Click, Predict, Trust: Clinician-Guided and Clinician-Evaluated AI Segmentation for Lung Cancer CT-Based Prognosis within the Knowledge-to-Action Framework
Authors: Mohammad R. Salmanpour, Sonya Falahati, Amir Hossein Pouria, Amin Mousavi, Somayeh Sadat Mehrnia, Morteza Alizadeh, Arman Gorji, Zeinab Farsangi, Alireza Safarian, Mehdi Maghsudi, Carlos Uribe, Arman Rahmim, Ren Yuan
Organizations: Department of Basic and Translational Research, BC Cancer Research Institute, Vancouver, BC, Canada · Department of Radiology, University of British Columbia, Vancouver, BC, Canada · Technological Virtual Collaboration (TECVICO Corp.), Vancouver, BC, Canada · Amirkabir University of Technology, Department of Computer Engineering, Tehran, Iran · Department of Computer, Abhar Branch, Islamic Azad University, Abhar, Iran · Department of Integrative Oncology, Breast Cancer Research Center, Motamed Cancer Institute, ACECR, Tehran, Iran · Department of Mathematics, University of Isfahan, Isfahan, Iran · NAIRG, Department of Neuroscience, Hamadan University of Medical Sciences, Hamadan, Iran · Rajaie Cardiovascular Medical and Research Center, Iran University of Medical Science, Tehran, Iran · Department of Nuclear Medicine, University Hospital, Paracelsus Medical University, Salzburg, Austria · BC Cancer, Vancouver Center, Vancouver, BC, Canada
Lung cancer remains the leading cause of cancer mortality, with CT imaging central to screening, prognosis, and treatment. Manual segmentation is variable and time-intensive, while deep learning (DL) offers automation but faces barriers to clinical adoption. Guided by the Knowledge-to-Action framework, this study develops a clinician-in-the-loop DL pipeline to enhance reproducibility, prognostic accuracy, and clinical trust. Multi-center CT data from 999 patients across 12 public datasets were analyzed using five DL models (3D Attention U-Net, ResUNet, VNet, ReconNet, SAM-Med3D), benchmarked against expert contours on whole and click-point cropped images. Segmentation reproducibility was assessed using 497 PySERA-extracted radiomic features via Spearman correlation, ICC, Wilcoxon tests, and MANOVA, while prognostic modeling compared supervised (SL) and semi-supervised learning (SSL) across 38 dimensionality reduction strategies and 24 classifiers. Six physicians qualitatively evaluated masks across seven domains, including clinical meaningfulness, boundary quality, prognostic value, trust, and workflow integration. VNet achieved the best performance (Dice = 0.83, IoU = 0.71), radiomic stability (mean correlation = 0.76, ICC = 0.65), and predictive accuracy under SSL (accuracy = 0.88, F1 = 0.83). SSL consistently outperformed SL across models. Radiologists favored VNet for peritumoral representation and smoother boundaries, preferring AI-generated initial masks for refinement rather than replacement. These results demonstrate that integrating VNet with SSL yields accurate, reproducible, and clinically trusted CT-based lung cancer prognosis, highlighting a feasible path toward physician-centered AI translation.
Medical image segmentation delineates organs, lesions, and other structures in scans to support measurement, diagnosis, treatment planning, and research.
Lung cancer is the leading cause of cancer-related mortality, with approximately 2.5 million new cases and 1.8 million deaths annually, making reliable diagnosis a clinical priority. Although deep learning models have achieved strong performance in lung cancer classification, evaluation has largely focused on predictive accuracy, leaving their decision-making processes insufficiently examined. This study compares three architecturally distinct models: a Convolutional Neural Network (CNN), a pretrained ResNet50, and a Vision Transformer (ViT), trained on the IQ-OTH/NCCD lung cancer CT dataset. Local Interpretable Model-Agnostic Explanations (LIME) were applied to investigate model reasoning. In addition to standard performance metrics, a dual-correlation framework was introduced to measure both prediction agreement and explanation agreement across model pairs. All three models achieved strong classification performance, with ResNet50 attaining 98.61% accuracy, CNN 97.91%, and ViT 93.75%, while all achieved ROC-AUC scores of 0.99. Prediction correlations exceeded 0.99 across all model pairs, indicating highly consistent outputs. However, LIME explanation correlations remained below 0.26, revealing substantial differences in the image regions used to reach those predictions. Analysis of misclassified samples further identified a consistent spatial pattern: incorrect predictions were associated with attention outside the lung parenchyma, whereas correct predictions focused primarily within lung regions. These findings demonstrate that prediction agreement is a poor proxy for reasoning consistency, and that interpretability evaluation must be treated as an independent validation criterion alongside predictive performance in clinical AI systems.
