Clinical Focus


  • Diagnostic Radiology
  • Genitourinary Imaging
  • Urology Imaging

Academic Appointments


  • Clinical Assistant Professor, Radiology

Professional Education


  • Fellowship: UCLA Radiology Fellowships (2018) CA
  • Residency: UCLA Radiology Residency (2017) CA
  • Internship: University of California Irvine Dept of Internal Medicine (2013) CA
  • Medical Education: University of Cincinnati College of Medicine Registrar (2012) OH
  • Board Certification: American Board of Radiology, Diagnostic Radiology (2018)

All Publications


  • Prevalence of Malignancy and Histopathologic Association of Bosniak Classification, Version 2019 Class III and IV Cystic Renal Masses. The Journal of urology Tse, J. R., Shen, L., Shen, J., Yoon, L., Kamaya, A. 2020: 101097JU0000000000001438

    Abstract

    PURPOSE: Bosniak Classification, version 2019 (v2019) describes two types of class III and IV masses each: 1) thick, wall/septa ≥4 mm (III-WS), 2) obtuse protrusion ≤3 mm (III-OP), 3) obtuse protrusion ≥4 mm (IV-OP), and 4) acute protrusion of any size (IV-AP). The purposes were to determine the prevalence of malignancy and histopathologic features of class III and IV masses and subclasses.MATERIALS AND METHODS: In this IRB-approved and HIPAA-compliant study, three fellowship-trained abdominal radiologists (R1-3) reviewed cystic renal masses that had tissue pathology and pre-operative renal mass protocol CT or MRI. Classes based on v2019 and prior classification systems were retrospectively re-assigned and associated with malignancy, aggressive histologic features (necrosis or high Fuhrman grade), and radiologic progression following resection.RESULTS: The final sample included 79 masses (59 malignant, 20 benign) from 74 patients. Based on v2019, prevalence of malignancy ranged from 56-61% (mean 60%) for class III and 83-83% (mean 83%) for class IV (p=0.036, 0.013, 0.036 for R1-3). Prevalence of malignancy within subclasses were: III-WS (47-53%); III-OP (71-85%); IV-OP (75-87%); IV-AP (87-95%; p=0.029, 0.001, 0.005). All readers were more likely to classify malignancies with aggressive histologic features as class IV (88-100%) rather than class III (0-12%; p=0.012, <0.001, 0.002), corresponding to a negative predictive value of 96-100%. Following treatment (mean follow-up length 1210 days), one patient developed metastases.CONCLUSIONS: Bosniak Classification, version 2019 can help risk stratification of class III-IV masses by identifying those likely to be malignant and have aggressive histologic features.

    View details for DOI 10.1097/JU.0000000000001438

    View details for PubMedID 33085925

  • Bosniak Classification of Cystic Renal Masses Version 2019: Comparison of Categorization using CT and MRI. AJR. American journal of roentgenology Tse, J. R., Shen, J., Shen, L., Yoon, L., Kamaya, A. 2020

    Abstract

    Please see the Author Video associated with this article. Background: Bosniak Classification, version 2019 recently proposed refinements for cystic renal mass characterization and now formally incorporates MRI, which may improve concordance with CT. Purpose: To compare concordance of CT and MRI in evaluation of cystic renal masses using Bosniak Classification, version 2019. Materials and Methods: In this IRB-approved and HIPAA compliant study, three abdominal radiologists (R1-R3) retrospectively reviewed 68 consecutive cystic renal masses from 45 patients assessed with both CT and MR renal mass protocols within a year between 2005-2019. CT and MRI were reviewed independently and in separate sessions, using both the original and version 2019 Bosniak Classification systems. Results: Using Bosniak Classification, version 2019, cystic renal masses were classified into 12 category I, 19 category II, 13 category IIF, 4 category III, and 20 category IV by CT and 8 category I, 15 category II, 23 category IIF, 9 category III, and 13 category IV by MRI. Among individual features, MRI depicted more septa (p<0.001, p=0.046, p=0.005 for R1-R3; McNemar's test) for all radiologists, though both CT and MRI showed a similar number of protrusions (p=0.823, 1.0, 0.302) and maximal septa/wall thickness (p=1.0, 1.0, 0.145). Of discordant cases with version 2019, MRI led to the higher category in 12 masses. Reason for upgrade was most commonly due to protrusions identified only on MRI (n=4), increased number of septa (n=3), and a new category of heterogeneously T1-hyperintense (n=3). Neither modality was more likely to lead to a category change for both version 2019 (p=0.502; McNemar's test) and the original Bosniak classification system (p=0.823). Overall inter-rater agreement was substantial for both CT (κ=0.745) and MRI (κ=0.655) using version 2019 and was slightly higher than that of the original system (CT κ=0.707; MRI κ=0.623). Conclusion: CT and MRI were concordant in the majority of cases using Bosniak Classification, version 2019 and category changes by modality were not statistically significant. Inter-rater agreements were substantial for both CT and MRI. Clinical Impact: Bosniak Classification, version 2019 applied to cystic renal masses has substantial inter-rater agreement and does not lead to systematic category upgrades with either CT or MRI.

    View details for DOI 10.2214/AJR.20.23656

    View details for PubMedID 32755181

  • Human-machine partnership with artificial intelligence for chest radiograph diagnosis. NPJ digital medicine Patel, B. N., Rosenberg, L., Willcox, G., Baltaxe, D., Lyons, M., Irvin, J., Rajpurkar, P., Amrhein, T., Gupta, R., Halabi, S., Langlotz, C., Lo, E., Mammarappallil, J., Mariano, A. J., Riley, G., Seekins, J., Shen, L., Zucker, E., Lungren, M. 2019; 2: 111

    Abstract

    Human-in-the-loop (HITL) AI may enable an ideal symbiosis of human experts and AI models, harnessing the advantages of both while at the same time overcoming their respective limitations. The purpose of this study was to investigate a novel collective intelligence technology designed to amplify the diagnostic accuracy of networked human groups by forming real-time systems modeled on biological swarms. Using small groups of radiologists, the swarm-based technology was applied to the diagnosis of pneumonia on chest radiographs and compared against human experts alone, as well as two state-of-the-art deep learning AI models. Our work demonstrates that both the swarm-based technology and deep-learning technology achieved superior diagnostic accuracy than the human experts alone. Our work further demonstrates that when used in combination, the swarm-based technology and deep-learning technology outperformed either method alone. The superior diagnostic accuracy of the combined HITL AI solution compared to radiologists and AI alone has broad implications for the surging clinical AI deployment and implementation strategies in future practice.

    View details for DOI 10.1038/s41746-019-0189-7

    View details for PubMedID 31754637

    View details for PubMedCentralID PMC6861262

  • Erratum: Author Correction: Human-machine partnership with artificial intelligence for chest radiograph diagnosis. NPJ digital medicine Patel, B. N., Rosenberg, L., Willcox, G., Baltaxe, D., Lyons, M., Irvin, J., Rajpurkar, P., Amrhein, T., Gupta, R., Halabi, S., Langlotz, C., Lo, E., Mammarappallil, J., Mariano, A. J., Riley, G., Seekins, J., Shen, L., Zucker, E., Lungren, M. P. 2019; 2: 129

    Abstract

    [This corrects the article DOI: 10.1038/s41746-019-0189-7.].

    View details for DOI 10.1038/s41746-019-0198-6

    View details for PubMedID 31840097

    View details for PubMedCentralID PMC6904441