Advertisement

Response

      We are honored by the interest of Njei et al
      • Njei B.
      • McCarty T.R.
      • Navaneethan U.
      Artificial intelligence-assisted cholangioscopy for automatic detection of malignant biliary strictures.
      in our proof-of-concept work.
      • Saraiva M.M.
      • Ribeiro T.
      • Ferreira J.P.S.
      • et al.
      Artificial intelligence for automatic diagnosis of biliary stricture malignancy status in single-operator cholangioscopy: a pilot study.
      We share their view of a potential role of artificial intelligence (AI) for significantly enhancing the diagnostic yield of digital cholangioscopy for malignant biliary strictures. In our view, the application of AI to cholangioscopy will, at this stage, work alongside current criterion standard techniques, as a complement rather than disrupting the current standard of care. Therefore, although we agree with Njei et al
      • Njei B.
      • McCarty T.R.
      • Navaneethan U.
      Artificial intelligence-assisted cholangioscopy for automatic detection of malignant biliary strictures.
      that AI-assisted cholangioscopy should ultimately achieve diagnosis without more invasive techniques, we believe that in the near future AI will be applied to increase the yield of current techniques, particularly cholangioscopy-guided biopsies, which remain the current criterion standard. This could be potentially accomplished through accurate detection of morphologic features intricately associated with malignant strictures, as is the case with tumor vessels.
      • Robles-Medranda C.
      • Oleas R.
      • Sánchez-Carriel M.
      • et al.
      Vascularity can distinguish neoplastic from non-neoplastic bile duct lesions during digital single-operator cholangioscopy.
      ,
      • Pereira P.
      • Mascarenhas M.
      • Ribeiro T.
      • et al.
      Automatic detection of tumor vessels in indeterminate biliary strictures in digital single-operator cholangioscopy.
      The potential of convolutional neural networks for the analysis of endoscopic images is vast. Notwithstanding, we subscribe that these promising results should be analyzed with consideration of current methodologic limitations and knowledge gaps. We believe that the path toward clinically applicable AI-assisted cholangioscopy has only just begun. In fact, the readiness level of AI technologies for cholangioscopy remains at early stages (Fig. 1), because most studies assess their performance in controlled settings.
      • Saraiva M.M.
      • Ribeiro T.
      • Ferreira J.P.S.
      • et al.
      Artificial intelligence for automatic diagnosis of biliary stricture malignancy status in single-operator cholangioscopy: a pilot study.
      ,
      • Ghandour B.
      • Hsieh H.-W.
      • Akshintala V.
      • et al.
      Machine learning for classification of indeterminate biliary strictures during cholangioscopy.
      Figure thumbnail gr1
      Figure 1Description of technology readiness levels (TRLs). Adapted from Martínez-Plumed F, Gómez E, Hernández-Orallo J. Futures of artificial intelligence through technology readiness levels. Telemat Inform 2021;58:101525.
      Trusting the clinical output of an AI algorithm will require moving from opaque black-box AI models toward explainable AI algorithms, in which users (preferably both healthcare practitioners and patients) understand AI recommendations. Attempts to improve explainability, for example through the application of heat maps, are ongoing.
      • Ghassemi M.
      • Oakden-Rayner L.
      • Beam A.L.
      The false hope of current approaches to explainable artificial intelligence in health care.
      ,
      • Mascarenhas M.
      • Ribeiro T.
      • Afonso J.
      • et al.
      Deep learning and colon capsule endoscopy: automatic detection of blood and colonic mucosal lesions using a convolutional neural network.
      The development of this technology will require several years, probably decades, and we should not expect a linear path. A nadir after this initial zenith of enthusiasm is predictable, but we expect that work on this field will ultimately lead to a heightened plateau of productivity.

      Disclosure

      All authors disclosed no financial relationships.

      References

        • Njei B.
        • McCarty T.R.
        • Navaneethan U.
        Artificial intelligence-assisted cholangioscopy for automatic detection of malignant biliary strictures.
        Gastrointest Endosc. 2022; 96: 1092-1093
        • Saraiva M.M.
        • Ribeiro T.
        • Ferreira J.P.S.
        • et al.
        Artificial intelligence for automatic diagnosis of biliary stricture malignancy status in single-operator cholangioscopy: a pilot study.
        Gastrointest Endosc. 2022; 95: 339-348
        • Robles-Medranda C.
        • Oleas R.
        • Sánchez-Carriel M.
        • et al.
        Vascularity can distinguish neoplastic from non-neoplastic bile duct lesions during digital single-operator cholangioscopy.
        Gastrointest Endosc. 2021; 93: 935-941
        • Pereira P.
        • Mascarenhas M.
        • Ribeiro T.
        • et al.
        Automatic detection of tumor vessels in indeterminate biliary strictures in digital single-operator cholangioscopy.
        Endosc Int Open. 2022; 10: E262-E268
        • Ghandour B.
        • Hsieh H.-W.
        • Akshintala V.
        • et al.
        Machine learning for classification of indeterminate biliary strictures during cholangioscopy.
        Am J Gastroenterol. 2021; : 116
        • Ghassemi M.
        • Oakden-Rayner L.
        • Beam A.L.
        The false hope of current approaches to explainable artificial intelligence in health care.
        Lancet Digital Health. 2021; 3: e745-e750
        • Mascarenhas M.
        • Ribeiro T.
        • Afonso J.
        • et al.
        Deep learning and colon capsule endoscopy: automatic detection of blood and colonic mucosal lesions using a convolutional neural network.
        Endosc Int Open. 2022; 10: E171-E177

      Linked Article

      • Artificial intelligence–assisted cholangioscopy for automatic detection of malignant biliary strictures
        Gastrointestinal EndoscopyVol. 96Issue 6
        • Preview
          We read with great interest the article by Mascarenhas Saraiva et al1 regarding optimal diagnosis of malignant biliary strictures by using an artificial intelligence (AI) algorithm. We concur with their findings and agree that the introduction of AI algorithms such as convolutional neural network (CNN) imaging may significantly increase our diagnostic repertoire in dealing with patients who have suspected bile duct cancers. In addition, owing to the high accuracy of AI-assisted cholangioscopy, patients with highly suggestive lesions on cholangioscopy who are suitable for surgery may be able to proceed to early surgery before cancer progresses.
        • Full-Text
        • PDF