DWDavid C. Wilbur

David C. Wilbur

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


David C. Wilbur has a total of 33 co-authors. Here are the top co-authors by number of shared publications:

  • Alexander Andryushkin

    Corista LLC Concord MA USA

    2 shared publications
  • Jason R. Pettus

    Department of Pathology and Laboratory Medicine Dartmouth‐Hitchcock Medical Center Lebanon MH USA

    2 shared publications
  • Lynn D. Cornell

    Department of Laboratory Medicine and Pathology Mayo Clinic – Rochester Rochester MN USA

    2 shared publications
  • Marluce Bibbo

    Thomas Jefferson University

    2 shared publications
  • Maxwell L. Smith

    Department of Laboratory Medicine and Pathology Mayo Clinic – Scottsdale Phoenix AZ USA

    2 shared publications
  • Alain P. Verhest

    From the Task Force on Computerized Screening Devices and Performance Assessment, Diagnostic Cytology Towards the 21st Century: An International Expert Conference and Tutorial, Kamuela, Hawaii, U.S.A., June 5–10, 1997

    1 shared publication

Latest Publications

  • Open Access

    A novel approach to integrating artificial intelligence into routine practice

    by Madelyn Lew, David C. Wilbur

    Although the benefits of artificial intelligence (AI) in clinical practice have been previously reported, its use has not been widely implemented for various reasons, including cost of investment, disruption of historical workflows, perceived negative impact on job security, and regulatory burden of bringing such technology to market. However, innovative AI applications can be implemented to leverage the advantages of technology while bypassing the need for expensive hardware and overhaul of existing workflows. Lessons gleaned from the use of such AI applications highlight how, with expanded testing and systematized approaches to regulatory approval, they can enhance workflow efficiency and quality improvement as well as assist in the training of future practitioners in resource‐limited regions and high‐volume institutions alike.


  • Automated identification of glomeruli and synchronised review of special stains in renal biopsies by machine learning and slide registration: a cross‐institutional study

    by David C. Wilbur, Maxwell L. Smith, Lynn D. Cornell, Alexander Andryushkin, Jason R. Pettus

    Aims: Machine learning in digital pathology can improve efficiency and accuracy via prescreening with automated feature identification. Studies using uniform histological material have shown promise. Generalised application requires validation on slides from multiple institutions. We used machine learning to identify glomeruli on renal biopsies and compared performance between single and multiple institutions.Methods and results: Randomly selected, adequately sampled renal core biopsy cases (71) consisting of four stains each (haematoxylin and eosin, trichrome, silver, periodic acid Schiff) from three institutions were digitised at ×40. Glomeruli were manually annotated by three renal pathologists using a digital tool. Cases were divided into training/validation (n = 52) and evaluation (n = 19) cohorts. An algorithm was trained to develop three convolutional neural network (CNN) models which tested case cohorts intra‐ and inter‐institutionally. Raw CNN search data from each of the four slides per case were merged into composite regions of interest containing putative glomeruli. The sensitivity and modified specificity of glomerulus detection (versus annotated truth) were calculated for each model/cohort. Intra‐institutional (3) sensitivity ranged from 90 to 93%, with modified specificity from 86 to 98%. Interinstitutional (1) sensitivity was 77%, with modified specificity 97%. Combined intra‐ and inter‐institutional (1) sensitivity was 86%, with modified specificity 92%.Conclusions: Feature detection sensitivity degrades when training and test material originate from different sites. Training using a combined set of digital slides from three institutions improves performance. Differing histology methods probably account for algorithm performance contrasts. Our data highlight the need for diverse training sets for the development of generalisable machine learning histology algorithms.


  • Open Access

    Using Image Registration and Machine Learning to Develop a Workstation Tool for Rapid Analysis of Glomeruli in Medical Renal Biopsies

    by David C. Wilbur, Jason R. Pettus, Maxwell L. Smith, Lynn D. Cornell, Alexander Andryushkin, Richard Wingard, Eric Wirch


  • Automation in Cervical Cytology

    by David C. Wilbur, Marluce Bibbo


  • Computerized Screening Devices and Performance Assessment: Development of a Policy Towards Automation

    by Peter H. Bartels, Marluce Bibbo, Martha L. Hutchinson, Thomas Gahm, Heinz K. Grohs, Elaine Gwi-Mak, Edward A. Kaufman, Raymond H. Kaufman, Bryan Kenyon Knight, Leopold G. Koss, et al.

    The extension of automation to the diagnostic assessment of clinical materials raises issues of professional responsibility, on the part of both the medical professional and designer of the device. The International Academy of Cytology (IAC) and other professional cytology societies should develop a policy towards automation in the diagnostic assessment of clinical cytologic materials. The following summarizes the discussion of the initial position statement at the International Expert Conference on Diagnostic Cytology Towards the 21st Century, Hawaii, June 1997. 1. The professional in charge of a clinical cytopathology laboratory continues to bear the ultimate medical responsibility for diagnostic decisions made at the facility, whether automated devices are involved or not. 2. The introduction of automated procedures into clinical cytology should under no circumstances lead to a lowering of standards of performance. A prime objective of any guidelines should be to ensure that an automated procedure, in principle, does not expose any patient to new risks, nor should it increase already-existing, inherent risks. 3. Automated devices should provide capabilities for the medical professional to conduct periodic tests of the appropriate performance of the device. 4. Supervisory personnel should continue visual quality control screening of a certain percentage of slides dismissed at primary screening as within normal limits (WNL), even when automated procedures are employed in the laboratory. 5. Specifications for the design of primary screening devices for the detection of cervical cancer issued by the IAC in 1984 were reaffirmed. 6. The setting of numeric performance criteria is the proper charge of regulatory agencies, which also have the power of enforcement. 7. Human expert verification of results represents the "gold standard" at this time. Performance characteristics of computerized cytology devices should be determined by adherence to defined and well-considered protocols. Manufacturers should not claim a new standard of care; this is the responsibility of the medical community and professional groups. 8. Cytology professionals should support the development of procedures that bring about an improvement in diagnostic decision making. Advances in technology should be adopted if they can help solve problems in clinical cytology. The introduction of automated procedures into diagnostic decision making should take place strictly under the supervision and with the active participation and critical evaluation by the professional cytology community. Guidelines should be developed for the communication of technical information about the performance of automated screening devices by the IAC to governmental agencies and national societies. Also, guidelines are necessary for the official communication of IAC concerns to industry, medicolegal entities and the media. Procedures and guidelines for the evaluation of studies pertaining to the performance of automated devices, performance metrics and definitions for evaluation criteria should be established.