Deep learning for automated contouring of neurovascular structures on magnetic resonance imaging for prostate cancer patients

Abstract

Background and purpose: Manual contouring of neurovascular structures on prostate magnetic resonance imaging (MRI) is labor-intensive and prone to considerable interrater disagreement. Our aim is to contour neurovascular structures automatically on prostate MRI by deep learning (DL) to improve workflow and interrater agreement.

Materials and methods: Segmentation of neurovascular structures was performed on pre-treatment 3.0 T MRI data of 131 prostate cancer patients (training [n = 105] and testing [n = 26]). The neurovascular structures include the penile bulb (PB), corpora cavernosa (CCs), internal pudendal arteries (IPAs), and neurovascular bundles (NVBs). Two DL networks, nnU-Net and DeepMedic, were trained for auto-contouring on prostate MRI and evaluated using volumetric Dice similarity coefficient (DSC), mean surface distances (MSD), Hausdorff distances, and surface DSC. Three radiation oncologists evaluated the DL-generated contours and performed corrections when necessary. Interrater agreement was assessed and the time required for manual correction was recorded.

Results: nnU-Net achieved a median DSC of 0.92 (IQR: 0.90-0.93) for the PB, 0.90 (IQR: 0.86-0.92) for the CCs, 0.79 (IQR: 0.77-0.83) for the IPAs, and 0.77 (IQR: 0.72-0.81) for the NVBs, which outperformed DeepMedic for each structure (p < 0.03). nnU-Net showed a median MSD of 0.24 mm for the IPAs and 0.71 mm for the NVBs. The median interrater DSC ranged from 0.93 to 1.00, with the majority of cases (68.9%) requiring manual correction times under two minutes.

Conclusions: DL enables reliable auto-contouring of neurovascular structures on pre-treatment MRI data, easing the clinical workflow in neurovascular-sparing MR-guided radiotherapy.

Keywords: Artificial intelligence (AI); Contouring; Deep learning (DL); Magnetic resonance-guided radiotherapy (MRgRT); Neurovascular-sparing; Prostate cancer (PCa).

Fig. 1

Representative case with the penile bulb (cyan), corpus cavernosum (green), internal pudendal artery (red), and neurovascular bundle (yellow) contours on T2-weighted MRI. Axial MR images obtained at the level of the penile bulb and corpora cavernosa (A), prostate apex level (B), prostate midgland level (C), and prostate base level (D). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

A representative example of the nnU-Net contour results after manual evaluation and correction for the internal pudendal arteries (red) and the neurovascular bundles (yellow) by three raters in the axial (A) and coronal (B) direction. Perfect interrater agreement was observed for the internal pudendal arteries between all raters and for the neurovascular bundles with rater 1 and 3. Rater 2 showed a Dice similarity coefficient of 0.90 for the left neurovascular bundle and 0.93 for the right neurovascular bundle with rater 1 and 3. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

A representative example of the nnU-Net contour results after manual evaluation and correction for the penile bulb (cyan) and the corpora cavernosa (green) by three raters in the axial (A) and coronal (B) direction. Perfect interrater agreement was observed for the corpora cavernosa between all raters and for the penile bulb between rater 1 and 3. Rater 2 showed a Dice similarity coefficient of 0.93 for the penile bulb with rater 1 and 3. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Pie chart with manual correction times for each DL-generated model of three radiation oncologists (3 × 15 patients = 45).