Dataset Information

Test-retest repeatability of a deep learning architecture in detecting and segmenting clinically significant prostate cancer on apparent diffusion coefficient (ADC) maps.

ABSTRACT:

Objectives

To evaluate short-term test-retest repeatability of a deep learning architecture (U-Net) in slice- and lesion-level detection and segmentation of clinically significant prostate cancer (csPCa: Gleason grade group >?1) using diffusion-weighted imaging fitted with monoexponential function, ADC_m.

Methods

One hundred twelve patients with prostate cancer (PCa) underwent 2 prostate MRI examinations on the same day. PCa areas were annotated using whole mount prostatectomy sections. Two U-Net-based convolutional neural networks were trained on three different ADC_m b value settings for (a) slice- and (b) lesion-level detection and (c) segmentation of csPCa. Short-term test-retest repeatability was estimated using intra-class correlation coefficient (ICC(3,1)), proportionate agreement, and dice similarity coefficient (DSC). A 3-fold cross-validation was performed on training set (N?=?78 patients) and evaluated for performance and repeatability on testing data (N?=?34 patients).

Results

For the three ADC_m b value settings, repeatability of mean ADC_m of csPCa lesions was ICC(3,1)?=?0.86-0.98. Two CNNs with U-Net-based architecture demonstrated ICC(3,1) in the range of 0.80-0.83, agreement of 66-72%, and DSC of 0.68-0.72 for slice- and lesion-level detection and segmentation of csPCa. Bland-Altman plots suggest that there is no systematic bias in agreement between inter-scan ground truth segmentation repeatability and segmentation repeatability of the networks.

Conclusions

For the three ADC_m b value settings, two CNNs with U-Net-based architecture were repeatable for the problem of detection of csPCa at the slice-level. The network repeatability in segmenting csPCa lesions is affected by inter-scan variability and ground truth segmentation repeatability and may thus improve with better inter-scan reproducibility.

Key points

• For the three ADC_m b value settings, two CNNs with U-Net-based architecture were repeatable for the problem of detection of csPCa at the slice-level. • The network repeatability in segmenting csPCa lesions is affected by inter-scan variability and ground truth segmentation repeatability and may thus improve with better inter-scan reproducibility.

SUBMITTER: Hiremath A

PROVIDER: S-EPMC7821380 | biostudies-literature | 2021 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Test-retest repeatability of a deep learning architecture in detecting and segmenting clinically significant prostate cancer on apparent diffusion coefficient (ADC) maps.

Hiremath Amogh A Shiradkar Rakesh R Merisaari Harri H Prasanna Prateek P Ettala Otto O Taimen Pekka P Aronen Hannu J HJ Boström Peter J PJ Jambor Ivan I Madabhushi Anant A

European radiology 20200723 1

<h4>Objectives</h4>To evaluate short-term test-retest repeatability of a deep learning architecture (U-Net) in slice- and lesion-level detection and segmentation of clinically significant prostate cancer (csPCa: Gleason grade group > 1) using diffusion-weighted imaging fitted with monoexponential function, ADC<sub>m</sub>.<h4>Methods</h4>One hundred twelve patients with prostate cancer (PCa) underwent 2 prostate MRI examinations on the same day. PCa areas were annotated using whole mount prostat ...[more]

PMID: 32700021

Similar Datasets

Project description:BackgroundBrain metastases are particularly common in patients with small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), with NSCLC showing a less aggressive clinical course and lower chemo- and radio sensitivity compared to SCLC. Early adequate therapy is highly desirable and depends on a reliable classification of tumor type. The apparent diffusion coefficient is a noninvasive neuroimaging marker with the potential to differentiate between major histological subtypes. Here we determine the sensitivity and specificity of the apparent diffusion coefficient to distinguish between NSCLC and SCLC.MethodsWe enrolled all NSCLC and SCLC patients diagnosed between 2008 and 2019 at the University Medical Center Göttingen. Cranial MR scans were visually inspected for brain metastases and the ratio of the apparent diffusion coefficient (ADC) was calculated by dividing the ADC measured within the solid part of a metastasis by a reference ADC extracted from an equivalent region in unaffected tissue on the contralateral hemisphere.ResultsOut of 411 enrolled patients, we detected 129 patients (83 NSCLC, 46 SCLC) with sufficiently large brain metastases with histologically classified lung cancer and no hemorrhage. We analyzed 185 brain metastases, 84 of SCLC and 101 of NSCLC. SCLC brain metastases showed an ADC ratio of 0.68 ± 0.12 SD, and NSCLC brain metastases showed an ADC ratio of 1.47 ± 0.31 SD. Receiver operating curve statistics differentiated brain metastases of NSCLC from SCLC with an area under the curve of 0.99 and a 95% CI of 0.98 to 1, p < 0.001. Youden's J cut-point is 0.97 at a sensitivity of 0.989 and a specificity of 0.988.ConclusionsIn patients with lung cancer and brain metastases with solid tumor parts, ADC ratio enables an ad hoc differentiation of SCLC and NSCLC, easily achieved during routine neuroradiological examination. Non-invasive MR imaging enables an early-individualized management of brain metastases from lung cancer.Trial registrationThe study was registered in the German Clinical Trials Register (DRKS00023016).

