Artificial intelligence for synovitis scoring of ultrasound images
Name:
Dr. Gurjit Kaeley
Email
Gurjit.kaeley@jax.ufl.edu
Phone
(904) 633-0914
Faculty Department/Division
Rheumatology & Clinical Immunology
This project is primarily:
Clinical
Research Project Description:
The overall purpose of this study is to develop a machine learning algorithm that will automate the arthritis grade scoring of grey scale and power Doppler musculoskeletal ultrasound images. Since Power Doppler Ultrasound (PDUS) and Gray-scale Ultrasound (GSUS) based scoring approaches are known to suffer with issues such as variability across sonographers, we hypothesize the development of such a novel machine learning approach will significantly improve reliability and improve feasibility of clinical trials using ultrasound synovitis scoring.
Does this project have an international component or travel?
Potentially
Medicare MMS ultrasound utilization trends in the United States
Name:
Dr. Gurjit Kaeley
Email
gurjit.kaeley@jax.ufl.edu
Phone
(904) 633-0071
Faculty Department/Division
Rheumatology & Clinical Immunology
This project is primarily:
CQI
Research Project Description:
There are sparse data regarding the aggregate payments to Rheumatology providers by Medicare. The release of Medicare Provider Utilization and Payment Data (PUF) from 2012 provides an opportunity to analyze the degree of Rheumatology provider participation in the care of Medicare patients as well as payment for services rendered. This retrospective cross sectional analysis may help define the role of Rheumatologists in the care of Medicare beneficiaries and compare it to other specialties involved in musculoskeletal care. It may also inform on the current and projected Rheumatology workforce involved in Rheumatologic care of Medicare beneficiaries. No patient-specific identifiers are included within these datasets, therefore we are seeking exempt review for this project.
Does this project have an international component or travel?
No
Redcap utilization for clinical database
Name:
Dr. Gurjit Kaeley
Email
gurjit.kaeley@jax.ufl.edu
Phone
(904) 633-0071
Faculty Department/Division
Rheumatology & Clinical Immunology
This project is primarily:
Clinical
Research Project Description:
Redcap stores its data and system project information in a single MySQL database. We are hoping to create a protocol that includes patient data in order to better provide clinical services.
Does this project have an international component or travel?
No
The effect of lipemia on the laboratory assessment of blood gases
Faculty Information
Name:
Dr. Bremansu Osa-Andrews
Email
b.osaandrews@ufl.edu
Phone
(605) 651-3989
Faculty Department/Division
Clinical Pathology and laboratory medicine
This project is primarily:
Clinical
Research Project Description:
Unlike the electrolyte exclusion phenomenon, little is known about the effects of lipemia on the laboratory assessment of total carbon dioxide (tCO2) and bicarbonate (HCO3-). Depending on the method used, lipemia-induced turbidity of a specimen can impact the outcome of the results. The photometric-enzymatic, rather than the indirect ion selective method is more easily interfered with by a turbid sample. Consequently, HCO3- and tCO2 results from a severely lipemic sample may be falsely diminished. A recent case report suggests an association between high lipids and HCO3-. A larger study to investigate this trend is imperative to guide laboratory and medical practice, since preanalytical component accounts for approximately 68% of all clinical laboratory errors.
Research goals: This study aims to investigate the analytical relationship between lipemia and blood gases analyses.
Methods: The methods for the preliminary project includes retrieval of patient data from electronic medical record and analyze data with statistical methods.
Student’s role: Students will be trained on how to retrieve patient results/data from the Electronic medical records (EPIC), filter and analyze the data. Students will also learn how to design, develop and write a manuscript including research articles. This is an exciting opportunity for students to engage in scholarly activity with the potential of publication.
The project does not currently have funding.
Does this project have an international component or travel?
