AI-based early detection of pancreatic ductal adenocarcinoma
AI-based early detection of pancreatic ductal adenocarcinoma
Current Research at Mayo Clinic
Current Research at Mayo Clinic
Embedded Files
Pancreatic ductal adenocarcinoma (PDAC) or pancreatic cancer is the third leading cause of cancer-related deaths in the USA and early detection of this cancer is an urgent unmet need to improve the survival time of the pancreatic cancer patients. For that reason, I developed an artificial intelligence (AI)-based algorithm that can detect patients with early pancreatic cancer with a high accuracy. For developing the algorithm, I utilized pre-diagnostic CT images that contain suggestive information about early pancreatic cancer. The CTs were obtained 3-36 months before actual diagnosis of pancreatic cancer. First, I computationally extracted various radiomics features from the manually segmented pancreas on these CT images. These features contain crucial information about cancerous tissues and are usually imperceptible by human radiologists on CT images. Second, I selected the most important features through a feature selection algorithm to train four different machine-learning (ML) models. I trained these models using 110 pre-diagnostic CTs with early pancreatic cancer and 182 control CTs with normal pancreas. Eventually, I tested the performance of these models on 45 prediagnostic and 83 control CTs. My ML-model obtained an accuracy of 92.2% for detecting early pancreatic cancer from these CT images. Additionally, I compared the performance of this ML-model against human radiologists on the same 128 (45 pre-diagnostic + 83 control) test CTs. As compared to the ML model that achieved an accuracy of 92.2%, human radiologists achieved an accuracy of only 66% for detecting pancreatic cancer on these CTs and therefore was outperformed by the ML model by a significant margin.
Finally, in order to assess the robustness of this ML model, I performed a variational study where I introduced varying degree of perturbations to the radiomics features e.g., noise, segmentation variability, etc. and tested the model’s performance against these perturbations. The model showed consistently good performance. For example, model’s accuracy was 92.2% and 89.1% where 1% and 5% noise was added to the radiomics features. The overall range of accuracy was 87%-92% with varying degree of perturbations that further showed the robustness of the model.
Representative publications:
Mukherjee S, Patra A, Khasawneh H, Korfiatis P, Rajamohan N, Suman G, Majumder S, Panda A, Johnson MP, Larson NB, Wright DE, Kline TL, Fletcher JG, Chari ST, Goenka AH. Radiomics-Based Machine-Learning Models Can Detect Pancreatic Cancer on Prediagnostic CTs at a Substantial Lead Time Prior to Clinical Diagnosis. Gastroenterology 163 (5), 1435-1446, 2022 (Impact Factor: 29.4). (https://pubmed.ncbi.nlm.nih.gov/35788343/)
Mukherjee S, Korfiatis P, Patnam NJ, Trivedi KH, Karbhari A, Suman G, Fletcher JG, Goenka AH. Assessing the Robustness of a Machine Learning Model for Early Detection of Pancreatic Adenocarcinoma (PDA): Evaluating Resilience to Variations in Image Acquisition and Radiomics Workflow Using Image Perturbation Methods, Abdominal Radiology 2024 (https://pubmed.ncbi.nlm.nih.gov/38175255/)
Mukherjee, S., Antony, A., Patnam, N.G. et al. Pancreas segmentation using AI developed on the largest CT dataset with multi-institutional validation and implications for early cancer detection. Sci Rep 15, 17096 (2025) (https://pubmed.ncbi.nlm.nih.gov/40379726/ )
