Towards reliable machine learning for ultrasound screening of aortic stenosis

Aortic stenosis (AS), a prevalent and serious heart valve disorder, requires early detection but remains difficult to diagnose in routine practice. Although echocardiography with Doppler imaging is the clinical standard, these assessments are typically limited to trained specialists. Point-of-care ultrasound (POCUS) offers an accessible alternative for AS screening but is restricted to basic 2D B-mode imaging, often lacking the analysis Doppler provides. This thesis introduces MultiASNet, a multimodal machine learning framework designed to enhance AS screening with POCUS by combining 2D B-mode videos with structured data from echocardiography reports, including Doppler parameters. Using contrastive learning, MultiASNet aligns video features with report features in tabular form from the same patient to improve interpretive quality. To address misalignment arising when a single report corresponds to multiple video views—each encoding partial or no information relevant to AS diagnosis—we employ cross-attention within a transformer-based video and tabular network to down-weight irrelevant report features. The model integrates structured data only during training, enabling independent use with B-mode videos during inference for broader accessibility. In addition, MultiASNet incorporates sample selection to counteract label noise from observer variability, yielding improved accuracy on two datasets. To mitigate bias against underrepresented cases, the model also undergoes additional training with deep feature reweighting to improve fairness across patient subgroups. We achieved balanced accuracy scores of 93.0% on a private dataset and 83.9% on the public TMED-2 dataset for AS detection. For severity classification, balanced accuracy scores were 80.4% and 59.4% on the private and public datasets, respectively. MultiASNet uses a multi-faceted approach to address barriers to reliable feature learning, involving multimodal enrichment, sample selection, and subgroup feature reweighting. This approach demonstrates the potential to develop suitable, yet more clinically robust AI systems in POCUS, bringing AS severity classification closer to the capabilities of the diagnostic standard using Doppler-based assessment.