Using Artificial Intelligence to Help Predict Treatment Response in Patients with Blood Cancer

癌症是世界范围内的主要死亡原因。血癌是英国第五大常见癌症。被诊断患有血癌的患者，如骨髓瘤和淋巴瘤，接受化疗和放疗等治疗，以帮助他们在癌症中生存。虽然大多数患者可以在不同程度上受益，但不同的患者对相同的治疗反应不同。没有一种治疗方法是完全适合每个人的。因此，必须为每个患者找到最佳治疗方法，以改善治疗效果，减少他们的疼痛和不适，并最大限度地减少NHS的成本。一个高度需求的解决方案是个性化的医疗决策工具，以应对这些挑战并简化临床服务。该项目的目标是开发和评估新的人工智能（AI）工具，以更准确地预测患者对治疗的反应。正电子发射断层扫描/计算机断层扫描（PET/CT）图像被认为在评估肿瘤预后方面具有预测意义。利用最先进的人工智能技术，这个跨学科项目将利用主要由Clatterbridge癌症中心NHS基金会信托基金（CCC）和利物浦大学医院NHS基金会信托基金（LUHFT）提供的实验数据。这些大规模的多维数据将包括PET/CT图像数据、临床症状和医疗合作伙伴收集的人口统计数据，并将利用这些独特的多模态数据开发和评估利用深度学习和统计学习的新AI工具，以实现有效和准确的临床决策，预测治疗结果。我们将确保新的人工智能工具既准确又可解释，使它们值得信赖。这些工具将由我们在CCC和LUHFT的临床合作伙伴进行评估。

Cancer is a leading cause of death worldwide. Blood cancer is the fifth most common cancer in the UK. Patients who are diagnosed with blood cancers, such as myeloma and lymphoma, undergo treatments like chemotherapy and radiotherapy to help them survive cancer. Although most patients can benefit from them to varying degrees, different patients respond differently to the same treatment. No single treatment is totally suitable for everyone.Therefore, it is essential to find optimal treatments for individual patients to improve treatment outcomes, reduce their pain and discomfort, and minimize costs for the NHS. A highly demanded solution is a personalized medicine decision-making tool to address these challenges and streamline clinical services.The project's objective is to develop and evaluate new artificial intelligence (AI) tools for predicting patients' responses to treatments more accurately than is currently possible. This test has the potential to guide doctors on which treatments are most likely to benefit individual cancer patients.Positron Emission Tomography/Computed Tomography (PET/CT) images are recognised as having predictive significance in evaluating a tumor's prognosis. Using state-of-the-art AI technology, this interdisciplinary project will leverage experimental data primarily provided by The Clatterbridge Cancer Centre NHS Foundation Trust (CCC) and the Liverpool University Hospitals NHS Foundation Trusts (LUHFT). The large-scale, multidimensional data will include PET/CT image data, clinical symptoms, and demographic data collected by healthcare partners.New AI tools, utilising both deep learning and statistical learning, will be developed and evaluated by using these unique multimodal data to achieve efficient and accurate clinical decision-making for predicting treatment outcomes. We will ensure that the new AI tools are both accurate and explainable, making them trustworthy. These tools will be evaluated by our clinical partners at CCC and LUHFT.