Evaluation of an AI system for the detection of diabetic retinopathy from images captured with a handheld portable fundus camera: the MAILOR AI study.

OBJECTIVES: To evaluate the performance of an artificial intelligence (AI) system (Pegasus, Visulytix Ltd., UK*) at the detection of diabetic retinopathy (DR) from images captured by a handheld portable fundus camera. METHODS: A cohort of 6404 patients (~80% with diabetes mellitus) was screened for retinal diseases using a handheld portable fundus camera (Pictor Plus, Volk Optical Inc., USA) at the Mexican Advanced Imaging Laboratory for Ocular Research. The images were graded for DR by specialists according to the Scottish DR grading scheme. The performance of the AI system was evaluated, retrospectively, in assessing referable DR (RDR) and proliferative DR (PDR) and compared with the performance on a publicly available desktop camera benchmark dataset. RESULTS: For RDR detection, Pegasus performed with an 89.4% (95% CI: 88.0-90.7) area under the receiver operating characteristic (AUROC) curve for the MAILOR cohort, compared with an AUROC of 98.5% (95% CI: 97.8-99.2) on the benchmark dataset. This difference was statistically significant. Moreover, no statistically significant difference was found in performance for PDR detection with Pegasus achieving an AUROC of 94.3% (95% CI: 91.0-96.9) on the MAILOR cohort and 92.2% (95% CI: 89.4-94.8) on the benchmark dataset. CONCLUSIONS: Pegasus showed good transferability for the detection of PDR from a curated desktop fundus camera dataset to real-world clinical practice with a handheld portable fundus camera. However, there was a substantial, and statistically significant, decrease in the diagnostic performance for RDR when using the handheld device.

Segmentation of phase contrast microscopy images based on multi-scale local Basic Image Features histograms.

Phase contrast microscopy (PCM) is routinely used for the inspection of adherent cell cultures in all fields of biology and biomedicine. Key decisions for experimental protocols are often taken by an operator based on typically qualitative observations. However, automated processing and analysis of PCM images remain challenging due to the low contrast between foreground objects (cells) and background as well as various imaging artefacts. We propose a trainable pixel-wise segmentation approach whereby image structures and symmetries are encoded in the form of multi-scale Basic Image Features local histograms, and classification of them is learned by random decision trees. This approach was validated for segmentation of cell versus background, and discrimination between two different cell types. Performance close to that of state-of-the-art specialised algorithms was achieved despite the general nature of the method. The low processing time ( < 4 s per 1280 × 960 pixel images) is suitable for batch processing of experimental data as well as for interactive segmentation applications.

Mutual interaction of special phytoestrogenic compounds, their synthetic carboxy-derivatives and the less-calcemic vitamin D analog activities in human derived female cultured osteoblasts.

Cultured female-derived human bone cells (hObs) responded by different parameters to different phytoestrogenic and vitamin D compounds. Pre- and post-menopausal hObs express ERα and ERß mRNA with higher abundance of ERα. Pre-treatment with the less-calcemic vitamin D analog JKF 1624F(2)-2 (JKF) upregulated responsiveness to estrogens via modulation of ERs expression. These estrogenic compounds induce the expression and activity of 25 hydroxy-vitamin D(3)-1α hydroxylase (1OHase). We now analyzed the effects of carboxy-genistein (cG), carboxy-biocainin A (cBA) and carboxy-daidzein (cD), of BA, D or G and of licorice derived compounds glabridin (Glb) and glabrene (Gla) and estradiol-17ß (E(2)) on DNA synthesis, creatine kinase specific activity (CK), intracellular and membranal E(2) binding and their modulations by JKF in hObs. We also analyzed modulation by phytoestrogenic compounds of 1OHase mRNA expression and activity. We showed that: (1) all compounds stimulated DNA synthesis and CK. (2) JKF and all estrogenic compounds modulated ERα and ERß mRNA expression. (3) Pre-treatment with JKF increased DNA synthesis and CK responses only to E(2), D, G and Gla. (4) JKF increased the intracellular competitive binding only of E(2), D and G. (5) JKF abolished the membranal binding of all protein-bound estrogens. (6) JKF and all estrogenic compounds except the protein-bound ones up-regulated 1OHase expression and activity. In conclusion phytoestrogens and their analogs increase DNA synthesis and CK, and lead to increased production of 1,25(OH)(2)D(3) in hObs, while pre-treatment with JKF modulates the effect of estrogenic compounds via regulation of ERs mRNA expression in a yet unclear mechanism.

