Report
View in EDS

Fully transformer-based biomarker prediction from colorectal cancer histology: a large-scale multicentric study

Bibliographic Details
Title: Fully transformer-based biomarker prediction from colorectal cancer histology: a large-scale multicentric study
Authors: Wagner, Sophia J., Reisenbüchler, Daniel, West, Nicholas P., Niehues, Jan Moritz, Veldhuizen, Gregory Patrick, Quirke, Philip, Grabsch, Heike I., Brandt, Piet A. van den, Hutchins, Gordon G. A., Richman, Susan D., Yuan, Tanwei, Langer, Rupert, Jenniskens, Josien Christina Anna, Offermans, Kelly, Mueller, Wolfram, Gray, Richard, Gruber, Stephen B., Greenson, Joel K., Rennert, Gad, Bonner, Joseph D., Schmolze, Daniel, James, Jacqueline A., Loughrey, Maurice B., Salto-Tellez, Manuel, Brenner, Hermann, Hoffmeister, Michael, Truhn, Daniel, Schnabel, Julia A., Boxberg, Melanie, Peng, Tingying, Kather, Jakob Nikolas
Publication Year: 2023
Collection: Computer Science
Subject Terms: Computer Science - Computer Vision and Pattern Recognition
More Details: Background: Deep learning (DL) can extract predictive and prognostic biomarkers from routine pathology slides in colorectal cancer. For example, a DL test for the diagnosis of microsatellite instability (MSI) in CRC has been approved in 2022. Current approaches rely on convolutional neural networks (CNNs). Transformer networks are outperforming CNNs and are replacing them in many applications, but have not been used for biomarker prediction in cancer at a large scale. In addition, most DL approaches have been trained on small patient cohorts, which limits their clinical utility. Methods: In this study, we developed a new fully transformer-based pipeline for end-to-end biomarker prediction from pathology slides. We combine a pre-trained transformer encoder and a transformer network for patch aggregation, capable of yielding single and multi-target prediction at patient level. We train our pipeline on over 9,000 patients from 10 colorectal cancer cohorts. Results: A fully transformer-based approach massively improves the performance, generalizability, data efficiency, and interpretability as compared with current state-of-the-art algorithms. After training on a large multicenter cohort, we achieve a sensitivity of 0.97 with a negative predictive value of 0.99 for MSI prediction on surgical resection specimens. We demonstrate for the first time that resection specimen-only training reaches clinical-grade performance on endoscopic biopsy tissue, solving a long-standing diagnostic problem. Interpretation: A fully transformer-based end-to-end pipeline trained on thousands of pathology slides yields clinical-grade performance for biomarker prediction on surgical resections and biopsies. Our new methods are freely available under an open source license.
Comment: Updated Figure 2 and Table A.5
Document Type: Working Paper
Access URL: http://arxiv.org/abs/2301.09617
Accession Number: edsarx.2301.09617
Database: arXiv
FullText Text:
  Availability: 0
CustomLinks:
  – Url: http://arxiv.org/abs/2301.09617
    Name: EDS - Arxiv
    Category: fullText
    Text: View this record from Arxiv
    MouseOverText: View this record from Arxiv
  – Url: https://resolver.ebsco.com/c/xy5jbn/result?sid=EBSCO:edsarx&genre=article&issn=&ISBN=&volume=&issue=&date=20230123&spage=&pages=&title=Fully transformer-based biomarker prediction from colorectal cancer histology: a large-scale multicentric study&atitle=Fully%20transformer-based%20biomarker%20prediction%20from%20colorectal%20cancer%20histology%3A%20a%20large-scale%20multicentric%20study&aulast=Wagner%2C%20Sophia%20J.&id=DOI:
    Name: Full Text Finder (for New FTF UI) (s8985755)
    Category: fullText
    Text: Find It @ SCU Libraries
    MouseOverText: Find It @ SCU Libraries
Header DbId: edsarx
DbLabel: arXiv
An: edsarx.2301.09617
RelevancyScore: 1045
AccessLevel: 3
PubType: Report
PubTypeId: report
PreciseRelevancyScore: 1044.68005371094
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Fully transformer-based biomarker prediction from colorectal cancer histology: a large-scale multicentric study
