A multimodal vision foundation model for clinical dermatology

In the paper

“On representative dermoscopy and clinical benchmarks such as HAM10000 (ref. 34) and PAD-UFES-20 (ref. 43), PanDerm surpassed the next-best models by 4.7% (P < 0.001) and 9.0% (P < 0.001), respectively”

What the sources say

• Supported

  The citation is attached to HAM10000, identifying the dataset used as a dermoscopy benchmark, rather than to the PanDerm performance figures. The Methods section describes HAM10000 as a collection of 10,015 dermatoscopic images, and Data Records identifies it as a dataset. This source is not being cited for the 4.7% or 9.0% results. (LLM, conf=0.92)

  cites [34] Tschandl, P., Rosendahl, C. & Kittler, H. The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Sci. Data 5, 180161 (2018). source
• Not checked yetcites [43] Pacheco, A. G. et al. PAD-UFES-20: a skin lesion dataset composed of patient data and clinical images collected from smartphones. Data Brief 32, 106221 (2020). source

In the paper

“Some efforts have focused on self-supervised learning specifically for dermatology using public datasets33,34 or web-sourced skin images35.”

What the sources say

• Not checked yetcites [33] Wang, Z., Lyu, J. & Tang, X. autoSMIM: automatic superpixel-based masked image modeling for skin lesion segmentation. IEEE Trans. Med. Imaging 42, 3501–3511 (2023). source
• Supported

  The citation is attached to “public datasets,” and this paper identifies and releases one: HAM10000. The Abstract (page not provided) states that its dermatoscopic images are publicly available through the ISIC archive for machine-learning use. The paper does not itself report a self-supervised-learning effort, but the citation supports the dataset identification. (LLM, conf=0.88)

  cites [34] Tschandl, P., Rosendahl, C. & Kittler, H. The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Sci. Data 5, 180161 (2018). source
• Not checked yetcites [35] Shen, Y. et al. Optimizing skin disease diagnosis: harnessing online community data with contrastive learning and clustering techniques. NPJ Digit. Med. 7, 28 (2024). source

In the paper

“Built on rich knowledge representations, these models have shown impressive performance across medical fields such as ophthalmology20, radiology21 and pathology22–25.”

What the sources say

• Not checked yetcites [20] Zhou, Y. et al. A foundation model for generalizable disease detection from retinal images. Nature 622, 156–163 (2023). source
• Supported

  Citation 21 is attached to radiology, not to ophthalmology or pathology. The Discussion reports robust performance of the paper’s cancer-imaging foundation model across three tasks and four cohorts, including superiority over baseline approaches for anatomical-site classification and malignancy prediction. (LLM, conf=0.92)

  cites [21] Pai, S. et al. Foundation model for cancer imaging biomarkers. Nat. Mach. Intell. 6, 354–367 (2024). source
• Supported

  The citation range 22–25 is attached to pathology, not ophthalmology or radiology. The Abstract reports that this pathology foundation model achieves state-of-the-art performance on various digital pathology tasks, supporting the attributed claim. (LLM, conf=0.93)

  cites [22] Xu, H. et al. A whole-slide foundation model for digital pathology from real-world data. Nature 630, 181–188 (2024). source
• Supported

  This source supports the pathology portion of the citing sentence. The Introduction reports that its foundation model outperformed previous state-of-the-art models across pathology tasks, and the Discussion describes competitive performance across diverse clinical applications in pathology. The citations attached to ophthalmology and radiology are separate. (LLM, conf=0.91)

  cites [23] Chen, R. J. et al. Towards a general-purpose foundation model for computational pathology. Nat. Med. 30, 850–862 (2024). source
• Supported

  This source supports the pathology portion of the multi-citation sentence. In the Results and Discussion, it reports strong performance by a pathology foundation model in cancer detection, including performance close to clinical-grade specialist models and better performance on some rare cancer variants. The ophthalmology and radiology portions are attached to other citations. (LLM, conf=0.93)

  cites [24] Vorontsov, E. et al. A foundation model for clinical-grade computational pathology and rare cancers detection. Nat. Med. 30, 2949–2935 (2024). source
• Supported

  This source is cited for pathology, not for the separately cited ophthalmology and radiology examples. Its Abstract reports that the CHIEF pathology foundation model outperformed state-of-the-art deep-learning methods by up to 36.1%, supporting the performance claim. (LLM, conf=0.92)

  cites [25] Wang, X. et al. A pathology foundation model for cancer diagnosis and prognosis prediction. Nature 634, 970–978 (2024). source

In the paper

“Using TBP images alone, PanDerm achieved a sensitivity of 0.893, outperforming the next-best model by 4.2% (Fig. 4j, left).”

