Mary Wong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Mary Wong has a total of 49 co-authors. Here are the top co-authors by number of shared publications:
- Kevin M. Waters
Department of Pathology and Laboratory Medicine Cedars‐Sinai Medical Center Los Angeles California USA
3 shared publications - Brent K. Larson
Department of Pathology and Laboratory Medicine Cedars‐Sinai Medical Center Los Angeles California USA
2 shared publications - Brian Cox
Pacific Rim Pathology San Diego California USA
2 shared publications Deepti Dhall
Cedars-Sinai Medical Center, Los Angeles, CA, USA
2 shared publications- Lysandra Voltaggio
Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, MD, USA
2 shared publications Stacey Kim
Cedars-Sinai Medical Center, Los Angeles, CA 90048, USA
2 shared publications
Latest Publications
- 23 April 2025
- Published by Wolters Kluwer Health in American Journal of Gastroenterology
- Vol. 121 (4) , 863-870
Deep Learning Model for Histologic Diagnosis of Dysplastic Barrett's Esophagus: Multisite Cohort External Validation
INTRODUCTION: The risk of progression to esophageal adenocarcinoma in Barrett's esophagus (BE) increases with advancing degrees of dysplasia. There is a critical need to improve the diagnosis of BE dysplasia, given substantial interobserver variability and overcalls of dysplasia during manual community pathologist reads. We aimed to externally validate a previously cross-validated BE dysplasia diagnosis deep learning model (BEDDLM) that predicts dysplasia grade on whole slide images (WSIs). METHODS: We digitized nondysplastic BE (NDBE), low-grade (LGD), and high-grade dysplasia (HGD) histology slides from 3 external academic centers. A consensus read by 2 expert study pathologists was used as the criterion standard. Slide stain characteristics were normalized using cycle-generative adversarial networks. WSIs were assessed by BEDDLM using an ensemble approach, combining a “You Only Look Once” model followed by a ResNet101 classifier model. RESULTS: We included 489 WSIs. Consensus histopathology revealed 232 NDBE, 117 LGD, and 140 HGD WSIs. The mean age (SD) was 66.9 (11.4) years; 413 (84.7%) were men. Using the BEDDLM ensemble model, sensitivity and specificity for NDBE were 73.3% (95% confidence interval [CI]: 67.09%–78.85%) and 93.4% (95% CI: 89.62%–96.10%); for LGD, 84.6% (95% CI: 76.78%–90.62%) and 80.6% (95% CI: 76.26%–84.54%); and for HGD, 80.7% (95% CI: 73.19%–86.89%) and 94.8% (95% CI: 91.97%–96.91%), respectively. The F1 score was 0.81, 0 0.69, and 0.83 for NDBE, LGD, and HGD, respectively. DISCUSSION: Our externally validated deep learning model demonstrates substantial accuracy for the diagnosis of BE dysplasia grade on WSIs.
- 23 April 2025
- Published by Wolters Kluwer Health in American Journal of Gastroenterology
- Vol. 121 (4) , 863-870
Deep Learning Model for Histologic Diagnosis of Dysplastic Barrett's Esophagus: Multisite Cohort External Validation
INTRODUCTION: The risk of progression to esophageal adenocarcinoma in Barrett's esophagus (BE) increases with advancing degrees of dysplasia. There is a critical need to improve the diagnosis of BE dysplasia, given substantial interobserver variability and overcalls of dysplasia during manual community pathologist reads. We aimed to externally validate a previously cross-validated BE dysplasia diagnosis deep learning model (BEDDLM) that predicts dysplasia grade on whole slide images (WSIs). METHODS: We digitized nondysplastic BE (NDBE), low-grade (LGD), and high-grade dysplasia (HGD) histology slides from 3 external academic centers. A consensus read by 2 expert study pathologists was used as the criterion standard. Slide stain characteristics were normalized using cycle-generative adversarial networks. WSIs were assessed by BEDDLM using an ensemble approach, combining a “You Only Look Once” model followed by a ResNet101 classifier model. RESULTS: We included 489 WSIs. Consensus histopathology revealed 232 NDBE, 117 LGD, and 140 HGD WSIs. The mean age (SD) was 66.9 (11.4) years; 413 (84.7%) were men. Using the BEDDLM ensemble model, sensitivity and specificity for NDBE were 73.3% (95% confidence interval [CI]: 67.09%–78.85%) and 93.4% (95% CI: 89.62%–96.10%); for LGD, 84.6% (95% CI: 76.78%–90.62%) and 80.6% (95% CI: 76.26%–84.54%); and for HGD, 80.7% (95% CI: 73.19%–86.89%) and 94.8% (95% CI: 91.97%–96.91%), respectively. The F1 score was 0.81, 0 0.69, and 0.83 for NDBE, LGD, and HGD, respectively. DISCUSSION: Our externally validated deep learning model demonstrates substantial accuracy for the diagnosis of BE dysplasia grade on WSIs.
