Research Highlights
Breast Cancer Heterogeneity Quantification
Breast cancers are subtyped and treated differently based on their molecular makeup. Even within a given subtype, some cancers are admixed with other subtypes. We proposed metrics to measure molecular heterogeneity in cancer and showed that these can predict survival. We also demonstrated that the digital pathology images of breast tumors show the coexistence of multiple morphological phenotypes in accordance with the molecular heterogeniety.
|
Learn to Segment Nuclei
We created the largest dataset of hand-annotated nuclei to advance state-of-the art in generalized nuclei segmentation. We organized an international competition to train deep neural networks to segment nuclei in digital pathology in MICCAI 2018, based on our previous work. Currently, we are organizing a multi-organ nuclei segmentation and classification challenge at ISBI 2020. We use nuclei segmentation algorithms for several downstream computational pathology tasks including cancer mutation prediction, cancer staging, treatment effectiveness and outcome prediction, etc.
|
Prostate Cancer Recurrence Prediction
Accurate prediction of the treatment outcome is important for cancer treatment planning. We developed a fully-automated deep learning algorithm to predict prostate cancer recurrence after radical prostatectomy from H&E stained tissue images of individual patients. We demonstrated that our algorithm outperforms the current clinical grade diagnostic protocol to assess mid-Gleason grade prostate cancer. Our approach might help in choosing between a combination of treatment options such as active surveillance, radical prostatectomy, radiation, and hormone therapy.
|
Single Image Super Resolution
We developed several techniques that pushed the state-of-the-art in single image super-resolution. We showed that single image super-resolution is a spatially local problem where simpler but insightful models work better than brute force deep learning. We also exploited the sparsity and localization properties of wavelets to develop simple and efficient but effective single image super resolution algorithms.
|
Publications
Journals
- D. Anand, N.C. Kurian, S. Dhage, Neeraj Kumar, S. Rane, P. H. Gann, A. Sethi, "Deep learning to estimate human epidermal growth factor receptor 2 status from hematoxylin and eosin-stained breast tissue images", in Journal of Pathology Informatics, Vol. 11, July 2020
- Neeraj Kumar, Ruchika Verma, et al. "A multi-organ nuclei segmentation challenge", in IEEE Transactions on Medical Imaging, vol. 39, no. 5, pp. 1380-1391, May 2020
- Neeraj Kumar, Dan Zhao, Dulal Bhaumik, Amit Sethi and Peter H. Gann, "Quantification of intrinsic subtype ambiguity in Luminal A breast cancer and its relationship to clinical outcomes", in BMC Cancer 19, 215 (2019)
- Neeraj Kumar, Phanikrishna Uppala, Karthik Duddu, Hari Sreedhar, Vishal Varma, Grace Guzman, Michael Walsh, and Amit Sethi, "Semi-supervised Segmentation of Hyperspectral Tissue Images", in IEEE Transactions on Medical Imaging vol. 38, no. 5, pp. 1304-1313, May 2019
- Neeraj Kumar and Amit Sethi, "Super Resolution by Comprehensively Exploiting Dependencies of Wavelet Coefficients", in IEEE Transactions on Multimedia, vol. 20, no. 2, pp. 298-309, February 2018
- Amit Sethi, Lingdao Sha, Neeraj Kumar, Virgilia Macias, Ryan J. Deaton, and Peter Gann, "Computer Vision Detects subtle Histological Effects of Dutasteride on Benign Prostate", in British Journal of Urology International, 122(1), pp.143-151, Feb. 2018
- Neeraj Kumar , Ruchika Verma, Sanuj Sharma, Surabhi Bhargava, Abhishek Vahadane and Amit Sethi, "A dataset and a technique for generalized nuclear segmentation in histological images for computational pathology", in IEEE Transactions on Medical Imaging, vol. 36, no. 7, pp 1550-1560, July 2017
- Neeraj Kumar and Amit Sethi, "Fast Learning-Based Single Image Super-Resolution", in IEEE Transactions on Multimedia, vol. 18, no. 8, pp. 1504-1515, Aug. 2016
- Neeraj Kumar, Ruchika Verma and Amit Sethi, "Convolutional neural networks for wavelet domain single image super resolution", in Pattern Recognition Letters, Volume 90 Issue C, Pages 65-71, April 2017
- Amit Sethi, Lingdao Sha, Abhishek Vahadane, Ryan J Deaton, Neeraj Kumar, Virgilia Macias, Peter H. Gann, "Empirical Comparison of Color Normalization Methods for Epithelial-Stromal Classification in H&E Images", in Journal of Pathology Informatics, January 2016
- Amit Sethi and Neeraj Kumar, "Systems and methods for Computational pathology using Points-of-Interest", US Patent No. 10,573,003 (Issued) and Indian Patent application no. 201711005034
- Hrushikesh Loya, Pranav Poduval, Deepak Anand, Neeraj Kumar, and Amit Sethi, “Uncertainity Estimation in Cancer Survival Prediction”, in International Conference on Learning Representations (ICLR) 2020, Addis Ababa, Ethiopia
- Neeraj Kumar, Cheng Lu, Joseph Willis, and Anant Madabhushi, “Computationally Derived Morphological Features of Cancer Nuclei from Colon Whole Slide Images Can Distinguish Stage 2 from Stage 4 Colon Cancers”, in 109𝑡ℎ annual meeting of United States and Canadian Academy of Pathology 2020, LA (CA), USA
