Achieving Expert-Level Interpretation of Serum Protein Electrophoresis through Deep Learning Driven by Human Reasoning.

BACKGROUND: Serum protein electrophoresis (SPE) is a common clinical laboratory test, mainly indicated for the diagnosis and follow-up of monoclonal gammopathies. A time-consuming and potentially subjective human expertise is required for SPE analysis to detect possible pitfalls and to provide a clinically relevant interpretation. METHODS: An expert-annotated SPE dataset of 159 969 entries was used to develop SPECTR (serum protein electrophoresis computer-assisted recognition), a deep learning-based artificial intelligence, which analyzes and interprets raw SPE curves produced by an analytical system into text comments that can be used by practitioners. It was designed following academic recommendations for SPE interpretation, using a transparent architecture avoiding the "black box" effect. SPECTR was validated on an external, independent cohort of 70 362 SPEs and challenged by a panel of 9 independent experts from other hospital centers. RESULTS: SPECTR was able to identify accurately both quantitative abnormalities (r ≥ 0.98 for fractions quantification) and qualitative abnormalities [receiver operating characteristic-area under curve (ROC-AUC) ≥ 0.90 for M-spikes, restricted heterogeneity of immunoglobulins, and beta-gamma bridging]. Furthermore, it showed highly accurate at both detecting (ROC-AUC ≥ 0.99) and quantifying (r = 0.99) M-spikes. It proved highly reproducible and resilient to minor variations and its agreement with human experts was higher (κ = 0.632) than experts between each other (κ = 0.624). CONCLUSIONS: SPECTR is an algorithm based on artificial intelligence suitable to high-throughput SPEs analyses and interpretation. It aims at improving SPE reproducibility and reliability. It is freely available in open access through an online tool providing fully editable validation assistance for SPE.

[Examples of accreditation of serum and urinary proteins electrophoresis]. / Expériences d'accréditation des électrophorèses sériques et urinaires.

Serum proteins and urinary proteins electrophoresis are useful biological tests. They are often prescribed for the screening of monoclonal gammopathys and also during follow-up for treatment response. This test can be accredited according to standard NF ISO 15189 since laboratories use analysers and adapted reagents, but there are numerous protocols for the method validation. This paper present the results of a survey proposed in 2019 to biologists by CNBH and SFBC. The aim of this survey is to give biologists a choice among several protocols that have been positively evaluated by COFRAC.

[Immunoanalytical characteristics of HE4 protein]. / Caractéristiques immunoanalytiques de la protéine HE4.

Besides structural and physiological data of HE4 protein, this paper points out the optimal conditions for sampling, assays and interpretation of results for the management of ovarian cancer.

Immunoanalytical characteristics of fecal calprotectin.

Besides structural and physiological data of calprotectin, this paper points out the optimal conditions for sampling, fecal assays and interpretation of results.

[Immunoanalytical characteristics of galectin-3]. / Caractéristiques immunoanalytiques de la galectine-3.

This paper points out the structural and physiological data of galectin-3, and emphasizes its role in cardiac fibrosis and heart failure pathophysiology. Then we summarize the optimal conditions for sampling, assays and we discuss the interpretation of results.

Human hair follicle pigmentary unit as a direct target for modulators of melanogenesis, as studied by [14C]-2-thiouracil incorporation.

The purpose of this study was to evaluate human hair follicle melanogenic activity using the [14C]-2-thiouracil, which was known to incorporate into nascent melanins. Results obtained on pigmented, grey and non-pigmented hair follicles demonstrated that [(14)C]-2-TU incorporation was restricted to the melanogenic compartment with a strong accumulation located around dermal papilla and within the fibre of pigmented hair follicles. Quantitative analysis of [(14)C]-2-TU incorporation showed a significant increase in pigmented hair follicles upon stimulation with 1 microm forskolin concomitant to an increase in tyrosinase levels. A strong significant decrease in [14C]-2-TU incorporation was noted, when hair follicles were incubated with the tyrosinase competitive inhibitor kojic acid (200 microm). Incubation with the MC1-R agonist alpha-MSH (0.2 microm) did not induce a significant stimulation of hair melanogenesis. The present model could thus represent a useful new tool to identify modulators of human hair pigmentation.

Expression of NAD+ dependent 15-hydroxyprostaglandin dehydrogenase and protection of prostaglandins in human hair follicle.

NAD(+) dependent 15-hydroxyprostaglandin dehydrogenate (15-PGDH) catalyses oxidation of 15(S)-hydroxyl group of prostaglandins and as a result inactivates their physiological potential. Positive effects of prostaglandins or prostaglandin analogues were reported on terminal hair, vellus hair or eyelash growth and a complex prostaglandin network was recently described in human hair follicle. In the present study, we showed that 15-PGDH was expressed in human hair follicle mainly in melanocytes and keratinocytes, which brought us to consider this enzyme as a possible target to sustain local prostaglandin production. Using a recombinant enzymatic strategy, specific 15-PGDH inhibitors were screened. We identified a thiazolidine dione derivative exhibiting efficacy on follicular outer root sheath keratinocytes, since it concomitantly decreased the production of deactivated 13,14 dihydro 15-ketoprostaglandin F(2alpha) and sustained prostaglandin F(2alpha)in vitro production. In the context of recent interest in prostaglandins and prostaglandin analogues as hair regrowth agents, we postulated that the use of selected 15-PGDH inhibitors could reinforce or prolong the effect of these physiological mediators on hair and skin.

