ArticleFrontiers in oncology2026
A comprehensive model with dual plane ultrasound based deep learning radiomics to distinguish parotid pleomorphic adenoma from Warthin tumor.
Article in Frontiers in oncology, 2026. The graph could read no effect estimate from its abstract, so it casts no vote on the map. Not yet cited in PubMed.
What it found
Each row is one number read from the abstract, on the scale the paper reported it, with its interval. Left of the dashed line favours the treatment, right favours the comparator. Under each row is the sentence it came from. New to these charts? A ten-minute tutorial.
The abstract states no effect estimate the extractor could read, or names no intervention and outcome on the map, so this paper lights no cell and moves no belief. It is still indexed, cited and linked below.
The trial behind it
Trials whose registry record cites this paper, or whose number appears in the abstract. A trial that started after this paper was published is citing it as background, not reporting it.
Neither the registry nor the abstract names a trial number. If this is a trial report, that itself is worth knowing.
Who cites it
0 citing papers in PubMed.
No citing paper in PubMed yet.
Corrections and comments
PubMed lists nothing against this paper. Absence here is not a guarantee, only a check that was made.
Authors and funding
8 authors.
Funding
No grant is acknowledged in the PubMed record.
Abstract
Objective: To develop and validate a dual-plane ultrasound (US)-based deep learning radiomics model for the noninvasive discrimination between parotid pleomorphic adenoma (PA) and Warthin tumor (WT). Methods: This retrospective two-center study enrolled 656 patients with pathologically confirmed parotid tumors (361 PA, 295 WT). Cases from Center 1 (n = 556) were randomly split into training and internal validation sets (7:3), while Center 2 cases (n = 100) formed the external test set. Radiomics features were extracted from long-axis and short-axis plane US images respectively to construct dual-plane radiomics models. For deep learning, pre-trained ResNet/DenseNet was adopted, and two fusion strategies were compared: multi-channel image fusion (long-axis, short-axis, and their pixel-wise mean) versus feature fusion (concatenation of deep features extracted from each plane). A comprehensive prediction model (nomogram) was further constructed by integrating clinical information and US features. Performance was evaluated using AUC, calibration curves, and DCA, and compared against three US physicians of varying experience. Diagnostic performance with and without model assistance was also assessed. Results: The multi-channel fusion deep learning model (AUC = 0.860, 95%CI: 0.790-0.931) outperformed feature fusion (AUC = 0.818, 95%CI: 0.735-0.904). Nomogram achieved the highest AUC: 0.907(95%CI: 0.850-0.964) in testing set, outperforming individual clinical, radiomics and deep learning models. It also exceeded the diagnostic performance of all physicians. Model assistance improved junior physicians' AUC by 0.062 and specificity by 0.220; senior and expert physicians showed higher sensitivity and accuracy. Conclusions: The proposed dual-plane US-based deep learning radiomics model effectively differentiates PA from WT noninvasively, enhances diagnostic performance-particularly for less experienced physicians-and demonstrates potential for clinical translation.
Indexed as
Identifiers
What OpenQuestion holds
Registered trials
Read under generation 80e0d062 · epoch 390. Bibliography from PubMed, PubMed Central and OpenAlex; grants from NIH RePORTER; trial links from ClinicalTrials.gov; estimates, votes and beliefs from the OpenQuestion graph.