PeopleofSuperintelligence

Researcher

Synho Do

Synho Do is a medical imaging researcher whose work applies machine learning to radiology. He coauthored a 2012 CT algorithm for measuring pneumothorax, studied training-set size for medical image classification in 2015, and helped develop a fully automated bone-age assessment system in 2017. In 2019, he coauthored SinoNet, which reads CT sinograms without image reconstruction; in February 2026, he coauthored a framework for setting operational safety policies for clinical AI. As of October 2026, he is an assistant professor of radiology at Harvard Medical School and directs MGH's Laboratory of Medical Imaging and Computation.

Works
105
Citations
5,817
h-index
31

Publication numbers from OpenAlex, as of 27 October 2026.

Timeline

  1. 2012

    Pneumothorax analysis on CT

    In 2012, he coauthors an automated algorithm to quantify pneumothorax on multidetector CT images. It combines two-dimensional and three-dimensional processing to reduce processing time by two-thirds.

  2. 2015

    Medical-image training data

    In 2015, he coauthors a study on how much data medical-image classification needs. The team trains a convolutional neural network on six training-set sizes and tests it on 6,000 CT images.

  3. 2017

    Automated bone-age assessment

    In 2017, he coauthors a fully automated deep-learning system for bone-age assessment. Its pipeline segments a region of interest, standardizes and preprocesses radiographs, and performs bone-age assessment.

  4. 2019

    SinoNet reads CT sinograms

    In 2019, he coauthors SinoNet, a convolutional neural network that processes CT sinograms without first reconstructing images. The team evaluates it for anatomy recognition and intracranial hemorrhage detection.

  5. February 2026

    Operational safety for clinical AI

    In February 2026, he coauthors the Safety-Aware Receiver Operating Characteristic (SA-ROC) framework, which combines analytical signals into an operational directive for clinical AI.

  6. As of October 2026

    Radiology and AI at Harvard

    As of October 2026, he is an Assistant Professor of Radiology at Harvard Medical School and directs the Laboratory of Medical Imaging and Computation at Massachusetts General Hospital.

Career

  1. NowHarvard UniversityAssistant Professor of Radiology
  2. NowMass General BrighamDirector, Laboratory of Medical Imaging and Computation

From the biography.

Sources and credits

Photo: We have no published photo of Synho Do yet. Send us a free-licensed photo.

  1. Automated Quantification of Pneumothorax in CT pmc.ncbi.nlm.nih.gov. Checked 27 October 2026.
  2. How much data is needed to train a medical image deep learning system to achieve necessary high accuracy? arxiv.org. Checked 27 October 2026.
  3. Fully Automated Deep Learning System for Bone Age Assessment pmc.ncbi.nlm.nih.gov. Checked 27 October 2026.
  4. Machine Friendly Machine Learning: Interpretation of Computed Tomography Without Image Reconstruction pmc.ncbi.nlm.nih.gov. Checked 27 October 2026.
  5. Defining operational safety in clinical artificial intelligence systems nature.com. Checked 27 October 2026.
  6. Synho Do, Ph.D. | Mass General Research Institute researchers.mgh.harvard.edu. Checked 27 October 2026.
  7. Affiliate Faculty - Kempner Institute kempnerinstitute.harvard.edu. Checked 27 October 2026.