Korfiatis P, Suman G, Patnam NJ, Trivedi KH, Karbhari A, Mukherjee S, Cook C, Klug JR, Patra A, Khasawneh H, Rajamohan N, Fletcher JG, Truty MJ, Majumder S, Bolan CW, Sandrasegaran K, Chari ST, Goenka AH. Automated Artificial Intelligence Model Trained on a Large Data Set Can Detect Pancreas Cancer on Diagnostic Computed Tomography Scans As Well As Visually Occult Preinvasive Cancer on Prediagnostic Computed Tomography Scans, Gastroenterology 165(6), 1533-1546, 2023. (https://pubmed.ncbi.nlm.nih.gov/37657758/)
Editorials on this work
Rosenthal, Michael, Schawkat, Khoschy, Wolpin, Brian, “A Growing Hope for Earlier Detection of Pancreatic Cancer”. Gastroenterology, Volume 163, Issue 5, 1170 – 1172. (https://pubmed.ncbi.nlm.nih.gov/35961377/ )
Rosenthal MH, Schawkat K. “Beyond the AJR: CT Radiomic Features of the Pancreas Predict Development of Pancreatic Cancer.” AJR. American Journal of Roentgenology. 2023 (https://pubmed.ncbi.nlm.nih.gov/36197051/ )
Søreide K, Ismail W, Roalsø M, Ghotbi J, Zaharia C, “Early Diagnosis of Pancreatic Cancer: Clinical Premonitions, Timely Precursor Detection and Increased Curative-Intent Surgery”. Cancer Control. 2023;30 (https://pubmed.ncbi.nlm.nih.gov/36916724/)
AI-based segmentation of pancreatic ductal adenocarcinoma
AI-based segmentation of pancreatic ductal adenocarcinoma
Current Research at Mayo Clinic
Current Research at Mayo Clinic
Pancreatic ductal adenocarcinoma (PDAC) has a dismal 5-year survival rate. Treatment outcomes vary substantially among patients who undergo surgical resection and response assessment to non-surgical modalities remains a challenge. To optimize treatment outcomes, there is a critical need of precision predictive and prognostic biomarkers, which can guide customization of treatment strategies to the unique biological profile of each patient's neoplasm.
Recently, quantitative imaging biomarkers have shown promising results for early detection of PDAC, its risk-stratification, therapy response assessment, and prognostication. When combined with computational methods, these biomarkers outperform traditional approaches for prediction of outcomes-of-interest. However, lack of robust tools for tumor segmentation is the main barrier for the clinical integration of these insights, for their prospective validation in adaptive oncologic trials, and for the discovery of next generation biomarkers. In addition, precise tumor segmentation is essential for applications such as treatment response assessment and 3D modeling for surgical and radiation therapy planning.
Manual PDAC segmentation is a tedious and time-consuming process, which is limited by high inter- and intra-reader variability due to variations in the shape, size, location, and tumor density (hypo- or isodense) coupled with the ill-defined boundaries of PDAC. Even deep learning (DL)-based PDAC segmentation has been consistently less accurate [range of dice similarity coefficients (DSCs): 0.52–0.61] vis-à-vis other solid tumors. Second, most DL models have been trained on small datasets or have not been validated on public datasets. Therefore, there are inevitable risks of over-fitting and over-estimation of model performances. Finally, DL-based segmentation of PDAC is an extremely imbalanced computational problem due to small tumor volume vis-à-vis large non-tumoral anatomy on imaging. We hypothesized that limiting the input to the peri-tumoral anatomy through a 3D bounding box coupled with the use of a large and diverse training dataset could overcome these challenges. Therefore, we developed a computationally efficient bounding box-based 3D CNN model that showed the highest reported accuracy [Dice Similarity Coefficient (DSC: 0.84 ± 0.06)] for user-guided volumetric PDA segmentation on standard-of-care CTs. Further, the model's high performance was generalizable on two multi-institutional open-source datasets – the MSD (DSC 0.82 ± 0.06) and the TCIA (DSC 0.84 ± 0.08).