- Open Access
- 2 August 2021
- Published by Taylor & Francis in Hepatic Oncology
- Vol. 8 (4) , HEP39
Evaluation of tumor mutational burden in small early hepatocellular carcinoma and progressed hepatocellular carcinoma
While researchers know that tumor mutational burden (TMB) is low in hepatocellular carcinoma (HCC), prior studies have not investigated TMB in cirrhosis, small early HCC and progressed HCC. HCC (n ...
- 2 August 2021
- Published by Taylor & Francis in Hepatic Oncology
- Vol. 8 (4) , HEP39
Open AccessEvaluation of tumor mutational burden in small early hepatocellular carcinoma and progressed hepatocellular carcinoma
While researchers know that tumor mutational burden (TMB) is low in hepatocellular carcinoma (HCC), prior studies have not investigated TMB in cirrhosis, small early HCC and progressed HCC. HCC (n ...
- 14 October 2020
- Published by Wiley in Histopathology
- Vol. 78 (3) , 453-458
Cell polarity (the ‘four lines’) distinguishes gastric dysplasia from epithelial changes in reactive gastropathy
Aims: Gastric dysplasia is a risk factor for synchronous and subsequent gastric carcinoma. Distinguishing gastric dysplasia from reactive changes is subject to interobserver disagreement and is a frequent reason for expert consultation. We previously used assessment of surface cell polarity (the ‘four lines’) as a key feature to decrease equivocal diagnoses in Barrett oesophagus. In the current study, we examined for the presence or absence of the four lines in gastric dysplasia and reactive gastropathy.Materials and methods: The study includes all (n = 91) in‐house biopsies with at least gastric dysplasia from the surgical pathology archives of two academic institutions during a 5‐year period from 2008 to 2012. A reactive gastropathy group (n = 60) was created for comparison.Results: The dysplasia/neoplasia group was comprised of 14 biopsies of gastric foveolar‐type dysplasia, 59 of intestinal‐type dysplasia, 14 with dysplasia in fundic gland polyps, three pyloric gland adenomas and one oxyntic gland adenoma. Loss of surface cell polarity was seen in all 88 dysplasia cases with evaluable surface epithelium. All 57 reactive gastropathy cases with evaluable surface epithelium showed intact surface cell polarity except in focal areas directly adjacent to erosions in 17 cases, where the thin wisp of residual surface mucin could not be appreciated on haematoxylin and eosin.Conclusion: Surface cell polarity (the four lines) was lost in all gastric dysplasia biopsies with evaluable surface epithelium and maintained in all biopsies of reactive gastropathy. Caution should be taken in using this feature adjacent to erosions in reactive gastropathy.
- 14 October 2020
- Published by Wiley in Histopathology
- Vol. 78 (3) , 453-458
Cell polarity (the ‘four lines’) distinguishes gastric dysplasia from epithelial changes in reactive gastropathy
Aims: Gastric dysplasia is a risk factor for synchronous and subsequent gastric carcinoma. Distinguishing gastric dysplasia from reactive changes is subject to interobserver disagreement and is a frequent reason for expert consultation. We previously used assessment of surface cell polarity (the ‘four lines’) as a key feature to decrease equivocal diagnoses in Barrett oesophagus. In the current study, we examined for the presence or absence of the four lines in gastric dysplasia and reactive gastropathy.Materials and methods: The study includes all (n = 91) in‐house biopsies with at least gastric dysplasia from the surgical pathology archives of two academic institutions during a 5‐year period from 2008 to 2012. A reactive gastropathy group (n = 60) was created for comparison.Results: The dysplasia/neoplasia group was comprised of 14 biopsies of gastric foveolar‐type dysplasia, 59 of intestinal‐type dysplasia, 14 with dysplasia in fundic gland polyps, three pyloric gland adenomas and one oxyntic gland adenoma. Loss of surface cell polarity was seen in all 88 dysplasia cases with evaluable surface epithelium. All 57 reactive gastropathy cases with evaluable surface epithelium showed intact surface cell polarity except in focal areas directly adjacent to erosions in 17 cases, where the thin wisp of residual surface mucin could not be appreciated on haematoxylin and eosin.Conclusion: Surface cell polarity (the four lines) was lost in all gastric dysplasia biopsies with evaluable surface epithelium and maintained in all biopsies of reactive gastropathy. Caution should be taken in using this feature adjacent to erosions in reactive gastropathy.
