Chemo (Spain)

Artificial intelligence (AI), especially through the development of deep learning techniques like convolutional neural networks (CNNs), has revolutionized numerous fields. CNNs, introduced by Yann LeCun in the 1990s (Hubbard, W.; Jackel, L. D. Backpropagation Applied to Handwritten Zip Code Recognition. Neural Comput. 1989, 1 (4), 541- 551. https://doi.org/10.1162/neco.1989.1.4.541), have found applications in healthcare for medical diagnostics, autonomous vehicles in transportation, stock market prediction in finance, and image recognition in computer vision to name just a few. Similarly, in analytical chemistry, deep learning has enhanced data analysis from techniques like MS spectrometry, NMR, fluorescence spectroscopy, and chromatography. Another AI branch, Natural Language Processing (NLP), has surged recently with the advent of Large Language Models (LLMs), such as OpenAI's ChatGPT. This paper demonstrates the application of an LLM via a smartphone to conduct multivariate data analyses, in an interactive conversational manner, of a hyper-spectral imaging data set from laser-induced breakdown spectroscopy (LIBS). We demonstrate the potential of LLMs to process and analyze data sets, which automatically generate and execute code in response to user queries, and anticipate their growing role in the future of analytical chemistry.

With the aim to shorten the time for diagnosis and accelerate access to correct management, a non-invasive diagnostic test for endometriosis was developed and validated. The IVD test combines an ELISA test kit to quantify CA125 and BDNF concentrations in serum and a data treatment algorithm hosted in medical software processing results from the ELISA test and responses to six clinical variables. Serum samples and clinical variables extracted from psychometric questionnaires from 77 patients were collected from the Oxford Endometriosis CaRe Centre biobank (UK). Case/control classification was performed based on laparoscopy and histological verification of the excised lesions. Biomarkers serum concentrations and clinical variables were introduced to the software, which generates the qualitative diagnostic result ("positive" or "negative"). This test allowed the detection of 32% of cases with superficial endometriosis, which is an added value given the limited efficacy of existing imaging techniques. Even in the presence of various confounding medical conditions, the test maintained a specificity of 100%, supporting its suitability for use in patients with underlying medical conditions.

Abstract With the aim to shorten the time for diagnosis and accelerate access to correct management, a non-invasive diagnostic test for endometriosis was developed and validated. The IVD test combines an ELISA test kit to quantify CA125 and BDNF concentrations in serum and a data treatment algorithm hosted in medical software processing results from the ELISA test and responses to six clinical variables. Serum samples and clinical variables extracted from psychometric questionnaires from 77 patients were collected from the Oxford Endometriosis CaRe Centre biobank (UK). Case/control classification was performed based on laparoscopy and histological verification of the excised lesions. Biomarkers serum concentrations and clinical variables were introduced to the software, which generates the qualitative diagnostic result (“positive” or “negative”). This test allowed the detection of 32% of cases with superficial endometriosis, which is an added value given the limited efficacy of existing imaging techniques. Even in the presence of various confounding medical conditions, the test maintained a specificity of 100%, supporting its suitability for use in patients with underlying medical conditions.

Mice are the most commonly used laboratory animal, yet there are limited studies which investigate the effects of repeated handling on their welfare and scientific outcomes. Furthermore, simple methods to evaluate distress in mice are lacking, and specialized behavioral or biochemical tests are often required. Here, two groups of CD1 mice were exposed to either traditional laboratory handling methods or a training protocol with cup lifting for 3 and 5 weeks. The training protocol was designed to habituate the mice to the procedures involved in subcutaneous injection, e.g., removal from the cage, skin pinch. This protocol was followed by two common research procedures: subcutaneous injection and tail vein blood sampling. Two training sessions and the procedures (subcutaneous injection and blood sampling) were video recorded. The mouse facial expressions were then scored, focusing on the ear and eye categories of the mouse grimace scale. Using this assessment method, trained mice expressed less distress than the control mice during subcutaneous injection. Mice trained for subcutaneous injection also had reduced facial scores during blood sampling. We found a clear sex difference as female mice responded to training faster than the male mice, they also had lower facial scores than the male mice when trained. The ear score appeared to be a more sensitive measure of distress than the eye score, which may be more indicative of pain. In conclusion, training is an important refinement method to reduce distress in mice during common laboratory procedures and this can best be assessed using the ear score of the mouse grimace scale.