Deep Learning-Fuzzy Inference (DeLFI): an interpretable model for detecting indirect immunofluorescence patterns associated with nasopharyngeal cancer


      The detection of serum antibodies to Epstein-Barr Virus proteins by immunofluorescence assay (IFA) is considered the gold standard for nasopharyngeal cancer (NPC) screening in high-risk populations. Given the high survival rates of early asymptomatic patients compared to the poor prognosis for late-stage NPC, IFA has tremendous life-saving potential for general population screening. However, IFA requires visual interpretation of cellular staining patterns by trained pathology staff, making it labor-intensive and hence unscalable. This study demonstrates that an automated Fuzzy Inference (FI) system achieved high agreement (κ=0.82) with a human IFA expert in identifying cellular patterns associated with NPC. Integrating a deep learning module into FI further improved FI’s performance (κ=0.90) and reduced the number of uncertain cases which required manual evaluation. The resulting hybrid model, termed deep learning-fuzzy inference (DeLFI), was then evaluated with a separate testing set of clinical samples. In this clinical validation, DeLFI outperformed human evaluation on the Area Under the Curve (0.926 vs 0.821) and closely matched human performance on Youden’s J index (0.81 vs 0.80). We conclude that DeLFI’s combination of deep learning with fuzzy inference has the potential to improve the scalability and accuracy of NPC detection.
      To read this article in full you will need to make a payment

      Purchase one-time access:

      Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online access
      One-time access price info
      • For academic or personal research use, select 'Academic and Personal'
      • For corporate R&D use, select 'Corporate R&D Professionals'


      Subscribe to The American Journal of Pathology
      Already a print subscriber? Claim online access
      Already an online subscriber? Sign in
      Institutional Access: Sign in to ScienceDirect


      1. Screening Test Review Committee: Report of the Screening Test Review Committee. Academy of Medicine, Singapore, 2019 Mar.

      2. Chan SH, Goh CHK, Wee JTS, Tam CH, Tan LKS, Heng D, Lee MT: MOH Clinical Practice Guidelines 1/2010. Ministry of Health, Singapore, 2010 Feb.

        • Middeldorp J.M.
        Epstein-Barr Virus-Specific Humoral Immune Responses in Health and Disease.
        Curr Top Microbiol Immunol. 2015; 391: 289-323
        • Tay J.K.
        • Chan S.H.
        • Lim C.M.
        • Siow C.H.
        • Goh H.L.
        • Loh K.S.
        The Role of Epstein-Barr Virus DNA Load and Serology as Screening Tools for Nasopharyngeal Carcinoma.
        Otolaryngol Head Neck Surg. 2016; 155: 274-280
        • Tay J.K.
        • Siow C.H.
        • Goh H.L.
        • Lim C.M.
        • Hsu P.P.
        • Chan S.H.
        • Loh K.S.
        A comparison of EBV serology and serum cell-free DNA as screening tools for nasopharyngeal cancer: Results of the Singapore NPC screening cohort.
        Int J Cancer. 2020; 146: 2923-2931
        • Chan S.H.
        • Soo M.Y.
        • Gan Y.Y.
        • Fones-Tan A.
        • Sim P.S.
        • Kaur A.
        • Chew C.T.
        Epstein Barr virus (EBV) antibodies in the diagnosis of NPC–comparison between IFA and two commercial ELISA kits.
        Singapore Med J. 1998; 39: 263-265
      3. Coghill AE, Hsu W-L, Pfeiffer RM, Juwana H, Yu KJ, Lou P-J, Wang C-P, Chen J-Y, Chen C-J, Middeldorp JM, Hildesheim A: Epstein-Barr virus serology as a potential screening marker for nasopharyngeal carcinoma among high-risk individuals from multiplex families in Taiwan. Cancer Epidemiol Biomarkers Prev 2014, 23:1213–1219.Pubmed Partial Author

        • Hutajulu S.H.
        • Fachiroh J.
        • Argy G.
        • Indrasari S.R.
        • Indrawati L.P.L.
        • Paramita D.K.
        • Jati T.B.R.
        • Middeldorp J.M.
        Seroprevalence of IgA anti Epstein-Barr virus is high among family members of nasopharyngeal cancer patients and individuals presenting with chronic complaints in head and neck area.
        PLoS One. 2017; 12: e0180683
        • Paramita D.K.
        • Fachiroh J.
        • Haryana S.M.
        • Middeldorp J.M.
        Evaluation of commercial EBV RecombLine assay for diagnosis of nasopharyngeal carcinoma.
        J Clin Virol. 2008; 42: 343-352
        • Chan K.C.A.
        • Woo J.K.S.
        • King A.
        • Zee B.C.Y.
        • Lam W.K.J.
        • Chan S.L.
        • Chu S.W.I.
        • Mak C.
        • Tse I.O.L.
        • Leung S.Y.M.
        • Chan G.
        • Hui E.P.
        • Ma B.B.Y.
        • Chiu R.W.K.
        • Leung S.-F.
        • van Hasselt A.C.
        • Chan A.T.C.
        • Lo Y.M.D.
        Analysis of Plasma Epstein-Barr Virus DNA to Screen for Nasopharyngeal Cancer. N Engl J Med.
        Massachusetts Medical Society. 2017; 377: 513-522
        • Nicholls J.M.
        • Lee V.H.-F.
        • Chan S.-K.
        • Tsang K.-C.
        • Choi C.-W.
        • Kwong D.L.-W.
        • Lam K.-O.
        • Chan S.-Y.
        • Tong C.-C.
        • So T.-H.
        • Leung T.-W.
        • Luk M.-Y.
        • Khong P.-L.
        • Lee A.W.-M.
        Negative plasma Epstein-Barr virus DNA nasopharyngeal carcinoma in an endemic region and its influence on liquid biopsy screening programmes.
        Br J Cancer. 2019; 121: 690-698
        • Ricchiuti V.
        • Adams J.
        • Hardy D.J.
        • Katayev A.
        • Fleming J.K.
        Automated Processing and Evaluation of Anti-Nuclear Antibody Indirect Immunofluorescence Testing.
        Front Immunol. 2018; 9: 927
        • Horsfall A.C.
        • Venables P.J.
        • Mumford P.A.
        • Maini R.N.
        Interpretation of the Raji cell assay in sera containing anti-nuclear antibodies and immune complexes.
        Clin Exp Immunol. 1981; 44: 405-415
        • Goh S.M.P.
        • Swaminathan M.
        • Lai J.U.-M.
        • Anwar A.
        • Chan S.H.
        • Cheong I.
        Increasing the accuracy and scalability of the Immunofluorescence Assay for Epstein Barr Virus by inferring continuous titers from a single sample dilution.
        J Immunol Methods. 2017; 440: 35-40
        • Landis J.R.
        • Koch G.G.
        The measurement of observer agreement for categorical data.
        Biometrics. 1977; 33: 159-174
      4. Bonanno D, Nock K, Smith L, Elmore P, Petry F: An approach to explainable deep learning using fuzzy inference. Next-Generation Analyst V, International Society for Optics and Photonics, 2017, p. 102070D.