Cardiovascular risk in chronic myeloid leukaemia: A multidisciplinary consensus on screening and management.

INTRODUCTION: Tyrosine kinase inhibitors (TKIs) have become the mainstay of treatment for chronic myeloid leukaemia (CML), but cardiovascular (CV) risk and exacerbation of underlying risk factors associated with TKIs have become widely debated. Real-world evidence reveals little application of CV risk factor screening or continued monitoring within UK CML management. This consensus paper presents practical recommendations to assist healthcare professionals in conducting CV screening/comorbidity management for patients receiving TKIs. METHODS: We conducted a multidisciplinary panel meeting and two iterative surveys involving 10 CML specialists: five haematologists, two cardio-oncologists, one vascular surgeon, one haemato-oncology pharmacist and one specialist nurse practitioner. RESULTS: The panel recommended that patients commencing second-/third-generation TKIs undergo formal CV risk assessment at baseline, with additional investigations and involvement of cardiologists/vascular surgeons for those with high CV risk. During treatment, patients should undergo CV monitoring, with the nature and frequency of testing dependent on TKI and baseline CV risk. For patients who develop CV adverse events, decision-making around TKI interruption, cessation or change should be multidisciplinary and balance CV and haematological risk. CONCLUSION: The panel anticipates these recommendations will support healthcare professionals in implementing CV risk screening and monitoring, broadly and consistently, and thereby help optimise TKI treatment for CML.

Deep learning models for predicting RNA degradation via dual crowdsourcing.

Medicines based on messenger RNA (mRNA) hold immense potential, as evidenced by their rapid deployment as COVID-19 vaccines. However, worldwide distribution of mRNA molecules has been limited by their thermostability, which is fundamentally limited by the intrinsic instability of RNA molecules to a chemical degradation reaction called in-line hydrolysis. Predicting the degradation of an RNA molecule is a key task in designing more stable RNA-based therapeutics. Here, we describe a crowdsourced machine learning competition ('Stanford OpenVaccine') on Kaggle, involving single-nucleotide resolution measurements on 6,043 diverse 102-130-nucleotide RNA constructs that were themselves solicited through crowdsourcing on the RNA design platform Eterna. The entire experiment was completed in less than 6 months, and 41% of nucleotide-level predictions from the winning model were within experimental error of the ground truth measurement. Furthermore, these models generalized to blindly predicting orthogonal degradation data on much longer mRNA molecules (504-1,588 nucleotides) with improved accuracy compared with previously published models. These results indicate that such models can represent in-line hydrolysis with excellent accuracy, supporting their use for designing stabilized messenger RNAs. The integration of two crowdsourcing platforms, one for dataset creation and another for machine learning, may be fruitful for other urgent problems that demand scientific discovery on rapid timescales.

Patient, Family Member and Physician Perspectives and Experiences with AML Treatment Decision-Making.

INTRODUCTION: Treatment decisions in older adults with acute myeloid leukemia (AML) are challenging, particularly for those who are not candidates for intensive chemotherapy (IC), and the trade-offs patients, their families and physicians consider when choosing a treatment option are not well understood. This qualitative research explored the value of extending survival and the treatment decision-making process from a multi-stakeholder perspective. METHODS: Overall, 28 patients with AML (≥ 65 years old, unsuitable for IC), 25 of their relatives and 10 independent physicians from the US, UK and Canada took part in one-on-one, 60-minute qualitative interviews. RESULTS: Across all stakeholders, improved health-related quality of life (HRQoL), extended survival and relief of AML symptoms were recognized as most important in AML treatment decision-making. However, extending survival in 'good health' was more important than extending survival alone, particularly because of the extra time it gives patients and their relatives together, and allows patients to achieve important goals. Patients' limited understanding of available treatment options, paired with incorrect perceptions of treatment side effects, impacted their involvement in the treatment decision-making process. Patients and physicians perceived physicians to have the most influence in the decision-making process despite their priorities not always aligning. CONCLUSION: These findings illustrate the importance of having structured discussions which explicitly assess patients' goals and their understanding and expectations of treatments and also the need for patient friendly resources about the lived experience of AML and available treatment options. These measures will help to ensure that patients are fully involved in the shared decision-making process.

Deep learning models for predicting RNA degradation via dual crowdsourcing.

Messenger RNA-based medicines hold immense potential, as evidenced by their rapid deployment as COVID-19 vaccines. However, worldwide distribution of mRNA molecules has been limited by their thermostability, which is fundamentally limited by the intrinsic instability of RNA molecules to a chemical degradation reaction called in-line hydrolysis. Predicting the degradation of an RNA molecule is a key task in designing more stable RNA-based therapeutics. Here, we describe a crowdsourced machine learning competition ("Stanford OpenVaccine") on Kaggle, involving single-nucleotide resolution measurements on 6043 102-130-nucleotide diverse RNA constructs that were themselves solicited through crowdsourcing on the RNA design platform Eterna. The entire experiment was completed in less than 6 months, and 41% of nucleotide-level predictions from the winning model were within experimental error of the ground truth measurement. Furthermore, these models generalized to blindly predicting orthogonal degradation data on much longer mRNA molecules (504-1588 nucleotides) with improved accuracy compared to previously published models. Top teams integrated natural language processing architectures and data augmentation techniques with predictions from previous dynamic programming models for RNA secondary structure. These results indicate that such models are capable of representing in-line hydrolysis with excellent accuracy, supporting their use for designing stabilized messenger RNAs. The integration of two crowdsourcing platforms, one for data set creation and another for machine learning, may be fruitful for other urgent problems that demand scientific discovery on rapid timescales.