PET-CT radiomics and artificial intelligence for predicting lymph node metastasis in cervical cancer: a systematic review and meta-analysis.
Xiaoyu Liu, Junqiang Lei
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and meta-analysis assessed PET/CT-based radiomics and artificial intelligence models for predicting lymph node metastasis in cervical cancer. Nine studies involving 1,268 patients were included, with diagnostic performance pooled using random-effects models. The models had pooled sensitivity of 0.90 (95% CI 0.84-0.93), specificity of 0.88 (95% CI 0.74-0.95), and an AUC of 0.92 (95% CI 0.90-0.94). The authors also reported modest methodological quality and identified methodological heterogeneity and limited high-quality evidence as barriers to clinical translation.
Key findings
- Across nine studies and 1,268 patients, pooled sensitivity was 0.90 (95% CI 0.84-0.93) and specificity was 0.88 (95% CI 0.74-0.95).
- The pooled AUC was 0.92 (95% CI 0.90-0.94) for PET/CT-based radiomics-AI models.
- Mean Radiomics Quality Score was 10.5; semi-automated segmentation and larger sample sizes were associated with improved diagnostic stability.
Why it matters
Preoperative assessment of lymph node metastasis can inform therapeutic planning in cervical cancer, and these pooled results suggest PET/CT radiomics-AI models may offer useful non-invasive diagnostic information. However, pooled accuracy does not establish reliability across settings: methodological quality and heterogeneity remain concerns, so the evidence supports further validation rather than assuming routine clinical performance.
What to keep in mind
The evidence comprised nine studies and had modest methodological quality (mean RQS: 10.5); the authors also identify methodological heterogeneity and a lack of high-quality studies.