Wen Foo

Positions:

Associate Professor of Pathology

Pathology
School of Medicine

Member of the Duke Cancer Institute

Duke Cancer Institute
School of Medicine

Education:

M.D. 2008

Duke University School of Medicine

Anatomic and Clinical Pathology, Pathology

Duke University School of Medicine

Cytopathology Fellowship, Pathology

Harvard Medical School

Grants:

Publications:

Structured approach to resolving discordance between PI-RADS v2.1 score and targeted prostate biopsy results: an opportunity for quality improvement.

BACKGROUND: Prostate multiparametric magnetic resonance imaging (mpMRI) can identify lesions within the prostate with characteristics identified in Prostate Imaging Reporting and Data System (PI-RADS) v2.1 associated with clinically significant prostate cancer (csPCa) or Gleason grade group (GGG) ≥ 2 at biopsy. OBJECTIVE: To assess concordance (PI-RADS 5 lesions with csPCa) of PI-RADS v2/2.1 with targeted, fusion biopsy results and to examine causes of discordance (PI-RADS 5 lesions without csPCa) with aim to provide a structured approach to resolving discordances and develop quality improvement (QI) protocols. METHODS: A retrospective study of 392 patients who underwent mpMRI at 3 Tesla followed by fusion biopsy. PI-RADS v2/2.1 scores were assigned to lesions identified on mpMRI and compared to biopsy results expressed as GGG. Positive predictive value (PPV) of PI-RADS v2/2.1 was calculated for all prostate cancer and csPCa. Discordant cases were re-reviewed by a radiologist with expertise in prostate mpMRI to determine reason for discordance. RESULTS: A total of 521 lesions were identified on mpMRI. 121/521 (23.2%), 310/524 (59.5%), and 90/521 (17.3%) were PI-RADS 5, 4, and 3, respectively. PPV of PI-RADS 5, 4, and 3 for all PCa and csPCa was 0.80, 0.55, 0.24 and 0.63, 0.33, and 0.09, respectively. 45 cases of discordant biopsy results for PI-RADS 5 lesions were found with 27 deemed "true" discordances or "unresolved" discordances where imaging re-review confirmed PI-RADS appropriateness, while 18 were deemed "false" or resolved discordances due to downgrading of PI-RADS scores based on imaging re-review. Adjusting for resolved discordances on re-review, the PPV of PI-RADS 5 lesions for csPCa was deemed to be 0.74 and upon adjusting for presence of csPCa found in cases of unresolved discordance, PPV rose to 0.83 for PI-RADS 5 lesions. CONCLUSION: Although PIRADS 5 lesions are considered high risk for csPCa, the PPV is not 100% and a diagnostic dilemma occurs when targeted biopsy returns discordant. While PI-RADS score is downgraded in some cases upon imaging re-review, a number of "false" or "unresolved" discordances were identified in which MRI re-review confirmed initial PI-RADS score and subsequent pathology confirmed presence of csPCa in these lesions. CLINICAL IMPACT: We propose a structured approach to resolving discordant biopsy results using multi-disciplinary re-review of imaging and archived biopsy strikes as a quality improvement pathway. Further work is needed to determine the value of re-biopsy in cases of unresolved discordance and to develop robust QI systems for prostate MRI.
Authors
Arcot, R; Sekar, S; Kotamarti, S; Krischak, M; Michael, ZD; Foo, W-C; Huang, J; Polascik, TJ; Gupta, RT
MLA Citation
Arcot, Rohith, et al. “Structured approach to resolving discordance between PI-RADS v2.1 score and targeted prostate biopsy results: an opportunity for quality improvement.Abdom Radiol (Ny), vol. 47, no. 8, Aug. 2022, pp. 2917–27. Pubmed, doi:10.1007/s00261-022-03562-w.
URI
https://scholars.duke.edu/individual/pub1523848
PMID
35674785
Source
pubmed
Published In
Abdom Radiol (Ny)
Volume
47
Published Date
Start Page
2917
End Page
2927
DOI
10.1007/s00261-022-03562-w