In the paper

“When clinical measurements were available for all models, PanDerm maintained its advantage with a 3.5% higher sensitivity (Fig. 4j, right), reaching a sensitivity of 0.893.”

In the paper

“PanDerm showed exceptional accuracy in distinguishing melanomas likely to metastasize, achieving an AUROC of 0.964 (95% CI 0.937–0.991), surpassing the next-best model by 2.0% (P = 0.073) (Fig. 3d).”

In the paper

“showing superior data scalability and training efficiency compared with existing self-supervised algorithms (Fig. 1f).”

In the paper

“The model achieves unified representation learning across total-body photography (TBP) and clinical, dermoscopic and dermatopathology images, enabling comprehensive patient analysis throughout diverse clinical workflows (Fig. 1d).”

In the paper

“Using the optimized pipeline for all models, PanDerm achieved AUROC improvements of 4.3% in SDDI1 (P < 0.001) and 3.7% in SDDI2 over the next-best model (Fig. 3c, middle).”

In the paper

“Even using computationally efficient methods, PanDerm maintained its advantages, outperforming both linear-probe and fine-tune versions of BiomedGPT by 14.3% and 5.1%, respectively, in linear probing (Supplementary Table 3).”

In the paper

“This processing system, combined with PanDerm’s advanced lesion change detection capabilities45, significantly improved change detection accuracy from 0.596 (95% confidence interval (CI) 0.567–0.624) to 0.706 (95% CI 0.686–0.725) in sequential digital dermoscopic imaging data (SDDI1) (Fig. 3a,c) (P < 0.001) and from 0.683 (95% CI 0.517–0.894) to 0.767 (95% CI 0.649–0.886) in SDDI2 (Fig. 3b,c, left”

What the sources say

• Not checked yetcites [45] Zhang, B. et al. Short-term lesion change detection for melanoma screening with novel Siamese neural network. IEEE Trans. Med. Imaging 40, 840–851 (2020). source

In the paper

“PanDerm showed strong performance even with limited training data, achieving comparable results to other models while using only 10–30% of the labeled images (Fig. 2b and Supplementary Tables 5–10).”

In the paper

“It maintained high predictive accuracy over extended follow-up periods, with time-dependent areas under the curve (AUCs) of 0.950 (95% CI 0.910–0.991), 0.931 (95% CI 0.887–0.976) and 0.909 (95% CI 0.880–0.937) at 3 years, 5 years and 7 years, exceeding multi-clinical variables by 6.8%, 2.9% and 5.0%, respectively (Fig. 3i).”

In the paper

“while showing emergent capabilities in dermatology similar to those of DINOv2 in natural images, with linear probing performance comparable to full-parameter fine-tuning (Supplementary Table 2).”

In the paper

“For lesions later diagnosed as malignant, PanDerm achieved an AUROC of 0.840 (95% CI 0.769–0.911), surpassing the next-best model by 15.0% (P < 0.01) (Fig. 3c, right).”

In the paper

“PanDerm achieved weighted F1 improvements of 3.2%, 7.1% and 8.2% on MMT-09, DermNet and MMT-74, respectively, compared with the next-best models (all P < 0.001).”

In the paper

“PanDerm also consistently outperformed other AI approaches (Fig. 3j), showing improvements of 2.3%, 3.0% and 2.5% at 3 years, 5 years and 7 years, respectively.”

In the paper

“In lesion-specific risk stratification, PanDerm also ranked first with an AUROC of 0.705 (95% CI 0.698–0.712) and balanced accuracy (BACC) of 0.657 (95% CI 0.6513–0.663), with all results statistically significant (P < 0.001; Fig. 4d,h).”

In the paper

“Significantly, it detected malignant lesions in 79 out of 80 patients while reducing unnecessary examinations by 60.8% compared with melanographers (3,498 versus 8,913 lesions recommended for detailed examination) (Fig. 4j,k and Supplementary Table 18).”

