Optimal radiological gallbladder lesion characterization by combining visual assessment with CT-based radiomics

Yin, Yunchao; Yakar, Derya; Slangen, Jules J. G.; Hoogwater, Frederik J.H.; Kwee, Thomas C.; Haas, Robbert J. de

doi:10.1007/s00330-022-09281-6

Cited by 6 publications

(8 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Various machine-learning models were built based on extracted radiomic features to differentiate GBC and benign gallbladder lesions. The specificity of the radiomic analysis achieved 80%, but the sensitivity was merely 64% [6]. In addition, when including both the gallbladder and adjacent liver parenchyma in the radiomic analysis, the diagnostic performance did not significantly improve.…”

Section: Introductionmentioning

confidence: 88%

“…Approval from the institutional review board was obtained, and the need for written informed consent was waived. The current study population was also part of two previous studies which were focused on different research questions [5,6].…”

Section: Study Populationmentioning

confidence: 99%

“…In a previous study, the best results for differentiation between GBC and benign gallbladder disease were observed when combining CT-based radiomic analysis with radiological visual interpretation [6]. To determine whether the results of the CNN could also improve the radiological visual interpretation, an additional analysis was performed combining the CNN prediction with radiological visual interpretation.…”

Section: Combining Convolutional Neural Network Prediction With Radio...mentioning

confidence: 99%

“…However, the specificity was merely 60% [5]. Another recent study adopted the quantitative approach of radiomic analysis to evaluate gallbladder lesions [6]. Various machine-learning models were built based on extracted radiomic features to differentiate GBC and benign gallbladder lesions.…”

Section: Introductionmentioning

confidence: 99%

“…In addition, when including both the gallbladder and adjacent liver parenchyma in the radiomic analysis, the diagnostic performance did not significantly improve. The Diagnostics 2023, 13, 704 2 of 10 best results were obtained when combining CT-based radiomics with visual radiological assessment [6].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

The Value of Deep Learning in Gallbladder Lesion Characterization

et al. 2023

Self Cite

View full text Add to dashboard Cite

Background: The similarity of gallbladder cancer and benign gallbladder lesions brings challenges to diagnosing gallbladder cancer (GBC). This study investigated whether a convolutional neural network (CNN) could adequately differentiate GBC from benign gallbladder diseases, and whether information from adjacent liver parenchyma could improve its performance. Methods: Consecutive patients referred to our hospital with suspicious gallbladder lesions with histopathological diagnosis confirmation and available contrast-enhanced portal venous phase CT scans were retrospectively selected. A CT-based CNN was trained once on gallbladder only and once on gallbladder including a 2 cm adjacent liver parenchyma. The best-performing classifier was combined with the diagnostic results based on radiological visual analysis. Results: A total of 127 patients were included in the study: 83 patients with benign gallbladder lesions and 44 with gallbladder cancer. The CNN trained on the gallbladder including adjacent liver parenchyma achieved the best performance with an AUC of 0.81 (95% CI 0.71–0.92), being >10% better than the CNN trained on only the gallbladder (p = 0.09). Combining the CNN with radiological visual interpretation did not improve the differentiation between GBC and benign gallbladder diseases. Conclusions: The CT-based CNN shows promising ability to differentiate gallbladder cancer from benign gallbladder lesions. In addition, the liver parenchyma adjacent to the gallbladder seems to provide additional information, thereby improving the CNN’s performance for gallbladder lesion characterization. However, these findings should be confirmed in larger multicenter studies.

show abstract

Section: Introductionmentioning

confidence: 88%

Section: Study Populationmentioning

confidence: 99%

Section: Combining Convolutional Neural Network Prediction With Radio...mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

The Value of Deep Learning in Gallbladder Lesion Characterization

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

The added value of the visual analysis of DWI in post-surgery follow-up of soft tissue sarcoma of the extremities: do we really need ADC?

et al. 2023

View full text Add to dashboard Cite

Introduction MRI has a fundamental role in the follow-up of soft tissue sarcomas (STSs). However, the differentiation of recurrences/residual disease from post-surgical changes is a complex task, with a central role for the radiologist. Materials and methods We retrospectively evaluated 64 post-surgery MRI for extremities STSs. MR protocol included DWI (b = 0, 1000). Two radiologists were asked to consensually evaluate: presence/absence of tumoral nodules, lesion conspicuity, imaging diagnostic confidence, ADC values, and DWI overall image quality. The gold standard was histology or MR follow-up. Results Thirty-seven lesions in 29/64 patients were confirmed as local recurrence or residual disease (n = 16 ≤ 1 cm) with 1 MR false positive. On DWI, the conspicuity of the proved tumor lesions resulted excellent in 29/37, good in 3/37 and low in 5/37, higher than conventional imaging. A statistically significant higher diagnostic confidence of DWI compared to conventional imaging (p < 0.001) and DCE (p = 0.009) was observed. In the 37 histologically confirmed lesions, mean ADC value was 1.31 × 10–9 m2/s. Overall scar tissues mean ADC was 1.70 × 10–9 m2/s. DWI quality resulted adequate in 81% and unsatisfactory in 5%. Conclusions In this highly heterogeneous group of tumors, the role of ADC seems to be limited. Based on our experience, looking at DWI images makes the lesions promptly and easily detectable. This technique gives less deceptive findings making the reader more confident in detecting/excluding tumoral tissue; the main drawback is the image quality and the lack of standardization.

show abstract