Skip to content
DraconiticResearch
Search Sources API
203,948 studies31 sourcesUpdated 39 min ago
Peer-reviewedCC BY 4.0Frontiers in Endocrinology

Incremental value of a cross-dataset ulcer-region prior for infection and ischaemia classification in diabetic foot ulcer image patches

  • Ajuan Gong1
  • Wanjin Fu2
  • Wenjing Li1
  • Xiaoqun Xu1
  • Xiaomin Huang1
  • Nuojin Wang1
  • Yue Wang1
  • Tianrong Pan1
  • Yijun Du1
  1. 1Department of Endocrinology, The Second Affiliated Hospital of Anhui Medical University
  2. 2Clinical Pharmacology Research Center, The Second Affiliated Hospital of Anhui Medical University
Open at sourceDOICite

Abstract

Objective: To assess whether a DFUC2022-derived soft ulcer-region prior improves four-class infection/ischaemia label prediction in ulcer-centred DFUC2021 patches, using comparators matched for classifier architecture or base-model outputs.

Methods: We analysed 5, 955 labelled patches using five fixed image-similarity-grouped outer folds and five training seeds. CNN checkpoint selection and stacker fitting and selection excluded each outer evaluation fold. Three incremental comparisons assessed Full lesion-prior transfer (LPT) against an architecture-matched neutral-prior control, reliability-aware LPT (RA-LPT) against its output-only stack, and strong RGB stacks with versus without explicit prior features. Pointwise 95% confidence intervals and approximate two-sided p values were obtained from 5, 000 paired crossed seed × similarity-group bootstrap resamples; p values were Holm-adjusted across the three comparisons.

Results: Full LPT achieved Macro-F1 of 0.8921 ± 0.0059 versus 0.8777 ± 0.0085 for neutral Full LPT (mean ± sample SD across seeds). The mean paired difference was +0.0144 (95% CI, + 0.0048 to +0.0244; Holm-adjusted approximate p = 0.0132), positive in all five seeds. RA-LPT achieved Macro-F1 of 0.9181 ± 0.0038 versus 0.9206 ± 0.0028 for its corresponding output-only stack, without demonstrating an incremental improvement (difference, −0.0026; 95% CI, −0.0066 to +0.0012; adjusted p = 0.3079). Adding explicit prior features to the strong RGB stack also did not demonstrate an improvement in Macro-F1 (−0.0016; 95% CI, −0.0057 to +0.0025; adjusted p = 0.4223). The strong RGB stack achieved 0.9370 ± 0.0050, above Full LPT. In a later exploratory training-size analysis, Full LPT and its neutral control improved with increasing labelled-training budgets; mean Full LPT-minus-neutral Macro-F1 differences were positive at all four budgets, but all pointwise 95% confidence intervals included zero.

Conclusion: In the main five-seed evaluation, sample-specific prior information improved Full LPT relative to its architecture-matched neutral control. Its incremental value was comparator-dependent: additional explicit prior features showed no confirmed improvement beyond the evaluated model-output stacks. Patient identifiers and target-domain lesion masks were unavailable; these internal development findings require patient-linked, multicentre external validation before clinical use.

Related

  1. CBAM-enhanced convolutional networks for diabetic retinopathy classification in the ODIR dataset
    Haewon Byeon
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textMedical imaging & radiologyCrossref
  2. Assessing the Positive Predictive Value of International Classification of Diseases Code K55.0 for Identifying Acute Mesenteric Ischaemia
    Meis Jawad, Yasmin Soltanzadeh-Naderi, Talha Butt +1
    Diagnostics (Basel, Switzerland)Peer-reviewedCC BYFull textGeneral medicineEurope PMC
  3. Beyond Molecular Classification: The Incremental Value of Prognostic Biomarkers in Endometrial Cancer
    Vlad Gabriel Tirnovanu, Iustina Petra Solomon Condriuc, Raluca Anamaria Mogos +6
    Journal of Clinical MedicinePeer-reviewedCC BY 4.0Full textObstetrics, gynaecology & paediatricsCrossrefEurope PMC
  4. Incremental Value of CSF Biomarker-Integrated Classification of Cerebral Amyloid Angiopathy
    M. Losa, M. Cotta Ramusino, I. Gandoglia +20
    medRxivPreprintCC BY-NC-NDFull textNeurologymedRxiv
  5. SiNFluD: A novel benchmark dataset for sindhi figurative language classification
    Wazir Ali, Adeeb Noor, Saifullah Tumrani
    Data in BriefPeer-reviewedCC BYFull textGeneral medicineDOAJ

