Classification of cancer at prostate MRI: deep learning versus clinical PI-RADS assessment
To compare the performance of clinical assessment to a deep learning system optimized for segmentation trained with T2-weighted and diffusion MRI in the task of detection and segmentation of lesions suspicious for sPC.
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
Oct 8 2019
|
| In: |
Radiology
Year: 2019, Volume: 293, Issue: 3, Pages: 607-617 |
| ISSN: | 1527-1315 |
| DOI: | 10.1148/radiol.2019190938 |
| Online Access: | Resolving-System, Volltext: https://doi.org/10.1148/radiol.2019190938 Verlag: https://pubs.rsna.org/doi/10.1148/radiol.2019190938 
|
| Author Notes: | $hPatrick Schelb, Simon Kohl, MSc, Jan Philipp Radtke, MD, Manuel Wiesenfarth, PhD, Philipp Kickingereder, MD, Sebastian Bickelhaupt, MD, Tristan Anselm Kuder, PhD,Albrecht Stenzinger, MD, Markus Hohenfellner, MD, Heinz-Peter Schlemmer, MD, PhD, Klaus H. Maier-Hein, PhD, David Bonekamp, MD |
MARC
| LEADER | 00000caa a2200000 c 4500 | ||
|---|---|---|---|
| 001 | 1678861839 | ||
| 003 | DE-627 | ||
| 005 | 20220817001419.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 191015s2019 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1148/radiol.2019190938 |2 doi | |
| 035 | |a (DE-627)1678861839 | ||
| 035 | |a (DE-599)KXP1678861839 | ||
| 035 | |a (OCoLC)1341246737 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 33 |2 sdnb | ||
| 100 | 1 | |a Schelb, Patrick |d 1994- |e VerfasserIn |0 (DE-588)1197061495 |0 (DE-627)1678864773 |4 aut | |
| 245 | 1 | 0 | |a Classification of cancer at prostate MRI |b deep learning versus clinical PI-RADS assessment |c $hPatrick Schelb, Simon Kohl, MSc, Jan Philipp Radtke, MD, Manuel Wiesenfarth, PhD, Philipp Kickingereder, MD, Sebastian Bickelhaupt, MD, Tristan Anselm Kuder, PhD,Albrecht Stenzinger, MD, Markus Hohenfellner, MD, Heinz-Peter Schlemmer, MD, PhD, Klaus H. Maier-Hein, PhD, David Bonekamp, MD |
| 264 | 1 | |c Oct 8 2019 | |
| 300 | |a 11 | ||
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Gesehen am 15.10.2019 | ||
| 520 | |a To compare the performance of clinical assessment to a deep learning system optimized for segmentation trained with T2-weighted and diffusion MRI in the task of detection and segmentation of lesions suspicious for sPC. | ||
| 520 | |a BackgroundMen suspected of having clinically significant prostate cancer (sPC) increasingly undergo prostate MRI. The potential of deep learning to provide diagnostic support for human interpretation requires further evaluation.PurposeTo compare the performance of clinical assessment to a deep learning system optimized for segmentation trained with T2-weighted and diffusion MRI in the task of detection and segmentation of lesions suspicious for sPC.Materials and MethodsIn this retrospective study, T2-weighted and diffusion prostate MRI sequences from consecutive men examined with a single 3.0-T MRI system between 2015 and 2016 were manually segmented. Ground truth was provided by combined targeted and extended systematic MRI-transrectal US fusion biopsy, with sPC defined as International Society of Urological Pathology Gleason grade group greater than or equal to 2. By using split-sample validation, U-Net was internally validated on the training set (80% of the data) through cross validation and subsequently externally validated on the test set (20% of the data). U-Net-derived sPC probability maps were calibrated by matching