Browsing by Author "Crump, Trafford"
Now showing 1 - 6 of 6
Results Per Page
Sort Options
- ItemOpen AccessDetecting Eye Diseases and Intraocular Lesions from Fundus Images Using Deep Learning Approaches(2023-12-20) Shakeri Hoosein Abad, Esmaeil; Far, Behrouz; Crump, Trafford; Mohammed, Emad; Kim, KangsooIn this study, the focus begins with addressing the critical issue of diabetic retinopathy (DR) detection, a leading cause of blindness globally, by using a combination of SHapley Additive exPlanations (SHAP) analysis and transfer learning ResNet50 model. Achieving impressive accuracy rates of 97% for binary and 81% for multi-class DR classification, the study demonstrates the potential of SHAP analysis to enhance interpretability and contextual understanding of prediction outcomes. Shifting the study to uveal melanoma (UM), an intraocular cancer with significant risks, the research used similar methodologies to predict UM, achieving a high binary classification accuracy of 82.5% in InceptionV3 model. The application of SHAP analysis once again highlights its value in shedding light on prediction rationales and improving result comprehension. The study further extends into the use of four distinct convolutional neural network (CNN)-based architectures for UM detection, emphasizing the manual collection and preprocessing of 854 RGB fundus images. Through transfer learning, DenseNet169 appears as the most accurate model, achieving 89% accuracy in binary classification of choroidal nevus (CN). Essentially, SHAP analysis continues to play an essential role in enhancing interpretability, offering detailed insights into the significant image regions influencing CN predictions. In conclusion, this study emphasises the power of combining deep transfer learning CNN-based models, and SHAP analysis to not only achieve robust predictive performance but also to address the critical challenge of interpretability in deep learning models, contributing significantly to the fields of medical image analysis and diagnostic decision-making.
- ItemOpen AccessPersonalized Recommendation Using Reinforcement Learning(2022-05) Afsar, Mohammad Mehdi; Far, Behrouz; Crump, Trafford; Yanushkevich, Svetlana; Murari, Kartikeya; Reformat, MarekThe massive volume of information available on the web leads to the problem of information overload, which makes it difficult for a decision maker to make right decisions. Recommender systems (RSs) are software tools and algorithms that have been developed with the idea of helping users find their items of interest through predicting their preferences or ratings on items. It has been shown that the problem of recommending items to the user could be considered as a sequential decision problem and be formulated as a Markov decision process, so reinforcement learning (RL) algorithms can be used to solve this problem. The primary aim of this dissertation is to investigate this topic and to propose new recommendation approaches using RL. The first part of this thesis, namely chapters 2 and 3, presents a healthcare use case of intelligent agents and RSs. In particular, chapter 2 presents a high-level design, called ALAN, to play the role of a patient decision aid for prostate cancer patients. ALAN is a multilayered, multi-agent system in which each agent is responsible to provide a specific service in order to facilitate shared decision making for these patients. Moreover, an article RS with learning ability is proposed in chapter 3 to represent the Learning agent in ALAN, which combines multi-armed bandits with knowledge-based RSs for the provision of information for cancer patients. Motivated by the first part, the second part of this thesis (chapters 4 and 5) deeply explores the topic of recommendation using RL algorithms. More precisely, chapter 4 provides a thorough literature review on reinforcement learning based recommender systems (RLRSs). The main goal of this chapter is to provide a deep analysis of almost all important RLRSs proposed and to depict a vista toward the field since the beginning. This chapter illustrates the importance of deep RL (DRL) in reviving the use of RL for RSs. Chapter 5 is an outcome of this chapter, which tries to address an important problem of DRL when applied to real applications like RSs, i.e., sample inefficiency. In this chapter, RL is combined with imitation learning in order to accelerate RL agent’s learning and to make it sample efficient. Finally, chapter 6 proposes a new recommendation approach from a totally new perspective. This chapter borrows ideas from Computer Networks field, clustering in wireless sensor networks in particular, and presents a probabilistic recommendation approach that can balance the similarity-diversity trade-off. The proposed approach is simple, scalable, and completely explainable.
- ItemOpen AccessThe feasibility of patient-reported outcomes, physical function, and mobilization in the care pathway for head and neck cancer surgical patients(2022-05-27) Daun, Julia T.; Twomey, Rosie; Capozzi, Lauren C.; Crump, Trafford; Francis, George J.; Matthews, T. W.; Chandarana, Shamir; Hart, Robert D.; Schrag, Christiaan; Matthews, Jennifer; McKenzie, C. D.; Lau, Harold; Dort, Joseph C.; Culos-Reed, S. N.Abstract Background Head and neck cancer (HNC) patients are an understudied population whose treatment often includes surgery, causing a wide range of side effects. Exercise prehabilitation is a promising tool to optimize patient outcomes and may confer additional benefits as a prehabilitation tool. The primary objective of this study was to assess the feasibility of measuring patient-reported outcomes (PROs), physical function, and in-hospital mobilization across the HNC surgical timeline in advance of a future prehabilitation trial. The secondary objective was to examine potential changes in these outcomes across the surgical timeline. Methods HNC patients scheduled to undergo oncologic resection with free-flap reconstruction completed assessments of PROs and physical function at three timepoints across the surgical timeline (baseline, in-hospital, and postsurgical/outpatient). Mobilization was measured during the in-hospital period. The feasibility of recruitment and measurement completion was tracked, as were changes in both PROs and physical function. Results Of 48 eligible patients, 16 enrolled (recruitment rate of 33%). The baseline and in-hospital PROs were completed by 88% of participants, while the outpatient assessments were completed by 81% of participants. The baseline and in-hospital assessment of physical function were completed by 56% of participants, and 38% completed the outpatient assessment. Measuring in-hospital mobilization was completed for 63% of participants. Conclusion Measuring PROs and in-hospital mobilization is feasible across the surgical timeline in HNC; however, the in-person assessment of physical function prior to surgery was not feasible. A multidisciplinary collaboration between exercise specialists and clinicians supported the development of new clinical workflows in HNC surgical care that will aid in the implementation of a future prehabilitation trial for this patient population.