- 11 August 2020
- Published by Wiley in Histopathology
- Vol. 78 (3) , 392-400
Lymphogranuloma venereum (LGV) of the anorectum: evaluation of clinicopathological associations and the utility of a novel RNA in‐situ hybridisation stain
Aims: Recent studies from multiple global regions have reported a resurgence of lymphogranuloma venereum (LGV) proctitis, which is caused by Chlamydia trachomatis (CT). LGV proctitis is histologically indistinguishable from other forms of sexually transmitted proctitis and is difficult to differentiate from inflammatory bowel disease. While immunohistochemical stains are available for syphilis, there is no commonly available stain for the tissue identification of CT.Materials and methods: From 200 positive CT nucleic acid tests (NAT) from anorectal swabs, we identified 12 patients with biopsies collected from the distal colorectum or anus within 90 days of the positive NAT. We collected basic demographic information and tabulated clinical and histological findings. We examined the performance of a novel RNA in‐situ hybridisation (ISH) stain targeting CT 23s rRNA on these 12 cases and 10 controls from the anorectum.Results: All 12 patients were male; nine were HIV+, two had concurrent gonococcal infection, one had concurrent syphilis and one had cytomegalovirus co‐infection. The majority of biopsies (11 of 12) showed mild or moderate acute inflammation, had a prominent lymphoplasmacytic infiltrate (eight of 11) and lacked marked crypt distortion (10 of 10). The RNA ISH stain was positive in 10 of 12 cases (sensitivity 83%). One case showed equivocal staining. No controls showed definitive positive staining (specificity 100%). One had equivocal staining.Conclusion: Our series showed that anorectal LGV had similar histological findings to those of prior STI proctitis series predominantly comprised of syphilis. The novel RNA ISH stain was sensitive and specific and may show utility in differentiating types of STI proctitis.
- 11 August 2020
- Published by Wiley in Histopathology
- Vol. 78 (3) , 392-400
Lymphogranuloma venereum (LGV) of the anorectum: evaluation of clinicopathological associations and the utility of a novel RNA in‐situ hybridisation stain
Aims: Recent studies from multiple global regions have reported a resurgence of lymphogranuloma venereum (LGV) proctitis, which is caused by Chlamydia trachomatis (CT). LGV proctitis is histologically indistinguishable from other forms of sexually transmitted proctitis and is difficult to differentiate from inflammatory bowel disease. While immunohistochemical stains are available for syphilis, there is no commonly available stain for the tissue identification of CT.Materials and methods: From 200 positive CT nucleic acid tests (NAT) from anorectal swabs, we identified 12 patients with biopsies collected from the distal colorectum or anus within 90 days of the positive NAT. We collected basic demographic information and tabulated clinical and histological findings. We examined the performance of a novel RNA in‐situ hybridisation (ISH) stain targeting CT 23s rRNA on these 12 cases and 10 controls from the anorectum.Results: All 12 patients were male; nine were HIV+, two had concurrent gonococcal infection, one had concurrent syphilis and one had cytomegalovirus co‐infection. The majority of biopsies (11 of 12) showed mild or moderate acute inflammation, had a prominent lymphoplasmacytic infiltrate (eight of 11) and lacked marked crypt distortion (10 of 10). The RNA ISH stain was positive in 10 of 12 cases (sensitivity 83%). One case showed equivocal staining. No controls showed definitive positive staining (specificity 100%). One had equivocal staining.Conclusion: Our series showed that anorectal LGV had similar histological findings to those of prior STI proctitis series predominantly comprised of syphilis. The novel RNA ISH stain was sensitive and specific and may show utility in differentiating types of STI proctitis.
