Bladder cancer is a common type of cancer that can cause substantial morbidity and mortality among both men and women. Bladder cancer causes over 15,210 deaths per year in the United States. It is estimated that 72,570 new bladder cancer cases will be diagnosed in 2013. Correct staging of the bladder cancer is crucial for the decision of neoadjuvant chemotherapy and minimizing the risk of under-treatment or over-treatment. A reliable assessment of the response to neoadjuvant therapy at an early stage is vital for identifying tumors that do not respond and allowing the patient a chance of alternative treatment. MRI and CT are important methods for pre-treatment staging or treatment response monitoring for a variety of bladder cancers. CT is an effective non-invasive modality for measuring primary site gross tumor volume (GTV) and the addition of MRI is on the rise. GTV has been used as a biomarker for predicting treatment outcome of bladder tumors. Other pathological information and diagnostic test (bimanual evaluation, cystoscopy) results and immunohistochemical biomarkers are also useful for staging and treatment response monitoring. The goal of this project is to develop effective decision support tools that merge image-based and non-image-based biomarkers to assist radiologists and oncologists in assessment of cancer stage and change as a result of treatment. We will (1) develop a quantitative image analysis tool (QIBC) for bladder GTV estimation on multi-modality (MM) images, (2) develop a computer decision support system (CDSS-S) to assist clinicians in cancer staging, (3) develop a computer decision support system (CDSS-T) to assist clinicians in evaluation of the change in the tumor characteristics as a result of neoadjuvant treatment, (4) evaluate the effects of QIBC and CDSS-T on inter-clinician variability and efficiency in estimation of GTV and treatment response, and (5) evaluate CDSS-S and CDSS-T as decision support tools in pilot clinical studies. We hypothesize that the use of QIBC, CDSS-S and CDSS-T can improve the clinicians' accuracy, consistency and efficiency in bladder GTV estimation on MM imaging exams, the assessment of bladder cancer stage and response to treatment. To test our hypothesis, we will perform the following specific tasks: (1) to collect a database of multi-modality MR, CT exams of bladder cancers for development, training and testing of the QIBC and CDSS algorithms; (2) to develop advanced computer vision techniques to quantitatively estimate bladder GTV and image characteristics; (3) to develop predictive models using machine learning techniques to combine MM image-based, pathological and immunohistochemical biomarkers for cancer staging and determination of non-responders; (4) to compare the inter-clinician variability and efficiency in clinicians' estimation of GTV and treatment response with and without the proposed QIBC and CDSS-T by observer studies; and (5) to evaluate the CDSS-S and CDSS-T as decision support tools in pilot clinical studies.
If successfully developed, the CDSS-S and CDSS-T can serve as non-invasive, objective, and reproducible clinical decision support systems for cancer staging and treatment response monitoring. Correct staging of the bladder cancer is crucial for the decision of chemotherapy treatment and minimizing the risk of under-treatment or over-treatment. If the response to chemotherapy can be estimated accurately it is possible to identify those patients that do not respond, stop the treatment early, and seek alternative treatment. In addition, although we will focus on the specific application to the bladder tumors in this proposed project, we will design the decision support tools in a modular, expandable, and re-trainable framework. The software package will be versatile and is adaptable to other tumor types or imaging modalities in the future by proper retraining with case samples of the tumor type of interest and expansion of the decision support tools as needed. Therefore, the development of the CDSS-S and CDSS-T will potentially benefit not only the bladder tumor patients but also patients with other tumor types that require staging and monitoring of treatment response.
|Gordon, Marshall N; Hadjiiski, Lubomir M; Cha, Kenny H et al. (2018) Deep-learning convolutional neural network: Inner and outer bladder wall segmentation in CT urography. Med Phys :|
|Balagurunathan, Yoganand; Beers, Andrew; Kalpathy-Cramer, Jayashree et al. (2018) Semi-automated pulmonary nodule interval segmentation using the NLST data. Med Phys 45:1093-1107|
|Cha, Kenny H; Hadjiiski PhD, Lubomir M; Cohan Md, Richard H et al. (2018) Diagnostic Accuracy of CT for Prediction of Bladder Cancer Treatment Response with and without Computerized Decision Support. Acad Radiol :|
|Cha, Kenny H; Hadjiiski, Lubomir; Chan, Heang-Ping et al. (2017) Bladder Cancer Treatment Response Assessment in CT using Radiomics with Deep-Learning. Sci Rep 7:8738|
|Garapati, Sankeerth S; Hadjiiski, Lubomir; Cha, Kenny H et al. (2017) Urinary bladder cancer staging in CT urography using machine learning. Med Phys 44:5814-5823|
|Cha, Kenny H; Hadjiiski, Lubomir; Samala, Ravi K et al. (2016) Urinary bladder segmentation in CT urography using deep-learning convolutional neural network and level sets. Med Phys 43:1882|
|Kalpathy-Cramer, Jayashree; Mamomov, Artem; Zhao, Binsheng et al. (2016) Radiomics of Lung Nodules: A Multi-Institutional Study of Robustness and Agreement of Quantitative Imaging Features. Tomography 2:430-437|
|Cha, Kenny H; Hadjiiski, Lubomir M; Samala, Ravi K et al. (2016) Bladder Cancer Segmentation in CT for Treatment Response Assessment: Application of Deep-Learning Convolution Neural Network-A Pilot Study. Tomography 2:421-429|
|Hadjiiski, Lubomir; Weizer, Alon Z; Alva, Ajjai et al. (2015) Treatment Response Assessment for Bladder Cancer on CT Based on Computerized Volume Analysis, World Health Organization Criteria, and RECIST. AJR Am J Roentgenol 205:348-52|