Amniotic fluid and plasma levels of parathyroid hormone-related protein and hormonal modulation of its secretion by amniotic fluid cells.

Parathyroid hormone-related (PTHrP), the major mediator of humoral hypercalcemia of malignancy, may also regulate placental calcium flux, uterine contraction and fetal tissue development. In the present study, we demonstrated that the mean immunoreactive PTHrP concentrations in amniotic fluid at mid-gestation (21.2 +/- 3.7 pmol/l) and at term (19.0 +/- 2.7 pmol/l) were 13-16-fold higher than levels measured in either fetal (1.6 +/- 0.1 pmol/l) or maternal plasma (1.4 +/- 0.3 pmol/l) at term and equal to levels found in plasma of patients with humoral hypercalcemia of malignancy. In vitro studies pointed to three possible sources of PTHrP in amniotic fluid: cultured amniotic fluid cells, cells derived from the amniotic membrane overlying the placenta and placental villous core mesenchymal cells. Treatment of cultured amniotic fluid cells with human prolactin, human placental lactogen (hPL) or human growth hormone (100 micrograms/l) increased PTHrP secretion after 24 h by 43%, 109% and 90%, respectively. Insulin-like growth factors I and II (100 micrograms/l), insulin (100 micrograms/l) and epidermal growth factor (EGF) (10 micrograms/l) increased PTHrP secretion by 53%, 46%, 68% and 118%, respectively. The stimulation of PTHrP secretion by EGF or by hPL was both time- and dose-dependent. In contrast, calcitriol and dexamethasone (10 nmol/l) decreased PTHrP secretion by 32% and 75%, respectively. Estradiol, progesterone, dihydrotestosterone and human chorionic gonadotropin had no effect on PTHrP secretion. These findings support the notion that PTHrP may play a physiological role in the uteroplacental unit and demonstrate that human amniotic fluid cells could be a useful model for studying the regulation of PTHrP production and secretion by hormones and growth factors.

Anti-idiotypic antibody as an oestrogen mimetic: removal of Fc fragment converts agonist to antagonist.

Previous studies indicated that the anti-idiotypic antibody (clone 1D5) caused an increase in uterine creatine kinase (CK) activity when administered in vivo to immature female rats, indicating that the antibody has oestrogenic-like activity. It was, therefore, of interest to investigate the structural requirements of clone 1D5 to act as an oestrogen mimetic in an in vitro model system. In the present study, the effect of clone 1D5 and its proteolytic fragments, F(ab')2, Fab' and Fc on CK activity was examined in cultured skeletal cells having functional oestrogen receptor (ER). Incubation of female-derived calvaria cells or epiphyseal cartilage cells with clone 1D5 (8.33 nM) or oestradiol (E2) (30 nM) for 24 h caused a significant increase in CK activity, indicating that clone 1D5 acted as an agonist. On the other hand, incubation of male-derived calvaria cells devoid of a functional ER with clone 1D5 or E2 did not have any effect on CK activity. Incubation of female-derived calvaria cells with clone 1D5 and E2 did not result in any further increase in CK activity, whereas dihydrotestosterone (DHT) did not alter the response to clone 1D5. The CK response to clone 1D5, in female-derived calvaria cells was time- and dose-dependent and could be inhibited in a dose-dependent manner by the oestrogen antagonist tamoxifen. In contrast, the proteolytic fragments of clone 1D5, the F(ab')2 dimer (12 nM) and the Fab' monomer (24 nM), and the Fc fragment (28 nM) did not have E2-like activity in these cells.(ABSTRACT TRUNCATED AT 250 WORDS)

Decreased bone density in elderly men treated with the gonadotropin-releasing hormone agonist decapeptyl (D-Trp6-GnRH).