- 8 July 2020
- Published by Wolters Kluwer Health in Journal of Clinical Gastroenterology
- Vol. 55 (3) , 244-249
Quantitative Analysis of Intramucosal Mast Cells in Irritable Bowel Syndrome
Goal: We aimed to study the density of intramucosal mast cells in histologically normal colonic mucosa biopsied from patients with a clinical diagnosis of irritable bowel syndrome (IBS). Background: Mast cell activation has been thought to implicate in the pathogenesis of inflammatory bowel disease (IBD). Whether it serves a role in the pathogenesis of IBS remains controversial. Study: A total of 127 colonoscopic mucosal biopsies were immunohistochemically stained, including 51 IBS, 66 IBD, and 10 normal control samples. Intact mast cells were quantified in 3 high power fields (HPF) in areas showing the highest density. Results: CD117 was sensitive in detecting mast cells in colonic mucosa. The mast cell counts in all biopsies ranged from 2 to 60 per HPF (mean=17.5±7.2). The density of intramucosal mast cells were similar among IBS, IBD and normal control groups (P=0.6733). IBD in remission versus IBS (17.1±8.0 vs. 18.1±7.0; P=0.4804), Crohn disease versus ulcerative colitis (17.1±10.4 vs. 17.2±5.2; P=0.9463), IBS with diarrhea versus without diarrhea (19.5±6.3 vs. 16.8±6.9; P=0.1404). Forty biopsies (31.5%) showing ≥20 mast cells per HPF appeared to equally distribute among various disease groups (P=0.7283). Conclusions: There is no significant difference in the number of intramucosal mast cells between IBS and IBD that show normal colonic biopsies. In IBS patients, the number of intramucosal mast cell does not correlate with symptoms. The mast cell count (≥20/HPF) is not a reliable criterion for the diagnosis of IBS or for the distinction between patients with IBS and those with IBD in remission.
- 8 July 2020
- Published by Wolters Kluwer Health in Journal of Clinical Gastroenterology
- Vol. 55 (3) , 244-249
Quantitative Analysis of Intramucosal Mast Cells in Irritable Bowel Syndrome
Goal: We aimed to study the density of intramucosal mast cells in histologically normal colonic mucosa biopsied from patients with a clinical diagnosis of irritable bowel syndrome (IBS). Background: Mast cell activation has been thought to implicate in the pathogenesis of inflammatory bowel disease (IBD). Whether it serves a role in the pathogenesis of IBS remains controversial. Study: A total of 127 colonoscopic mucosal biopsies were immunohistochemically stained, including 51 IBS, 66 IBD, and 10 normal control samples. Intact mast cells were quantified in 3 high power fields (HPF) in areas showing the highest density. Results: CD117 was sensitive in detecting mast cells in colonic mucosa. The mast cell counts in all biopsies ranged from 2 to 60 per HPF (mean=17.5±7.2). The density of intramucosal mast cells were similar among IBS, IBD and normal control groups (P=0.6733). IBD in remission versus IBS (17.1±8.0 vs. 18.1±7.0; P=0.4804), Crohn disease versus ulcerative colitis (17.1±10.4 vs. 17.2±5.2; P=0.9463), IBS with diarrhea versus without diarrhea (19.5±6.3 vs. 16.8±6.9; P=0.1404). Forty biopsies (31.5%) showing ≥20 mast cells per HPF appeared to equally distribute among various disease groups (P=0.7283). Conclusions: There is no significant difference in the number of intramucosal mast cells between IBS and IBD that show normal colonic biopsies. In IBS patients, the number of intramucosal mast cell does not correlate with symptoms. The mast cell count (≥20/HPF) is not a reliable criterion for the diagnosis of IBS or for the distinction between patients with IBS and those with IBD in remission.
- 24 October 2019
- Published by SAGE Publications in International Journal of Surgical Pathology
- Vol. 28 (3) , 252-258
Low-Grade Appendiceal Mucinous Neoplasms: A Single Institution Experience of 64 Cases With Clinical Follow-up and Correlation With the Current (Eighth Edition) AJCC Staging
Background. In this single-institution study, we applied the current (eighth edition) American Joint Committee on Cancer pathologic staging criteria to 64 low-grade mucinous neoplasms of the appendix (LAMNs), examined their histopathologic features, and studied the patients’ clinical outcomes. Design. Sixty-four LAMNs, with a median follow-up of 52 months, were reviewed. Results. The distribution of pathologic stages was pTis (n = 39), pT3 (n = 1), pT4a (n = 5), pT4aM1a (n = 8), and pT4aM1b (n = 11). Recurrence was observed in only 2 patients (both with pT4aM1b disease), one of whom died of disease. All remaining patients were disease-free after a median clinical follow-up of 60 months. Conclusions. Our study confirms that pTis LAMNs have an excellent prognosis and suggests that pT4a and pT4aM1a LAMNs may also have a low risk of developing progressive disease.