Representative publication:
Mukherjee S, Patra A, Khasawneh H, Korfiatis P, Rajamohan N, Suman G, Majumder S, Panda A, Johnson MP, Larson NB, Wright DE, Kline TL, Fletcher JG, Chari ST, Goenka AH. Bounding Box-Based 3D AI Model for User-Guided Volumetric Segmentation of Pancreatic Ductal Adenocarcinoma on Standard-of-Care CTs, Pancreatology 23(5), 522-529, 2023 (https://pubmed.ncbi.nlm.nih.gov/37296006/ ).
AI-based detection of type-2 diabetes
AI-based detection of type-2 diabetes
Current Research at Mayo Clinic
Current Research at Mayo Clinic
Diabetes is the most common chronic metabolic disease globally, with approximately 1.5 million Americans diagnosed each year, 90% of whom have type 2 diabetes. Despite accessible glucose testing and risk-assessment tools, about 30% of type 2 diabetes cases are undiagnosed, and globally, 50% of people with diabetes (over 185 million) remain undiagnosed. This is because early-stage hyperglycemia often lacks severe symptoms. Undiagnosed patients still face increased risk of complications, highlighting the need for new biomarkers to aid early detection.
In patients with diabetes, the pancreas undergoes microstructural changes, such as reduced volume and increased fat content, which are difficult to detect through imaging at early stages. Objective tools for quantifying these changes could help identify undiagnosed diabetes in routine CT scans. Radiomics, a method that extracts quantitative biomarkers from imaging, combined with machine learning (ML), can identify subclinical changes. This technique has been used for detecting early pancreatic cancer and other conditions, and could potentially be applied to detect type 2 diabetes through pancreatic imaging. The goal of the study was to explore whether ML models using pancreatic radiomics from CT scans could identify the imaging signature of type 2 diabetes.
Representative publication:
Wright DE#, Mukherjee S#, Patra A, Khasawneh H, Korfiatis P, Suman G, Chari ST, Kudva YC, Kline TL, Goenka AH. Radiomics-based machine learning (ML) classifier for detection of type 2 diabetes on standardof-care abdomen CTs: a proof-of-concept study. Abdominal Radiology, 47, 3806-3816 (2022) (#co- first author) (https://pubmed.ncbi.nlm.nih.gov/38175255/ ).
Implementation of Dose Volume Constraint (DVC)
Implementation of Dose Volume Constraint (DVC)
Past Research at Memorial Sloan-Kettering Cancer Center (MSKCC)
Past Research at Memorial Sloan-Kettering Cancer Center (MSKCC)
Fig.1. Dose Volume Histogram (DVH) before (solid) and after (dashed) implementing the proposed algorithm. Dose-volume constraints (DVCs) are represented by 'x' .
Dose volume constraints (DVCs) play an important role in modern IMRT treatment. DVC allows only a certain percentage of volume of an organ to be overdosed and therefore can provide a clinically desirable dose distribution. DVCs are usually expressed as V (>d Gy) <v% which indicates that organ volume (V) that receives more than d Gy (Gray, unit of dose) is restricted to v% of volume V. Although, DVCs are crucial in controlling the amount of radiation dose in various organs, implementing DVC into intensity modulated radiation therapy (IMRT) optimization is a daunting task as it makes the optimization objective a non-convex function. DVC based optimization problem can be solved by a method named mixed integer programming (MIP). Although MIP can provide a ground truth solution, it is computationally expensive and takes many hours to solve. In my research, I adopt a "convex relaxation" method that simplifies MIP to provide an initial solution. I use this initial solution and apply an efficient heuristic to solve DVC based optimization problem. The proposed method is computationally faster and able to satisfy DVC as shown in figure 1.
Representative publication
Mukherjee S, Hong L, Deasy J, and Zarepisheh M, “Integrating soft and hard dose-volume constraints into hierarchical constrained IMRT optimization” Medical Physics, 47(2), 414-421, 2019 (Editor’s Choice). (https://pubmed.ncbi.nlm.nih.gov/31742731/ )
Fig.2. a) Image acquired by CBCT with 20ms pulse duration (ground truth) b) Image with 4 ms pulse (noisy image) c, d) Denoised images using Total-Variation with Split-Bregman (TVSB) and Total-variation with Nesterov's (TVN) algorithm respectively.
