Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Oct;25(10):e14461.
doi: 10.1002/acm2.14461. Epub 2024 Aug 2.

Geometric and dosimetric evaluation for breast and regional nodal auto-segmentation structures

Tiffany Tsui  1   2 , Alexander Podgorsak  3 , John C Roeske  1   2 , William Small Jr  1   2 , Tamer Refaat  1   2 , Hyejoo Kang  1   2
Affiliations

Geometric and dosimetric evaluation for breast and regional nodal auto-segmentation structures

Tiffany Tsui et al. J Appl Clin Med Phys. 2024 Oct.

Abstract

The accuracy of artificial intelligence (AI) generated contours for intact-breast and post-mastectomy radiotherapy plans was evaluated. Geometric and dosimetric comparisons were performed between auto-contours (ACs) and manual-contours (MCs) produced by physicians for target structures. Breast and regional nodal structures were manually delineated on 66 breast cancer patients. ACs were retrospectively generated. The characteristics of the breast/post-mastectomy chestwall (CW) and regional nodal structures (axillary [AxN], supraclavicular [SC], internal mammary [IM]) were geometrically evaluated by Dice similarity coefficient (DSC), mean surface distance, and Hausdorff Distance. The structures were also evaluated dosimetrically by superimposing the MC clinically delivered plans onto the ACs to assess the impact of utilizing ACs with target dose (Vx%) evaluation. Positive geometric correlations between volume and DSC for intact-breast, AxN, and CW were observed. Little or anti correlations between volume and DSC for IM and SC were shown. For intact-breast plans, insignificant dosimetric differences between ACs and MCs were observed for AxNV95% (p = 0.17) and SCV95% (p = 0.16), while IMNV90% ACs and MCs were significantly different. The average V95% for intact-breast MCs (98.4%) and ACs (97.1%) were comparable but statistically different (p = 0.02). For post-mastectomy plans, AxNV95% (p = 0.35) and SCV95% (p = 0.08) were consistent between ACs and MCs, while IMNV90% was significantly different. Additionally, 94.1% of AC-breasts met ΔV95% variation <5% when DSC > 0.7. However, only 62.5% AC-CWs achieved the same metrics, despite AC-CWV95% (p = 0.43) being statistically insignificant. The AC intact-breast structure was dosimetrically similar to MCs. The AC AxN and SC may require manual adjustments. Careful review should be performed for AC post-mastectomy CW and IMN before treatment planning. The findings of this study may guide the clinical decision-making process for the utilization of AI-driven ACs for intact-breast and post-mastectomy plans. Before clinical implementation of this auto-segmentation software, an in-depth assessment of agreement with each local facilities MCs is needed.

Keywords: AI‐algorithm; Auto‐contour; auto‐segmentation; breast radiation treatment.

PubMed Disclaimer

Conflict of interest statement

The authors have no affiliations with or involvement in any organization or entity with any financial interest.

Figures

FIGURE 1
FIGURE 1
Overall study workflow that includes data selection in AURA, data preparation in Eclipse Contouring and Radformation Auto Contour, data export in Varian Eclipse, and data analysis in Eclipse, Velocity, Excel, and Python. AURA, Aria reporting system.
FIGURE 2
FIGURE 2
Physician‐dependent performance of AC for the SC. The volumes of MC are systematically larger than the AC volumes for MD1, which is not observed for MD2. The identity line is plotted in black. AC, automatically‐segmented contours; MC, manually‐segmented contours; SC, supraclavicular nodes.
FIGURE 3
FIGURE 3
DSC versus the MC volumes in cubic centimeters for the considered target structures. Correlation is denoted by the PCC. Positive correlation between the two quantities was observed for (a) the breast, (b) CW, and (c) AxN. Anti‐correlation and weak correlation were observed for the (d) SC and (e) IMNs, respectively. AxN, axillary nodes; CW, chestwall; DSC, Dice similarity coefficient; IMN, internal mammary nodes; MC, manually‐segmented contour; Pearson correlation coefficient; SC, supraclavicular nodes.
FIGURE 4
FIGURE 4
DSC versus the lengths in three dimensions of target structures. (a)–(c) AxN in x‐, y‐, and z‐directions are denoted by AxN(X), AxN(Y), and AxN(Z), respectively. (d)–(f) SC in x‐, y‐, and z‐directions are denoted by SC(X), SC(Y), and SC(Z), respectively. (g)–(i) IMNs in x‐, y‐, z‐directions are denoted by IMN(X), IMN(Y), and IMN(Z), respectively. AxN, axillary nodes; DSC, Dice similarity coefficient; IMN, internal mammary nodes; SC, supraclavicular nodes.
FIGURE 5
FIGURE 5
ΔV95% versus DSC for (a) breast and CW and (b) AxN in intact‐breast and post‐mastectomy breast plans. AxN, axillary nodes; CW, chestwall; DSC, Dice similarity coefficient.
FIGURE 6
FIGURE 6
Two examples of out‐of‐plane discrepancies between MC and AC. (a) Out‐of‐plane discrepancies between MC (red) and AC (cyan) of AxN; (b) Out‐of‐plane discrepancies between MC (green) and AC (burgundy) of breast. These cases would have reduced dosimetric agreement in terms of the V95% relative to in‐plane accuracy, as the field aperture for a 3DCRT would be defined by the superior and inferior borders of the target. V95% represents the percentage of the PTV that received at least 95% of its prescribed dose. 3DCRT, 3D conformal radiotherapy; AC, automatically segmented contour; AxN, axillary node; MC, manually‐segmented contour; PTV, planning target volume.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Chen X, Sun S, Bai N, et al. A deep learning‐based auto‐segmentation system for organs‐at‐risk on whole‐body computed tomography images for radiation therapy. Radiother Oncol J Eur Soc Ther Radiol Oncol. 2021;160:175‐184.doi: 10.1016/j.radonc.2021.04.019 - DOI - PubMed
    1. Gao L, Yusufaly TI, Williamson CW, Mell LK. Optimized atlas‐based auto‐segmentation of bony structures from whole‐body computed tomography. Pract Radiat Oncol. 2023;13(5):e442‐e450. doi:10.1016/j.prro.2023.03.013 - DOI - PubMed
    1. Kosmin M, Ledsam J, Romera‐Paredes B, et al. Rapid advances in auto‐segmentation of organs at risk and target volumes in head and neck cancer. Radiother Oncol J Eur Soc Ther Radiol Oncol. 2019;135:130‐140. doi:10.1016/j.radonc.2019.03.004 - DOI - PubMed
    1. Wu Y, Kang K, Han C, et al. A blind randomized validated convolutional neural network for auto‐segmentation of clinical target volume in rectal cancer patients receiving neoadjuvant radiotherapy. Cancer Med. 2022;11(1):166‐175. doi:10.1002/cam4.4441 - DOI - PMC - PubMed
    1. Almberg SS, Lervåg C, Frengen J, et al. Training, validation, and clinical implementation of a deep‐learning segmentation model for radiotherapy of loco‐regional breast cancer. Radiother Oncol J Eur Soc Ther Radiol Oncol. 2022;173:62‐68. doi:10.1016/j.radonc.2022.05.018 - DOI - PubMed