Unsupervised domain adaptation for histopathology image segmentation with incomplete labels.

Stain variations pose a major challenge to deep learning segmentation algorithms in histopathology images. Current unsupervised domain adaptation methods show promise in improving model generalization across diverse staining appearances but demand abundant accurately labeled source domain data. This paper assumes a novel scenario, namely, unsupervised domain adaptation based segmentation task with incompletely labeled source data. This paper propose a Stain-Adaptive Segmentation Network with Incomplete Labels (SASN-IL). Specifically, the algorithm consists of two stages. The first stage is an incomplete label correction stage, involving reliable model selection and label correction to rectify false-negative regions in incomplete labels. The second stage is the unsupervised domain adaptation stage, achieving segmentation on the target domain. In this stage, we introduce an adaptive stain transformation module, which adjusts the degree of transformation based on segmentation performance. We evaluate our method on a gastric cancer dataset, demonstrating significant improvements, with a 10.01% increase in Dice coefficient compared to the baseline and competitive performance relative to existing methods.

CrossU-Net: Dual-modality cross-attention U-Net for segmentation of precancerous lesions in gastric cancer.

Gastric precancerous lesions (GPL) significantly elevate the risk of gastric cancer, and precise diagnosis and timely intervention are critical for patient survival. Due to the elusive pathological features of precancerous lesions, the early detection rate is less than 10%, which hinders lesion localization and diagnosis. In this paper, we provide a GPL pathological dataset and propose a novel method for improving the segmentation accuracy on a limited-scale dataset, namely RGB and Hyperspectral dual-modal pathological image Cross-attention U-Net (CrossU-Net). Specifically, we present a self-supervised pre-training model for hyperspectral images to serve downstream segmentation tasks. Secondly, we design a dual-stream U-Net-based network to extract features from different modal images. To promote information exchange between spatial information in RGB images and spectral information in hyperspectral images, we customize the cross-attention mechanism between the two networks. Furthermore, we use an intermediate agent in this mechanism to improve computational efficiency. Finally, we add a distillation loss to align predicted results for both branches, improving network generalization. Experimental results show that our CrossU-Net achieves accuracy and Dice of 96.53% and 91.62%, respectively, for GPL lesion segmentation, providing a promising spectral research approach for the localization and subsequent quantitative analysis of pathological features in early diagnosis.

Laser renal denervation: A comprehensive evaluation of microstructural renal artery lesions.

Renal artery denervation (RDN) has been proposed for resistant arterial hypertension. Beyond conventional radiofrequency (RF) ablation, there are emerging RDN technologies, including laser catheter ablation. We aimed at evaluamting the local effects of laser ablation on the renal artery and perivascular nerve injury in comparison with radiofrequency ablation. Thirteen pigs (mean weight 36.7 ± 4.7 kg, age 3 months) were divided into three groups: (1) laser ablation in normotensive pigs (LA; n = 3), (2) bipolar RF ablation in normotensive pigs (RF; n = 7), and (3) a sham group (SHAM; n = 3). Transcatheter laser and RF ablations were performed under general anesthesia. After euthanasia, pathology and immunohistochemical studies were performed. Artery wall and perivascular nerve lesions were found in the LA and RF groups. A lower rate of intimal microdissections was evaluated after laser ablation when compared with RF ablation (0 vs. 1.5 [1;2.75] per artery; p < .05). There was a numerical prevalence of nerves with tyrosine hydroxylase expression loss after laser RDN when compared with RF ablation (16.7 [4; 18] vs. 9.1 [0;18] per specimen). Laser ablation is noninferior to RF ablation regarding perivascular nerve injury and is associated with less intimal layer damage.

Transcatheter radiofrequency pulmonary artery denervation in swine: the evaluation of lesion degree, hemodynamics and pulmonary hypertension inducibility.

