Academic Journal
Προβολή σε EDS

From black-box to clinical reasoning: CARE-MD a novel framework for Self-eXplainable skin lesions.

Λεπτομέρειες βιβλιογραφικής εγγραφής
Τίτλος: From black-box to clinical reasoning: CARE-MD a novel framework for Self-eXplainable skin lesions.
Συγγραφείς: Kandpal KC; Department of Computer Science and Engineering, Harcourt Butler Technical University, Kanpur, India. Electronic address: kailashkandpal@vsef.org., Verma P; Department of Computer Science and Engineering, Harcourt Butler Technical University, Kanpur, India. Electronic address: pverma@hbtu.ac.in.
Πηγή: Computer methods and programs in biomedicine [Comput Methods Programs Biomed] 2026 Oct; Vol. 285, pp. 109478. Date of Electronic Publication: 2026 Jun 05.
Τύπος έκδοσης: Journal Article
Γλώσσα: English
Στοιχεία περιοδικού: Publisher: Elsevier Scientific Publishers Country of Publication: Ireland NLM ID: 8506513 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1872-7565 (Electronic) Linking ISSN: 01692607 NLM ISO Abbreviation: Comput Methods Programs Biomed Subsets: MEDLINE
Imprint Name(s): Publication: Limerick : Elsevier Scientific Publishers
Original Publication: Amsterdam : Elsevier Science Publishers, c1984-
Ιατρικοί όροι (MeSH): Skin Diseases*/diagnosis , Skin Diseases*/diagnostic imaging , Deep Learning* , Clinical Reasoning*, Humans ; Algorithms ; Sensitivity and Specificity ; Reproducibility of Results
Περίληψη: Background and Objective: Deep learning models have achieved remarkable diagnostic accuracy in medical imaging, yet their lack of interpretability limits clinical trust and deployment. This study presents CARE-MD (Clinical Algorithm for Reasoning-Enhanced Medical Diagnosis), a self-explainable framework that mirrors the clinical reasoning process to provide transparent, concept-driven, and case-referable diagnosis of skin lesions.
Methods: CARE-MD follows a four-phase reasoning structure: Observation, Interpretation, Reference, and Validation. In Phase 1, attention-guided localization isolates the lesion region using a modified U-Net backbone. Phase 2 employs a Concept Bottleneck Model (CBM) model to predict human-interpretable dermatological concepts. Phase 3 introduces prototype-based reasoning to compare latent features with previously learned prototypes for case-level interpretability. Phase 4 validates explanations through consistency analysis between attention maps, concept activations, and clinician-annotated lesion regions. The framework was evaluated on two publicly available datasets-ISIC 2018 and HAM10000-using metrics such as accuracy, precision, sensitivity, specificity, Dice coefficient, and explanation consistency score.
Results: CARE-MD achieved 87.2% accuracy, 86.8% precision, 85.6% sensitivity, and 88.1% specificity on the ISIC 2018 dataset. Cross-dataset testing on HAM10000 showed stable performance with minimal decline across all metrics, indicating strong generalization capability. The attention-guided module improved Dice score by 3% compared with baseline Attention U-Net, while the prototype referencing module enhanced explanation consistency by 6%. Qualitative results confirmed that CARE-MD produces coherent alignment between salient lesion regions, predicted dermatological concepts, and prototype-based references.
Conclusions: CARE-MD provides a structured approach toward aligning deep learning predictions with clinical reasoning by embedding interpretability into each diagnostic phase. While the framework shows improved interpretability and competitive performance relative to the evaluated baselines, it introduces additional architectural complexity and relies on concept-level supervision. These findings suggest that CARE-MD represents a promising direction for self-explainable medical diagnosis, while also highlighting inherent trade-offs between interpretability, model complexity, and annotation requirements. A notable strength of CARE-MD is its unified integration of multiple interpretability mechanisms within a clinically motivated reasoning structure.
(Copyright © 2026 Elsevier B.V. All rights reserved.)
Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Contributed Indexing: Keywords: Concept Bottleneck Model (CBM); Interpretability; Medical imaging; Prototype-based reasoning; Self-explainable AI; Skin lesion classification
Entry Date(s): Date Created: 20260611 Date Completed: 20260806 Latest Revision: 20260806
Update Code: 20260807
DOI: 10.1016/j.cmpb.2026.109478
PMID: 42275844
Βάση Δεδομένων: MEDLINE
FullText Text:
  Availability: 0
CustomLinks:
  – Url: https://www.doi.org/10.1016/j.cmpb.2026.109478?
    Name: ScienceDirect (all content) (s7799221)
    Category: fullText
    Text: View record from ScienceDirect
    MouseOverText: View record from ScienceDirect
  – Url: https://resolver.ebsco.com/c/fiv2js/result?sid=EBSCO:cmedm&genre=article&issn=18727565&ISBN=&volume=285&issue=&date=20261001&spage=109478&pages=109478&title=Computer methods and programs in biomedicine&atitle=From%20black-box%20to%20clinical%20reasoning%3A%20CARE-MD%20a%20novel%20framework%20for%20Self-eXplainable%20skin%20lesions.&aulast=Kandpal%20KC&id=DOI:10.1016/j.cmpb.2026.109478
    Name: Full Text Finder (for New FTF UI) (ns324271)
    Category: fullText
    Text: Full Text Finder
    MouseOverText: Full Text Finder
Header DbId: cmedm
DbLabel: MEDLINE
An: 42275844
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: From black-box to clinical reasoning: CARE-MD a novel framework for Self-eXplainable skin lesions.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AU" term="%22Kandpal+KC%22">Kandpal KC</searchLink>; Department of Computer Science and Engineering, Harcourt Butler Technical University, Kanpur, India. Electronic address: kailashkandpal@vsef.org.<br /><searchLink fieldCode="AU" term="%22Verma+P%22">Verma P</searchLink>; Department of Computer Science and Engineering, Harcourt Butler Technical University, Kanpur, India. Electronic address: pverma@hbtu.ac.in.
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%228506513%22">Computer methods and programs in biomedicine</searchLink> [Comput Methods Programs Biomed] 2026 Oct; Vol. 285, pp. 109478. <i>Date of Electronic Publication: </i>2026 Jun 05.
– Name: TypePub
  Label: Publication Type
  Group: TypPub
  Data: Journal Article
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: TitleSource
  Label: Journal Info
  Group: Src
  Data: <i>Publisher: </i><searchLink fieldCode="PB" term="%22Elsevier+Scientific+Publishers%22">Elsevier Scientific Publishers </searchLink><i>Country of Publication: </i>Ireland <i>NLM ID: </i>8506513 <i>Publication Model: </i>Print-Electronic <i>Cited Medium: </i>Internet <i>ISSN: </i>1872-7565 (Electronic) <i>Linking ISSN: </i><searchLink fieldCode="IS" term="%2201692607%22">01692607 </searchLink><i>NLM ISO Abbreviation: </i>Comput Methods Programs Biomed <i>Subsets: </i>MEDLINE
– Name: PublisherInfo
  Label: Imprint Name(s)
  Group: PubInfo
  Data: <i>Publication</i>: Limerick : Elsevier Scientific Publishers<br /><i>Original Publication</i>: Amsterdam : Elsevier Science Publishers, c1984-
– Name: SubjectMESH
  Label: MeSH Terms
  Group: Su
  Data: <searchLink fieldCode="MM" term="%22Skin+Diseases%22">Skin Diseases*</searchLink>/<searchLink fieldCode="MM" term="%22Skin+Diseases+diagnosis%22">diagnosis</searchLink> <br /><searchLink fieldCode="MM" term="%22Skin+Diseases%22">Skin Diseases*</searchLink>/<searchLink fieldCode="MM" term="%22Skin+Diseases+diagnostic+imaging%22">diagnostic imaging</searchLink> <br /><searchLink fieldCode="MM" term="%22Deep+Learning%22">Deep Learning*</searchLink> <br /><searchLink fieldCode="MM" term="%22Clinical+Reasoning%22">Clinical Reasoning*</searchLink><br /><searchLink fieldCode="MH" term="%22Humans%22">Humans</searchLink> ; <searchLink fieldCode="MH" term="%22Algorithms%22">Algorithms</searchLink> ; <searchLink fieldCode="MH" term="%22Sensitivity+and+Specificity%22">Sensitivity and Specificity</searchLink> ; <searchLink fieldCode="MH" term="%22Reproducibility+of+Results%22">Reproducibility of Results</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Background and Objective: Deep learning models have achieved remarkable diagnostic accuracy in medical imaging, yet their lack of interpretability limits clinical trust and deployment. This study presents CARE-MD (Clinical Algorithm for Reasoning-Enhanced Medical Diagnosis), a self-explainable framework that mirrors the clinical reasoning process to provide transparent, concept-driven, and case-referable diagnosis of skin lesions.