Project description:Purpose To assess the repeatability of apparent diffusion coefficient (ADC) estimates in extracranial soft-tissue diffusion-weighted magnetic resonance imaging across a wide range of imaging protocols and patient populations. Materials and Methods Nine prospective patient studies and one prospective volunteer study, performed between 2006 and 2016 with research ethics committee approval and written informed consent from each subject, were included in this single-institution study. A total of 141 tumors and healthy organs were imaged twice (interval between repeated examinations, 45 minutes to 10 days, depending the on study) to assess the repeatability of median and mean ADC estimates. The Levene test was used to determine whether ADC repeatability differed between studies. The Pearson linear correlation coefficient was used to assess correlation between coefficient of variation (CoV) and the year the study started, study size, and volumes of tumors and healthy organs. The repeatability of ADC estimates from small, medium, and large tumors and healthy organs was assessed irrespective of study, and the Levene test was used to determine whether ADC repeatability differed between these groups. Results CoV aggregated across all studies was 4.1% (range for each study, 1.7%-6.5%). No correlation was observed between CoV and the year the study started or study size. CoV was weakly correlated with volume (r = -0.5, P = .1). Repeatability was significantly different between small, medium, and large tumors (P < .05), with the lowest CoV (2.6%) for large tumors. There was a significant difference in repeatability between studies-a difference that did not persist after the study with the largest tumors was excluded. Conclusion ADC is a robust imaging metric with excellent repeatability in extracranial soft tissues across a wide range of tumor sites, sizes, patient populations, and imaging protocol variations. Online supplemental material is available for this article.

Project description:ObjectivesThe aim of this study was to evaluate the repeatability of a region of interest (ROI) volume and mean apparent diffusion coefficient (ADC) in standard-of-care 3 T multiparametric magnetic resonance imaging (mpMRI) of the prostate obtained with the use of endorectal coil.Materials and methodsThis prospective study was Health Insurance Portability and Accountability Act compliant, with institutional review board approval and written informed consent. Men with confirmed or suspected treatment-naive prostate cancer scheduled for mpMRI were offered a repeat mpMRI within 2 weeks. Regions of interest corresponding to the whole prostate gland, the entire peripheral zone (PZ), normal PZ, and suspected tumor ROI (tROI) on axial T2-weighted, dynamic contrast-enhanced subtract, and ADC images were annotated and assessed using Prostate Imaging Reporting and Data System (PI-RADS) v2. Repeatability of the ROI volume for each of the analyzed image types and mean ROI ADC was summarized with repeatability coefficient (RC) and RC%.ResultsA total of 189 subjects were approached to participate in the study. Of 40 patients that gave initial agreement, 15 men underwent 2 mpMRI examinations and completed the study. Peripheral zone tROIs were identified in 11 subjects. Tumor ROI volume was less than 0.5 mL in 8 of 11 subjects. PI-RADS categories were identical between baseline-repeat studies in 11/15 subjects and differed by 1 point in 4/15. Peripheral zone tROI volume RC (RC%) was 233 mm (71%) on axial T2-weighted, 422 mm (112%) on ADC, and 488 mm (119%) on dynamic contrast-enhanced subtract. Apparent diffusion coefficient ROI mean RC (RC%) were 447 × 10 mm/s (42%) in PZ tROI and 471 × 10 mm/s (30%) in normal PZ. Significant difference in repeatability of the tROI volume across series was observed (P < 0.005). The mean ADC RC% was lower than volume RC% for tROI ADC (P < 0.05).ConclusionsPI-RADS v2 overall assessment was highly repeatable. Multiparametric magnetic resonance imaging sequences differ in volume measurement repeatability. The mean tROI ADC is more repeatable compared with tROI volume in ADC. Repeatability of prostate ADC is comparable with that in other abdominal organs.

Dataset Information

Test-retest repeatability of a deep learning architecture in detecting and segmenting clinically significant prostate cancer on apparent diffusion coefficient (ADC) maps.

Objectives

Methods

Results

Conclusions

Key points

Publications

Test-retest repeatability of a deep learning architecture in detecting and segmenting clinically significant prostate cancer on apparent diffusion coefficient (ADC) maps.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets