Lihua Tan

Objective Perineural invasion (PNI) is an adverse feature in cervical cancer and may influence nerve-sparing surgery. We compared conventional radiomics and VGG-SAM deep-learning strategies for MRI-based preoperative prediction of PNI. Methods A retrospective cohort of 103 patients with cervical cancer, including 82 PNI-negative and 21 PNI-positive patients, was analyzed. PyRadiomics features were extracted from raw DICOM MRI and paired tumor masks. Four deep variants were evaluated: VGG only, SAM only, VGG + SAM naive fusion, and VGG + SAM learnable fusion. Patient-level stratified five-fold cross-validation was used. Primary metrics were sensitivity, specificity, balanced accuracy, F1-score, ROC-AUC, and PR-AUC. Uncertainty for the naive-versus-learnable comparison was assessed with 10,000 paired bootstrap resamples and exact McNemar tests. Results The strongest radiomics baseline, PyRadiomics + XGBoost, achieved balanced accuracy 0.6057, F1-score 0.3684, ROC-AUC 0.6731, and PR-AUC 0.4460. Naive VGG + SAM fusion achieved the best sensitivity (0.5714), balanced accuracy (0.7491), F1-score (0.6154), ROC-AUC (0.7854), and PR-AUC (0.6231). Learnable fusion achieved the highest specificity (0.9634) and precision (0.7500) while matching the highest accuracy (0.8544). Paired bootstrap comparisons showed wide confidence intervals across clinically relevant metrics. Conclusion Hybrid VGG-SAM modeling outperformed conventional radiomics alone, but greater fusion complexity did not yield uniform benefit. Naive fusion favored sensitivity-oriented discrimination, whereas learnable fusion mainly shifted the operating point toward higher specificity. Larger multi-institutional cohorts and clinical comparators are needed.

Background: Perineural invasion (PNI) is associated with aggressive tumor behavior in several malignancies, but its independent prognostic value in early-stage cervical cancer remains uncertain. We evaluated the clinical significance of PNI and explored molecular and immune features associated with PNI. Methods: We retrospectively analyzed 499 patients with FIGO 2009 stage IB–IIA cervical cancer treated with radical hysterectomy and pelvic lymphadenectomy. Associations between PNI, clinicopathological variables, recurrence-free survival, and overall survival were assessed using Kaplan–Meier methods and Cox regression. An independent cohort of 286 cervical cancers from The Cancer Genome Atlas (TCGA) was analyzed to characterize PNI-associated transcriptomic patterns, pathway enrichment, immune cell composition, and microRNA profiles. Results: PNI was identified in 11.6% of cases and was associated with larger tumor size, deep stromal invasion, and lymphovascular space invasion. PNI was not an independent prognostic factor in the overall cohort; however, it was associated with increased recurrence risk in the subgroup without high-risk factors and not meeting Sedlis criteria, with a modest improvement in 5-year recurrence discrimination when incorporated into Sedlis-based models. In TCGA, PNI was associated with differential gene expression and enrichment of oncogenic and immune-related pathways, an increased estimated abundance of resting mast cells, and six differentially expressed microRNAs. Conclusions: In early-stage cervical cancer, PNI is strongly correlated with established adverse pathological features and shows a subgroup-specific association with recurrence in an otherwise low-risk postoperative population. The multi-omics findings are exploratory and support biological hypotheses regarding tumor–nerve–immune interactions; external validation is needed before PNI can be used to guide postoperative management.