In the paper

“Notably, it maintained high performance even on clinical photographs that were not used during training, with AUROC gains of 4.0%, 2.6% and 2.1% on the three external clinical datasets (all P < 0.001).”

In the paper

“PanDerm consistently outperformed all other models, achieving significant improvements on 9 of 10 datasets, with average gains of 5.1%, 8.0%, 4.2% and 0.9% on dermoscopic, clinical, TBP and pathology datasets, respectively (Fig. 2a).”

In the paper

“In a cohort of 480 patients with 196,933 lesions from Australia, PanDerm achieved a weighted F1 score of 0.896 (95% CI 0.879–0.913) for photodamage assessment and an AUROC of 0.983 (95% CI 0.979–0.987) for nevus counting, surpassing all other models (P < 0.05 and P < 0.001, respectively; Fig. 4b,c,g).”

In the paper

“PanDerm showed significant superiority over all pretrained models, achieving higher AUROC scores across all external datasets (Fig. 2c).”

In the paper

“These advantages enabled PanDerm to improve upon both traditional models42,60 and recent generalist medical models such as BiomedGPT41.”

What the sources say

• Not checked yetcites [42] Dosovitskiy, A. et al. An image is worth 16×16 words: transformers for image recognition at scale. In Proc. 9th International Conference on Learning Representations https://openreview.net/forum?id=YicbFdNTTy (OpenReview.net, 2021).
• Not checked yetcites [60] He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In Proc. IEEE Conference on Computer Vision and Pattern Recognition 770–778 (IEEE, 2016). source
• Not checked yetcites [41] Zhang, K. et al. A generalist vision–language foundation model for diverse biomedical tasks. Nat. Med. 30, 3129–3141 (2024). source

In the paper

“In addition, PanDerm offers practical advantages in clinical settings, processing images about four to five times faster than MedSAM while using less computational resources (Supplementary Table 23).”

In the paper

“In terms of overall accuracy, PanDerm outperformed the average human reviewer by 10.2% and surpassed the best-performing human by 3.6%.”

In the paper

“Furthermore, when used independently, PanDerm achieved higher diagnostic accuracy than both unassisted readers (top 1 scores: 3.6 versus 2.83; P < 0.001) and human–AI collaboration (top 1 scores: 3.6 versus 3.08; P < 0.001).”

In the paper

“Our model with CLIP-large teachers significantly improved performance and outperformed CLIP-large itself.”

In the paper

“The model showed robust performance, achieving an 80.4% mean (macro-averaged) recall, with notably high recall rates for critical skin lesions: 87.2% for melanoma and 86.0% for BCC.”

In the paper

“Notable diagnostic variability between pathologists was observed.”

In the paper

“Recent advances in foundation models have emerged as a promising direction to address this challenge15,16.”

What the sources say

• Not checked yetcites [15] Gui, H., Omiye, J. A., Chang, C. T. & Daneshjou, R. The promises and perils of foundation models in dermatology. J. Invest. Dermatol. 144, 1440–1448 (2024). source
• Not checked yetcites [16] Moor, M. et al. Foundation models for generalist medical artificial intelligence. Nature 616, 259–265 (2023). source

In the paper

“In clinical practice, diagnosing and treating skin conditions involves a range of tasks, including total-body skin cancer detection and risk assessment2–5, differential diagnosis of hundreds of dermatological conditions such as inflammatory dermatoses and pigmentary disorders6, multimodal image analysis7,8, pathology interpretation9,10, monitoring lesion changes11,12 and predicting outcomes13,14.”

What the sources say

• Not checked yetcites [2] Primiero, C. A. et al. A narrative review: opportunities and challenges in artificial intelligence skin image analyses using total body photography. J. Invest. Dermatol. 144, 1200–1207 (2024). source
• Not checked yetcites [3] Primiero, C. A. et al. A protocol for annotation of total body photography for machine learning to analyze skin phenotype and lesion classification. Front. Med. 11, 1380984 (2024). source
• Not checked yetcites [4] Olsen, C. M. et al. Risk stratification for melanoma: models derived and validated in a purpose-designed prospective cohort. J. Natl Cancer Inst. 110, 1075–1083 (2018). source
• Not checked yetcites [5] Usher-Smith, J. A., Emery, J., Kassianos, A. P. & Walter, F. M. Risk prediction models for melanoma: a systematic review. Cancer Epidemiol. Biomark. Prev. 23, 1450–1463 (2014). source
• Not checked yetcites [6] Liu, Y. et al. A deep learning system for differential diagnosis of skin diseases. Nat. Med. 26, 900–908 (2020). source
• Not checked yetcites [7] Yap, J., Yolland, W. & Tschandl, P. Multimodal skin lesion classification using deep learning. Exp. Dermatol. 27, 1261–1267 (2018). source
• Not checked yetcites [8] Luo, N. et al. Artificial intelligence-assisted dermatology diagnosis: from unimodal to multimodal. Comput. Biol. Med. https://doi.org/10.1016/j.compbiomed.2023.107413 (2023). source