BACKGROUND: Mechanisms of positive effects of pulmonary artery (PA) denervation (PADN) remain poorly understood. The study aimed to evaluate pulmonary hemodynamic changes after PADN and their association with the extent of PA wall damage in an acute thromboxane A2 (TXA2)-induced pulmonary hypertension (PH) model in swine. METHODS: In this experimental sham-controlled study, 17 normotensive male white Landrace pigs (the mean weight 36.2 ± 4.5 kg) were included and randomly assigned to group I (n = 9)-PH modeling before and after PADN, group II (n = 4)-PADN only, or group III (n = 4)-PH modeling before and after a sham procedure. Radiofrequency (RF) PADN was performed in the PA trunk and at the proximal parts of the right and left PAs. PA wall lesions were characterized at the autopsy study using histological and the immunohistochemical examination. RESULTS: In groups I and II, no statistically significant changes in the mean pulmonary arterial pressure nor systemic blood pressure were found after PADN (-0.8 ± 3.4 vs 4.3 ± 8.6 mmHg, P = 0.47; and 6.0 ± 15.9 vs -8.3 ± 7.5 mmHg, P = 0.1; correspondingly). There was a trend towards a lower diastolic pulmonary arterial pressure after PADN in group I when compared with group III during repeat PH induction (34.4 ± 2.9 vs 38.0 ± 0.8; P = 0.06). Despite the presence of severe PA wall damage at the RF application sites, S100 expression was preserved in the majority of PA specimens. The presence of high-grade PA lesions was associated with HR acceleration after PADN (ρ = 0.68, p = 0.03). No significant correlation was found between the grade of PA lesion severity and PA pressure after PADN with or without PH induction. CONCLUSIONS: Extended PADN does not affect PH induction using TXA2. Significant PA adventitia damage is associated with HR acceleration after PADN. Possible delayed effects of PADN on perivascular nerves and pulmonary hemodynamics require further research in chronic experiments.

Pulmonary and Systemic Hemodynamics following Multielectrode Radiofrequency Catheter Renal Denervation in Acutely Induced Pulmonary Arterial Hypertension in Swine.

OBJECTIVE: We aimed to assess the effects of renal denervation (RDN) on systemic and pulmonary hemodynamics in a swine model of thromboxane A2- (TXA2-) induced pulmonary arterial hypertension (PAH). METHODS: The study protocol comprised two PAH inductions with a target mean pulmonary artery pressure (PAP) of 40 mmHg at baseline and following either the RDN or sham procedure. Ten Landrace pigs underwent the first PAH induction; then, nine animals were randomly allocated in 1 : 1 ratio to RDN or sham procedure; the second PAH induction was performed in eight animals (one animal died of pulmonary embolism during the first PAH induction, and one animal died after RDN). In the RDN group, ablation was performed in all available renal arteries, and balloon inflation within artery branches was performed in controls. An autopsy study of the renal arteries was performed. RESULTS: At baseline, the target mean PAP was achieved in all animals with 25.0 [20.1; 25.2] mcg of TXA2. The second PAH induction required the same mean TXA2 dose and infusion time. There was no statistically significant difference in the mean PAP at second PAH induction between the groups (39.0 ± 5.3 vs. 39.75 ± 0.5 mmHg, P > 0.05). In the RDN group, the second PAH induction resulted in a numerical but insignificant trend toward a decrease in the mean systemic blood pressure and systemic vascular resistance, when compared with the baseline induction (74 ± 18.7 vs. 90.25 ± 28.1 mmHg and 1995.3 ± 494.3 vs. 2433.7 ± 1176.7 dyn∗sec∗cm -5, P > 0.05, respectively). No difference in hemodynamic parameters was noted in the sham group between the first and second PAH induction. Autopsy demonstrated artery damage in both groups, but RDN resulted in more severe lesions. CONCLUSIONS: According to our results, RDN does not result in significant acute pulmonary or systemic hemodynamic changes in the TXA2-induced PAH model. The potential chronic effects of RDN on PAH require further research.

An Ovine Model for Percutaneous Pulmonary Artery Laser Denervation: Perivascular Innervation and Ablation Lesion Characteristics.

BACKGROUND: Pulmonary artery denervation (PADN) is an evolving interventional procedure capable to reduce pulmonary artery (PA) pressure. We aimed to compare PA nerve distribution in different specimens and assess the feasibility of an ovine model for a denervation procedure and evaluate the acute changes induced by laser energy. METHODS: The experiment was divided into two phases: (1) the analysis of PA nerve distribution in sheep, pigs, and humans using histological and immunochemical methods; (2) fiberoptic PADN in sheep and postmortem laser lesion characteristics. RESULTS: PA nerve density and distribution in sheep differ from humans, although pigs and sheep share similar characteristics, nerve fibers are observed in the media layer, adventitia, and perivascular tissue in sheep. Necrosis of the intima and focal hemorrhages within the media, adventitia, and perivascular adipose tissue were evidenced post laser PADN. Among the identified lesions, 40% reached adventitia and could be classified as effective for PADN. The use of 20 W ablation energy was safer and 30 W-ablation led to collateral organ damage. CONCLUSIONS: An ovine model is suitable for PADN procedures; however, nerve distribution in the PA bifurcation and main branches differ from human PA innervation. Laser ablation can be safely used for PADN procedures.

