Diagnostic accuracy of deep learning for the invasiveness assessment of ground-glass nodules with fine segmentation: a systematic review and meta-analysis

Abstract

Background: Accurate recognition of invasive lung adenocarcinoma (IAC) presenting as ground-glass nodules (GGNs) is crucial for guiding clinical decision-making and timely surgical intervention. This study aimed to systematically evaluate the diagnostic accuracy of deep learning (DL) models via fine nodule segmentation in assessing the invasiveness of lung adenocarcinoma.

Methods: Literature from the inception of the PubMed, Embase, Cochrane Library, and Web of Science databases was searched. Studies related to DL and nodule segmentation in diagnosing IAC were evaluated and included. Titles and abstracts were screened, and the Quality Assessment of Diagnostic Accuracy Studies 2 was used to assess the quality of the selected studies. The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) criteria of diagnostic tests were used to assess the certainty of evidence.

Results: Eight studies involving 5,281 nodules and 4,676 patients were included and analyzed. Meta-analysis showed that the combined sensitivity of DL for the diagnosis of IAC was 0.81 [95% confidence interval (CI): 0.73-0.87], while the specificity was 0.86 (95% CI: 0.80-0.90). The area under the summary receiver operating characteristic (SROC) curve was 0.90 (95% CI: 0.88-0.93), but the overall quality of the evidence was suboptimal.

Conclusions: DL and nodule segmentation demonstrated high accuracy in assessing lung adenocarcinoma invasiveness, but the certainty of the associated evidence was low. More large-scale, multicenter, high-quality diagnostic accuracy studies are needed to validate the performance and usefulness of DL in the assessment of lung adenocarcinoma invasiveness.

Keywords: Artificial intelligence (AI); computed tomography (CT); deep learning (DL); invasive lung adenocarcinoma (IAC); pulmonary nodules.

Figure 1

PRISMA 2020 flow diagram of the selection process of the included studies. PRISMA, Preferred Reporting Systematic Assessment and Meta-Analysis.

Figure 2

Application of DL with precise nodule segmentation in lung adenocarcinoma imaging. CT, computed tomography; HE, hematoxylin-eosin; 3D, three-dimensional; IAC, invasive adenocarcinoma; MIA, microinvasive adenocarcinoma; AAH, adenomatous atypical hyperplasia; AIS, adenocarcinoma in situ; ReLU, rectified linear unit; MPL, multilayer perceptron; ROC, receiver operating characteristic; DL, deep learning.

Figure 3

A summary of the risk of bias and applicability concerns for each QUADAS-2 domain presented as percentages across the eight studies included in the analysis. QUADAS-2, Quality Assessment of Diagnostic Accuracy Studies 2.

Figure 4

Deeks’ funnel plot of publication bias. ESS, effective sample size.

Figure 5

Forest plots of sensitivity and specificity in the eight included studies. FN, false negative; FP, false positive; TP, true positive; TN, true negative; CI, confidence interval.

Figure 6

Forest plots of deep learning-based imaging for invasiveness assessment of ground-glass nodules with fine segmentation. CI, confidence interval.

Figure 7

Summary receiver operating characteristic curves of the included studies. The solid black circle represents the summary point, and the outlined circles represent the individual studies. The dotted and dashed lines indicate the 95% confidence and 95% prediction regions, respectively. SENS, sensitivity; SPEC, specificity; SROC, summary receiver operating characteristic; AUC, area under the receiver operating characteristic curve.

Figure 8

Fagan’s nomogram for the results of clinical application. LR, likelihood ratio; Prob, probability; Pos, positive; Neg, negative.

Figure 9

Sensitivity analysis of the leave-one-out method. CI, confidence interval.

Figure 10

Subgroup analysis. (A) Nodule segmentation method. Group 0: manual; Group 1: automatic. (B) Use of public dataset. Group 0: no; Group 1: yes. (C) Sample size of the validation and test set. Group 0: sample size <150; Group 1: sample size ≥150. (D) External validation. Group 0: no; Group 1: yes. Weights and the between-subgroup heterogeneity test are from the Mantel-Haenszel model. CI, confidence interval; MH, Mantel-Haenszel; RR, risk ratio.