In the paper

“These results show PanDerm’s potential to improve patient care across diverse clinical scenarios and serve as a model for developing multimodal foundation models in other medical specialties, potentially accelerating the integration of artificial intelligence support in healthcare.”

In the paper

“In general computer vision, foundation models are pretrained on massive datasets such as ImageNet31 or JFT-300M (ref. 32) and most existing dermatology AI models still rely on these models for downstream adaptation.”

What the sources say

• Not checked yetcites [31] Russakovsky, O. et al. ImageNet large scale visual recognition challenge. Int. J. Comput. Vis. 115, 211–252 (2015). source
• Not checked yetcites [32] Sun, C., Shrivastava, A., Singh, S. & Gupta, A. Revisiting unreasonable effectiveness of data in deep learning era. In Proc. IEEE International Conference on Computer Vision 843–852 (IEEE, 2017). source

In the paper

“The performance of foundation models is inherently linked to the scale of their parameters and training data28–30.”

What the sources say

• Not checked yetcites [28] Achiam, J. et al. Gpt-4 technical report. Preprint at https://arxiv.org/abs/2303.08774 (2023).
• Not checked yetcites [29] Zhai, X., Kolesnikov, A., Houlsby, N. & Beyer, L. Scaling vision transformers. In Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition 12104–12113 (IEEE, 2022). source
• Not checked yetcites [30] Kaplan, J. et al. Scaling laws for neural language models. Preprint at https://arxiv.org/abs/2001.08361 (2020).

In the paper

“Monitoring suspicious melanocytic lesions over a 3-month period is a widely accepted procedure for early melanoma detection, as changes often prompt excision to rule out melanoma, while stability can be reassuring12.”

What the sources say

• Not checked yetcites [12] Altamura, D., Avramidis, M. & Menzies, S. W. Assessment of the optimal interval for and sensitivity of short-term sequential digital dermoscopy monitoring for the diagnosis of melanoma. Arch. Dermatol. 144, 502–506 (2008). source

In the paper

“the evaluation mainly reflects overall accuracy rather than disease-specific analysis”

In the paper

“These findings suggest promising directions for developing foundation models in other medical specialties in which the integration of diverse imaging modalities and complex clinical workflows is crucial for patient care.”

In the paper

“Our development approach, combining systematic data curation, advanced self-supervised learning and rigorous clinical validation, provides a framework for developing medical AI systems that can adapt to varying levels of clinical expertise and healthcare settings.”

In the paper

“These capabilities stem from its rich knowledge representation, developed through pretraining on varied dermatological image modalities and conditions, leading to consistent and significant performance improvements across tasks and modalities.”

In the paper

“our results suggest that specialty-specific foundation models designed with clinical workflows in mind may offer more practical solutions for specialties in which multiple imaging modalities are crucial.”

In the paper

“Our strategy minimizes this risk, ensuring that benchmark evaluations accurately reflect real-world model performance.”

In the paper

“comprehensive bias assessment requires metrics beyond overall accuracy.”

In the paper

“This approach contrasts with previous efforts, such as SwAVDerm35, which relied on web-sourced skin data, inadvertently incorporating images from commonly used benchmarks such as ISIC55 and DermNet44, increasing the risk of data leakage and compromising evaluation validity.”

What the sources say

• Not checked yetcites [35] Shen, Y. et al. Optimizing skin disease diagnosis: harnessing online community data with contrastive learning and clustering techniques. NPJ Digit. Med. 7, 28 (2024). source
• Not checked yetcites [55] Codella, N. C. et al. Skin lesion analysis toward melanoma detection: a challenge at the 2017 International Symposium on Biomedical Imaging (ISBI), hosted by the International Skin Imaging Collaboration (ISIC). In Proc. IEEE 15th International Symposium on Biomedical Imaging (eds Meijering, E. & Summers, R.) 168–172 (IEEE, 2018). source
• Not checked yetcites [44] Dermnet. Dermnet https://dermnet.com/ (2023).