Stimulation Mapping of the Pulmonary Artery for Denervation Procedures: an Experimental Study.

Transcatheter pulmonary artery denervation (PADN) has been developed for the correction of pulmonary hypertension. We investigated pulmonary artery stimulation mapping and its role in PADN procedures. Artery stimulation was performed in 17 Landrace pigs. Low-frequency stimulation defined areas of ventricular and atrial capture. High-frequency stimulation evoked the following responses: sinus rhythm slowing and/or atrial rhythm acceleration in 59% of animals, phrenic nerve capture in 100%, and laryngeal recurrent nerve capture in 23%. The sites with evoked heart rate responses were marked by discrete radiofrequency ablations (RFA). An autopsy showed nerves in the adventitia and perivascular fat under the RFA sites, and the lack of muscarinic-1, tyrosine hydroxylase, and dopamine-5 receptors' expression. During PADN, areas adjacent to the course of phrenic and recurrent laryngeal nerves should be avoided. RFA at points with heart rate responses leads to the non-reproducibility of evoked reactions and the disappearance of neural markers' expression. Graphical abstract.

Extended Renal Artery Denervation Is Associated with Artery Wall Lesions and Acute Systemic and Pulmonary Hemodynamic Changes: A Sham-Controlled Experimental Study.

OBJECTIVES: We sought to assess acute changes in systemic and pulmonary hemodynamics and microscopic artery lesions following extended renal artery denervation (RDN). BACKGROUND: RDN has been proposed to reduce sympathetic nervous system hyperactivation. Although the effects of RDN on systemic circulation and overall sympathetic activity have been studied, data on the impact of RDN on pulmonary hemodynamics is lacking. METHODS: The study comprised 13 normotensive Landrace pigs. After randomization, 7 animals were allocated to the group of bilateral RDN and 6 animals to the group of a sham procedure (SHAM). Hemodynamic measures, cannulation, and balloon-based occlusion of the renal arteries were performed in both groups. In the RDN group, radiofrequency ablation was performed in all available arteries and their segments. An autopsy study of the renal arteries was carried out in both groups. RESULTS: The analysis was performed on 12 pigs (6 in either group) since pulmonary thromboembolism occurred in one case. A statistically significant drop in the mean diastolic pulmonary artery pressure (PAP) was detected in the RDN group when compared with the SHAM group (change by 13.0 ± 4.4 and 10.0 ± 3.0 mmHg, correspondingly; P = 0.04). In 5 out of 6 pigs in the RDN group, a significant decrease in systemic systolic blood pressure was found, when compared with baseline (98.8 ± 17.8 vs. 90.2 ± 12.6 mmHg, P = 0.04), and a lower mean pulmonary vascular resistance (PVR) (291.0 ± 77.4 vs. 228.5 ± 63.8 dyn∗sec∗cm-5, P = 0.03) after ablation was found. Artery dissections were found in both groups, with prevalence in animals after RDN. CONCLUSIONS: Extensive RDN leads to a rapid and significant decrease in PAP. In the majority of cases, RDN is associated with an acute lowering of systolic blood pressure and PVR. Extended RDN is associated with artery wall lesions and thrombus formation underdiagnosed by angiography.

Telocytes in the human sinoatrial node.

The sinoatrial node (SAN) is composed mostly of pacemaker, transitional and Purkinje-like cells. Pacemaker cells, especially in the centre of the SAN, are surrounded by dense fibrous tissue and do not have any contact with transitional cells. We hypothesize that the SAN contains telocytes that have contacts with pacemaker cells and contractile myocardium. Immunohistochemistry using antibodies against HCN4 and antibody combinations against CD34 and HCN4 was carried out on 12 specimens. Confocal laser scanning microscopy (CLSM) with two mixtures of primary antibodies, namely CD34/S100 and vimentin/S100, was performed in three cases. In two cases, CLSM was carried out with CD117 antibody. Specimens for electron microscopy and immunocytochemistry with HCN4 immunogold labelling were taken from another three patients. In our study, we found cells with the immunophenotype of telocytes in the SAN. There were twice as many of these cells in the centre of the SAN as in the periphery (20.3 ± 4.8 versus 10.8 ± 4.4 per high-power field). They had close contact with pacemaker cells and contractile cardiomyocytes and expressed HCN4. The ultrastructural characteristics of these cells are identical to those of telocytes observed earlier in other organs. Our study provides evidence that telocytes are present in the SAN.