Total-variation based nose-reduction algorithms for low-dose cone-beam computed tomography (CBCT)
Total-variation based nose-reduction algorithms for low-dose cone-beam computed tomography (CBCT)
Past Research at St Jude Children's Research Hospital, Memphis, TN
Past Research at St Jude Children's Research Hospital, Memphis, TN
My research at St Jude primarily focused on developing total-variation based noise reduction algorithms for low-dose cone-beam CT imaging. Cone-beam computed tomography (CBCT) has been extensively used in radiation therapy as a medical imaging technique to acquire a high-resolution volumetric image of a patient for treatment positioning. However, the repeated use of CBCT during the treatment course increases the risk of extra radiation dose delivered to patients. The extra radiation exposure to normal tissue during CBCT increases the risk of cancer and genetic defects. Therefore, the unwanted CBCT radiation dose must be minimized in order for the patients to truly benefit from the modern medical imaging techniques.
One way to reduce the CBCT dose is to acquire the CBCT projection data by reducing either the X-ray source tube current or the pulse duration. However, for lower current (mA) level or shorter pulse duration (ms), the projection image is contaminated with excessive quantum noise.
I was interested in reducing the noise level on the projected image that is acquired by low mAs CBCT protocols. A total-variation (TV) based noise reduction algorithm was studied and applied to a computer–simulated phantom, physical phantom and patient data. The algorithm had shown to have the potential in reducing the noise level for low-dose CBCT images without compromising the contrast and resolution of the images. Figure-2 shows the denoised images using TV-based algorithm.
Representative publication:
Mukherjee S, Farr JB and Yao W, “A study of total-variation based noise reduction algorithms for low-dose cone-beam computed tomography”, International Journal of Image Processing 10(4), 188-204, 2016.
Fig.3. A) original conductivity profile for a breast-like geometry B) corresponding reconstructed conductivity profile using thermo-acoustic tomography (TAT) C) Line profile of original and reconstructed conductivity distribution
Thermo-acoustic tomography (TAT) based simulation of breast and prostate cancer imaging
Thermo-acoustic tomography (TAT) based simulation of breast and prostate cancer imaging
Past Research at Oklahoma State University, Stillwater, OK
Past Research at Oklahoma State University, Stillwater, OK
My Ph.D. research mainly focused on simulation of microwave-induced thermo-acoustic tomography, abbreviated as MI-TAT, for breast and prostate cancer imaging. MI-TAT is an emerging medical imaging modality which combines both microwave imaging and ultrasound imaging for providing high-contrast and high-resolution images of cancerous tissues. In MI-TAT, biological cells are irradiated by short-pulsed microwave energy in the frequency range of 434 MHz-3 GHz. Absorption of this energy by tissues causes a thermo-elastic expansion of the cells and produces an acoustic/pressure wave. Detection of this acoustic wave by ultrasonic detectors can give crucial information about the dielectric properties (e.g. relative permittivity and electrical conductivity) of the tissues. It has been well proved that malignant tissues usually exhibit different electrical conductivity from normal tissues because of different concentrations of ions and water and absorb more microwave energy. MI-TAT exploits this different microwave energy absorption characteristic to give an imaging contrast between malignant and normal tissues.
I developed a finite element method (FEM) based forward and inverse modeling of thermo-acoustic tomography to reconstruct conductivity distribution in a computer-simulated breast and prostate like geometry. The algorithm was written in C++/MATLAB. Since cancerous tissues possess different conductivity from normal tissues, reconstructing conductivity distribution using MI-TAT can give a precise location of cancerous tissues. Figure-3 shows a simulated reconstructed image using TAT for a breast-like geometry.