- Hrushikesh Loya, Deepak Anand, P. Poduval, Neeraj Kumar, and Amit Sethi, “A Bayesian framework to quantify survival uncertainty”, in Molecular Analysis for Personalised Therapy Congress 2019, London, UK
- Neeraj Kumar, Yash Dharmamer, Amit Sethi and Peter Gann, “Quantification of intratumoral heterogeneity in individual luminal A breast cancers from whole transcriptome data through semi-supervised learning”, in annual meeting of American Association of Cancer Research (AACR) 2019, Atlanta (GA), USA
- Neeraj Kumar, Dan Zhao, Dulal Bhaumik, Amit Sethi and Peter Gann, “Quantifying intrinsic subtype admixture in luminal A breast cancer and its relationship to clinical outcomes”, in San Antonio Breast Cancer Symposium (SABCS) 2018, San Antonio (TX), USA
- Shubham Dhage, Deepak Anand, Neeraj Kumar, Peter Gann and Amit Sethi, “Computer vision detects morphological correlates of HER2 positive breast cancer in HE stained histological images”, in San Antonio Breast Cancer Symposium (SABCS) 2018, San Antonio (TX), USA
- Neeraj Kumar, Dan Zhao, Amit Sethi and Peter Gann, “PAM50 Subtype Admixture in Individual Breast Cancers and the Relationship of this Intratumoral Heterogeneity to Clinical Variables”, in annual meeting of American Association of Cancer Research (AACR) 2018, Chicago (IL), USA
- Peter Gann, Lingdao Sha, Neeraj Kumar, Virgilia Macias, and Amit Sethi, "Computer Vision Detects Subtle Histological Effects of Dutasteride on Benign Prostate", in Experimental Biology 2017 , Chicago, USA
- Neeraj Kumar, Ruchika Verma, Ashish Arora, Abhay Kumar, Sanchit Gupta and Amit Sethi, "Convolutional neural networks for Prostate cancer recurrence prediction", in SPIE Medical Imaging
- Ruchika Verma, Neeraj Kumar, Amit Sethi and Peter H Gann, "Detecting multiple sub-types of breast cancer in a single patient", in IEEE International Conference on Image Processing (ICIP)-2016 , Phoenix, USA
- A. Vahadane, Neeraj Kumar and Amit Sethi, "Super resolution of histological images", in International Symposium on Biomedical Imaging (ISBI)-2016 , Prague, Czech Republic
- Neeraj Kumar and Amit Sethi, "On spatial neighborhood of patch based Super Resolution", in IEEE International Conference on Image Processing (ICIP)-2015 , Montreal Canada
- Neeraj Kumar, Said Mossouai, Jerome Idier and David Brie, "Impact of sparse representation on admissible solutions of spectral unmixing by Nonnegative Matrix Factorization", in IEEE Workshop on Hyperspectral Image and Signal Processing (WHISPERS) 2015 , Tokyo, Japan
- Neeraj Kumar, Naveen Kumar Rai and Amit Sethi, "Learning to predict Super Resolution wavelet coefficients", in 21st IEEE International conference on Pattern Recognition (ICPR) , 11-15 November, 2012, Tsukuba, Japan
- Neeraj Kumar and Amit Sethi, "On Image driven choice of wavelet basis for image super resolution", 9th IEEE International conference on Signal Processing and Communications (SPCOM) , 22-25 JULY, 2012, IISc Bangalore, India
- Neeraj Kumar, Amit Sethi and Rahul Nallamothu, "Neurel Network based image deblurring", 11th IEEE Symposium on Neural Network Applications in Electrical Engineering (NEUREL) , 23-25 September, 2012, Belgrade, Serbia
- Neeraj Kumar, Pankaj Kumar Deswal, Jatin Mehta and Amit Sethi, "Neurel Network based single image super resolution", in 11th IEEE Symposium on Neural Network Applications in Electrical Engineering (NEUREL) , 23-25 September, 2012, Belgrade, Serbia
- Amit Sethi, Neeraj Kumar and Naveen Kumar Rai, "Spatial Neighborhood based learning set-up for Super Resolution", in IEEE International Conference on Innovations in Social and Humanitarian Engineering (INDICON) , 07-09 December 2012, Kochi, Kerala, India
- Neeraj Kumar and Amit Sethi, "Image Interpolation based on Inter-scale Dependency of Wavelet Coefficients", in proceedings of National Workshop on Wavelets, Multi-Resolution and Multi-Fractal Analysis in Earth, Ocean and Atmospheric Sciences-Current Trends , February 2012, IIT Bombay
- Neeraj Kumar, Said Moussaoui, and Jerome Idier, "On solution space reduction of non-negative matrix factorizations through sparse representation", Ecole Centrale Nantes, France, March 2014
- "Sub-type admixture leads to poor clinical outcome among breast cancer cases classified as Luminal A", Neeraj Kumar, Dan Zhao, Peter Gann and Amit Sethi,(Journal paper)
- "Uniqueness of two-sided sparsified Non-Negative Matrix Factorizations", Neeraj Kumar, Said Mossouai, Jerome Idier and David Brie (Journal paper)
- "On reducing the number of admissible solutions of non-negative matrix factorization", Neeraj Kumar, Amit Sethi, Said Moussaoui, David Brie, Jerome Idier (Journal paper)
Press and Videos
- "AI Pathologist Helps Zero in on Correct Cancer Diagnosis", blog entry as our group ended up in the five finalists for NVIDIA Global Impact Award 2017
- Deep learning for computational pathology by Dr. Neeraj Kumar for HasGeek Deep Learning Conference 2016
- Computational Pathology-An Introduction by Dr. Neeraj Kumar for India Deep Learning Initiative 2017