Our research lab focuses on developing and evaluating new computer-aided quantitative medical image processing algorithms and machine learning models with easy-to-use graphic user interface (GUI) tools to help predict disease (i.e., cancer and stroke) risk and prognosis, which include to develop and evaluate new computer-aided detection and/or diagnosis (CAD) schemes to:

• Predict (a) short-term breast cancer risk using bilateral digital mammograms [1], (b) radiological complete response (rCR) of breast cancer patients to NACT using pre-NACT breast MR images [2], (c) risk of lung cancer recurrence after surgery among the stage I NSCLC patients [3] using the diagnostic chest CT images, (d) progression-free survival (PFS) of ovarian cancer patients after chemotherapy using longitudinal abdomen CT images acquired from pre-therapy and 4-6 weeks on-therapy [4], and (e) risk of metastasis of cervical cancer [5] and gastric cancer patients using CT images [6].
• Classify between (a) malignant and benign breast lesions using conventional digital mammograms [7], contrast-enhanced digital mammograms [8] and breast MRI [9], (b) malignant and benign lung nodules using CT images [10], and (c) tumor epithelium and stroma using digital histopathology images [11].
• Improve performance and potential clinical utility of CAD schemes to detect suspicious breast lesions using digital mammograms [12] and lung nodules using chest CT images [13] using new adaptive optimization and result cuing methods.
• Apply, modify and test deep transfer learning models to (a) segment suspicious tumors and other regions of interest (ROIs) depicting on CT and digital histopathology images [14], and (b) classify between malignant and benign tumors including breast masses using digital mammograms [15] and lung nodules using chest CT images [16].
• Quantify severity and/or predict treatment efficacy of other diseases including acute ischemic or hemorrhagic stroke [17], and COVID-19 infected pneumonia [18]).