Non-Cochrane · DTA · PM41828050
Artificial Intelligence for Preoperative Prediction of Lymph Node Metastasis and Depth of Invasion in Oral Tongue Squamous Cell Carcinoma: A Systematic Review and Meta-Analysis.
Ho Yi-Yun, Hsu Chun-Wei, Chu Ta-Yi, et al. · Diagnostics (Basel, Switzerland) · 2026 · PubMed · doi:10.3390/diagnostics16050774
Findings
Population Patients with oral tongue squamous cell carcinoma (OTSCC)
32 findings
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Summary AUC for lymph node metastasis Preoperative | AUC 0.786 GRADE not reported 9 studies · 538 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Positive likelihood ratio for lymph node metastasis Preoperative | PLR 2.85 GRADE not reported 9 studies · 538 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative | Sensitivity 0.679 [0.604, 0.745] GRADE not reported 9 studies · 538 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative | Specificity 0.762 [0.705, 0.811] GRADE not reported 9 studies · 538 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Diagnostic odds ratio for lymph node metastasis Preoperative | DOR 6.76 GRADE not reported 9 studies · 538 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Negative likelihood ratio for lymph node metastasis Preoperative | NLR 0.42 GRADE not reported 9 studies · 538 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Outcome definition: OLNM (cN0-restricted) | Sensitivity 0.69 [0.575, 0.784] GRADE not reported 4 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Outcome definition: OLNM (cN0-restricted) | Specificity 0.785 [0.668, 0.868] GRADE not reported 4 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Outcome definition: general LNM (mixed cN0/cN+) | Specificity 0.769 [0.696, 0.829] GRADE not reported 5 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Outcome definition: general LNM (mixed cN0/cN+) | Sensitivity 0.678 [0.571, 0.769] GRADE not reported 5 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · AI paradigm: deep learning or hybrid models | Sensitivity 0.765 GRADE not reported 4 studies · 160 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · AI paradigm: deep learning or hybrid models | Specificity 0.832 GRADE not reported 4 studies · 160 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · AI paradigm: traditional machine learning | Sensitivity 0.652 GRADE not reported 5 studies · 378 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · AI paradigm: traditional machine learning | Specificity 0.75 GRADE not reported 5 studies · 378 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Imaging modality: MRI | Specificity 0.802 GRADE not reported 5 studies · 249 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Imaging modality: MRI | Sensitivity 0.643 GRADE not reported 5 studies · 249 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Imaging modality: CT | Sensitivity 0.806 GRADE not reported 3 studies · 249 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Imaging modality: CT | Specificity 0.749 GRADE not reported 3 studies · 249 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Input type: multimodal | Sensitivity 0.714 [0.569, 0.826] GRADE not reported 6 studies · 221 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Input type: multimodal | Specificity 0.815 [0.744, 0.871] GRADE not reported 6 studies · 221 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Input type: unimodal | Specificity 0.712 [0.635, 0.778] GRADE not reported 3 studies · 317 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Input type: unimodal | Sensitivity 0.666 [0.558, 0.76] GRADE not reported 3 studies · 317 participants |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Summary AUC for lymph node metastasis Preoperative · Sensitivity analysis: extended analysis adding Committeri and Csury (k = 11) | AUC 0.828 GRADE not reported 11 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Sensitivity analysis: extended analysis adding Committeri and Csury (k = 11) | Sensitivity 0.688 GRADE not reported 11 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Sensitivity analysis: extended analysis adding Committeri and Csury (k = 11) | Specificity 0.809 GRADE not reported 11 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Sensitivity analysis: OLNM subgroup excluding Kubo (delayed-detection definition) | Sensitivity 0.685 GRADE not reported 3 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Sensitivity analysis: OLNM subgroup excluding Kubo (delayed-detection definition) | Specificity 0.787 GRADE not reported 3 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Sensitivity analysis: excluding Lan (included oropharyngeal SCC) | Sensitivity 0.664 GRADE not reported 8 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Sensitivity analysis: excluding Lan (included oropharyngeal SCC) | Specificity 0.77 GRADE not reported 8 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Summary AUC for lymph node metastasis Preoperative · Sensitivity analysis: excluding Lan (included oropharyngeal SCC) | AUC 0.766 GRADE not reported 8 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Sensitivity for lymph node metastasis Preoperative · Sensitivity analysis: leave-one-out (nine iterations) | stable pooled estimates across all nine iterations (sensitivity range: 0.657-0.697; specificity range: 0.743-0.787) GRADE not reported 8 studies |