<br />Methods: CARE-MD follows a four-phase reasoning structure: Observation, Interpretation, Reference, and Validation. In Phase 1, attention-guided localization isolates the lesion region using a modified U-Net backbone. Phase 2 employs a Concept Bottleneck Model (CBM) model to predict human-interpretable dermatological concepts. Phase 3 introduces prototype-based reasoning to compare latent features with previously learned prototypes for case-level interpretability. Phase 4 validates explanations through consistency analysis between attention maps, concept activations, and clinician-annotated lesion regions. The framework was evaluated on two publicly available datasets-ISIC 2018 and HAM10000-using metrics such as accuracy, precision, sensitivity, specificity, Dice coefficient, and explanation consistency score.<br />Results: CARE-MD achieved 87.2% accuracy, 86.8% precision, 85.6% sensitivity, and 88.1% specificity on the ISIC 2018 dataset. Cross-dataset testing on HAM10000 showed stable performance with minimal decline across all metrics, indicating strong generalization capability. The attention-guided module improved Dice score by 3% compared with baseline Attention U-Net, while the prototype referencing module enhanced explanation consistency by 6%. Qualitative results confirmed that CARE-MD produces coherent alignment between salient lesion regions, predicted dermatological concepts, and prototype-based references.<br />Conclusions: CARE-MD provides a structured approach toward aligning deep learning predictions with clinical reasoning by embedding interpretability into each diagnostic phase. While the framework shows improved interpretability and competitive performance relative to the evaluated baselines, it introduces additional architectural complexity and relies on concept-level supervision. These findings suggest that CARE-MD represents a promising direction for self-explainable medical diagnosis, while also highlighting inherent trade-offs between interpretability, model complexity, and annotation requirements. A notable strength of CARE-MD is its unified integration of multiple interpretability mechanisms within a clinically motivated reasoning structure.<br /> (Copyright © 2026 Elsevier B.V. All rights reserved.)
– Name: Abstract
  Label: Competing Interests
  Group: Ab
  Data: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
– Name: SubjectMinor
  Label: Contributed Indexing
  Group:
  Data: <i>Keywords: </i>Concept Bottleneck Model (CBM); Interpretability; Medical imaging; Prototype-based reasoning; Self-explainable AI; Skin lesion classification
– Name: DateEntry
  Label: Entry Date(s)
  Group: Date
  Data: <i>Date Created: </i>20260611 <i>Date Completed: </i>20260806 <i>Latest Revision: </i>20260806
– Name: DateUpdate
  Label: Update Code
  Group: Date
  Data: 20260807
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1016/j.cmpb.2026.109478
– Name: AN
  Label: PMID
  Group: ID
  Data: 42275844
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=cmedm&AN=42275844
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1016/j.cmpb.2026.109478
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        StartPage: 109478
    Subjects:
      – SubjectFull: Humans
        Type: general
      – SubjectFull: Algorithms
        Type: general
      – SubjectFull: Sensitivity and Specificity
        Type: general
      – SubjectFull: Reproducibility of Results
        Type: general
      – SubjectFull: Skin Diseases diagnosis
        Type: general
      – SubjectFull: Skin Diseases diagnostic imaging
        Type: general
      – SubjectFull: Deep Learning
        Type: general
      – SubjectFull: Clinical Reasoning
        Type: general
    Titles:
      – TitleFull: From black-box to clinical reasoning: CARE-MD a novel framework for Self-eXplainable skin lesions.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Kandpal KC
      – PersonEntity:
          Name:
            NameFull: Verma P
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 10
              Text: 2026 Oct
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-electronic
              Value: 1872-7565
          Numbering:
            – Type: volume
              Value: 285
          Titles:
            – TitleFull: Computer methods and programs in biomedicine
              Type: main
ResultId 1