(3if) Chemical Imaging with Artificial Intelligence enables Multiscale Analysis of Complex Biological Systems

My research interest is to develop models and workflows at the patient level that can integrate molecular imaging to computational approaches for early cancer detection and improved patient outcomes. As a faculty member, my research program will integrate concepts from molecular imaging, deep learning, and genomics for building multiscale models for characterizing tumors and their progression. It is now well established that cancer is not just a single disease but is rather heterogeneous in terms of tumor subtype, its manifestation, and progression. There is a need to optimally delineate patient-specific identifiers and markers for comprehensive molecular mapping of the disease. My prior work on combining engineering technologies and clinical pathology positions me to establish a laboratory that engages in an interdisciplinary approach to enable technology-driven healthcare.

My lab will actively engage/collaborate with clinicians to develop molecular and digital approaches for i) immune profiling of the tumor and its microenvironment, ii) couple imaging with genomics using supervised and unsupervised data analysis and, iii) extend the developed protocols to multiple cancer subtypes for a comprehensive cancer care portal. The overarching goal of these aims is to enable individualized diagnostic profiling of patients leading to personalized therapy and disease management. Additionally, I plan to collaborate with genomics, proteomics and computational labs to identify novel indicators for patient meta-analysis to understand the underlying etiology of the disease. Students and postdoctoral researchers in my group will draw expertise in chemical engineering, bioengineering, computer science, medical imaging, and pathology.

Keywords

Molecular imaging, spectroscopy, artificial intelligence, cancer diagnosis, digital health

Successful Proposals:

• Beckman Postdoctoral Fellow (2019- Present)
• Partnership 2020: Leveraging US-India Cooperation in Higher Education Grant (2019 -2020)
• Baxter Young Investigator Award (2018)
• Beckman Graduate Fellow (2018 -2019)

Awards:

• Tomas B. Hirschfeld Scholar Award, FACSS (2019)
• Coblentz Graduate Student Award (2019)
• William G. Fateley Award, Outstanding Contribution to Vibrational Spectroscopy (2019)
• Invited Speaker and Winner (out of the 13 graduate students selected from across the world), Annual Engineering Ph.D. Summit, EPFL Switzerland (2018)
• Eastern Analytical Symposium Graduate Research Award (2018)
• Nadine Barrie Smith Memorial Fellowship (2018)
• Big Data Summer Fellow for computational medicine, Weill Cornell Medicine, New York (2017)
• Illinois Distinguished Fellowship (2014-17)

Research Experience

My research has focused on developing digital protocols to capture changes in cellular chemistry related to tumor progression in patients. Artificial Intelligence (AI)-based digital workflows for guiding clinical assessment is an emerging field for enhanced patient healthcare and efficient clinical management. The current pipeline of cancer diagnosis involves staining the tissue sample with different molecular markers followed by a manual interpretation of these stained images, placing an unrealistic burden on pathologists to provide a precise diagnosis that is universally concordant for this heterogeneous and varied disease. There is a need for multiplexed measurements to map different cellular environments within the tumor and around it.

We developed a microscopy-based molecular imaging approach and combined it with modern computing and deep learning for quantitative and holistic patient analysis. In this work, principles of spectroscopic imaging, molecular sensing, and pattern recognition were put together on one platform to develop a fast, accurate, and easy to use clinical toolset for characterizing morphological and biochemical signature of the tumor and its associated microenvironment. Infrared spectroscopic imaging can extract both biochemical and structural details of the tissue samples in a high throughput fashion. The measured biochemical fingerprint of the patient sample is encoded in 1000s of vibrational frequencies, making the use of sophisticated data mining tools essential for clinical applications. Currently, I am working with clinicians to develop standardized deep learning approaches for stratifying breast biopsies based on different risk categories of tumor development. The treatment and management profiles of all these categories are different, making it essential to identify these lesions precisely. The developed digital pipeline will address the long-standing need for precise triaging quantitatively, improving patient outcomes.

Postdoctoral Project

Multi-institutional study for risk stratification of breast cancer patients using deep learning and molecular imaging

Mentor: Prof. Rohit Bhargava (Department of Bioengineering, University of Illinois at Urbana-Champaign, Urbana, IL)

Clinical Collaborators: Dr. Malvika Solanki (Mayo Clinic, Rochester, MN), Dr. Andre Balla (University of Illinois at Chicago Medical School, Chicago, IL), Dr. Sunil Badve (Indiana University Medical School, Indianapolis, IN), Dr. Anna Higham (Carle Hospital, Urbana, IL)

Ph.D. Dissertation
Breast cancer diagnosis using Fourier transform infrared imaging and statistical learning

Advisor: Prof. Rohit Bhargava (Department of Bioengineering, University of Illinois at Urbana Champaign, Urbana, IL)

Teaching Experience and Interests

In addition to research, I have taken many guest lectures at the undergraduate and the graduate level. I was invited as a guest lecturer for the Advanced Bioinstrumentation course in 2019. I delivered two guest lectures in the same class again in 2020 due to encouraging responses from the students. I had also given guest lectures in a couple of other courses focusing on medical imaging and tumor microenvironment (Special Topics: Tissue Microenvironment Analysis and Biological Measurement). I have also mentored several high school students, undergraduate and graduate students from Bioengineering, Computer Science, Chemical Engineering, and Medicine departments. Recently, I have also been mentoring a postdoctoral student at Indian Institute of Technology,Delhi (India) aspart of our recent Indo-US collaborative grant. I took numerous courses in the chemical engineering department during my undergraduate education, therefore, I will be comfortable in teaching core chemical engineering undergraduate and graduate courses such as transport phenomena, fluid dynamics and chemical kinetics. One of the areas of my passion for teaching is to develop a curriculum that can bridge gaps between engineering technologies and medical science. We can do this by teaching machine learning/data science and molecular imaging approaches in the context of changing healthcare regimes. I am excited to develop courses and training programs in that direction.

Selected Publications (6 out of 15)

1. Schnell, , Mittal, S., Falahkheirkhah, K., Mittal, A., Yeh, K., Kenkel, S., Kadjacsy-Balla, A., Carney, S., and Bhargava, R. (2020). All-digital Histopathology by infrared-optical hybrid microscopy. Proceedings of the National Academy of Sciences, 117 (7), 3388-3396
2. Kenkel, S., Mittal, S., and Bhargava, R. (2020). Molecular Characterization of Nanomaterials by Closed-Loop Atomic Force Microscopy-Infrared Spectroscopic Nature Communications, 11 (1), 1-10.
3. Mittal, S and Bhargava, R. (2019). A comparison of mid-infrared spectral regions on the accuracy of tissue classification, Analyst, 144, 2635-2642.
4. Mittal, S., Stoean, , Kajdacsy-Balla, and Bhargava, R. (2019). Digital Assessment of Stained Images for Comprehensive Tumor Analysis. Frontiers in Bioengineering and Biotechnology, 7, 246.
5. Gupta, S.*, Mittal, S.*, Kadjacsy-Balla, A., Bhargava, R., and Bajaj, C. (2019). A Fully Automated, Faster Noise Rejection Approach to Increasing the Analytical Capability of Chemical Imaging for Digital Histopathology. PloS One, 14(4), e0205219. (*equal contribution)
6. Mittal, S., Yeh, , Leslie, L. S., Kenkel, S., Kajdacsy-Balla, A., and Bhargava, R. (2018). Simultaneous cancer and tumor microenvironment subtyping using confocal infrared microscopy for all-digital molecular Histopathology. Proceedings of the National Academy of Sciences, 115 (25).