Brigham Research Institute Poster Session Site logo-1
Close this search box.

Andrew Song, PhD

Job Title

Postdoctoral fellow

Academic Rank

Fellow or Postdoc




Andrew H. Song, Mane Williams, Drew Williamson, Guillaume Jaume, Andrew Zhang, Bowen Chen

Principal Investigator

Faisal Mahmood



AI-driven efficient patient prognosis based on 3D pathology samples

Scientific Abstract

Human tissue forms a microenvironment that is fundamentally three-dimensional (3D). However, the standard-of-care in pathologic diagnosis involves selecting a few two-dimensional (2D) sections for microscopic evaluation, risking sampling bias and misdiagnosis. Diverse methods for capturing 3D tissue morphologies have been developed, with minimal success in translation to clinical practice; manual and computational evaluations of such large 3D data have been impractical and/or unable to provide patient-level clinical insights. Here we present Modality-Agnostic Multiple instance learning for volumetric Block Analysis (MAMBA), a deep-learning-based platform for processing 3D tissue images and predicting patient outcomes. Archived prostate cancer specimens were imaged with open-top light-sheet microscopy or microcomputed tomography and the resulting 3D datasets were used to train risk-stratification networks based on 5-year biochemical recurrence outcomes via MAMBA. With the 3D block-based approach, MAMBA achieves an area under the receiver operating characteristic curve (AUC) of 0.86 and 0.74, superior to 2D traditional single-slice-based prognostication (AUC of 0.79 and 0.57), suggesting superior prognostication with 3D morphological features. Further analyses reveal that the incorporation of greater tissue volume improves prognostic performance and mitigates risk prediction variability from sampling bias, suggesting the value of capturing larger extents of heterogeneous 3D morphology.

Lay Abstract

Human tissues have a three-dimensional structure, but when pathologists examine them for diagnosis, they rely on flat, two-dimensional tissue slices cut from the excised biopsy. This 2D pathology routine can lead to mistakes in diagnosis, which the emerging paradigm of 3D pathology is trying to address. While many imaging methods have been developed to capture the 3D morphological characteristics of tissues, they have not been practical for regular clinical use. This is due to the lack of a software platform that can process large amounts of 3D data. We introduce an AI-based framework for 3D pathology (MAMBA). It’s a smart system based on AI, that examines 3D images of tissues and predicts how patients’ clinical outcomes. We tested it using samples of prostate cancer tissue imaged with state-of-the-art 3D imaging machines, and MAMBA outperformed conventional 2D pathology methods at telling us which patients might have cancer recurrence.
Overall, this approach is better than the usual 2D method and helps us avoid mistakes caused by only looking at portion of the tissue. MAMBA can be a tremendous asset to doctors when making decisions about treatment, and also help us find new ways to predict how diseases will behave in the future

Clinical Implications

With the rapid growth and adoption of 3D spatial biology and pathology, MAMBA provides a general and efficient framework for 3D pathology for clinical decision support and can help to reveal novel 3D morphological biomarkers for prognosis and therapeutic response.