- ItemOpen AccessThe impact of chronic airway disease on symptom severity and global suffering in Canadian rhinosinusitis patients(2018-05-29) Luu, Kimberly; Sutherland, Jason; Crump, Trafford; Liu, Giuping; Janjua, ArifAbstract Background Patients with Chronic Rhinosinusitis (CRS) can suffer from a significant decline in their quality of life. CRS patients have a high prevalence of comorbid conditions and it is important to understand the impact of these conditions on their CRS-related quality of life. This study measures the impacts of chronic pulmonary comorbidities on quality of life, pain, and depression scores among patients with CRS awaiting Endoscopic Sinus Surgery (ESS). Methods This study is based on cross-sectional analysis of prospectively collected patient-reported outcome data collected pre-operatively from patients waiting for ESS. Surveys were administered to patients to assess sino-nasal morbidity (SNOT-22), depression and pain. The impact of pulmonary comorbidity on SNOT-22 scores, pain and depression was measured. Results Two hundred fifthy-three patients were included in the study, 91 with chronic pulmonary comorbidity. The mean SNOT-22 scores were significantly higher among patients with chronic pulmonary comorbidities than among patients without (37 and 48, respectively). This difference is large enough to be clinically significant. Patients with chronic pulmonary comorbidities reported slightly higher depression scores than those without. Conclusions This study found that among CRS patients waiting for ESS, chronic pulmonary comorbidities are strongly associated with significantly higher symptom burden.
- ItemOpen AccessThe impact of comorbid depression in chronic rhinosinusitis on post-operative sino-nasal quality of life and pain following endoscopic sinus surgery(2019-04-30) Ospina, Javier; Liu, Guiping; Crump, Trafford; Sutherland, Jason M; Janjua, ArifAbstract Background Depression and chronic pain are debilitating disorders that co-exist with many chronic diseases. Chronic rhinosinusitis (CRS) is no exception. Nonetheless, little is known about the association between these co-related conditions and the treatment of CRS. The objective of this study is to measure outcomes following endoscopic sinus surgery (ESS) in CRS patients reporting significant pre-operative depression and pain. Methods This is a prospective longitudinal cohort study examining patients with CRS who had failed maximal medical therapy and subsequently underwent ESS. Participants completed a several patient-reported outcome (PRO) instruments pre-operatively and 6 months post-operatively. The PROs included the Sinonasal Outcome Test-22 (SNOT-22), the Patient Health Questionnaire (PHQ-9) measuring symptoms of depression and an assessment of chronic pain using the pain intensity (P), interference with enjoyment of life (E) and general (G) activity instrument, the PEG instrument. Results The study had 142 participants complete their pre-operative and post-operative surveys. The participation rate was 40.1% among eligible patients. The prevalence of at least moderate depression was 22 patients (15.5%) among participants. Compared with non-depressed participants, the pre-operative sino-nasal disease burden and pain scores were higher among depressed participants (p < 0.001) and the gain in health following surgery was smaller (p < 0.001). Conclusions Pre-operative disease burden is higher among depressed patients. Post-operative gains in sino-nasal quality of life attributable to endoscopic sinus surgery were significantly smaller among depressed participants. Pre-operative screening for depression could identify opportunities for medical intervention and improve outcomes among CRS patients.
- ItemOpen AccessWait lists and adult general surgery: is there a socioeconomic dimension in Canada?(2019-03-13) Sutherland, Jason M.; Kurzawa, Zuzanna; Karimuddin, Ahmer; Duncan, Katrina; Liu, Guiping; Crump, TraffordAbstract Background Little is known about whether patients’ socioeconomic status influences their access to elective general surgery in Canada. The purpose of this study was to assess the association between socioeconomic status and wait times for elective general surgery. Methods Analysis of prospectively recruited participants’ data. The setting was six hospitals in the Vancouver Coastal Health Authority, a geographically defined region that includes Vancouver, British Columbia, Canada. Participants had elective general surgery between October 2013 and April 2017, community dwelling, aged 19 years or older and could complete survey forms. The outcome measure was wait time, defined as the number of weeks between being registered for elective general surgery and surgery date. Results One thousand three hundred twenty elective general surgery participants were included in the study. The response rate among eligible patients was 53%. Regression analyses found no statistically significant association between patients’ wait time with SES, adjusting for health status, cancer status, surgical priority level, comorbidity burden and demographic characteristics. Participants with proven or suspected cancer status had shorter waits relative to participants waiting for surgery for benign conditions. Participants with at least one comorbidity tended to experience shorter waits of approximately 5 weeks (p < 0.01). Pre-operative pain or depression/anxiety were not associated with shorter wait times. Conclusions Although this study found no relationship between SES and surgical wait time for elective general surgeries in the study hospitals, patients in lower SES categories reported worse health when assigned to the surgical queue.