Indiana University Medical Student Program for Research and Scholarship (IMPRS)

Permanent URI for this community

https://hdl.handle.net/1805/29656

Browse

Browsing Indiana University Medical Student Program for Research and Scholarship (IMPRS) by Subject "artificial intelligence"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Clinical Features Distinguishing Diabetic Retinopathy Severity Using Artificial Intelligence
    (2022-07-29) Happe, Michael; Gill, Hunter; Salem, Doaa Hassan; Janga, Sarath Chandra; Hajrasouliha, Amir
    BACKGROUND AND HYPOTHESIS: 1 in 29 American diabetics suffer from diabetic retinopathy (DR), the weakening of blood vessels in the retina. DR goes undetected in nearly 50% of diabetics, allowing DR to steal the vision of many Americans. We hypothesize that increasing the rate and ease of diagnosing DR by introducing artificial intelligence-based methods in primary medical clinics will increase the long-term preservation of ocular health in diabetic patients. PROJECT METHODS: This retrospective cohort study was conducted under approval from the Institutional Review Board of Indiana University School of Medicine. Images were deidentified and no consent was taken due to the nature of this retrospective study. We categorized 676 patient files based upon HbA1c, severity of non-proliferative diabetic retinopathy (NPDR), and proliferative diabetic retinopathy (PDR). Retinal images were annotated to identify common features of DR: microaneurysms, hemorrhages, cotton wool spots, exudates, and neovascularization. The VGG Image Annotator application used for annotations allowed us to save structure coordinates into a separate database for future training of the artificial intelligence system. RESULTS: 228 (33.7%) of patients were diagnosed with diabetes, and 143 (62.7%) of those were diagnosed with DR. Two-sample t tests found significant differences between the HbA1c values of all diabetics compared to diabetics without retinopathy (p<0.007) and between all severities of DR versus diabetics without retinopathy (p<0.002). 283 eyes were diagnosed with a form of DR in this study: 37 mild NPDR, 42 moderate NPDR, 56 severe NPDR, and 148 PDR eyes. POTENTIAL IMPACT: With the dataset of coordinates and HbA1c values from this experiment, we aim to train an artificial intelligence system to diagnose DR through retinal imaging. The goal of this system is to be conveniently used in primary medical clinics to increase the detection rate of DR to preserve the ocular health of millions of future Americans.
  • Thumbnail Image
    Item
    Comparison of Artificial Intelligence and Eyeball Method in the Detection of Fatty Liver Disease
    (2023-07-26) Catron, Evan J.; Passarelli, Robert P.; Danielle, Wilmes; Wei, Barry; Le, Thi M.U.; Li, Ping; Zhang, Wenjun; Lin, Jingmei; Melcher, Mark L.; Mihaylov, Plamen V.; Kubal, Chandrashekhar A.; Mangus, Robert S.; Ekser, Burcin
    Background: Quantification of liver fat content relies on visual microscopic inspection of liver biopsies by pathologists. Their percent macrosteatosis (%MaS) estimation is vital in determining donor liver transplantability; however, the eyeball method may vary between observers. Overestimations of %MaS can potentially lead to the discard of viable donor livers. We hypothesize that artificial intelligence (AI) could be helpful in providing a more objective and accurate measurement of %MaS. Methods: Literature review identified HALO (image analysis) and U-Net (deep-learning) as high-accuracy AI programs capable of calculating %MaS in liver biopsies. We compared (i) an experienced pathologist’s and (ii) a transplant surgeon’s eyeball %MaS estimations from de-novo liver transplant (LT) biopsy samples taken 2h post-reperfusion to (iii) the HALO-calculated %MaS (Fig.1). 250 patients had undergone LT at Indiana University between 2020-2021, and 211 had sufficient data for inclusion. Each biopsy was digitized into 5 random non-overlapping tiles at 20x magnification (a total of 1,055 images). We used HALO software for analysis and set the minimum vacuole area to 10μm² to avoid the inclusion of microsteatosis. Microsteatosis was excluded by the pathologist and the surgeon by the eyeball method using the same 1,055 images. Each %MaS estimation was compared with early allograft dysfunction (EAD). EAD is defined by the presence of at least one of the following: INR >1.6 on postoperative day (POD) 7, total bilirubin >10mg/dL on POD7, or AST/ALT >2000IU/L within the first 7 days following LT. Results: Of 211 LTs, 42 (19.9%) had EAD. The mean %MaS estimation of pathologist and transplant surgeon were 6.3% (SD: 11.9%) and 3.2% (SD: 6.4%), respectively. HALO yielded a significantly lower mean %MaS of 2.6% (SD: 2.6%) than the pathologist’s eyeball method (p<0.001). The mean %MaS calculated by HALO was higher in EAD patients than in non-EAD (p=0.032), but this difference did not reach statistical significance in the pathologist’s estimation (p=0.069). Conclusions: Although mean %MaS measurements from all parties were mild (<10%), human eyeball estimations of %MaS were significantly higher than HALO’s %MaS. The HALO-calculated %MaS differed significantly between the EAD and non-EAD LTs which might suggest a possible correlation between the AI’s steatosis analysis and EAD outcomes. However, pathologic variables other than %MaS (necrosis or cholestasis) should be included in future analyses to determine whether %MaS is the dominant parameter predicting EAD. AI is a promising tool to quantify liver steatosis and will help pathologists and transplant surgeons predict liver transplant viability.