| AI-based (machine learning or deep learning) prediction models applied to imaging or histopathology | Histopathologic reference standard (single-arm diagnostic accuracy) | Specificity for lymph node metastasis Preoperative · Sensitivity analysis: leave-one-out (nine iterations) | stable pooled estimates across all nine iterations (sensitivity range: 0.657-0.697; specificity range: 0.743-0.787) GRADE not reported 8 studies |
Included primary studies 12
Open in Studies| Study | Year | Also included by | Outcomes it informs here | Trial data |
|---|---|---|---|---|
| Wang F 2022 Magnetic Resonance Imaging-Based Radiomics Features Associated with Depth of Invasion Predicted Lymph Node Metastasis and Prognosis in… PMID:34888985 | 2022 | 2 other reviews | not stated | not extracted yet |
| Yuan 2021 Machine learning-based MRI texture analysis to predict occult lymph node metastasis in early-stage oral tongue squamous cell carcinoma. PMID:33569617 | 2021 | 2 other reviews | not stated | not extracted yet |
| Committeri 2022 Radiomics Metrics Combined with Clinical Data in the Surgical Management of Early-Stage (cT1-T2 N0) Tongue Squamous Cell Carcinomas: A… PMID:35336841 | 2022 | 1 other review | not stated | not extracted yet |
| Kubo 2022 Development of a radiomics and machine learning model for predicting occult cervical lymph node metastasis in patients with tongue cancer. PMID:35431177 | 2022 | 1 other review | not stated | not extracted yet |
| Kudoh 2023 Radiomics analysis of [18F]-fluoro-2-deoxyglucose positron emission tomography for the prediction of cervical lymph node metastasis in… PMID:35254609 | 2023 | 1 other review | not stated | not extracted yet |
| Lan 2024 MRI-based deep learning and radiomics for prediction of occult cervical lymph node metastasis and prognosis in early-stage oral and… PMID:38729119 | 2024 | 1 other review | not stated | not extracted yet |
| Vidiri 2023 Magnetic resonance imaging-based prediction models for tumor stage and cervical lymph node metastasis of tongue squamous cell carcinoma. PMID:37701020 | 2023 | 1 other review | not stated | not extracted yet |
| Wang D 2024 Magnetic resonance imaging-based radiomics and deep learning models for predicting lymph node metastasis of squamous cell carcinoma of the… PMID:38378316 | 2024 | 1 other review | not stated | not extracted yet |
| Zhong 2022 Tumor radiomics signature for artificial neural network-assisted detection of neck metastasis in patient with tongue cancer. PMID:34358534 | 2022 | 1 other review | not stated | not extracted yet |
| Csury 2024 The modified Polsby-Popper score, a novel quantitative histomorphological biomarker and its potential to predict lymph node positivity and… PMID:38132808 | 2024 | this review only | not stated | not extracted yet |
| Han 2024 A CT-based integrated model for preoperative prediction of occult lymph node metastasis in early tongue cancer. PMID:38685941 | 2024 | this review only | not stated | not extracted yet |
| Liu 2024 Prediction of Prognosis of Tongue Squamous Cell Carcinoma Based on Clinical MR Imaging Data Modeling. PMID:37872687 | 2023 | this review only | not stated | not extracted yet |
Overlap with other reviews
Reviews that include the same primary studies are not independent evidence. Shared counts match studies by PMID, DOI or NCT, and otherwise by author and year (a probable match). Only the non-Cochrane dental reviews have study lists, so overlap with Cochrane reviews is not counted yet.
Review question
| Element | Planned (methods) | Found (results) |
|---|---|---|
| Population | Patients with oral tongue squamous cell carcinoma | OTSCC patients (n=25 to 319 per study; 12 studies, 9 pooled) |
| Intervention | Machine learning or deep learning models applied to MRI, CT, PET/CT or histopathological whole-slide imaging | Traditional machine learning (n=7) and deep learning or hybrid models (n=5) on MRI (n=5), CT (n=3), PET/CT (n=1), histopathology (n=1) |
| Comparison | Histopathologic reference standard for depth of invasion and/or nodal status | Postoperative histopathology (12 studies) |
| Outcomes | Depth of invasion or lymph node metastasis prediction; sensitivity, specificity, AUC | Lymph node metastasis (OLNM k=4, general LNM k=5 pooled); DOI/pT narrative (1 study) |
| Study designs | Original studies; English only; excluded case reports, editorials, letters, reviews, meta-analyses, conference abstracts | Retrospective diagnostic accuracy studies (12; 9 with extractable 2x2 data) |
| Timing | Preoperative | Preoperative |