- 24 October 2019
- Published by SAGE Publications in International Journal of Surgical Pathology
- Vol. 28 (3) , 252-258
Low-Grade Appendiceal Mucinous Neoplasms: A Single Institution Experience of 64 Cases With Clinical Follow-up and Correlation With the Current (Eighth Edition) AJCC Staging
Background. In this single-institution study, we applied the current (eighth edition) American Joint Committee on Cancer pathologic staging criteria to 64 low-grade mucinous neoplasms of the appendix (LAMNs), examined their histopathologic features, and studied the patients’ clinical outcomes. Design. Sixty-four LAMNs, with a median follow-up of 52 months, were reviewed. Results. The distribution of pathologic stages was pTis (n = 39), pT3 (n = 1), pT4a (n = 5), pT4aM1a (n = 8), and pT4aM1b (n = 11). Recurrence was observed in only 2 patients (both with pT4aM1b disease), one of whom died of disease. All remaining patients were disease-free after a median clinical follow-up of 60 months. Conclusions. Our study confirms that pTis LAMNs have an excellent prognosis and suggests that pT4a and pT4aM1a LAMNs may also have a low risk of developing progressive disease.
- 10 October 2018
- Published by Wiley in Histopathology
- Vol. 74 (3) , 415-423
Neuroendocrine proliferations in inflammatory bowel disease: differentiating neuroendocrine tumours from neuroendocrine cell micronests
Aims: A wide spectrum of well‐differentiated neuroendocrine proliferations (NEPs) are observed in inflammatory bowel disease (IBD), ranging from neuroendocrine tumours (NETs) to microscopic neuroendocrine cell clusters, best described as neuroendocrine cell micronests (NCMs). Finding NCMs in surveillance biopsies of IBD patients often poses a diagnostic conundrum. While such lesions may have been referred to as ‘microcarcinoids’ in the literature, it is unclear whether these represent early neoplasms. The study was undertaken to characterise NCMs and to differentiate NCMs from NETs.Methods and results: Institutional surgical pathology archives were searched to identify cases of NEPs in IBD patients. Clinicopathological features were examined. NCMs were defined as scattered, indistinct neuroendocrine cell clusters without confluent growth or new stroma formation, located in the lamina propria and muscularis mucosae. NETs were defined as discrete, mass‐forming lesions. Seventeen NEPs were identified, including eight NCMs and nine NETs. All NEPs were incidentally discovered. While NETs were commonly found in the rectum and appendix, NCMs were only noted in the rectosigmoid area. Unlike NETs, NCMs could not be measured as a discrete lesion as these clusters were non‐confluent and scattered. None of the patients with NCMs developed NETs after a mean follow‐up of 4.1 years (range = 0.5–21.0 years). None of the NETs showed NCMs in the background mucosa.Conclusions: NCMs have distinct pathological features, are not associated with NETs in IBD patients and should not be misinterpreted as ‘microcarcinoids’. Identification of NCMs in surveillance biopsies may not require further clinical work‐up or invasive procedures usually performed for NETs.
- 10 October 2018
- Published by Wiley in Histopathology
- Vol. 74 (3) , 415-423
Neuroendocrine proliferations in inflammatory bowel disease: differentiating neuroendocrine tumours from neuroendocrine cell micronests
Aims: A wide spectrum of well‐differentiated neuroendocrine proliferations (NEPs) are observed in inflammatory bowel disease (IBD), ranging from neuroendocrine tumours (NETs) to microscopic neuroendocrine cell clusters, best described as neuroendocrine cell micronests (NCMs). Finding NCMs in surveillance biopsies of IBD patients often poses a diagnostic conundrum. While such lesions may have been referred to as ‘microcarcinoids’ in the literature, it is unclear whether these represent early neoplasms. The study was undertaken to characterise NCMs and to differentiate NCMs from NETs.Methods and results: Institutional surgical pathology archives were searched to identify cases of NEPs in IBD patients. Clinicopathological features were examined. NCMs were defined as scattered, indistinct neuroendocrine cell clusters without confluent growth or new stroma formation, located in the lamina propria and muscularis mucosae. NETs were defined as discrete, mass‐forming lesions. Seventeen NEPs were identified, including eight NCMs and nine NETs. All NEPs were incidentally discovered. While NETs were commonly found in the rectum and appendix, NCMs were only noted in the rectosigmoid area. Unlike NETs, NCMs could not be measured as a discrete lesion as these clusters were non‐confluent and scattered. None of the patients with NCMs developed NETs after a mean follow‐up of 4.1 years (range = 0.5–21.0 years). None of the NETs showed NCMs in the background mucosa.Conclusions: NCMs have distinct pathological features, are not associated with NETs in IBD patients and should not be misinterpreted as ‘microcarcinoids’. Identification of NCMs in surveillance biopsies may not require further clinical work‐up or invasive procedures usually performed for NETs.