Teaching interventional cytopathology

Interventional cytopathology is a unique area of pathology, where cytopathologists play a primary role in obtaining fine needle aspiration biopsies and/or making determinations through rapid on-site evaluations to guide sample procurement in real-time. Unsurprisingly, experience and skill are directly related to success in these endeavors, and both can be fostered with formal instruction. There is a wealth of resources available to aid in teaching interventional cytopathology, including instructional videos, courses, and model phantoms which can help to build familiarity and confidence. These tools can provide a basic framework upon which skills can be developed through in-person guidance, real-time feedback and practice. This article reviews the tools available to enhance training, details the authors’ institutional experience in teaching interventional cytopathology at a tertiary care center, and provides recommendations and pearls for success in this endeavor.
Authors
MLA Citation
Jiang, X. S., and W. C. Foo. “Teaching interventional cytopathology.” Seminars in Diagnostic Pathology, Jan. 2022. Scopus, doi:10.1053/j.semdp.2022.01.002.
URI
https://scholars.duke.edu/individual/pub1507548
Source
scopus
Published In
Seminars in Diagnostic Pathology
Published Date
DOI
10.1053/j.semdp.2022.01.002

A Hybrid Human-Machine Learning Approach for Screening Prostate Biopsies Can Improve Clinical Efficiency Without Compromising Diagnostic Accuracy.

CONTEXT.—: Prostate cancer is a common malignancy, and accurate diagnosis typically requires histologic review of multiple prostate core biopsies per patient. As pathology volumes and complexity increase, new tools to improve the efficiency of everyday practice are keenly needed. Deep learning has shown promise in pathology diagnostics, but most studies silo the efforts of pathologists from the application of deep learning algorithms. Very few hybrid pathologist-deep learning approaches have been explored, and these typically require complete review of histologic slides by both the pathologist and the deep learning system. OBJECTIVE.—: To develop a novel and efficient hybrid human-machine learning approach to screen prostate biopsies. DESIGN.—: We developed an algorithm to determine the 20 regions of interest with the highest probability of malignancy for each prostate biopsy; presenting these regions to a pathologist for manual screening limited the initial review by a pathologist to approximately 2% of the tissue area of each sample. We evaluated this approach by using 100 biopsies (29 malignant, 60 benign, 11 other) that were reviewed by 4 pathologists (3 urologic pathologists, 1 general pathologist) using a custom-designed graphical user interface. RESULTS.—: Malignant biopsies were correctly identified as needing comprehensive review with high sensitivity (mean, 99.2% among all pathologists); conversely, most benign prostate biopsies (mean, 72.1%) were correctly identified as needing no further review. CONCLUSIONS.—: This novel hybrid system has the potential to efficiently triage out most benign prostate core biopsies, conserving time for the pathologist to dedicate to detailed evaluation of malignant biopsies.
Authors
Dov, D; Assaad, S; Syedibrahim, A; Bell, J; Huang, J; Madden, J; Bentley, R; McCall, S; Henao, R; Carin, L; Foo, W-C
MLA Citation
Dov, David, et al. “A Hybrid Human-Machine Learning Approach for Screening Prostate Biopsies Can Improve Clinical Efficiency Without Compromising Diagnostic Accuracy.Arch Pathol Lab Med, vol. 146, no. 6, June 2022, pp. 727–34. Pubmed, doi:10.5858/arpa.2020-0850-OA.
URI
https://scholars.duke.edu/individual/pub1498438
PMID
34591085
Source
pubmed
Published In
Arch Pathol Lab Med
Volume
146
Published Date
Start Page
727
End Page
734
DOI
10.5858/arpa.2020-0850-OA

Clinicopathological features to distinguish malignant solitary fibrous tumors of the prostate from prostatic stromal tumors.

