🔥Telegram İfşa

We find no advantage of b-values higher than 1000-1500 or advanced modeling for assessment of sPC by MRI. See our new publication pubmed.ncbi.nlm.nih.gov/32930560/

Direct comparison of prostate MRI lesion segmentations performed by a deep learning system and multiple radiologists, allowing adjustment of Dice score for human to machine comparisons thieme-connect.com/products/ejour…

Pubmed link of our newest article, currently pdf link not shown, see previous tweet for pdf link: Comparison of Prostate MRI Lesion Segmentation Agreement Between Multiple Radiologists and a Fully Automatic Deep L… pubmed.ncbi.nlm.nih.gov/33212541/

Very nice work, apparently initially not accepted at MICCAI, truly surprising performance. Setting the heuristics and parameters in such effective way reflects the impact of your multi-domain expertise.

Mean Dice coefficient of 4 independent manual prostate MRI lesion assessments was only moderate 0.48-0.52, providing a benchmark for DL systems. A fully automatic DL system achieved Dice of 0.22, a secondary quality criterion? pubmed.ncbi.nlm.nih.gov/33212541/

Continued sliding-window re-calibration of a DL system leads to more stable performance. Have you tried this with your deep learning systems ?pubmed.ncbi.nlm.nih.gov/32767102/

Our Radiology Publication entitled "Classification of Cancer at Prostate MRI: Deep Learning versus Clinical PI-RADS Assessment" is in the top 5% of Altmetric output, thank you all very much for your interest in our work and your support pubs.rsna.org/doi/10.1148/ra…

ProsRegNet: A deep learning framework for registration of MRI and histopathology images of the prostate sciencedirect.com/science/articl… "Very important step to understand power of any imaging: histology-imaging registration. A #deeplearning based approach is presented in this paper"

Unser frischer Reviewartikel zu ML in mpMRT der Prostata #DeepLearning #MachineLearning #review #urology #prostatecancer #mri springermedizin.de/maschinelles-l…

Improvement of PI-RADS-dependent prostate cancer classification by quantitative image assessment using radiomics or… pubmed.ncbi.nlm.nih.gov/34147597/

In this retrospective analysis of multiparametric prostate MRI, quantitative or qualitative dynamic contrast-enhanced MRI did not assist with PI-RADS 3 lesion risk stratification. DKFZ Universitätsklinikum Heidelberg bit.ly/3Psgei4

pubmed.ncbi.nlm.nih.gov/37331287/ We systematically assessed repeatability of ADC in PI-RADS lesions before and after repositioning of the patient in the MRI scanner.

In this work by Cedric Weißer et al. we examine weakly supervised MRI Slice‐Level Deep Learning Classification of Prostate Cancer and find it approximates Full Voxel‐ and Slice‐Level Annotation with increasing training set size onlinelibrary.wiley.com/doi/full/10.10…

In this work by Netzer et al. we further investigate the generalizability of prostate MRI deep learning for prostate cancer detection, while utilizing the previously proposed dynamic threshold calibration approach link.springer.com/article/10.100…

Just published: Adrian Schrader evaluates risk modeling of Prostata cancer using AI, nomograms and PI-RADS. link.springer.com/article/10.100…