1
|
Bosma JS, Saha A, Hosseinzadeh M, Slootweg I, de Rooij M, Huisman H. Semisupervised Learning with Report-guided Pseudo Labels for Deep Learning-based Prostate Cancer Detection Using Biparametric MRI. Radiol Artif Intell 2023; 5:e230031. [PMID: 37795142 PMCID: PMC10546362 DOI: 10.1148/ryai.230031] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/04/2023] [Revised: 06/07/2023] [Accepted: 06/22/2023] [Indexed: 10/06/2023]
Abstract
Purpose To evaluate a novel method of semisupervised learning (SSL) guided by automated sparse information from diagnostic reports to leverage additional data for deep learning-based malignancy detection in patients with clinically significant prostate cancer. Materials and Methods This retrospective study included 7756 prostate MRI examinations (6380 patients) performed between January 2014 and December 2020 for model development. An SSL method, report-guided SSL (RG-SSL), was developed for detection of clinically significant prostate cancer using biparametric MRI. RG-SSL, supervised learning (SL), and state-of-the-art SSL methods were trained using 100, 300, 1000, or 3050 manually annotated examinations. Performance on detection of clinically significant prostate cancer by RG-SSL, SL, and SSL was compared on 300 unseen examinations from an external center with a histopathologically confirmed reference standard. Performance was evaluated using receiver operating characteristic (ROC) and free-response ROC analysis. P values for performance differences were generated with a permutation test. Results At 100 manually annotated examinations, mean examination-based diagnostic area under the ROC curve (AUC) values for RG-SSL, SL, and the best SSL were 0.86 ± 0.01 (SD), 0.78 ± 0.03, and 0.81 ± 0.02, respectively. Lesion-based detection partial AUCs were 0.62 ± 0.02, 0.44 ± 0.04, and 0.48 ± 0.09, respectively. Examination-based performance of SL with 3050 examinations was matched by RG-SSL with 169 manually annotated examinations, thus requiring 14 times fewer annotations. Lesion-based performance was matched with 431 manually annotated examinations, requiring six times fewer annotations. Conclusion RG-SSL outperformed SSL in clinically significant prostate cancer detection and achieved performance similar to SL even at very low annotation budgets.Keywords: Annotation Efficiency, Computer-aided Detection and Diagnosis, MRI, Prostate Cancer, Semisupervised Deep Learning Supplemental material is available for this article. Published under a CC BY 4.0 license.
Collapse
Affiliation(s)
- Joeran S. Bosma
- From the Diagnostic Image Analysis Group, Department of Medical
Imaging, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA
Nijmegen, the Netherlands
| | - Anindo Saha
- From the Diagnostic Image Analysis Group, Department of Medical
Imaging, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA
Nijmegen, the Netherlands
| | - Matin Hosseinzadeh
- From the Diagnostic Image Analysis Group, Department of Medical
Imaging, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA
Nijmegen, the Netherlands
| | - Ivan Slootweg
- From the Diagnostic Image Analysis Group, Department of Medical
Imaging, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA
Nijmegen, the Netherlands
| | - Maarten de Rooij
- From the Diagnostic Image Analysis Group, Department of Medical
Imaging, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA
Nijmegen, the Netherlands
| | - Henkjan Huisman
- From the Diagnostic Image Analysis Group, Department of Medical
Imaging, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA
Nijmegen, the Netherlands
| |
Collapse
|
2
|
Baffour FI, Huber NR, Ferrero A, Rajendran K, Glazebrook KN, Larson NB, Kumar S, Cook JM, Leng S, Shanblatt ER, McCollough CH, Fletcher JG. Photon-counting Detector CT with Deep Learning Noise Reduction to Detect Multiple Myeloma. Radiology 2023; 306:229-236. [PMID: 36066364 PMCID: PMC9771909 DOI: 10.1148/radiol.220311] [Citation(s) in RCA: 19] [Impact Index Per Article: 19.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/09/2022] [Revised: 06/15/2022] [Accepted: 07/18/2022] [Indexed: 12/24/2022]
Abstract
Background Photon-counting detector (PCD) CT and deep learning noise reduction may improve spatial resolution at lower radiation doses compared with energy-integrating detector (EID) CT. Purpose To demonstrate the diagnostic impact of improved spatial resolution in whole-body low-dose CT scans for viewing multiple myeloma by using PCD CT with deep learning denoising compared with conventional EID CT. Materials and Methods Between April and July 2021, adult participants who underwent a whole-body EID CT scan were prospectively enrolled and scanned with a PCD CT system in ultra-high-resolution mode at matched radiation dose (8 mSv for an average adult) at an academic medical center. EID CT and PCD CT images were reconstructed with Br44 and Br64 kernels at 2-mm section thickness. PCD CT images were also reconstructed with Br44 and Br76 kernels at 0.6-mm section thickness. The thinner PCD CT images were denoised by using a convolutional neural network. Image quality was objectively quantified in two phantoms and a randomly selected subset of participants (10 participants; median age, 63.5 years; five men). Two radiologists scored PCD CT images relative to EID CT by using a five-point Likert scale to detect findings reflecting multiple myeloma. The scoring for the matched reconstruction series was blinded to scanner type. Reader-averaged scores were tested with the null hypothesis of equivalent visualization between EID and PCD. Results Twenty-seven participants (median age, 68 years; IQR, 61-72 years; 16 men) were included. The blinded assessment of 2-mm images demonstrated improvement in viewing lytic lesions, intramedullary lesions, fatty metamorphosis, and pathologic fractures for PCD CT versus EID CT (P < .05 for all comparisons). The 0.6-mm PCD CT images with convolutional neural network denoising also demonstrated improvement in viewing all four pathologic abnormalities and detected one or more lytic lesions in 21 of 27 participants compared with the 2-mm EID CT images (P < .001). Conclusion Ultra-high-resolution photon-counting detector CT improved the visibility of multiple myeloma lesions relative to energy-integrating detector CT. © RSNA, 2022 Online supplemental material is available for this article.
