The paper introduces a clinical‑dictionary radiomic bottleneck model for classifying thyroid nodules from ultrasound images. It combines Family‑Aware Radiomic Learning to estimate measurements across feature families with Fidelity‑Guided Bottleneck Modeling to jointly optimize measurement estimation and classification while preserving measurement fidelity. The model decomposes decision evidence through a linear classifier, mapping feature contributions to clinical dictionary categories, and demonstrates competitive performance on ThyroidXL and TN3K datasets with interpretable case studies.
By Guibin Zhang, Weier Luo, Lei Cao
arXiv:2607. 12075v1 Announce Type: cross Abstract: Background: Deep learning models can classify thyroid nodules on ultrasound, but reliable clinical decision support also requires calibrated probabilities, uncertainty estimation, and selective referral, particularly under dataset shift.
By Md. Sadibul Hasan Sadib, Md. Mohayminul Mukit, Rahmatul Kabir Rasel Sarker, Tahmid Alam Tamim, Md. Monir Hossain Shimul
TAM-Chain is a multi‑scale thyroid cytology classification framework that uses Absorbing Markov Chains and Shannon Entropy to quantify uncertainty and dynamically decide when to stop processing and refer to a specialist. It processes images at 10×, 20×, and 40× magnifications, achieving a Macro F1 score of 0.9741 on an internal test set with a 0 % false‑negative rate, and maintains a Macro F1 of 0.7026 on an external validation set with severe domain shift. The method outperforms single‑magnification baselines by adaptively adjusting stopping steps and triggering specialist referrals, thereby reducing critical diagnostic errors.
By Hai Pham Ngoc
arXiv:2608. 10505v1 Announce Type: new Abstract: Automated radiology report generation is advancing rapidly in response to the shortage of radiologists, yet unlike a perception model, existing generation models offer no control over the sensitivity-specificity trade-off of their diagnostic content.
By Ying Jin, Noel C. F. Codella, John Corring, Mu Wei, Dinei Florencio, Eric Horvitz
This review examines the use of artificial intelligence and radiomics for diagnosing thyroid cancer, summarizing 42 studies that demonstrate the effectiveness of ultrasound-based radiomics in identifying malignancies. It highlights challenges such as interpretability, limited datasets, and operator dependence, and calls for standardization, prospective multicenter trials, and advances in explainable AI and personalized medicine. The authors emphasize that multidisciplinary collaboration and further algorithm refinement could enhance diagnostic accuracy and patient outcomes.
By Milad Yousefi, Hadi Vahedi, Shadi Farabi Maleki, Mahya Ahmadpour Youshanlui, Aida Jafari, Parisa Rostami, Kais I. Abdul-Lateef Al-Abdullah, Ryszard Tadeusiewicz, Pawel Plawiak, Roohallah Alizadehsani, Siamak Pedrammehr
Automated radiology report generation is advancing rapidly in response to the shortage of radiologists, yet unlike a perception model, existing generation models offer no control over the sensitivity-specificity trade-off of their diagnostic content. Such control is essential because clinical scenarios diverge: emergency triage prioritizes sensitivity to reduce missed findings, whereas confirmatory interpretation emphasizes specificity to limit unnecessary interventions.