Diagnostic performance of machine learning and deep learning algorithms for thyroid cancer metastasis: a systematic review and meta-analysis.

This moderator had a strong and highly significant effect (F (1,160) = 11.19, p = 0.0010).

also in 132,142 other papers

While the analysis was borderline significant (F (1) = 2.93, p = 0.0867), Radiomics-only models showed a trend toward reduced performance (DOR: 8.254, p = 0.089), with a slight reduction in between-study variance (σ2 = 0.328).

also in 11,409 other papers

However, this model did not reach statistical significance (F (5,156) = 1.60, p = 0.164), and the individual coefficients showed wide confidence intervals with considerable overlap, indicating substantial heterogeneity even within cancer subtypes.

also in 111,027 other papers

DL) The model category variable, differentiating between machine learning (ML) and deep learning (DL), did not yield a significant moderating effect (F (1) = 1.86, p = 0.1722), although machine learning models showed a trend toward better performance (DOR: 10.625, p:0.174).

also in 53,322 other papers