Kang Wang
University of California-San Francisco, CA·Medicine & Clinical Research
Colorectal cancer is the third most common cancer in the United States and the second leading cause of cancer death. About 20% of patients already have metastatic disease, meaning cancer that has spread, when they are diagnosed.
- Fields
Medicine & Clinical Research
Computer Science
Biomedical Engineering
- Topics
colorectal cancer
metastatic disease
liver metastases
- Grants
- 1 known grant, through May 2029
What they work on
- Build a large standardized database of abdominal CT images and radiology reports, using large language models to automatically extract findings from reports and select data for AI training.
- Train an anatomy-aware vision-language AI model to detect metastases from colorectal cancer in the liver, lymph nodes and peritoneum. Training will use report supervision, synthetic tumor generation and active learning to limit manual labeling.
- Develop AI methods that track metastatic lesions across multiple CT time points and generate structured reports documenting changes in disease burden.
- Test, in a prospective three-reader study, whether the AI system improves radiologists' sensitivity for detecting metastatic disease while reducing their reading time and cognitive workload.
More about the research
The cancer often spreads to the liver, lymph nodes, bones and the peritoneum, the lining of the abdomen. Computed tomography (CT) scans are the main tool for staging the disease and tracking treatment. Reading them well is hard. Small liver tumors can look like harmless spots, cancerous lymph nodes may be normal-sized, and peritoneal tumors can hide among other tissues. Radiologists must compare scans across many visits, take measurements and write detailed reports. This repetitive but demanding work adds to heavy workloads, burnout and diagnostic errors.
Why it matters
Colorectal cancer patients could benefit from earlier detection of subtle tumor spread on routine CT scans. Earlier detection may support more timely treatment decisions. The AI system could also reduce the time and mental strain of reading follow-up scans. This could help address radiologist burnout and lower diagnostic errors in busy cancer imaging practices.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Jie Xu
University of Florida, FL·Computer Science·Early-career
Develop transformer-based vision-language models that outline kidney masses on CT scans, extract clinical concepts from notes, and generate radiology reports, validated against expert annotations.
Rui Wang
Case Western Reserve University, OH·Pharmacology
Determine whether LRG1 secreted by liver endothelial cells drives colorectal cancer liver metastases to start and grow, using mice lacking endothelial LRG1 and an LRG1-neutralizing antibody.
Jeanne Shen
Stanford University, CA·Medicine & Clinical Research
Build a publicly available, expert-annotated multimodal colorectal cancer dataset. It will combine pathology whole-slide images, preoperative CT and MRI scans, and structured electronic medical record data.
The school
University of California-San Francisco: 89 labs on PI Pages, $2.1B a year on research, 164 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.