No
The utility of laboratory measurement of folate in relation to the risk of cancer
Name:
Dr. Bremansu Osa-Andrews
Email
b.osaandrews@ufl.edu
Phone
(605) 651-3989
Faculty Department/Division
Clinical Pathology and Laboratory medicine
This project is primarily:
Clinical
Research Project Description:
Following the FDA’s mandated folate-fortification of foods in the USA, the Choosing Wisely Campaign recommended that laboratories discontinue the measurement of blood folate levels. The Campaign instead suggested that clinicians should supplement folate in patients with suspected macrocytic anemia, without prior evaluation their folate status. However, the proposition to supplement folate without first testing may be challenged by recent reports associating high blood folate with the risk of certain cancers. In macrocytic anemia, the red blood cells are larger in size than normal with an MCV of >100 fL compared with normal and reduced size in normocytic and microcytic anemia respectively. Megaloblastic macrocytic anemia is caused by deficiency of either vitamin B12 or folate or both. Folate is required by cells for the synthesis of nucleic acids in cells and erythropoiesis. The physiological function of folate and vitamin B12 are connected in the sense that methyl folate is required for the activation of vitamin B12 to convert homocysteine to methionine, relieving the cardio-destructive effect of homocysteine. A recent retrospective study by our team underpinned the need to test for folate status in anemic patients. For proper patient care and patient safety, measurement of folate is a clinically useful test and this project among other things, seeks to provide evidence for this.
Research goals: This project will explore the utility of blood folate testing in the investigation of anemia, prevention of folate toxicity and vitamin B12 deficiency masking and the risks of cancers.
Student’s role: Students will be trained on how to retrieve patient results/data from the Electronic medical records (EPIC), filter and analyze the data. Students will also learn how to design, develop and write a manuscript including research articles. This is an exciting opportunity for students to engage in scholarly activity with the potential of publication.
The project does not currently have funding.
Relevant publication: Osa-Andrews B, Sanchez M, Hashim IA. The Continued Need for the Routine Assessment of Folate Status. Lab Med. 2023 Jul 5;54(4):424-428. doi: 10.1093/labmed/lmac148. PMID: 36637228.
Does this project have an international component or travel?
No
Developing Machine learning algorithms to assist in testing utilization and stewardship in the clinical laboratory
Name:
Maximo Marin
Email
maximo.marin@ufl.edu
Phone
(773) 704-5556
Faculty Department/Division
Pathology, Immunology and Laboratory Medicine
This project is primarily:
Translational
Research Project Description:
Thrombotic thrombocytopenic purpura (TTP) is a rare thrombotic microangiopathy (TMA) and a medical emergency. The ADAMTS13 (AS13) activity assay is needed to confirm the diagnosis. Most healthcare institutions/hospitals do not run this test onsite due to the lack of cost effectiveness and the technical complexity of performing the assay.
Fortunately, at UF Health Shands Core Laboratory, we have this test available onsite. However, it is inappropriately overutilized. Within the context of limited laboratory resources, a machine learning (ML) model (decision tree algorithm) was developed to potentially reduce overutilization of AS13 testing and aid clinical colleagues as a decision support tool in the future. This work was recently published, July 2023, (https://jlpm.amegroups.org/article/view/8019/html).
As an extension of the previous work, the current project is validating the ML model by
collecting data points in the medical chart from both control and diseased confirmed
patients. The dataset will likely consist of 300 to 500 patients with up to approximately 10,000 data points. The variables will include but are not limited to laboratory tests such as platelet count and haptoglobin, general demographics, and medical history of transplant or cancer. The ML model is optimized for the negative predictive value. In other words, if the ML model predicts negative for TTP, we are now able to redirect inappropriate AS13 testing requests, in addition to aiding our clinical colleagues when faced with a potential TMA diagnostic dilemma.
If the validation of the model is acceptable, the next phase of this project will include parallel testing with the current laboratory workflow processes and potentially interfacing the algorithm within the electronic medical chart or middleware.
Does this project have an international component or travel?
No
Diagnosis of multiple myeloma: Comparative studies on serum quantitative immunoglobulin test and serum protein electrophoresis.