Comparative proteomic analysis of extracellular matrix proteins secreted by two types of skin fibroblasts.

The hair follicle dermal papilla is composed primarily of extracellular matrix (ECM) proteins secreted by resident fibroblasts. Dermal papilla is endowed with hair morphogenic properties, yet its composition is poorly characterized. In an attempt to understand its specificity better, we compared the protein composition of ECM secreted by cultured dermal papilla fibroblasts with that of dermal fibroblasts. ECM proteins are generally large, difficult to solubilize, and abundantly post-translationally modified. We thus implemented an original protocol for analyzing them: ECM samples were enzymatically digested directly in the culture flasks and analyzed by LC-MS/MS. Sequencing of proteolytic peptides by MS/MS yielded protein identification. The relative abundance of a given protein in dermal fibroblast versus dermal papilla samples was estimated by comparing proteolytic peptide intensities detected by MS. Using this approach, several matrix proteins were found to be present at markedly different levels in each ECM type; in particular, thrombospondin 1 and fibronectin appeared to be overrepresented in the dermal papilla fibroblast ECM. MS results were supported by Western blot and immunostaining experiments. In addition, peptide intensities were processed in two ways, which proved to favor either the quantification accuracy or the information precision at the sequence level.

Absence of TRP-2 in melanogenic melanocytes of human hair.

Skin and hair colour mostly depend on the activity of melanogenic melanocytes. Numerous proteins involved in melanocyte function have been identified including pMel-17, Mitf-M, Sox10, tyrosinase, tyrosinase related proteins-1 (TRP-1) and -2 (TRP-2). In the hair, melanogenic activity occurs only during the anagen phase of the hair cycle. In order to evaluate the implications of some known melanogenic proteins in human hair pigmentation, we performed immunohistochemical studies to reveal the expression of pMel-17, Mitf-M, tyrosinase, TRP-1 and TRP-2 in active bulb melanocytes of eumelanic brown and black anagen hairs of different ethnic origins, e.g. brown Caucasian, black Asian and African hairs. The labelling was compared with that observed in Caucasian and African scalp epidermis (interfollicular epidermis) melanocytes. We found that while pMel-17, TRP-1 and TRP-2 were expressed in epidermal melanocytes irrespective of ethnic origin and melanin content of the scalp epidermis, Mitf-M and tyrosinase expression were clearly evidenced only in pigmented epidermis, e.g. African scalps. Regarding human hair, pMel-17, Mitf-M, tyrosinase and TRP-1 were detected in a similar manner in active bulb melanocytes of brown and black hairs. In contrast and unexpectedly, TRP-2 could not be detected in hair bulb melanocytes, whatever the hair colour and ethnic origin. The lack of TRP-2 was further confirmed by western blot analyses. Reverse transcriptase-polymerase chain reaction (RT-PCR) performed on hair bulb mRNA demonstrated that Mitf-M, tyrosinase and TRP-1 amplimer signals were easily detected, whereas the TRP-2 amplimer signal was barely detectable. Furthermore Sox10 was not detected in hair bulb. Altogether our results suggest that the absence of detectable level of TRP-2 is due to transcriptional control in active melanocytes of human eumelanic hair bulbs. According to the absence of TRP-2 in melanin-producing melanocytes of brown and black hair bulbs, one must consider that eumelanogenesis as well as brown and black colour do not require TRP-2 expression in human hair.

How important is the impact of pressure recovery on routine evaluation of aortic stenosis? A clinical study in 91 patients.

BACKGROUND AND AIM OF THE STUDY: Experimental investigations and invasive studies conducted in small series of patients using specially designed high-fidelity micromanometer tip catheters have suggested that downstream pressure recovery (PR) within the aorta may significantly affect transvalvular pressure gradient (PG) measurement. The study aims were to evaluate in a large cohort of patients the extent of PR when transvalvular PGs are routinely measured by fluid-filled pigtail side-hole catheters (FPC) using pullback from the left ventricle to the ascending aorta (AO), and to analyze factors influencing PR. The influence of PR on the correlation between catheter and Doppler PG measurements was also assessed in a subset of patients. METHODS: Transvalvular PG were measured in 91 patients with aortic stenosis using FPC pullback with the catheter positioned at different sites within the ascending aorta. In 71 patients, Doppler echocardiography was obtained within 24 h of catheterization. RESULTS: Mean PR ranged from 0 to 20 mmHg, corresponding to a PR index (percent of maximal PG) ranging from 0 to 31%. PG was < 50 mmHg in nine of 61 patients (15%) with a PG > 50 mmHg at the origin of the aorta when further measurements were conducted with the catheter positioned more distally in the ascending aorta. PR index better correlated with the ratio of valve area to ascending AO cross-sectional area (r = 0.61, p = 0.001) than with valve area (r = 0.37, p = 0.001) and ascending AO cross-sectional area (0.27, p = 0.02) alone. Differences between Doppler- and catheter-predicted PG were minimized when correcting Doppler by non-invasively calculated PR (p < 0.0001). CONCLUSION: The magnitude of PR recorded in aortic stenosis by FPC, as used in most clinical catheterization laboratories, is low in the vast majority of patients. As predicted from fluid mechanics theory, the ratio of valve area to ascending AO cross-sectional area is the central determinant of PR. PR may affect the Doppler-catheter correlation in some patients.