Representative publications:
Mukherjee S, Bunting C and Piao D, “Trans-rectal microwave-induced thermo-acoustic Computed tomography: An initial in-silico study,” X-Acoustics Imaging & Sensing, 1(1), 1-15, 2014.
Mukherjee S, Bunting C and Piao D, “Finite-element-method based reconstruction of heterogeneous conductivity distribution under point-illumination in trans-rectal imaging geometry for thermo-acoustic tomography”, Biomedical optics and 3D imaging, paper BSu.3A44, 2012.
Mukherjee S, Bunting C and Piao D, “Forward model of thermo-acoustic signal specific to intra-lumenal detection geometry”, Proc. SPIE, 7899, 36-42, 2011.
Selected Publications
Selected Publications
Peer-reviewed papers
1. Mukherjee, S., Antony, A., Patnam, N.G. et al. Pancreas segmentation using AI developed on the largest CT dataset with multi-institutional validation and implications for early cancer detection. Sci Rep 15, 17096 (2025). https://doi.org/10.1038/s41598-025-01802-9
2. Antony, A., Mukherjee, S., Bhinder, K., Murlidhar, M., Zarrintan, A., Goenka, AH.; Artificial Intelligence-Augmented Imaging for Early Pancreatic Cancer Detection. Visc Med 2025; https://doi.org/10.1159/000546603
3. Antony, A., Mukherjee, S., Bi, Y. et al. AI-Driven insights in pancreatic cancer imaging: from pre-diagnostic detection to prognostication. Abdominal Radiolology (2024). https://doi.org/10.1007/s00261-024-04775-x
4. Mukherjee S, Korfiatis P, Patnam NJ, Trivedi KH, Karbhari A, Suman G, Fletcher JG, Goenka AH. Assessing the Robustness of a Machine Learning Model for Early Detection of Pancreatic Adenocarcinoma (PDA): Evaluating Resilience to Variations in Image Acquisition and Radiomics Workflow Using Image Perturbation Methods, Abdominal Radiology 49(3), 964-974, 2024.
5. Oza, V. M., Mittal, N., Antony, A., Mukherjee, S., Goenka, A.H. (2024). "Advancements and Challenges in Artificial Intelligence-Driven Diagnosis and Management of Pancreatic Cancer." AI in Precision Oncology 1(4): 207-215.
6. Mukherjee S, Patra A, Khasawneh H, Korfiatis P, Rajamohan N, Suman G, Majumder S, Panda A, Johnson MP, Larson NB, Wright DE, Kline TL, Fletcher JG, Chari ST, Goenka AH. Bounding Box-Based 3D AI Model for User-Guided Volumetric Segmentation of Pancreatic Ductal Adenocarcinoma on Standard-of-Care CTs, Pancreatology 23(5), 522-529, 2023.
7. Korfiatis P, Suman G, Patnam NJ, Trivedi KH, Karbhari A, Mukherjee S, Cook C, Klug JR, Patra A, Khasawneh H, Rajamohan N, Fletcher JG, Truty MJ, Majumder S, Bolan CW, Sandrasegaran K, Chari ST, Goenka AH. Automated Artificial Intelligence Model Trained on a Large Data Set Can Detect Pancreas Cancer on Diagnostic Computed Tomography Scans As Well As Visually Occult Preinvasive Cancer on Prediagnostic Computed Tomography Scans, Gastroenterology 165(6), 1533-1546, 2023.
8. Mukherjee S, Patra A, Khasawneh H, Korfiatis P, Rajamohan N, Suman G, Majumder S, Panda A, Johnson MP, Larson NB, Wright DE, Kline TL, Fletcher JG, Chari ST, Goenka AH. Radiomics-Based Machine-Learning Models Can Detect Pancreatic Cancer on Prediagnostic CTs at a Substantial Lead Time Prior to Clinical Diagnosis. Gastroenterology 163 (5), 1435-1446, 2022 (Impact Factor: 29.4).