In the paper

“The HIBA55 dataset is curated from the HIBA data from the ISIC archive55, containing 1,635 dermoscopic images with melanoma and other classes.”

What the sources say

• Not checked yetcites [55] Codella, N. C. et al. Skin lesion analysis toward melanoma detection: a challenge at the 2017 International Symposium on Biomedical Imaging (ISBI), hosted by the International Skin Imaging Collaboration (ISIC). In Proc. IEEE 15th International Symposium on Biomedical Imaging (eds Meijering, E. & Summers, R.) 168–172 (IEEE, 2018). source

In the paper

“The HOP study has received approval from the Human Research Ethics Committee (HREC) from Metro South Health HREC (HREC/17/QPAH/816) and the University of Queensland HREC (2018000074).”

In the paper

“The naevus surveillance study images (NSSI) dataset is part of the Brisbane Naevus Morphology Study, circa 2009–2014. The study followed the Declaration of Helsinki protocols and was approved by the Princess Alexandra Hospital human research ethics committee.”

In the paper

“To compare models, we implement pairwise permutation tests, conducting 1,000 permutations per pair and recalculating performance metrics after each permutation.”

In the paper

“In addition, we perform t-tests to assess the statistical significance of inter-model performance variations.”

In the paper

“The dataset contains 647 clinical images from the United States.”

What the sources say

• Not checked yetcites [63] Daneshjou, R. et al. Disparities in dermatology AI performance on a diverse, curated clinical image set. Sci. Adv. 8, eabq6147 (2022). source

In the paper

“The Med-Node82 dataset contains 170 clinical images.”

What the sources say

• Not checked yetcites [82] Giotis, I. et al. MED-NODE: a computer-assisted melanoma diagnosis system using non-dermoscopic images. Expert Syst. Appl. 42, 6578–6585 (2015). source

In the paper

“We ensure robust evaluation through fivefold cross-validation, stratifying by both case and label attributes.”

In the paper

“The BCN20000 (ref. 79) dataset comprises 12,413 dermoscopic images in 9 categories: nevus, melanoma, basal cell carcinoma, seborrheic keratosis, actinic keratosis, solar lentigo, squamous cell carcinoma, dermatofibroma and vascular lesions, including lesions in hard-to-diagnose locations.”

What the sources say

• Not checked yetcites ref. 79

In the paper

“The Fitzpatrick17K (ref. 62) dataset comprises 16,577 clinical images annotated with both dermatological diagnoses and Fitzpatrick skin types (I–VI). It encompasses 114 distinct conditions”

What the sources say

• Not checked yetcites ref. 62

In the paper

“The HAM10000 (ref. 34) dataset contains 10,015 dermoscopic images across 7 classes: actinic keratoses, basal cell carcinoma, benign keratosis, dermatofibroma, melanocytic nevi, melanoma and vascular lesions.”

What the sources say

• Not checked yetcites ref. 34

In the paper

“probably owing to the limited data scale of skin images in medical-domain CLIP models.”

In the paper

“For skin tumor patch classification, melanoma slide classification, reader studies, metastasis prediction and skin lesion segmentation, we conduct k-fold cross-validation owing to either a relatively small sample size or following conventional practice.”

In the paper

“To assess statistical significance, we conduct two-sided t-tests comparing PanDerm’s performance against the baseline model for each task.”

In the paper

“The SDDI2 dataset has been approved by the Ethics Review Board of the Medical University of Vienna.”

In the paper

“The ACEMID pathology (ACEMID_path) pilot study has received approval from the Alfred Hospital Ethics Committee (approval number: 746/23)”

In the paper

“PH2 (ref. 81) is a clinical image dataset from Portugal with 200 images and 3 classes.”

What the sources say

• Not checked yetcites ref. 81

In the paper

“Before the evaluation, four experienced dermatologists collaboratively developed a standard ontology to systematically categorize the 128 skin conditions and facilitate expert evaluation (Extended Data Fig. 8).”

In the paper

“To ensure a fair comparison, all images are resized to 224 × 224.”