sextant-based cross-validation performance to clinical performance of Prostate Imaging Reporting and Data System (PI-RADS). Performance of PI-RADS and U-Net were compared by using sensitivities, specificities, predictive values, and Dice coefficient.ResultsA total of 312 men (median age, 64 years; interquartile range [IQR], 58-71 years) were evaluated. The training set consisted of 250 men (median age, 64 years; IQR, 58-71 years) and the test set of 62 men (median age, 64 years; IQR, 60-69 years). In the test set, PI-RADS cutoffs greater than or equal to 3 versus cutoffs greater than or equal to 4 on a per-patient basis had sensitivity of 96% (25 of 26) versus 88% (23 of 26) at specificity of 22% (eight of 36) versus 50% (18 of 36). U-Net at probability thresholds of greater than or equal to 0.22 versus greater than or equal to 0.33 had sensitivity of 96% (25 of 26) versus 92% (24 of 26) (both P > .99) with specificity of 31% (11 of 36) versus 47% (17 of 36) (both P > .99), not statistically different from PI-RADS. Dice coefficients were 0.89 for prostate and 0.35 for MRI lesion segmentation. In the test set, coincidence of PI-RADS greater than or equal to 4 with U-Net lesions improved the positive predictive value from 48% (28 of 58) to 67% (24 of 36) for U-Net probability thresholds greater than or equal to 0.33 (P = .01), while the negative predictive value remained unchanged (83% [25 of 30] vs 83% [43 of 52]; P > .99).ConclusionU-Net trained with T2-weighted and diffusion MRI achieves similar performance to clinical Prostate Imaging Reporting and Data System assessment.© RSNA, 2019Online supplemental material is available for this article.See also the editorial by Padhani and Turkbey in this issue. | ||
| 700 | 1 | |a Radtke, Jan Philipp |d 1985- |e VerfasserIn |0 (DE-588)1020328290 |0 (DE-627)688118089 |0 (DE-576)360178170 |4 aut | |
| 700 | 1 | |a Vollmuth, Philipp |d 1987- |e VerfasserIn |0 (DE-588)1043270086 |0 (DE-627)771319177 |0 (DE-576)394600738 |4 aut | |
| 700 | 1 | |a Stenzinger, Albrecht |e VerfasserIn |0 (DE-588)139606106 |0 (DE-627)703395238 |0 (DE-576)312432755 |4 aut | |
| 700 | 1 | |a Hohenfellner, Markus |d 1958- |e VerfasserIn |0 (DE-588)133862518 |0 (DE-627)557857988 |0 (DE-576)300155263 |4 aut | |
| 700 | 1 | |a Maier-Hein, Klaus H. |d 1980- |e VerfasserIn |0 (DE-588)1100551875 |0 (DE-627)85946461X |0 (DE-576)333771222 |4 aut | |
| 700 | 1 | |a Bonekamp, David |d 1977- |e VerfasserIn |0 (DE-588)128868104 |0 (DE-627)383668581 |0 (DE-576)297371797 |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t Radiology |d Oak Brook, Ill. : Soc., 1923 |g 293(2019), 3, Seite 607-617 |h Online-Ressource |w (DE-627)320487253 |w (DE-600)2010588-5 |w (DE-576)094056706 |x 1527-1315 |7 nnas |a Classification of cancer at prostate MRI deep learning versus clinical PI-RADS assessment |
| 773 | 1 | 8 | |g volume:293 |g year:2019 |g number:3 |g pages:607-617 |g extent:11 |a Classification of cancer at prostate MRI deep learning versus clinical PI-RADS assessment |
| 856 | 4 | 0 | |u https://doi.org/10.1148/radiol.2019190938 |x Resolving-System |x Verlag |3 Volltext |
| 856 | 4 | 0 | |u https://pubs.rsna.org/doi/10.1148/radiol.2019190938 |x Verlag |
| 951 | |a AR | ||
| 992 | |a 20191015 | ||
| 993 | |a Article | ||
| 994 | |a 2019 | ||
| 998 | |g 128868104 |a Bonekamp, David |m 128868104:Bonekamp, David |d 50000 |e 50000PB128868104 |k 0/50000/ |p 12 |y j | ||