9. Wright DE#, Mukherjee S#, Patra A, Khasawneh H, Korfiatis P, Suman G, Chari ST, Kudva YC, Kline TL, Goenka AH. Radiomics-based machine learning (ML) classifier for detection of type 2 diabetes on standardof-care abdomen CTs: a proof-of-concept study. Abdominal Radiology, 47, 3806-3816 (2022) (#co- first author)
10. Mukherjee S, Paquette N, Wang Y, Panigrahy A, and Lepore N, “Deep-learning based tractography for neonates’”, SIPAIM-MICCAI, 2020.
11. Mukherjee S, Hong L, Deasy J, and Zarepisheh M, “Integrating soft and hard dose-volume constraints into hierarchical constrained IMRT optimization” Medical Physics, 47(2), 414-421, 2019 (Editor’s Choice).
12. Mukherjee S, Farr JB and Yao W, “A study of total-variation based noise reduction algorithms for low-dose cone-beam computed tomography”, International Journal of Image Processing 10(4), 188-204, 2016.
13. Mukherjee S, Bunting C and Piao D, “Trans-rectal microwave-induced thermo-acoustic Computed tomography: An initial in-silico study,” X- Acoustics Imaging & Sensing, 1(1), 1-15, 2014.
14. Mukherjee S, Bunting C and Piao D, “Finite-element-method based reconstruction of heterogeneous conductivity distribution under point- illumination in trans-rectal imaging geometry for thermo-acoustic tomography”, Biomedical optics and 3D imaging, paper BSu.3A44, 2012.
15. Mukherjee S, Bunting C and Piao D, “Forward model of thermo-acoustic signal specific to intra-lumenal detection geometry”, Proc. SPIE, 7899, 36-42, 2011.
16. Jiang Y, Mukherjee S, Stine JE, Bunting C, and Piao D, “FPGA-assisted strategy toward efficient reconstruction (FAStER) in diffuse optical tomography,” Biomedical optics and 3D imaging, paper BSu.D18, 2010.
Conference abstracts
1. Mukherjee S, Korfiatis P, Patnam NJ, Trivedi KH, Karbhari A, Fletcher JG, Goenka AH, Fully Automated Volumetric Segmentation of Pancreatic Ductal Adenocarcinoma on CTs, Radiological Society of North America (RSNA), Chicago 2023.
2. Mukherjee S, Khasawneh H, Rajamohan N, Suman G, Singh A, Korfiatis P, Goenka AH. Convolutional Neural Network (CNN) for Volumetric Segmentation of Pancreatic Ductal Adenocarcinoma (PDA) on CTs Radiological Society of North America (RSNA), Chicago 2022.
3. Mukherjee S, Korfiatis P, Khasawneh H, Rajamohan N, Chari ST, Majumder S, Goenka AH. Radiomics-Based Machine Learning Model for Pancreatic Cancer Detection on Prediagnostic CTs: Assessment of Robustness through Image Perturbations SIIM-CMIMI, Baltimore, 2022.
4. Mukherjee S, Khasawneh H, Rajamohan N, Suman G, Singh A, Korfiatis P, Goenka AH. Convolutional Neural Network (CNN) for Volumetric Segmentation of Pancreatic Ductal Adenocarcinoma (PDA) on CTs Radiological Society of North America (RSNA), Chicago 2022.
5. Mukherjee S, Khasawneh H, Rajamohan N, Suman G, Singh A, Korfiatis P, and Goenka AH, “Convolutional Neural Network (CNN) for Volumetric Segmentation of Pancreatic Ductal Adenocarcinoma (PDA) on CTs”, Society of Advanced Body Imaging (SABI), New Orleans, LA 2022.