Samarpan Poudel, Vladislav D Veksler
Caldwell University School of Business and Computer Science Caldwell, NJ 07006
Laryngeal cancer imaging research lacks standardised public datasets to enable reproducible deep learning (DL) model development. We present LaryngealCT, a curated benchmark of 1,029 computed tomography (CT) scans aggregated from six collections from The Cancer Imaging Archive (TCIA). Uniform 1 mm isotropic volumes of interest encompassing the larynx were extracted using a weakly supervised parameter search framework validated by clinical experts. Six 3D DL architectures (custom 3D CNN, ResNet18,50,101, DenseNet121 and MedicalNet-pretrained ResNet50) were benchmarked on (i) early (Tis,T1,T2) vs. advanced (T3,T4) and (ii) T4 vs. non-T4 classification tasks. On the independent test set, the 3D CNN achieved the strongest overall performance across global and per-class metrics (Accuracy 0.854, F1-macro 0.841) in early vs. advanced classification. In the T4 task, AU-ROC values exceeded 0.82 for most models, but sensitivity for T4 disease remained limited (less than or equal to 0.412), with ResNet101 showing the most promising calibrated T4 recall (0.706. Model explainability assessed using GradCAMpp with thyroid cartilage overlays for T4 classification task revealed anatomically plausible peri-cartilage activations, although spatial overlap was modest. Through open-source data, pretrained models, and integrated explainability tools, LaryngealCT offers a reproducible foundation for AI-driven research to support future clinical decision-making in laryngeal oncology.
Nivea Roy, Son Tran, Atul Sajjanhar +4
School of Information Technology, Deakin University, Burwood, Australia · Department of Head and Neck Surgery, Kasturba Medical College, Manipal Academy of Higher Education, Manipal, India · Department of Radiodiagnosis and Imaging, Kasturba Medical College, Manipal Academy of Higher Education, Manipal, India +1
Zero-shot vision-language models (VLMs) offer a promptable alternative to task-specific training for gross tumor volume (GTV) delineation in non-small-cell lung cancer (NSCLC), but the prompt dimensions that govern their spatial behavior remain poorly understood. We study this question by probing alignment directions in VoxTell on a held-out internal NSCLC tumor dataset through sub-prompt decomposition into diagnosis, demographic, staging, anatomical, generic, and irrelevant controls; attribute-wise perturbation robustness; specificity ladders; and cross-case prompt swaps, while benchmarking against fine-tuned and zero-shot baselines using the Dice Similarity Coefficient (DSC) with Wilcoxon signed-rank tests and Benjamini-Hochberg correction. Alignment analyses revealed that anatomical location is the dominant driver of VoxTell's spatial attention: 63.4 percent of location perturbations caused catastrophic drops, prompt specificity improved from generic to full descriptions except for diagnosis-only prompts, irrelevant prompts correctly yielded zero segmentation, and cross-case prompt swaps confirmed patient-specific conditioning (matched DSC 0.906 vs. mismatched 0.406). Histology and stage substitutions had minimal effect, indicating that the model prioritizes "where to look" over "what to look for." In this context, VoxTell, operating fully zero-shot, achieved a mean DSC of 0.613, statistically indistinguishable from nnUNet (0.690, adjusted p = 0.156) and Ahmed et al. (0.675, adjusted p = 0.679), while significantly outperforming all other zero-shot models. Together, these findings argue that segmentation VLMs should be evaluated not only by Dice, but also by the prompt dimensions to which they align.
Suraj Pai, Thibault Heintz, Cosmin Ciausu +3
Artificial Intelligence in Medicine Program, Mass General Brigham, Harvard Medical School, USA