Suk Wai Lam

LUMC

In short, the research of Suk Wai Lam entails the following:

BONE-CLASS - Bone Tumor Classification with Multimodal Deep Learning

Bone tumors are a challenging group of ultrarare cancers, accounting for less than 1% of all cancers and often affecting a young patient population. These tumors have a broad morphological spectrum, with overlapping features between reactive, benign and malignant bone tumors. Therefore, gaining expertise to diagnose these ultra-rare tumors— which pathologists in non-expert centers might encounter only once in a lifetime— is not possible. Additionally, correlating the morphology of bone tumors with clinical and radiographic features is essential for accurate diagnosis, as tumors with similar morphology can have different clinical behavior, making bone tumor diagnostics even more complex.

To address the diagnostic challenges of bone tumors, we aim to develop a decision support model, ‘the BONE-CLASS”, using a unique, high-quality database as the foundation for deep learning studies. Given the importance of integrating radiological, pathological, and clinical data in the diagnostic process, our goal is to create a multimodal deep learning model that assist in bone tumor diagnosis by combining H&E stained whole-slide images, plain radiographs, and clinical information (the “triad”). By using these simple and widely accessible modalities - avoiding costly techniques - the BONE-CLASS can be of benefit for medical institutions with limited resources and lacking domain-specific expertise worldwide, ensuring that even centers without advanced equipment can improve diagnostic accuracy, shorten referral times, and ultimately enhance patient outcomes.