AI-based software for lung nodule detection in chest X-rays -- Time for a second reader approach?

Susanne Ohlmann-Knafo; Naglis Ramanauskas; Sebastian Huettinger; Emil Johnson Jeyakumar; Darius Barušauskas; Neringa Bielskienė; Vytautas Naujalis; Jonas Bialopetravičius; Jonas Ražanskas; Artūras Samuilis; Jūratė Dementavičienė; Dirk Pickuth

arXiv · 2206.10912

AI-based software for lung nodule detection in chest X-rays -- Time for a second reader approach?

Susanne Ohlmann-Knafo·Naglis Ramanauskas·Sebastian Huettinger·Emil Johnson Jeyakumar·Darius Barušauskas·Neringa Bielskienė·Vytautas Naujalis·Jonas Bialopetravičius·Jonas Ražanskas·Artūras Samuilis·Jūratė Dementavičienė·Dirk Pickuth

View original source ↗Read source document ↗Export metadata ↓

Abstract

Objectives: To compare artificial intelligence (AI) as a second reader in detecting lung nodules on chest X-rays (CXR) versus radiologists of two binational institutions, and to evaluate AI performance when using two different modes: automated versus assisted (additional remote radiologist review). Methods: The CXR public database (n = 247) of the Japanese Society of Radiological Technology with various types and sizes of lung nodules was analyzed. Eight radiologists evaluated the CXR images with regard to the presence of lung nodules and nodule conspicuity. After radiologist review, the AI software processed and flagged the CXR with the highest probability of missed nodules. The calculated accuracy metrics were the area under the curve (AUC), sensitivity, specificity, F1 score, false negative case number (FN), and the effect of different AI modes (automated/assisted) on the accuracy of nodule detection. Results: For radiologists, the average AUC value was 0.77 $\pm$ 0.07, while the average FN was 52.63 $\pm$ 17.53 (all studies) and 32 $\pm$ 11.59 (studies containing a nodule of malignant etiology = 32% rate of missed malignant nodules). Both AI modes -- automated and assisted -- produced an average increase in sensitivity (by 14% and 12%) and of F1-score (5% and 6%) and a decrease in specificity (by 10% and 3%, respectively). Conclusions: Both AI modes flagged the pulmonary nodules missed by radiologists in a significant number of cases. AI as a second reader has a high potential to improve diagnostic accuracy and radiology workflow. AI might detect certain pulmonary nodules earlier than radiologists, with a potentially significant impact on patient outcomes.

Explore related subjects

eess.IV cs.CV cs.LG

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Susanne Ohlmann-Knafo, Naglis Ramanauskas, Sebastian Huettinger, Emil Johnson Jeyakumar, Darius Barušauskas, Neringa Bielskienė, Vytautas Naujalis, Jonas Bialopetravičius, Jonas Ražanskas, Artūras Samuilis, Jūratė Dementavičienė, Dirk Pickuth. 2022-06-22. AI-based software for lung nodule detection in chest X-rays -- Time for a second reader approach?. https://arxiv.org/abs/2206.10912

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

IViT: A Novel Interpretable Visual Transformer for Skin Disease Detection

The clinical diagnosis of skin diseases is susceptible to interference from inter-class similarity of skin lesions, and over-reliance on clinicians'experience easily leads to subjective bias. Although existing deep learning aided diagnosis methods achieve competitive accuracy, they suffer from the black-box opacity of Vision Transformer (ViT) and poor adaptability to medical few-shot scenarios. Moreover, mainstream explainable algorithms generally face the bottleneck of significant accuracy degradation when improving interpretability. This paper proposes an interpretable ViT (IViT) constrained by Quadratic Programming (QP). The introduced pre-trained transfer learning adapts to few-shot feature extraction. A discrete QP feature selection framework is constructed to screen generic and discriminative features consistent with clinical diagnostic logic. A multi-objective loss function is designed to reduce feature redundancy and optimize activation distribution while preserving classification performance. Experimental results on six standard skin disease datasets show that IViT achieves an accuracy of 93.80%, only 0.21% lower than the baseline, with feature redundancy reduced by 29.5%. Its core activation regions are consistent with clinically concerned lesion areas. The proposed model balances accuracy and interpretability, providing a reliable solution for the clinical deployment of few-shot intelligent skin disease diagnosis.

eess.IV↗

arXiv2026

Automated Distinction of Intimal and Medial Intracranial Arterial Calcification from CT Head

Intracranial arterial calcifications (IACs) are a common finding on clinical non-contrast enhanced head CT scans and are associated with neurovascular disease. Calcifications can occur in the intimal or medial layer of the arterial wall, subtypes that differ in aetiology and may have distinct clinical relevance. These subtypes can be visually distinguished by radiologists based on the shape of the calcifications. We investigate three automated approaches for subtype classification of IAC from head CT-derived segmentation masks: (1) an automated adaptation of the established radiological visual score, (2) a sphericity-based method, and (3) a method based on shape embeddings extracted by a medical shape foundation model. All approaches use the same lightweight classification pipeline on top of the features they compute and are evaluated using 5-fold cross-validation. The three methods achieved comparable performance, with the embedding-based approach yielding the best overall results with a weighted F1 (mean $\pm$ SD) of up to 71.5 $\pm$ 3.7 for a single artery and 59.8 $\pm$ 1.7 for the joint artery classification. Performance was largely preserved when using automated instead of manual IAC segmentation masks, and we found the difference in weighted F1 not significant. Our results show that fully automated IAC subtype quantification from head CT is feasible and remains robust to the use of manual and automated IAC segmentation masks. Code at https://github.com/bjin96/iac-subtyping.

eess.IV↗

arXiv2026

Recasting the Destroy Step of Large Neighborhood Search as Dense Segmentation

Large neighborhood search improves an incumbent by releasing selected variables and repairing the resulting subproblem under a time limit. FOVEA casts variable selection as dense semantic segmentation on a fixed $128\times128$ canvas. A small U-Net reads twelve semantic channels and predicts cell scores, which are decoded into a variable set subject to a fixed variable budget. Family-specific layouts connect the search state to the canvas, while one set of network weights serves four problem families. The FP32 network input occupies $786$\,kB across instance sizes. Our analysis bounds constraint coupling for spatially coherent regions, gives a rank certificate for rounds with zero possible improvement, and bounds the coupling advantage attainable on an expander family. On the tested families and hardware, FOVEA trails the strongest graph encoder at $10^{4}$ variables and overtakes it near $1.5\times10^{4}$; the cost model provides an approximate crossover estimate. At larger sizes, where the graph-encoder baselines exceed device memory, FOVEA reduces the primal integral by $12$ to $16\%$ relative to the best runnable baselines. On the expander family, FOVEA performs comparably to random selection, with a coupling advantage close to one.

eess.IV↗