CI 62-80 days), respectively. The thickness map indicated a significantly reduced pRNFL in the superior temporal sectors or temporal sectors after 9 weeks, and retinal damage corresponded to the superior hemisphere's mRNFL and mGCIPL. En face images showed that the RNFL thinning area gradually expanded along the retinal nerve fiber direction and progressed toward the optic nerve head. CONCLUSIONS: The patterns of RGC damage in the macular and peripapillary areas of NAION eyes can be revealed by SS-OCT. Objective measurement of SS-OCT is valuable in characterizing NAION.

[The clinical features and value of macular ganglion cell complex thickness patterns in patients with optic chiasma lesion].

OBJECTIVE: To evaluate patterns of macular retinal ganglion cell complex thickness (mRGCCT) in patients with two types of neurologic lesions in optic chiasm region. METHODS: Retrospective case review study. Visual field test was conducted with automated perimetry (HAAG-STREIT OCTOPUS101 V6.07g Seven-in-One and 2010(HFA Ⅱ750-40220-5.1.1/5.1.1) Carl Zeiss Meditec), and visual field in the central 30 degree region was measured. Three types of thickness mapping including macular retinal thickness (MRT), macular ganglion cell complex thickness (mGCCT) and peripapillary retinal nerve fiber layer (pRNFL) were recorded with TOPCON 3D-OCT 2000 spectral domain OCT. The results of OCT were compared with visual field. RESULTS: Sixteen cases were included in this study, male 10 cases, female 6 cases. Age 5 to 77 years (average 48.2). Neurologic lesions including pituitary tumor (9 cases), craniopharyngioma (3 cases), optic canal dysplasia in optic chiasm region (1 case), meningioma (1 case), hemangioma (1 case) and aneurysm (1 case). 14 cases have chronic, progressive, long course of disease, OCT showed characteristic perpendicular bisector delimitation, bilateral nasal mGCCT atrophy and corresponding temporal hemianopsia. 2 cases (both were pituitary tumor ) have a chronic, short duration, accelerated attack, visual acuity and visual field rapidly declined, mGCCT is swelling or approach high limit of normal range. In pRNFL thickness map, atrophy of mRNFL in temporal peripapillary and nasal macular region could be observed. CONCLUSIONS: Chronic, progressive, long course of disease, combined with characteristic perpendicular bisector delimitation, bilateral nasal mGCCT atrophy and corresponding temporal hemianopsia is helpful to predict lesion of disease. mGCCT map is of equal importance with visual field test for diagnosis of lesion within optic chiasm. In cases with a chronic, short disease course, and accelerated attack, mGCCT could be swelling, but pRNFL could be thinner in temporal peripapillary and nasal macular region of single or both eyes, and visual field test is thus necessary in these cases.

Focal choroidal excavation: a preliminary interpretation based on clinic and review.

AIM: To describe the clinical and imaging characteristics associated with focal choroidal excavation (FCE), analyze the possible complication, and interpret its probable etiopathogenesis. METHODS: Retrospective descriptive case series of 37 eyes of 32 patients with FCE. Findings of spectral-domain optical coherence tomography (SD-OCT), fluorescein angiography, indocyanine green angiography, and clinical features were analyzed. RESULTS: All patients were Chinese. Five patients (15.6%) were bilaterally involved. Patients' ages ranged from 7 to 66y. Refractive error ranged between +2.0 D and -11.0 D. Mean best-corrected visual acuity was 0.6 (range, 0.1 to 1.2). Fundus examinations exhibited mild-moderate localized pigmentary disturbances in the corresponding area of 17 eyes. Fluorescein angiography performed in 18 patients showed varying degrees of hyperfluorescence and hypofluorescence related to a range of retinal pigment epithelium (RPE) alterations. Indocyanine green angiography performed in 7 patients showed hypofluorescence at the excavation. SD-OCT demonstrated choroidal excavation in all 37 eyes. Twenty-nine eyes showed a single lesion of FCE, and three eyes showed 2-3 separated lesions. Fifteen eyes showed separation between the photoreceptor tips and RPE consistent with nonconforming FCE. Central serous chorioretinopathy (CSC, n=1) and choroidal neovascularization (CNV, n=1) developed during follow-up. CONCLUSION: FCE could be interpreted as congenital focal choroidal dysplasia involving the RPE, choriocapillaris, and photoreceptor associated with the faulty anatomy. The abnormal anatomy of FCE was similar to anatomy at risk of CSC and CNV.