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Wagner%2C+Sophia+J%2E%22">Wagner, Sophia J.</searchLink><br /><searchLink fieldCode="AR" term="%22Reisenbüchler%2C+Daniel%22">Reisenbüchler, Daniel</searchLink><br /><searchLink fieldCode="AR" term="%22West%2C+Nicholas+P%2E%22">West, Nicholas P.</searchLink><br /><searchLink fieldCode="AR" term="%22Niehues%2C+Jan+Moritz%22">Niehues, Jan Moritz</searchLink><br /><searchLink fieldCode="AR" term="%22Veldhuizen%2C+Gregory+Patrick%22">Veldhuizen, Gregory Patrick</searchLink><br /><searchLink fieldCode="AR" term="%22Quirke%2C+Philip%22">Quirke, Philip</searchLink><br /><searchLink fieldCode="AR" term="%22Grabsch%2C+Heike+I%2E%22">Grabsch, Heike I.</searchLink><br /><searchLink fieldCode="AR" term="%22Brandt%2C+Piet+A%2E+van+den%22">Brandt, Piet A. van den</searchLink><br /><searchLink fieldCode="AR" term="%22Hutchins%2C+Gordon+G%2E+A%2E%22">Hutchins, Gordon G. A.</searchLink><br /><searchLink fieldCode="AR" term="%22Richman%2C+Susan+D%2E%22">Richman, Susan D.</searchLink><br /><searchLink fieldCode="AR" term="%22Yuan%2C+Tanwei%22">Yuan, Tanwei</searchLink><br /><searchLink fieldCode="AR" term="%22Langer%2C+Rupert%22">Langer, Rupert</searchLink><br /><searchLink fieldCode="AR" term="%22Jenniskens%2C+Josien+Christina+Anna%22">Jenniskens, Josien Christina Anna</searchLink><br /><searchLink fieldCode="AR" term="%22Offermans%2C+Kelly%22">Offermans, Kelly</searchLink><br /><searchLink fieldCode="AR" term="%22Mueller%2C+Wolfram%22">Mueller, Wolfram</searchLink><br /><searchLink fieldCode="AR" term="%22Gray%2C+Richard%22">Gray, Richard</searchLink><br /><searchLink fieldCode="AR" term="%22Gruber%2C+Stephen+B%2E%22">Gruber, Stephen B.</searchLink><br /><searchLink fieldCode="AR" term="%22Greenson%2C+Joel+K%2E%22">Greenson, Joel K.</searchLink><br /><searchLink fieldCode="AR" term="%22Rennert%2C+Gad%22">Rennert, Gad</searchLink><br /><searchLink fieldCode="AR" term="%22Bonner%2C+Joseph+D%2E%22">Bonner, Joseph D.</searchLink><br /><searchLink fieldCode="AR" term="%22Schmolze%2C+Daniel%22">Schmolze, Daniel</searchLink><br /><searchLink fieldCode="AR" term="%22James%2C+Jacqueline+A%2E%22">James, Jacqueline A.</searchLink><br /><searchLink fieldCode="AR" term="%22Loughrey%2C+Maurice+B%2E%22">Loughrey, Maurice B.</searchLink><br /><searchLink fieldCode="AR" term="%22Salto-Tellez%2C+Manuel%22">Salto-Tellez, Manuel</searchLink><br /><searchLink fieldCode="AR" term="%22Brenner%2C+Hermann%22">Brenner, Hermann</searchLink><br /><searchLink fieldCode="AR" term="%22Hoffmeister%2C+Michael%22">Hoffmeister, Michael</searchLink><br /><searchLink fieldCode="AR" term="%22Truhn%2C+Daniel%22">Truhn, Daniel</searchLink><br /><searchLink fieldCode="AR" term="%22Schnabel%2C+Julia+A%2E%22">Schnabel, Julia A.</searchLink><br /><searchLink fieldCode="AR" term="%22Boxberg%2C+Melanie%22">Boxberg, Melanie</searchLink><br /><searchLink fieldCode="AR" term="%22Peng%2C+Tingying%22">Peng, Tingying</searchLink><br /><searchLink fieldCode="AR" term="%22Kather%2C+Jakob+Nikolas%22">Kather, Jakob Nikolas</searchLink>
– Name: DatePubCY
  Label: Publication Year
  Group: Date
  Data: 2023
– Name: Subset
  Label: Collection
  Group: HoldingsInfo
  Data: Computer Science
– Name: Subject
  Label: Subject Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Computer+Science+-+Computer+Vision+and+Pattern+Recognition%22">Computer Science - Computer Vision and Pattern Recognition</searchLink>
– Name: Abstract
  Label: Description
  Group: Ab
  Data: Background: Deep learning (DL) can extract predictive and prognostic biomarkers from routine pathology slides in colorectal cancer. For example, a DL test for the diagnosis of microsatellite instability (MSI) in CRC has been approved in 2022. Current approaches rely on convolutional neural networks (CNNs). Transformer networks are outperforming CNNs and are replacing them in many applications, but have not been used for biomarker prediction in cancer at a large scale. In addition, most DL approaches have been trained on small patient cohorts, which limits their clinical utility. Methods: In this study, we developed a new fully transformer-based pipeline for end-to-end biomarker prediction from pathology slides. We combine a pre-trained transformer encoder and a transformer network for patch aggregation, capable of yielding single and multi-target prediction at patient level. We train our