Administration of GnRH agonist analogs to women may result in a hypoestrogenic state and bone mass reduction. In the present study we examined bone mineral density (BMD) and parameters of mineral metabolism in elderly men with benign prostatic hyperplasia before and during a hypoandrogenic state induced by the long-acting GnRH agonist D-Trp6-LHRH (decapeptyl). Our results showed that decapeptyl treatment caused a significant decrease in serum testosterone concentrations in all patients and resulted in a significant decrease in individual vertebral BMD in 10 of 17 patients. A significant decrease in BMD was observed in 5 patients after 6 months of treatment. Another 5 patients showed a decreased BMD only after 12 months. The mean serum concentrations of osteocalcin, phosphorus, and alkaline phosphatase activity increased after 6-12 months of treatment with decapeptyl. Serum calcium, vitamin D metabolites, and PTH concentrations remained unchanged during treatment. Urinary calcium excretion was slightly, but not significantly, increased after 6 months of treatment. These results demonstrate that long-acting GnRH agonist treatment may cause high turnover accelerated bone loss in some men during the first year of GnRH agonist treatment, as has been previously shown in women.

Prolonged treatment with finasteride (a 5 alpha-reductase inhibitor) does not affect bone density and metabolism.

OBJECTIVE: Since it is not clear whether testosterone or dihydrotestosterone is the active hormone in bone metabolism, we wished to assess the effect of finasteride, a 5 alpha-reductase inhibitor, or vertebral bone mineral density and parameters of bone and mineral metabolism. DESIGN: Patients were treated in a randomized, double-blind controlled study with either placebo, 1 or 5 mg/day finasteride. PATIENTS: Twenty-three men with benign prostatic hyperplasia (BPH) were included in this study; eight received placebo, seven were allocated to treatment with 1 mg/day, and eight to 5 mg/day finasteride for 12 months. MEASUREMENTS: Vertebral bone mineral density was measured at the lumbar spine by dual energy X-ray bone densitometry. Serum calcium, phosphorus, parathyroid hormone, osteocalcin and vitamin D metabolites were measured regularly. Urinary calcium and creatinine excretion were monitored as well. RESULTS: Finasteride caused a significant decrease in serum dihydrotestosterone after 6 and 12 months, but no effect on serum testosterone. Vertebral bone mineral density remained unaltered. None of the other parameters monitored were affected except for a small unexplained increase in 1.25-dihydroxyvitamin D in the group receiving 5 mg finasteride/day. CONCLUSIONS: Testosterone is probably the active hormone in bone metabolism. However, oestradiol, the product of testosterone aromatization (which remains unaltered under finasteride) may yet be another possible responsible steroid in the maintenance of bone density. We can also not rule out that the small amount of dihydrotestosterone remaining under finasteride administration is sufficient for maintaining normal bone metabolism.

Reciprocal modulation by sex steroid and calciotrophic hormones of skeletal cell proliferation.

We have demonstrated previously that 17 beta-estradiol (E2) stimulates cell proliferation in skeletal tissues, as measured by increased DNA synthesis and creatine kinase (CK) specific activity, and that calciotrophic hormones modulate E2 activity in rat osteoblastic sarcoma cells (ROS 17/2.8). Moreover, E2 failed to stimulate DNA synthesis in vitamin D-depleted female rat bone in the absence of prior i.p. injections of 1.25(OH)2D3. We have, therefore, studied the effects of pretreatment of cells by one hormone on their response to challenge by a second hormone. We now report reciprocal interactions of sex steroids and other hormones modulating bone formation on cell proliferation parameters in primary bone and cartilage cell cultures: these interactions can selectively augment or diminish cell responsiveness to a given hormone. Pretreatment of rat epiphyseal cartilage cell cultures with 1.25(OH)2D3, 24.25(OH)2D3 or parathyroid hormone (PTH) for 5 days, followed by E2 treatment for 24h, resulted in increased DNA synthesis compared to cultures pretreated with vehicle. Prostaglandin (PGE2) pretreatment blocked further response to E2. In the reciprocal case, rat epiphyseal cartilage cells, pretreated with E2, showed an increased response to PTH, a loss of the response to PGE2 or 24.25(OH)2D3 and an inhibition of CK activity and DNA synthesis by 1.25(OH)2D3, similar to the characteristic inhibitory action of 1.25(OH)2D3 in osteoblasts. By contrast, rat epiphyseal cartilage cells pretreated with testosterone showed no changes in response to PTH, 24.25(OH)2D3 or PGE2 and a decreased response to E2, but were stimulated by 1.25(OH)2D3. Rat embryo calvaria cell cultures behaved similarly to epiphyseal cartilage cultures except that 24.25(OH)2D3 pretreatment did not increase the response to E2. Reciprocally, pretreatment with E2 before exposure to calciotrophic hormones did not change the responses of rat embryo calvaria cell cultures to 1.25(OH)2D3 or 24.25(OH)2D3. These findings suggest that the mutual interactions between calciotrophic hormones and E2, demonstrated here in vitro, could selectively affect the responses of bone and cartilage cells to E2 by several mechanisms. These possibilities include increased E2 receptors and E2-stimulated differentiation of cartilage cells to more E2 responsive cells showing some characteristics of osteoblasts.

