Breast (Feb 2024)

Assessment of deep learning-based auto-contouring on interobserver consistency in target volume and organs-at-risk delineation for breast cancer: Implications for RTQA program in a multi-institutional study

  • Min Seo Choi,
  • Jee Suk Chang,
  • Kyubo Kim,
  • Jin Hee Kim,
  • Tae Hyung Kim,
  • Sungmin Kim,
  • Hyejung Cha,
  • Oyeon Cho,
  • Jin Hwa Choi,
  • Myungsoo Kim,
  • Juree Kim,
  • Tae Gyu Kim,
  • Seung-Gu Yeo,
  • Ah Ram Chang,
  • Sung-Ja Ahn,
  • Jinhyun Choi,
  • Ki Mun Kang,
  • Jeanny Kwon,
  • Taeryool Koo,
  • Mi Young Kim,
  • Seo Hee Choi,
  • Bae Kwon Jeong,
  • Bum-Sup Jang,
  • In Young Jo,
  • Hyebin Lee,
  • Nalee Kim,
  • Hae Jin Park,
  • Jung Ho Im,
  • Sea-Won Lee,
  • Yeona Cho,
  • Sun Young Lee,
  • Ji Hyun Chang,
  • Jaehee Chun,
  • Eung Man Lee,
  • Jin Sung Kim,
  • Kyung Hwan Shin,
  • Yong Bae Kim

Journal volume & issue
Vol. 73
p. 103599

Abstract

Read online

Purpose: To quantify interobserver variation (IOV) in target volume and organs-at-risk (OAR) contouring across 31 institutions in breast cancer cases and to explore the clinical utility of deep learning (DL)-based auto-contouring in reducing potential IOV. Methods and materials: In phase 1, two breast cancer cases were randomly selected and distributed to multiple institutions for contouring six clinical target volumes (CTVs) and eight OAR. In Phase 2, auto-contour sets were generated using a previously published DL Breast segmentation model and were made available for all participants. The difference in IOV of submitted contours in phases 1 and 2 was investigated quantitatively using the Dice similarity coefficient (DSC) and Hausdorff distance (HD). The qualitative analysis involved using contour heat maps to visualize the extent and location of these variations and the required modification. Results: Over 800 pairwise comparisons were analysed for each structure in each case. Quantitative phase 2 metrics showed significant improvement in the mean DSC (from 0.69 to 0.77) and HD (from 34.9 to 17.9 mm). Quantitative analysis showed increased interobserver agreement in phase 2, specifically for CTV structures (5–19 %), leading to fewer manual adjustments. Underlying IOV differences causes were reported using a questionnaire and hierarchical clustering analysis based on the volume of CTVs. Conclusion: DL-based auto-contours improved the contour agreement for OARs and CTVs significantly, both qualitatively and quantitatively, suggesting its potential role in minimizing radiation therapy protocol deviation.

Keywords