Cite

BibTeXRIS
BibTeX
@article{gong2026incremental,
  title = {{Incremental value of a cross-dataset ulcer-region prior for infection and ischaemia classification in diabetic foot ulcer image patches}},
  author = {Ajuan Gong and Wanjin Fu and Wenjing Li and Xiaoqun Xu and Xiaomin Huang and Nuojin Wang and Yue Wang and Tianrong Pan and Yijun Du},
  journal = {Frontiers in Endocrinology},
  year = {2026},
  month = nov,
  doi = {10.3389/fendo.2026.1944475},
  url = {https://www.frontiersin.org/articles/10.3389/fendo.2026.1944475/full},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Incremental value of a cross-dataset ulcer-region prior for infection and ischaemia classification in diabetic foot ulcer image patches
AU  - Ajuan Gong
AU  - Wanjin Fu
AU  - Wenjing Li
AU  - Xiaoqun Xu
AU  - Xiaomin Huang
AU  - Nuojin Wang
AU  - Yue Wang
AU  - Tianrong Pan
AU  - Yijun Du
PY  - 2026
DA  - 2026/11/06
JO  - Frontiers in Endocrinology
DO  - 10.3389/fendo.2026.1944475
UR  - https://www.frontiersin.org/articles/10.3389/fendo.2026.1944475/full
AB  - Objective: To assess whether a DFUC2022-derived soft ulcer-region prior improves four-class infection/ischaemia label prediction in ulcer-centred DFUC2021 patches, using comparators matched for classifier architecture or base-model outputs.  Methods: We analysed 5, 955 labelled patches using five fixed image-similarity-grouped outer folds and five training seeds. CNN checkpoint selection and stacker fitting and selection excluded each outer evaluation fold. Three incremental comparisons assessed Full lesion-prior transfer (LPT) against an architecture-matched neutral-prior control, reliability-aware LPT (RA-LPT) against its output-only stack, and strong RGB stacks with versus without explicit prior features. Pointwise 95% confidence intervals and approximate two-sided p values were obtained from 5, 000 paired crossed seed × similarity-group bootstrap resamples; p values were Holm-adjusted across the three comparisons.  Results: Full LPT achieved Macro-F1 of 0.8921 ± 0.0059 versus 0.8777 ± 0.0085 for neutral Full LPT (mean ± sample SD across seeds). The mean paired difference was +0.0144 (95% CI, + 0.0048 to +0.0244; Holm-adjusted approximate p = 0.0132), positive in all five seeds. RA-LPT achieved Macro-F1 of 0.9181 ± 0.0038 versus 0.9206 ± 0.0028 for its corresponding output-only stack, without demonstrating an incremental improvement (difference, −0.0026; 95% CI, −0.0066 to +0.0012; adjusted p = 0.3079). Adding explicit prior features to the strong RGB stack also did not demonstrate an improvement in Macro-F1 (−0.0016; 95% CI, −0.0057 to +0.0025; adjusted p = 0.4223). The strong RGB stack achieved 0.9370 ± 0.0050, above Full LPT. In a later exploratory training-size analysis, Full LPT and its neutral control improved with increasing labelled-training budgets; mean Full LPT-minus-neutral Macro-F1 differences were positive at all four budgets, but all pointwise 95% confidence intervals included zero.  Conclusion: In the main five-seed evaluation, sample-specific prior information improved Full LPT relative to its architecture-matched neutral control. Its incremental value was comparator-dependent: additional explicit prior features showed no confirmed improvement beyond the evaluated model-output stacks. Patient identifiers and target-domain lesion masks were unavailable; these internal development findings require patient-linked, multicentre external validation before clinical use.
SN  - 1664-2392
ER  -