Name:
Dr. Bremansu Osa-Andrews
Email
b.osaandrews@ufl.edu
Phone
(605) 651-3989
Faculty Department/Division
Pathology, Immunology and Laboratory medicine
This project is primarily:
Clinical
Research Project Description:
Background: Various clinical laboratory tests are valuable in enhancing the diagnosis of multiple myeloma, a cancer that forms in plasma cells. The nephelometric-based serum quantitative immunoglobulin test provides a great utility to understand the nature and magnitude of antibody-mediated adaptive responses. A linear increase in all the immunoglobulins classes (IgA, IgM, and IgG. IgD and IgE are not often analyzed, being rare) may be an indication of a polyclonal response, the physiological reaction to infectious states. Elevation of one type of immunoglobulin in relation to the other antibodies may be a function of malignancy. Multiple myeloma may be characterized by increased quantitated IgA or IgG while elevated levels of IgM is a common feature in Waldenström’s macroglobulinemia. Therefore, initial suspicion of monoclonality may be identified through the quantitative serum immunoglobulin test. This may be followed by a request for serum protein electrophoresis (SPE). SPE patronizes the electrophoretic movement of proteins from one end of an electrode (the cathode) to the other (the anode) by the application of voltage. Depending on the results of the SPE, an immunofixation gel electrophoresis (IFE) may be ordered to confirm the specific monoclonal protein responsible for the spike.
Hypothesis: In clinical practice, several SPEs are ordered without a preliminary order of quantitative serum immunoglobulin test. In other cases, after reviewing an SPE result, pathologists would recommend the ordering of quantitative serum immunoglobulin test for comparison. This order may be either to confirm polyclonal pattern or embolden a normal SPE electrophoretogram. However, there is seldom any data in the literature describing the correlative relationship between the two tests. It is therefore imperative to create an evidenced-based system of test ordering for the laboratory investigation of monoclonal protein (M-protein) or to rule out the presence of an M-protein.
Objectives: The project’s goal is to compare the results of quantitative serum immunoglobulin assay and SPE to determine their level of agreement. The utility of correlation studies is to provide vital information regarding the need to order.
Methodology: In the preliminary phase of this project, retrospective results of quantitative serum immunoglobulin assay and SPEs performed simultaneously will be pooled from the electronic medical records. The span of data will extend to two years prior. If there is an additional immunofixation order associated with the SPE, the IFE report will also be retrieved for comparison with the quantitative serum immunoglobulin test. All the data will be entered into and filtered in excel spreadsheet program and subsequently analyzed with GraphPad prism.
Conclusion: The present research will directly impact the diagnosis of multiple myeloma and enhance appropriate test utilization. It may not be clinically useful to order an SPE in the background of a normal quantitative serum immunoglobulin test; this project will show the essence of suitable testing order. The research will improve the development of a better ordering policy for laboratory operations with respect to the work up of multiple myeloma. Gaining better understanding of the comparative relationship between the nephelometric serum immunoglobulin test and the electrophoretic assays is expected to improve billing discrepancies. Only the necessary tests may be required to be ordered and in the appropriate order of importance that fits the comparative criteria.
What students/residents will learn and their responsibility: Medical students will be trained on how to design a project, statistical analysis, retrieval of patient reports from electronic medical record, and prepare a scientific abstract for submission to a conference. Students will also be trained in how to present their research results orally and visually in local forums and at external conferences. Medical students will subsequently be responsible for retrieving patient reports from electronic record data, manage, filter, and analyze the dater using appropriate statistical software and technologies. Students will make significant contributions to the diagnosis, laboratory investigation and treatment of not only multiple myeloma and its related conditions, but also of Waldenström’s macroglobulinemia. Students will learn about the principles behind the clinical pathology techniques used in the clinical laboratory to investigate multiple myeloma. Overall, students will gain experience of how to conduct clinical cancer research with translational outcomes.
Funding: The project is not currently grant funded, but funding opportunities are both welcomed and being sought.
Does this project have an international component or travel?
No
Spleen Volume in Type 1 Diabetes
Name:
Martha Campbell-Thompson
Email
mct@ufl.edu
Phone
(352) 273-6129
Faculty Department/Division
Pathology, Immunology and Lab. Medicine
This project is primarily:
Translational
Research Project Description:
Our studies are focused on understanding how the loss of pancreas size and changes in organ architecture are impacted by genetics and environment related to the risk of type 1 diabetes. We have two clinical trials using pancreas MRI scans of which the second is ongoing. The student engaged in this project will be trained to manually segment the spleen from pancreas MRI scans using ITKSnap software. Spleen volume data will be used to evaluate a new normalization factor for pancreas to spleen volumes per subject. It is expected that the student will be co-author on abstracts and manuscripts that include their spleen volume data. This project would be ideal for someone considering becoming a radiologist.
Does this project have an international component or travel?
No