Ping Cao

STUDY QUESTION: Can artificial intelligence (AI) and large language models (LLMs) effectively simplify patient education materials (PEMs) to advance reproductive genetic literacy and health equity? SUMMARY ANSWER: LLMs offer a promising approach to support healthcare professionals in generating effective, and simplified PEMs. WHAT IS KNOWN ALREADY: Reproductive genetic testing and counseling holds the potential to support a personalized approach to reduce the burden of genetic disorders. However, its uptake remains limited due to the complexity of the tests and the way that PEMs have been designed. This is more prominent in reproductive genetic testing, as vulnerability of patients may lead to over- or under-use of genetic testing technologies. STUDY DESIGN, SIZE, DURATION: We carried out a comparative observational study to evaluate the capacity of four AI/LLMs to simplify PEMs (n = 30) in reproductive genetics and assessing the clinical accuracy of simplified versions (n = 120) by experts (n = 30). Additionally, we devised a graphical user interface (GUI) to support real-time text simplification and readability analysis. PARTICIPANTS/MATERIALS, SETTING, METHODS: We collected 30 PEMs covering six topics in reproductive genetics from well-recognized platforms, such as WHO, MedlinePlus, and Johns Hopkins. Each PEM was processed by four AI/LLMs (GPT-3.5, GPT-4, Copilot, Gemini) using a fixed prompt, resulting in 120 simplified outputs. We measured readability improvements using five validated metrics, such as simple measure of gobbledygook, each capturing distinct textual characteristics such as sentence length and word complexity. To evaluate clinical reliability of the simplified outputs, a panel of experts (n = 30) in reproductive genetics independently scored each text (3 per text). MAIN RESULTS AND THE ROLE OF CHANCE: All four LLMs significantly improved the readability of the PEMs (P-values <0.001), reducing text complexity to an average 6th–7th grade reading level. While Gemini and Copilot achieved the highest improvement in readability scores, GPT-4 received the highest expert rating across all criteria—accuracy (4.1 ± 0.9), completeness (4.2 ± 0.8), and relevance of omissions (4.0 ± 0.9; P < 10−8). These findings highlight the importance of balancing readability with content integrity to support informed decision-making, as excessive simplification may compromise essential medical information. We devised an open-access GUI that provides real-time PEM simplification and readability analysis to support the integration of AI-assisted approaches in clinical practice (https://huggingface.co/spaces/CellularGenomicMedicine/HealthLiteracyEvaluator). LIMITATIONS, REASONS FOR CAUTION: Careful evaluation of LLM-simplified PEMs is required to ensure that simplification does not lead to omission of critical information. In addition, in this study, we report only the readability improvements of AI-generated texts and expert evaluations. To truly assess the potential of these tools in advancing reproductive genetic literacy and promoting health equity, real-world patient feedback is essential. WIDER IMPLICATIONS OF THE FINDINGS: Integrating AI/LLM into patient education strategies may advance health equity by improving understanding and facilitating informed decision-making. Thereby, more effective engagement of patients in reproductive genetic testing programs by assisting them with well-informed decision-making. STUDY FUNDING/COMPETING INTEREST(S): The EVA specialty program (KP111513) of MUMC+, the Horizon-Europe (NESTOR-101120075), the Estonian Research Council (PRG1076), the Horizon-2020 innovation (ERIN-EU952516) grants of the European Commission, the Swedish Research Council (grant no. 2024-02530), and the Novo Nordisk Foundation (grant no. NNF24OC0092384). The authors declare no conflict of interest relevant to this study. TRIAL REGISTRATION NUMBER: N/A.

We evaluated the efficacy of large language models (LLMs), specifically, generative pre‐trained transformer‐4 (GPT‐4), in predicting pregnancy following in vitro fertilization (IVF) treatment and compared its accuracy with results from an original published study. Our findings revealed that GPT‐4 can autonomously develop and refine advanced machine learning models for pregnancy prediction with minimal human intervention. The prediction accuracy was 0.79, and the area under the receiver operating characteristic curve (AUROC) was 0.89, exceeding or being at least equivalent to the metrics reported in the original study, that is, 0.78 for accuracy and 0.87 for AUROC. The results suggest that LLMs can facilitate data processing, optimize machine learning models in predicting IVF success rates, and provide data interpretation methods. This capacity can help bridge the knowledge gap between data scientists and medical personnel to solve the most pressing clinical challenges. However, more experiments on diverse and larger datasets are needed to validate and promote broader applications of LLMs in assisted reproduction.

High-throughput sequencing technologies have increasingly led to discovery of disease-causing genetic variants, primarily in postnatal multi-cell DNA samples. However, applying these technologies to preimplantation genetic testing (PGT) in nuclear or mitochondrial DNA from single or few-cells biopsied from in vitro fertilised (IVF) embryos is challenging. PGT aims to select IVF embryos without genetic abnormalities. Although genotyping-by-sequencing (GBS)-based haplotyping methods enabled PGT for monogenic disorders (PGT-M), structural rearrangements (PGT-SR), and aneuploidies (PGT-A), they are labour intensive, only partially cover the genome and are troublesome for difficult loci and consanguineous couples. Here, we devise a simple, scalable and universal whole genome sequencing haplarithmisis-based approach enabling all forms of PGT in a single assay. In a comparison to state-of-the-art GBS-based PGT for nuclear DNA, shallow sequencing-based PGT, and PCR-based PGT for mitochondrial DNA, our approach alleviates technical limitations by decreasing whole genome amplification artifacts by 68.4%, increasing breadth of coverage by at least 4-fold, and reducing wet-lab turn-around-time by ~2.5-fold. Importantly, this method enables trio-based PGT-A for aneuploidy origin, an approach we coin PGT-AO, detects translocation breakpoints, and nuclear and mitochondrial single nucleotide variants and indels in base-resolution.