Prenatal diagnosis of vitamin D-dependent rickets, type II: response to 1,25-dihydroxyvitamin D in amniotic fluid cells and fetal tissues.

Vitamin D-dependent rickets type II (VDDR-II; hereditary resistance to 1,25-dihydroxyvitamin D3 [1,25(OH)2D3]), an autosomal recessive genetic disease that results from a failure to respond to 1,25-(OH)2D3, is characterized by severe rickets, hypocalcemia, growth retardation, and high prevalence of alopecia. We used amniotic fluid cells in the 17th week of gestation to detect VDDR-II in fetuses at risk for the defect. First, we demonstrated in cells obtained from 15 control pregnancies the presence of a specific high affinity 1,25-(OH)2D3 receptor (Kd = 0.3 x 10(-11) mol/L; maximal number of binding sites, 6.1 fmol/mg protein) and 1,25-(OH)2D3-induced 25-hydroxyvitamin D3-24-hydroxylase activity (up to 30-fold increase). Amniotic fluid cells from a woman who had already given birth to a child with VDDR-II contained receptors that bound [3H]1,25-(OH)2D3 normally and responded to 1,25-(OH)2D3 stimulation with a 10-fold increase in 24-hydroxylase activity. The fetus was, therefore, judged unaffected, and a normal baby girl was born. At the age of 16 months she did not demonstrate clinical or biochemical features of VDDR-II. Amniotic fluid cells from another mother of a child with VDDR-II were unable to bind [3H]1,25-(OH)2D3, and the hormone failed to stimulate 24-hydroxylase activity. VDDR-II in this fetus was confirmed after termination of pregnancy by the total inability of 1,25-(OH)2D3 to stimulate 24-hydroxylase activity in tissue explants and cell cultures prepared from the fetus's kidney and skin. In contrast, tissues from dead control fetuses responded to stimulation by 1,25-(OH)2D3 with a 3- to 10-fold increase in 24-hydroxylase activity. Fetal kidney and skin explants and cell cultures also synthesized a [3H]1,25-(OH)2D3-like metabolite from [3H]25-OHD3 as early as the 17th week of gestation. 1,25-(OH)2D3 (10 nM) decreased the in vitro synthesis of the [3H]1,25-(OH)2D3-like metabolite in tissues from dead control fetuses, but not from the affected fetus. Thus, human fetuses at midgestation already have the regulatory mechanisms responsive to 1,25-(OH)2D3 present postnatally. The prenatal diagnosis of VDDR-II is now possible and is indicated in a high risk family.

Vitamin D deficiency in elderly patients in a general hospital.

Serum 25-hydroxyvitamin D (25-OHD) levels were measured in 338 elderly patients admitted to the Geriatric Medicine Departments of a general hospital in Israel in the course of one year. The mean (+/- SD) serum 25-OHD levels were significantly lower (P less than .01) in the elderly patients (13.5 +/- 8.9 ng/mL) than in healthy young controls (24.7 +/- 6.1 ng/mL). One hundred ten patients (35.5%) were either vitamin D deficient (25-OHD less than 5 ng/mL) or had borderline serum levels of 25-OHD (5-9 ng/mL). The mean (+/- SD) serum 25-OHD concentration of patients who were completely mobile before hospitalization was 15.5 +/- 8.8 ng/mL (n = 239). In patients mainly immobilized but able to leave the house occasionally, it was 10.2 +/- 6.3 ng/mL (n = 84) and of bed-ridden patients, it was 5.2 +/- 3.2 ng/mL (n = 15). No correlation was found between serum 25-OHD levels and the patients' age or serum calcium, phosphorus, alkaline phosphatase, and albumin values. Thus, in order to detect vitamin D deficiency in the elderly, it is necessary to measure serum 25-OHD concentration. The results demonstrate that vitamin D deficiency is common among elderly patients even in sunny climates and indicate the need for development of effective programs of prevention and treatment.