Application of a diagnosis flow draft based on appearance impression for detection of vulvar disease

Qi Zhou, Fang Chen, Yan Wang, Wenjie Qu, Yingxin Gong, Yuankui Cao, Hongwei Zhang, Qing Wang, Limei Chen, Qing Cong, Lin Lin, Jiayin Mo, Tianyi Bi, Jingxin Ding, Long Sui, Yanyun LiDiagnosis 2023 · DOI

Abstract

Based on the findings in ISSVD and IFCPC terminology systems, we explored a Diagnosis Flow Draft (DFD) for observers with different experience levels on the detection of vulvar disease.

Introduction

Diagnosing vulvar diseases presents a significant clinical challenge, as it requires expertise spanning gynecology, dermatology, and pathology. Currently, clinicians rely on two primary systems: the International Society for the Study of Vulvovaginal Disease (ISSVD) system, which focuses on pathophysiology, and the International Federation for Cervical Pathology and Colposcopy (IFCPC) system, which emphasizes colposcopic patterns. However, neither provides a standardized visual "decision tree" for real-time examinations. This study introduces a Diagnosis Flow Draft (DFD) designed to bridge this gap by ranking clinical findings based on their "Overall Diagnostic Value" (ODV), providing a data-driven visual hierarchy for diagnosis.

Methodology

This retrospective study analyzed 1,068 patients (5,340 qualified images) from the Obstetrics and Gynecology Hospital of Fudan University between 2020 and 2022. The research pipeline involved:

  1. Evaluation: Images were assessed using both 2011 ISSVD and 2011 IFCPC terminologies.
  2. Metric Calculation: Sensitivity, specificity, and the Youden Index were calculated for various clinical findings (e.g., atypical vessels, papules).
  3. ODV Scoring: A custom "Overall Diagnostic Value" score was developed to identify pathognomonic visual cues.
  4. DFD Construction: Findings were mapped into a visual flowchart where arrow thickness and brightness indicate the strength of diagnostic direction.
  5. Validation: Nine observers with varying experience levels tested the DFD against histopathological gold standards.

Results and Clinical Insights

The study identified 15 distinct types of vulvar diseases. Key demographic and clinical characteristics were mapped to provide a comprehensive overview of the patient population.

Patient demographics and clinical characteristics: (A) mean age, (B) accompaniment with CIN/VaIN, and (C) lesion location across 15 vulvar diseases.

One of the most striking findings was the high correlation between specific conditions. For instance, 83.3% of patients with vulvar Paget’s disease also presented with concurrent cervical or vaginal intraepithelial neoplasia (CIN/VaIN), suggesting that a Paget’s diagnosis should prompt immediate lower genital tract screening.

The DFD achieved a 68.8% agreement rate with histopathology, significantly outperforming the ISSVD (63.5%) and IFCPC (65.1%) systems alone (p=0.028). Crucially, the DFD effectively "democratized" expertise; there was no statistically significant difference in accuracy between experts with over 10 years of experience (68.8%) and novices with under 5 years (65.4%).

The Diagnosis Flow Draft (DFD) network diagram illustrating the visual hierarchy and diagnostic relationships between the 15 studied vulvar diseases.

Discussion

The DFD provides a robust framework for standardized vulvar examination, particularly by using visual weight to guide clinicians through a "possible disease diagnosis order." By identifying the "Top Five" visual findings for each disease—such as "atypical vessels" for Squamous Cell Carcinoma (SCC)—the tool simplifies complex diagnostic decisions.

While the DFD shows promise in leveling the playing field for junior clinicians, the authors note that visual assessment remains inherently subjective. Future work will focus on integrating the DFD with adjunct technologies, such as artificial intelligence and fluorescence labeling, to further refine diagnostic precision and reduce human error in vulvar disease detection.

Reference: Zhou, Q., et al. (2024). "Application of a diagnosis flow draft based on appearance impression for detection of vulvar disease." Diagnosis, 11(2), 151–163. https://doi.org/10.1515/dx-2023-0146

Summary updated

Abstract

Based on the findings in ISSVD and IFCPC terminology systems, we explored a DFD for observers with different experience on the detection of vulvar disease.
BibTeX
@inproceedings{zhou2023application,
  author = {Qi Zhou and Fang Chen and Yan Wang and Wenjie Qu and Yingxin Gong and Yuankui Cao and Hongwei Zhang and Qing Wang and Limei Chen and Qing Cong and Lin Lin and Jiayin Mo and Tianyi Bi and Jingxin Ding and Long Sui and Yanyun Li},
  title = {{Application of a diagnosis flow draft based on appearance impression for detection of vulvar disease}},
  booktitle = {Diagnosis},
  year = {2023},
  doi = {10.1515/dx-2023-0146},
}
Figure