Please visit our lab webpage for updated research projects

Response monitoring in lung immunotherapy using machine learning on serial CT

Currently, there are no predictive biomarkers to point to whether non-small cell lung cancer (NSLC) patients will benefit from immune-checkpoint inhibitor therapy, a new form of cancer therapy that helps the body’s immune system fight cancer more effectively. We use radiomic techniques to find previously unseen changes in patterns, both intra- and peri-nodular, in CT scans taken when the lung cancer is first diagnosed compared to scans taken after the first 2-3 cycles of immunotherapy treatment.

Using multiple datasets for training and validation, our models were able to identify responders from non-responders as well as determine early response to immunotherapy. We also identified that the radiological features were associated with greater infiltration of immune cells into lung tissue, based on diagnostic biopsies performed on some of the patients in the study.

Radiomics for Brain Tumor Characterization

The most pressing challenges thwarting prognosis and treatment management in Glioblastoma Multiforme (GBM) include:

(a) inability to estimate survival at a pre-treatment stage in order to identify candidates for specialized trials (b) inability to avoid highly-invasive surgeries in patients with radiation necrosis (a delayed treatment change) that mimics appearance of tumor recurrence (c) avoid “wait-and-watch” in recurrence patients currently misdiagnosed as pseudo-progression (an early radiation change).

We have developed novel computer-extracted radiographic features (radiomics) to comprehensively characterize GBM behavior and response in a non-invasive manner, and, thereby, capture morphologic diversity in patients. My projects in this space include:

• Development of a computer-aided diagnostic system to distinguish recurrent brain tumors from radiation induced effects using multi-parametric MRI

• Image analytics of texture descriptors on treatment-naive MRI for survival prediction of patients with GBM

• Using radiomics to track the progression of suspicious artifacts post-surgery, and distinguishing true-progression from pseudo-progression

Radiogenomics for Breast Cancer management

Adjuvant systemic therapies for breast cancer historically have been administered following definitive breast surgery. Preoperative or neoadjuvant systemic chemotherapy is the current standard of care for many patients with Stages I-III breast cancer when the need for combination chemotherapy is clear. Pre-op therapy enables the oncologist to evaluate tumor response and discontinue ineffective therapy or substitute an alternative systemic therapy. Further, a patient’s response to pre-op chemotherapy may provide prognostic information that can supplement conventional prognostic data, such as initial staging, tumor grade, and receptor status. In fact, patients who achieve a pathological complete response (pCR), typically defined as the absence of residual invasive disease in the breast and axillary nodes at surgery, have improved long-term recurrence-free and overall survival compared to those who do not. Unfortunately, the current clinical workflow allows for the assessment of clinical response only at the end of pre-op therapy, thus precluding the switch to more aggressive pre-op therapy. The development of better predictive analytic techniques to assess response earlier in the treatment regimen would allow an early intervention to switch to another regimen if pCR is not likely, improving the chance of a better response to therapy. Towards this end, we have developed radiogenomic pipelines that can provide insight into the underlying tumor biology as reflected on radiologic imaging. Projects in this space include:

• Identifying radiomic correlates of different molecular subtypes (enriched, basal, luminal) of HER2+ breast cancer
• Identifying radiomic correlates of various mutational status (TP53, PIK3CA) in breast cancer on baseline imaging

Pattern Recognition in Diabetic Retinopathy Analysis

Developed an automated diabetic retinopathy detection technique to assist ophthalmologists in analyzing various retinopathy stages

Automated bridge health monitoring system

Worked on the Long Term Bridge Performance (LTBP) Robotics program awarded by the Federal Highway Administration (FHWA) for the computer vision component. The project involved:

• Implementation of large scale image stitching to compose a coherent spatial mosaic from individual digital surface images
• Implementation of automated crack classification based on feature analysis from training images; employed statistical inference algorithms to identify cracks on bridge deck
• Designing the surface imaging system on the bridge inspection robot (RABIT)
• Designing the panoramic imaging system for context localization

Other projects:

• Micro-crack detection using a texture camera
• Application control using real-time hand gesture and blink detection
• Unravel before you travel: A decision system for air-ticket purchase