pipeline on over 9,000 patients from 10 colorectal cancer cohorts. Results: A fully transformer-based approach massively improves the performance, generalizability, data efficiency, and interpretability as compared with current state-of-the-art algorithms. After training on a large multicenter cohort, we achieve a sensitivity of 0.97 with a negative predictive value of 0.99 for MSI prediction on surgical resection specimens. We demonstrate for the first time that resection specimen-only training reaches clinical-grade performance on endoscopic biopsy tissue, solving a long-standing diagnostic problem. Interpretation: A fully transformer-based end-to-end pipeline trained on thousands of pathology slides yields clinical-grade performance for biomarker prediction on surgical resections and biopsies. Our new methods are freely available under an open source license.<br />Comment: Updated Figure 2 and Table A.5
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Working Paper
– Name: URL
  Label: Access URL
  Group: URL
  Data: <link linkTarget="URL" linkTerm="http://arxiv.org/abs/2301.09617" linkWindow="_blank">http://arxiv.org/abs/2301.09617</link>
– Name: AN
  Label: Accession Number
  Group: ID
  Data: edsarx.2301.09617
PLink https://login.libproxy.scu.edu/login?url=https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&scope=site&db=edsarx&AN=edsarx.2301.09617
RecordInfo BibRecord:
  BibEntity:
    Subjects:
      – SubjectFull: Computer Science - Computer Vision and Pattern Recognition
        Type: general
    Titles:
      – TitleFull: Fully transformer-based biomarker prediction from colorectal cancer histology: a large-scale multicentric study
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Wagner, Sophia J.
      – PersonEntity:
          Name:
            NameFull: Reisenbüchler, Daniel
      – PersonEntity:
          Name:
            NameFull: West, Nicholas P.
      – PersonEntity:
          Name:
            NameFull: Niehues, Jan Moritz
      – PersonEntity:
          Name:
            NameFull: Veldhuizen, Gregory Patrick
      – PersonEntity:
          Name:
            NameFull: Quirke, Philip
      – PersonEntity:
          Name:
            NameFull: Grabsch, Heike I.
      – PersonEntity:
          Name:
            NameFull: Brandt, Piet A. van den
      – PersonEntity:
          Name:
            NameFull: Hutchins, Gordon G. A.
      – PersonEntity:
          Name:
            NameFull: Richman, Susan D.
      – PersonEntity:
          Name:
            NameFull: Yuan, Tanwei
      – PersonEntity:
          Name:
            NameFull: Langer, Rupert
      – PersonEntity:
          Name:
            NameFull: Jenniskens, Josien Christina Anna
      – PersonEntity:
          Name:
            NameFull: Offermans, Kelly
      – PersonEntity:
          Name:
            NameFull: Mueller, Wolfram
      – PersonEntity:
          Name:
            NameFull: Gray, Richard
      – PersonEntity:
          Name:
            NameFull: Gruber, Stephen B.
      – PersonEntity:
          Name:
            NameFull: Greenson, Joel K.
      – PersonEntity:
          Name:
            NameFull: Rennert, Gad
      – PersonEntity:
          Name:
            NameFull: Bonner, Joseph D.
      – PersonEntity:
          Name:
            NameFull: Schmolze, Daniel
      – PersonEntity:
          Name:
            NameFull: James, Jacqueline A.
      – PersonEntity:
          Name:
            NameFull: Loughrey, Maurice B.
      – PersonEntity:
          Name:
            NameFull: Salto-Tellez, Manuel
      – PersonEntity:
          Name:
            NameFull: Brenner, Hermann
      – PersonEntity:
          Name:
            NameFull: Hoffmeister, Michael
      – PersonEntity:
          Name:
            NameFull: Truhn, Daniel
      – PersonEntity:
          Name:
            NameFull: Schnabel, Julia A.
      – PersonEntity:
          Name:
            NameFull: Boxberg, Melanie
      – PersonEntity:
          Name:
            NameFull: Peng, Tingying
      – PersonEntity:
          Name:
            NameFull: Kather, Jakob Nikolas
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 23
              M: 01
              Type: published
              Y: 2023
ResultId 1