Collapse
Affiliation(s)
- Francis I. Baffour
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Nathan R. Huber
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Andrea Ferrero
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Kishore Rajendran
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Katrina N. Glazebrook
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Nicholas B. Larson
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Shaji Kumar
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Joselle M. Cook
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Shuai Leng
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Elisabeth R. Shanblatt
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Cynthia H. McCollough
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| | - Joel G. Fletcher
- From the Department of Radiology (F.I.B., N.R.H., A.F., K.R., K.N.G.,
S.L., C.H.M., J.G.F.), Division of Biomedical Statistics and Informatics,
Department of Quantitative Health Sciences (N.B.L.), and Division of Hematology,
Department of Medicine (S.K., J.M.C.), Mayo Clinic, 200 First St SW, Rochester,
MN 55905; and Siemens Medical Solutions USA, Malvern, Pa (E.R.S.)
| |
Collapse
|
3
|
Schwartz FR, Vinson EN, Spritzer CE, Colglazier R, Samei E, French RJ, Said N, Waldman L, McCrum E. Prospective Multireader Evaluation of Photon-counting CT for Multiple Myeloma Screening. Radiol Imaging Cancer 2022; 4:e220073. [PMID: 36399038 PMCID: PMC9713593 DOI: 10.1148/rycan.220073] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/25/2022] [Revised: 08/31/2022] [Accepted: 10/26/2022] [Indexed: 05/27/2023]
Abstract
Purpose To determine whether photon-counting CT (PCCT) acquisition of whole-body CT images provides similar quantitative image quality and reader satisfaction for multiple myeloma screening at lower radiation doses than does standard energy-integrating detector (EID) CT. Materials and Methods Patients with monoclonal gammopathy of undetermined significance prospectively underwent clinical noncontrast whole-body CT with EID and same-day PCCT (August-December 2021). Five axial scan locations were evaluated by seven radiologists, with 11% (eight of 70) of images including osteolytic lesions. Images were shown in randomized order, and each reader rated the following: discernibility of the osseous cortex and osseous trabeculae, perceived image noise level, and diagnostic confidence. Presence of lytic osseous lesions was indicated. Contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR) were calculated. Comparisons were made using paired t tests and mixed linear effects models. Results Seven participants (four women) were included (mean age, 66 years ± 9 [SD]; body mass index, 30.1 kg/m2 ± 5.2). Mean cortical definition, trabecular definition, image noise, and image quality scores were 83, 67, 75, and 78 versus 84, 66, 74, and 76 for EID and PCCT, respectively (P = .65, .11, .26, and .11, respectively). PCCT helped identify more lesions (79% [22 of 28]) than did EID (64% [18 of 28]). CNRs and SNRs were similar between modalities. PCCT had lower radiation doses than EID (volume CT dose index: EID, 11.37 ± 2.8 vs PCCT, 1.8 ± 0.6 [P = .06]; dose-length product: EID, 1654.1 ± 409.6 vs PCCT, 253.4 ± 89.6 [P = .05]). Conclusion This pilot investigation suggests that PCCT affords similar quantitative and qualitative scores as EID at significantly lower radiation doses. Keywords: CT, CT-Spectral, Skeletal-Axial, Spine, Hematologic Diseases, Whole-Body Imaging, Comparative Studies Supplemental material is available for this article. © RSNA, 2022.
Collapse
|