A deep learning masked segmentation alternative to manual segmentation in biparametric MRI prostate cancer radiomics

Abstract

Objectives: To determine the value of a deep learning masked (DLM) auto-fixed volume of interest (VOI) segmentation method as an alternative to manual segmentation for radiomics-based diagnosis of clinically significant (CS) prostate cancer (PCa) on biparametric magnetic resonance imaging (bpMRI).

Materials and methods: This study included a retrospective multi-center dataset of 524 PCa lesions (of which 204 are CS PCa) on bpMRI. All lesions were both semi-automatically segmented with a DLM auto-fixed VOI method (averaging < 10 s per lesion) and manually segmented by an expert uroradiologist (averaging 5 min per lesion). The DLM auto-fixed VOI method uses a spherical VOI (with its center at the location of the lowest apparent diffusion coefficient of the prostate lesion as indicated with a single mouse click) from which non-prostate voxels are removed using a deep learning-based prostate segmentation algorithm. Thirteen different DLM auto-fixed VOI diameters (ranging from 6 to 30 mm) were explored. Extracted radiomics data were split into training and test sets (4:1 ratio). Performance was assessed with receiver operating characteristic (ROC) analysis.

Results: In the test set, the area under the ROC curve (AUCs) of the DLM auto-fixed VOI method with a VOI diameter of 18 mm (0.76 [95% CI: 0.66-0.85]) was significantly higher (p = 0.0198) than that of the manual segmentation method (0.62 [95% CI: 0.52-0.73]).

Conclusions: A DLM auto-fixed VOI segmentation can provide a potentially more accurate radiomics diagnosis of CS PCa than expert manual segmentation while also reducing expert time investment by more than 97%.

Key points: • Compared to traditional expert-based segmentation, a deep learning mask (DLM) auto-fixed VOI placement is more accurate at detecting CS PCa. • Compared to traditional expert-based segmentation, a DLM auto-fixed VOI placement is faster and can result in a 97% time reduction. • Applying deep learning to an auto-fixed VOI radiomics approach can be valuable.

Keywords: Biomarkers; Data curation; Deep learning; Multi-center study; Prostatic neoplasms.

Fig. 1

A Apparent diffusion coefficient map image (3× zoom factor) in a 76-year-old man with a suspicious lesion in the peripheral zone indicated by the arrow (PI-RADS 4) that proved to be an ISUP grade 2 PCa (based on MRI-TRUS fusion). B 18-mm auto-fixed VOI placed around the voxel with the lowest ADC value; due to the location of the lesion, a large number of voxels outside the prostate are included (red outline). C Result of auto-fixed VOI combined with deep learning–based segmentation to remove unwanted voxels outside the prostate

Fig. 2

Patient and lesion selection flowchart

Fig. 3

Axial T2-weighted images in a 73-year-old man with a suspicious lesion in the peripheral zone (PI-RADS 4, mostly based on DWI [third slice of ADC map containing the lesion attached for reference with a white cross indicating the single-click voxel with the quantitatively acquired lowest ADC value]) that proved to be an ISUP grade 3 PCa (based on MRI-TRUS fusion, confirmed by prostatectomy). A T2-weighted images without any segmentation. B Results of slice-by-slice manual lesion segmentation by an expert uroradiologist. C Results of 18-mm auto-fixed VOI lesion segmentation without DLM adjustment. D Deep learning–based total prostate segmentation. E Results of 18-mm auto-fixed VOI lesion segmentation with DLM adjustment

Fig. 4

Axial T2-weighted images in a 74-year-old man with a suspicious lesion in the transition zone (PI-RADS 3, mostly based on DWI [third slice of ADC map containing the lesion attached for reference with a white cross indicating the single-click voxel with the quantitatively acquired lowest ADC value]) that proved to be non-significant PCa (based on MRI-TRUS fusion). A T2-weighted images without any segmentation. B Results of slice-by-slice manual lesion segmentation by an expert uroradiologist. C Results of 18-mm auto-fixed VOI lesion segmentation without DLM adjustment. D Deep learning–based total prostate segmentation. E Results of 18-mm auto-fixed VOI lesion segmentation with DLM adjustment

Fig. 5

Deep learning masked auto-fixed VOI to extract bpMRI radiomics features for CS PCa: comparison of the performance of 13 different initial diameters of the auto-fixed VOI in the test set, expressed with AUCs and 95% confidence intervals (error bars)

Fig. 6

Test set smoothed ROC curves for the optimal DLM auto-fixed VOI model (initial 18–mm VOI diameter) and the model based on the expert manually segmented VOI as input