In the paper

“For the remaining datasets, we use nonparametric bootstrapping with 1,000 replicates to estimate 95% CIs for each model’s performance.”

In the paper

“Only de-identified retrospective data were used for research, without the active involvement of patients.”

In the paper

“Subsequently, we fine-tuned these models using identical hyperparameter setups to ensure a fair comparison.”

In the paper

“The ComBineMel dataset is part of the Computer Biomarkers Evaluation of Invasive Melanoma (ComBine Mel) study. The study was approved by the Alfred Hospital Ethics Committee on 8 August 2023 (approval number: HREC/98200/Alfred-2023).”

In the paper

“To mitigate bias, we included ten single time-point melanoma images, preventing reviewers from assuming that the first image in a series was benign.”

In the paper

“The melanoma slide classification task83 from the National Cancer Institute’s Clinical Proteomic Tumor Analysis Consortium Cutaneous Melanoma (CPTAC-CM) cohort consists of histopathology WSIs for cancer detection.”

What the sources say

• Not checked yetcites [83] Clark, K. et al. The Cancer Imaging Archive (TCIA): maintaining and operating a public information repository. J. Digit. Imaging 26, 1045–1057 (2013). source

In the paper

“To enhance the robustness and mitigate potential data selection bias, we perform five iterations of dataset splitting into training and testing sets, stratified by melanoma stage.”

In the paper

“DermNet44 contains 19,559 clinical images; this dataset consists of images of 23 types of skin diseases and captures common clinical presentations including inflammatory conditions (eczema, psoriasis), infections (bacterial, viral, fungal) and neoplastic diseases.”

What the sources say

• Not checked yetcites [44] Dermnet. Dermnet https://dermnet.com/ (2023).

In the paper

“The ISIC2018 dataset52 comprises 3,694 dermoscopic images with 2,594 images for training, 100 for validation and 1,000 for testing.”

What the sources say

• Not checked yetcites [52] Milton, M. A. A. Automated skin lesion classification using ensemble of deep neural networks in ISIC 2018: skin lesion analysis towards melanoma detection challenge. Preprint at https://arxiv.org/abs/1901.10802 (2019).

In the paper

“The Memorial Sloan Kettering Cancer Center (MSKCC)55 dataset is curated from the MSKCC data from the ISIC archive55, containing 8,984 dermoscopic images with melanoma and other classes.”

What the sources say

• Not checked yetcites [55] Codella, N. C. et al. Skin lesion analysis toward melanoma detection: a challenge at the 2017 International Symposium on Biomedical Imaging (ISBI), hosted by the International Skin Imaging Collaboration (ISIC). In Proc. IEEE 15th International Symposium on Biomedical Imaging (eds Meijering, E. & Summers, R.) 168–172 (IEEE, 2018). source

In the paper

“The MMT data study is part of a research agreement study with Monash eResearch Centre and was approved through the Monash University Human Research Ethics Committee.”

In the paper

“All other experiments used the clean version from a previous study78, which prevents data leakage by ensuring that lesions from the same patient are not split across sets.”

What the sources say

• Not checked yetcites previous study78

In the paper

“The PAD-UFES-20 (ref. 43) dataset from Brazil contains 2,298 close-up clinical images with 6 classes, including actinic keratosis, basal cell carcinoma of the skin, malignant melanoma, melanocytic nevus of the skin, squamous cell carcinoma and seborrheic keratosis.”

What the sources say

• Not checked yetcites ref. 43

In the paper

“The MYM study was approved by the Metro South Health Human Research Ethics Committee on 21 April 2016 (approval number: HREC/16/QPAH/125).”

In the paper

“The SDDI_Alfred study has received approval from the Alfred Hospital Ethics Committee (approval number: 198/19) for the use of sequential dermoscopic imaging data.”

In the paper

“To ensure robust assessment, each case was evaluated by three assessors, with cases showing significant scoring discordance (differences between 3/4 and 1/2) reviewed in consensus meetings to establish final scores.”

In the paper

“ISIC2024 (ref. 47) is a multicenter dataset with skin lesion crops from TBP. We chose holdout data with 49,025 crop images with three institutions (FNQH Cairns, Alfred Hospital, Melanoma Institute Australia) as the evaluation dataset.”

What the sources say

• Not checked yetcites ref. 47