STUDY QUESTION: Can generative artificial intelligence (AI) models produce high-fidelity images of human blastocysts? SUMMARY ANSWER: Generative AI models exhibit the capability to generate high-fidelity human blastocyst images, thereby providing substantial training datasets crucial for the development of robust AI models. WHAT IS KNOWN ALREADY: The integration of AI into IVF procedures holds the potential to enhance objectivity and automate embryo selection for transfer. However, the effectiveness of AI is limited by data scarcity and ethical concerns related to patient data privacy. Generative adversarial networks (GAN) have emerged as a promising approach to alleviate data limitations by generating synthetic data that closely approximate real images. STUDY DESIGN, SIZE, DURATION: Blastocyst images were included as training data from a public dataset of time-lapse microscopy (TLM) videos (n = 136). A style-based GAN was fine-tuned as the generative model. PARTICIPANTS/MATERIALS, SETTING, METHODS: We curated a total of 972 blastocyst images as training data, where frames were captured within the time window of 110–120 h post-insemination at 1-h intervals from TLM videos. We configured the style-based GAN model with data augmentation (AUG) and pretrained weights (Pretrained-T: with translation equivariance; Pretrained-R: with translation and rotation equivariance) to compare their optimization on image synthesis. We then applied quantitative metrics including Fréchet Inception Distance (FID) and Kernel Inception Distance (KID) to assess the quality and fidelity of the generated images. Subsequently, we evaluated qualitative performance by measuring the intelligence behavior of the model through the visual Turing test. To this end, 60 individuals with diverse backgrounds and expertise in clinical embryology and IVF evaluated the quality of synthetic embryo images. MAIN RESULTS AND THE ROLE OF CHANCE: During the training process, we observed consistent improvement of image quality that was measured by FID and KID scores. Pretrained and AUG + Pretrained initiated with remarkably lower FID and KID values compared to both Baseline and AUG + Baseline models. Following 5000 training iterations, the AUG + Pretrained-R model showed the highest performance of the evaluated five configurations with FID and KID scores of 15.2 and 0.004, respectively. Subsequently, we carried out the visual Turing test, such that IVF embryologists, IVF laboratory technicians, and non-experts evaluated the synthetic blastocyst-stage embryo images and obtained similar performance in specificity with marginal differences in accuracy and sensitivity. LIMITATIONS, REASONS FOR CAUTION: In this study, we primarily focused the training data on blastocyst images as IVF embryos are primarily assessed in blastocyst stage. However, generation of an array of images in different preimplantation stages offers further insights into the development of preimplantation embryos and IVF success. In addition, we resized training images to a resolution of 256 × 256 pixels to moderate the computational costs of training the style-based GAN models. Further research is needed to involve a more extensive and diverse dataset from the formation of the zygote to the blastocyst stage, e.g. video generation, and the use of improved image resolution to facilitate the development of comprehensive AI algorithms and to produce higher-quality images. WIDER IMPLICATIONS OF THE FINDINGS: Generative AI models hold promising potential in generating high-fidelity human blastocyst images, which allows the development of robust AI models as it can provide sufficient training datasets while safeguarding patient data privacy. Additionally, this may help to produce sufficient embryo imaging training data with different (rare) abnormal features, such as embryonic arrest, tripolar cell division to avoid class imbalances and reach to even datasets. Thus, generative models may offer a compelling opportunity to transform embryo selection procedures and substantially enhance IVF outcomes. STUDY FUNDING/COMPETING INTEREST(S): This study was supported by a Horizon 2020 innovation grant (ERIN, grant no. EU952516) and a Horizon Europe grant (NESTOR, grant no. 101120075) of the European Commission to A.S. and M.Z.E., the Estonian Research Council (grant no. PRG1076) to A.S., and the EVA (Erfelijkheid Voortplanting & Aanleg) specialty program (grant no. KP111513) of Maastricht University Medical Centre (MUMC+) to M.Z.E. TRIAL REGISTRATION NUMBER: Not applicable.

High-throughput sequencing technologies have increasingly led to discovery of disease-causing genetic variants, primarily in postnatal multi-cell DNA samples. However, applying these technologies to preimplantation genetic testing (PGT) in nuclear or mitochondrial DNA from single or few-cells biopsied from in vitro fertilised (IVF) embryos is challenging. PGT aims to select IVF embryos without genetic abnormalities. Although genotyping-by-sequencing (GBS)-based haplotyping methods enabled PGT for monogenic disorders (PGT-M), structural rearrangements (PGT-SR), and aneuploidies (PGT-A), they are labour intensive, only partially cover the genome and are troublesome for difficult loci and consanguineous couples. Here, we devised a simple, scalable and universal whole genome sequencing haplarithmisis-based approach enabling all forms of PGT in a single assay. In a comparison to state-of-the-art GBS-based PGT for nuclear DNA (37 embryos, 18 families, 25 indications), shallow sequencing-based PGT (10 embryos, 3 families), and PCR-based PGT for mitochondrial DNA (10 embryos, 2 families), our approach alleviates technical limitations by decreasing whole genome amplification artifacts by 68.4%, increasing breadth of coverage by 4-fold, and reducing wet-lab turn-around-time by 2.5-fold. Importantly, this method enables trio-based PGT-A for aneuploidy origin, an approach we coin PGT-AO, detects translocation breakpoints, and nuclear and mitochondrial single nucleotide variants and indels in base-resolution.

ChemInform is a weekly Abstracting Service, delivering concise information at a glance that was extracted from about 100 leading journals. To access a ChemInform Abstract of an article which was published elsewhere, please select a “Full Text” option. The original article is trackable via the “References” option.