6. Mukherjee S, Khasawneh H, Rajamohan N, Suman G, Singh A, Korfiatis P, and Goenka AH, “A Three-dimensional (3D) Convolutional Neural Network (CNN) for Semi-Automated Volumetric Segmentation of Pancreatic Ductal Adenocarcinoma (PDA) on CTs”, European Society of Radiology (ESR), Vienna, Austria, 2022.
7. Yang L, Mukherjee S, Khasawneh H, Patra A, Rajamohan N, Suman G, Singh A, and Goenka AH, “Radiomic Features Extracted from Volumetrically Segmented Pancreatic Ductal Adenocarcinoma (PDA): Robustness to a Combination of Image Perturbations Across Multiple CT Datasets”, European Society of Radiology (ESR), Vienna, Austria, 2022.
8. Singh A, Khasawneh H, Rajamohan N, Mukherjee S, Yang L, Suman G, Korfiatis P, and Goenka AH, “Impact of phase of CT acquisition on Convolutional Neural Network (CNN)-derived volumetric pancreas segmentations”, European Society of Radiology (ESR), Vienna, Austria, 2022.
9. Suman G, Korfiatis P, Patra A, Khasawneh H, Rajamohan N, Cook C, Mukherjee S, Fletcher J, Majumder S, Chari ST, Goenka AH. “Artificial Intelligence for Early Detection of Pancreatic Ductal Adenocarcinoma (PDA) on CTs”, Society of Abdominal Radiology (SAR), Austin, TX, 2022.
10. Mukherjee S, Korfiatis P, Suman G, Patra A, Khasawneh H, Rajamohan N, Panda A, Wright D, Majumder S and Goenka AH, “Machine Learning (ML) model based on CT radiomics signature for detection of pancreatic ductal adenocarcinoma (PDA) on prediagnostic CT scans”, Radiological Society of North America (RSNA), Chicago, IL, 2021.
11. Mukherjee S, Korfiatis P, Suman G, Patra A, Khasawneh H, Rajamohan N, Panda A, Wright D, Majumder S and Goenka AH, “Machine Learning (ML) model based on CT radiomics signature for detection of pancreatic ductal adenocarcinoma (PDA) on prediagnostic CT scans”, Society of Advanced Body Imaging (SABI), Washington DC, 2021.
12. Iheme L, Patra A, Suman G, Khasawneh H, Mukherjee S, Korfiatis P, and Goenka AH, “Impact Of Simulated Marginal Erosions of Volumetric Segmentation of Pancreatic Adenocarcinoma (PDA) on the Robustness of Radiomics Features”, American Roentgen Ray Society (ARRS), New Orleans, LA 2021.
13. Mukherjee S, Hong L, Deasy J, and Zarepisheh M, “A novel computationally tractable algorithm for integrating soft and hard dose-volume constraints into IMRT fluence optimization”, AAPM 2019, San Antonio, TX.
14. Mukherjee S, Hong L, Deasy J, and Zarepisheh M, “A computationally efficient algorithm for integrating dose-volume constraints into IMRT fluence optimization for automated treatment planning”, AAPM 2018, Nashville, TN.
15. Zhao H, Mukherjee S, Saraswat S, Tak J, Liang M, Witte R, and Xin H, “Full-wave Numerical Model for Thermoacoustic Imaging of the Human Breast and Detection of Breast Cancer”, IEEE APS/URSI 2018, BOSTON, MA.
16. Karunakaran C, Mukherjee S, Tak J, Saraswat S, Xin H, and Witte R, “Real-time thermoacoustic imaging and thermometry using a linear ultrasound array”, IEEE APS/URSI 2018, BOSTON, MA.
17. Mukherjee S, Farr JB, Merchant TE and Yao W, “Improvement of Image Registration using total-variation based noise reduction algorithms for low-dose cone-beam computed tomography”, AAPM 2016, Washington DC.
Mukherjee S, Yao W, “A comparative study of noise-reduction algorithms for low-dose cone-beam CT imaging”, AAPM 2015, Anaheim, CA.
Google Sites
Report abuse