| 998 | |g 1100551875 |a Maier-Hein, Klaus H. |m 1100551875:Maier-Hein, Klaus H. |d 910000 |d 911400 |d 50000 |e 910000PM1100551875 |e 911400PM1100551875 |e 50000PM1100551875 |k 0/910000/ |k 1/910000/911400/ |k 0/50000/ |p 11 | ||
| 998 | |g 133862518 |a Hohenfellner, Markus |m 133862518:Hohenfellner, Markus |d 910000 |d 910200 |e 910000PH133862518 |e 910200PH133862518 |k 0/910000/ |k 1/910000/910200/ |p 9 | ||
| 998 | |g 139606106 |a Stenzinger, Albrecht |m 139606106:Stenzinger, Albrecht |d 910000 |d 912000 |e 910000PS139606106 |e 912000PS139606106 |k 0/910000/ |k 1/910000/912000/ |p 8 | ||
| 998 | |g 1043270086 |a Vollmuth, Philipp |m 1043270086:Vollmuth, Philipp |d 910000 |d 911100 |e 910000PV1043270086 |e 911100PV1043270086 |k 0/910000/ |k 1/910000/911100/ |p 5 | ||
| 998 | |g 1020328290 |a Radtke, Jan Philipp |m 1020328290:Radtke, Jan Philipp |d 910000 |d 910200 |e 910000PR1020328290 |e 910200PR1020328290 |k 0/910000/ |k 1/910000/910200/ |p 3 | ||
| 998 | |g 1197061495 |a Schelb, Patrick |m 1197061495:Schelb, Patrick |d 50000 |e 50000PS1197061495 |k 0/50000/ |p 1 |x j | ||
| 999 | |a KXP-PPN1678861839 |e 3522785479 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"relHost":[{"title":[{"title":"Radiology","title_sort":"Radiology"}],"physDesc":[{"extent":"Online-Ressource"}],"note":["Fortsetzung der Druck-Ausgabe","Gesehen 07.11.22"],"language":["eng"],"origin":[{"publisher":"Soc.","publisherPlace":"Oak Brook, Ill.","dateIssuedDisp":"1923-","dateIssuedKey":"1923"}],"disp":"Classification of cancer at prostate MRI deep learning versus clinical PI-RADS assessmentRadiology","pubHistory":["1.1923 -"],"type":{"media":"Online-Ressource","bibl":"periodical"},"recId":"320487253","name":{"displayForm":["The Radiological Society of North America"]},"part":{"issue":"3","pages":"607-617","year":"2019","text":"293(2019), 3, Seite 607-617","volume":"293","extent":"11"},"corporate":[{"role":"isb","display":"Radiological Society of North America"}],"id":{"zdb":["2010588-5"],"issn":["1527-1315"],"eki":["320487253"]}}],"type":{"media":"Online-Ressource","bibl":"article-journal"},"origin":[{"dateIssuedKey":"2019","dateIssuedDisp":"Oct 8 2019"}],"language":["eng"],"id":{"doi":["10.1148/radiol.2019190938"],"eki":["1678861839"]},"note":["Gesehen am 15.10.2019"],"recId":"1678861839","name":{"displayForm":["$hPatrick Schelb, Simon Kohl, MSc, Jan Philipp Radtke, MD, Manuel Wiesenfarth, PhD, Philipp Kickingereder, MD, Sebastian Bickelhaupt, MD, Tristan Anselm Kuder, PhD,Albrecht Stenzinger, MD, Markus Hohenfellner, MD, Heinz-Peter Schlemmer, MD, PhD, Klaus H. Maier-Hein, PhD, David Bonekamp, MD"]},"physDesc":[{"extent":"11 S."}],"title":[{"subtitle":"deep learning versus clinical PI-RADS assessment","title":"Classification of cancer at prostate MRI","title_sort":"Classification of cancer at prostate MRI"}],"person":[{"display":"Schelb, Patrick","role":"aut","given":"Patrick","family":"Schelb"},{"role":"aut","display":"Radtke, Jan Philipp","family":"Radtke","given":"Jan Philipp"},{"role":"aut","display":"Vollmuth, Philipp","given":"Philipp","family":"Vollmuth"},{"display":"Stenzinger, Albrecht","role":"aut","given":"Albrecht","family":"Stenzinger"},{"role":"aut","display":"Hohenfellner, Markus","given":"Markus","family":"Hohenfellner"},{"given":"Klaus H.","family":"Maier-Hein","display":"Maier-Hein, Klaus H.","role":"aut"},{"given":"David","family":"Bonekamp","role":"aut","display":"Bonekamp, David"}]} | ||
| SRT | |a SCHELBPATRCLASSIFICA8201 | ||