Plurihormonal cells of normal anterior pituitary: Facts and conclusions.

INTRODUCTION: plurihormonality of pituitary adenomas is an ability of adenoma cells to produce more than one hormone. After the immunohistochemical analysis had become a routine part of the morphological study, a great number of adenomas appeared to be multihormonal in actual practice. We hypothesize that the same cells of a normal pituitary gland releases several hormones simultaneously. OBJECTIVE: To analyse a possible co-expression of hormones by the cells of the normal anterior pituitary of adult humans in autopsy material. MATERIALS AND METHODS: We studied 10 pituitary glands of 4 women and 6 men with cardiovascular and oncological diseases. Double staining immunohistochemistry using 11 hormone combinations was performed in all the cases. These combinations were: prolactin/thyroid-stimulating hormone (TSH), prolactin/luteinizing hormone (LH), prolactin/follicle-stimulating hormone (FSH), prolactin/adrenocorticotropic hormone (ACTH), growth hormone (GH)/TSH, GH/LH, GH/FSH, GH/ACTH, TSH/LH, TSH/FSH, TSH/ACTH. Laser Confocal Scanning Microscopy with a mixture of primary antibodies was performed in 2 cases. These mixtures were ACTH/prolactin, FSH/prolactin, TSH/prolactin, ACTH/GH, and FSH/GH. RESULTS: We found that the same cells of the normal adenohypophysis can co-express prolactin with ACTH, TSH, FSH, LH; GH with ACTH, TSH, FSH, LH, and TSH with ACTH, FSH, LH. The comparison of the average co-expression coefficients of prolactin, GH and TSH with other hormones showed that the TSH co-expression coefficient was significantly the least (9,5±6,9%; 9,6±7,8%; 1,0±1,3% correspondingly). CONCLUSION: Plurihormonality of normal adenohypophysis is an actually existing phenomenon. Identification of different hormones in pituitary adenomas enables to find new ways to improve both diagnostic process and targeted treatment.

Histological evidence of inflammatory reaction associated with fibrosis in the atrial and ventricular walls in a case-control study of patients with history of atrial fibrillation.

AIMS: Chronic inflammation in the atrial myocardium was shown to play an important role in the development of atrial fibrosis in patients with atrial fibrillation (AF). However, it is not clear to what extent atrial inflammatory reaction associated with AF extends on the ventricular myocardium. Our aim was to assess the extent of fibrosis and lymphomononuclear infiltration in human ventricular myocardium and explore its association with AF. METHODS AND RESULTS: Medical records from consecutive autopsies were checked for presence of AF. Heart specimens from 30 patients died from cardiovascular causes (64 ± 12 years, 17 men) were collected in three equal groups: no AF, paroxysmal AF, and permanent AF. Tissue samples were taken from the Bachmann's bundle, crista terminalis, posterior left atrium, left ventricle and right ventricle free walls and stained with Masson's trichrome for analysis of fibrosis extent. Immunohistochemistry was performed using antibodies against CD3- and CD45-antigens and quantified as number of antigen-positive cells per 1 mm2. Fibrosis extent, CD3+ and CD45+ cell counts were elevated in AF patients at all sites (P < 0.001 for all). Fibrosis extent demonstrated correlation with both CD3+ and CD45+ cell counts in the right (r = 0.781, P < 0.001 for CD45+ and r = 0.720, P < 0.001 for CD3+) and the left (r = 0.515, P = 0.004 for CD45+ and r = 0.573, P = 0.001 for CD3+) ventricles. Neither fibrosis nor inflammatory cell count showed association with either age or comorbidities. CONCLUSION: Histological signs of chronic inflammation affecting ventricular myocardium are strongly associated with AF and demonstrate significant correlation with fibrosis extent that cannot be explained by cardiovascular comorbidities otherwise.

Biatrial tachycardia following linear anterior wall ablation for the perimitral reentry: incidence and electrophysiological evaluations.