Mesenchymal tumors of the prostate are rare but encompass a wide differential diagnosis. In our study, we aimed to investigate the clinicopathological features that can be used to differentiate malignant solitary fibrous tumors (mSFTs) occurring in the prostate from prostatic stromal tumors. A total of 15 patients with mesenchymal tumors of the prostate were identified in Nanjing Drum Tower Hospital from 2009 to 2019, including 3 mSFTs, 9 stromal tumors of uncertain malignant potential (STUMPs), and 3 prostatic stromal sarcomas (PSSs). Immunohistochemical stains for signal transducer and activator of transcription 6 (STAT6), aldehyde dehydrogenase 1 (ALDH1), CD34, desmin, smooth muscle actin (SMA), progesterone receptor (PR), CD117, and cytokeratin (CK) were performed on representative sections from each tumor, and the clinical features, histology, and immunophenotype of these three groups were analyzed. There was no significant difference in mean patient age of patients diagnosed with mSFTs, STUMPs, and PSSs. mSFTs and PSSs showed significantly increased tumor size (p < 0.05), Ki-67 proliferation index (p < 0.0001), and mitotic activity (p < 0.05) when compared with STUMPs. mSFTs showed significantly higher expression of STAT6 compared with both PSSs and STUMPs (p < 0.0001, p < 0.0001). PR showed significantly more expression in STUMPs than in mSFTs or PSSs (p < 0.0001, p < 0.0001). Desmin and SMA showed significantly more expression in STUMPs than in mSFTs (p < 0.05). ALDH1, CD117, CK, and CD34 showed no significant difference in staining between mSFTs, STUMPs, and PSSs. Therefore, a limited panel of STAT6, PR, and Ki-67 may be useful in distinguishing between mSFTs, STUMPs, and PSSs.
Authors
Xu, Y; Li, Z; Shi, J; Fu, Y; Zhu, L; Fan, X; Foo, W-C
MLA Citation
Xu, Yuemei, et al. “Clinicopathological features to distinguish malignant solitary fibrous tumors of the prostate from prostatic stromal tumors.Virchows Arch, vol. 478, no. 4, Apr. 2021, pp. 619–26. Pubmed, doi:10.1007/s00428-020-02909-2.
URI
https://scholars.duke.edu/individual/pub1456763
PMID
32820389
Source
pubmed
Published In
Virchows Arch
Volume
478
Published Date
Start Page
619
End Page
626
DOI
10.1007/s00428-020-02909-2

Multiparametric Ultrasound for Targeting Prostate Cancer: Combining ARFI, SWEI, QUS and B-Mode.

Diagnosing prostate cancer through standard transrectal ultrasound (TRUS)-guided biopsy is challenging because of the sensitivity and specificity limitations of B-mode imaging. We used a linear support vector machine (SVM) to combine standard TRUS imaging data with acoustic radiation force impulse (ARFI) imaging data, shear wave elasticity imaging (SWEI) data and quantitative ultrasound (QUS) midband fit data to enhance lesion contrast into a synthesized multiparametric ultrasound volume. This SVM was trained and validated using a subset of 20 patients and tested on a second subset of 10 patients. Multiparametric US led to a statistically significant improvements in contrast, contrast-to-noise ratio (CNR) and generalized CNR (gCNR) when compared with standard TRUS B-mode and SWEI; in contrast and CNR when compared with MF; and in CNR when compared with ARFI. ARFI, MF and SWEI also outperformed TRUS B-mode in contrast, with MF outperforming B-mode in CNR and gCNR as well. ARFI, although only yielding statistically significant differences in contrast compared with TRUS B-mode, captured critical qualitative features for lesion identification. Multiparametric US enhanced lesion visibility metrics and is a promising technique for targeted TRUS-guided prostate biopsy in the future.
Authors
Morris, DC; Chan, DY; Lye, TH; Chen, H; Palmeri, ML; Polascik, TJ; Foo, W-C; Huang, J; Mamou, J; Nightingale, KR
MLA Citation
Morris, D. Cody, et al. “Multiparametric Ultrasound for Targeting Prostate Cancer: Combining ARFI, SWEI, QUS and B-Mode.Ultrasound Med Biol, vol. 46, no. 12, Dec. 2020, pp. 3426–39. Pubmed, doi:10.1016/j.ultrasmedbio.2020.08.022.
URI
https://scholars.duke.edu/individual/pub1461735
PMID
32988673
Source
pubmed
Published In
Ultrasound Med Biol
Volume
46
Published Date
Start Page
3426
End Page
3439
DOI
10.1016/j.ultrasmedbio.2020.08.022