INTRODUCTION: A left atrial (LA) anterior ablation line (AnL), connecting the mitral annulus and right pulmonary veins or a roof line, has been suggested as an alternative to mitral isthmus (MI) ablation for perimitral flutter (PMF). Theoretically, the AnL can exclude the LA septal wall from the reentrant circle, and lead to involvement of the right atrium (RA) in a tachycardia (AT) mechanism. METHODS AND RESULTS: Among 807 patients undergoing atrial fibrillation ablation, PMF was diagnosed in 28 subjects, and AnL was performed in 13, and MI ablation in 15 cases. In 4 (31%) patients, AnL resulted in abrupt AT cycle length prolongation, which was associated with the development of a clockwise biatrial tachycardia (bi-AT). The bi-AT propagated along the lateral and posterior mitral annulus, entered the RA via the coronary sinus, and after activating the RA septum reentered the LA over the Bachmann's bundle. The bi-AT was terminated by ablation in Bachmann's bundle insertion areas in the RA or LA. No bi-AT was documented in the MI group. One patient in the AnL group died of stroke in 10 days following the procedure. Anatomic evaluation showed that at the level of the AnL the RA anteroseptal area was separated from the LA by the aortic root, and was free from ablation damage. CONCLUSION: A bi-AT can develop when an AnL is created for PMF termination. Biatrial entrainment mapping facilitates diagnosis. Termination of the bi-AT is feasible when ablated from either RA or LA.

Evidence of specialized tissue in human interatrial septum: histological, immunohistochemical and ultrastructural findings.

BACKGROUND: There is a paucity of information on structural organization of muscular bundles in the interatrial septum (IAS). The aim was to investigate histologic and ultrastructural organization of muscular bundles in human IAS, including fossa ovalis (FO) and flap valve. METHODS: Macroscopic and light microscopy evaluations of IAS were performed from postmortem studies of 40 patients. Twenty three IAS specimens underwent serial transverse sectioning, and 17--longitudinal sectioning. The transverse sections from 10 patients were immunolabeled for HCN4, Caveolin3 and Connexin43. IAS specimens from 6 other patients underwent electron microscopy. RESULTS: In all IAS specimens sections the FO, its rims and the flap valve had muscle fibers consisting of working cardiac myocytes. Besides the typical cardiomyocytes there were unusual cells: tortuous and horseshoe-shaped intertangled myocytes, small and large rounded myocytes with pale cytoplasm. The cells were aggregated in a definite structure in 38 (95%) cases, which was surrounded by fibro-fatty tissue. The height of the structure on transverse sections positively correlated with age (P = 0.03) and AF history (P = 0.045). Immunohistochemistry showed positive staining of the cells for HCN4 and Caveolin3. Electron microscopy identified cells with characteristics similar to electrical conduction cells. CONCLUSIONS: Specialized conduction cells in human IAS have been identified, specifically in the FO and its flap valve. The cells are aggregated in a structure, which is surrounded by fibrous and fatty tissue. Further investigations are warranted to explore electrophysiological characteristics of this structure.

Structural abnormalities in atrial walls are associated with presence and persistency of atrial fibrillation but not with age.

OBJECTIVES: The purpose of this study was to assess the association between structural changes in human atria, age, and history of atrial fibrillation (AF). BACKGROUND: Development of fibrosis in atrial walls is associated with deterioration of atrial conduction and predisposes to AF in experiment. Human data, however, are scarce, and whether fibrosis is a cause or consequence of AF is not known. METHODS: Medical records for consecutive autopsies were checked for AF history and duration. Atrial specimens from 30 patients (ages 64 ± 12 years) were collected in 3 equal age-matched groups as patients without AF history, with paroxysmal AF, or with permanent AF. Tissue samples were obtained at the level of superior pulmonary veins, inferior pulmonary veins, center of posterior left atrial wall, terminal crest, and Bachmann's bundle. Histology sections were assessed for extent of fibrosis, fatty tissues, and inflammatory infiltration at each location. RESULTS: No correlation was observed between age and fibrosis at any location. Fibrosis extent and fatty infiltration were twofold to threefold higher at all locations in patients with history of AF and correlated with lymphomononuclear infiltration. Patients with permanent AF had greater fibrosis extent than did patients with paroxysmal AF. CONCLUSIONS: In post-mortem material, structural changes in the atria were not associated with age, but were significantly correlated with presence of AF and its severity. Our findings suggest that age-related changes per se are unlikely to be the sole cause of advanced fibrosis underlying AF.