KEY POINTS
- This retrospective study included 247 patients with FIGO 2018 stage IIA–IVA cervical cancer treated between 2020 and 2025. IMRT was used in 91.5%, and 59.9% received concurrent chemotherapy.
- Pretreatment lymphocyte-to-monocyte ratio was calculated from routine blood counts. Median LMR was 4.68, median initial maximum tumour diameter was 4.3 cm, and median diameter-based shrinkage after EBRT was 60%.
- Tumour response was measured on MRI within one week after EBRT and before brachytherapy using the change in maximum diameter. A reduction of at least 70% was considered a marked response and occurred in 78 of 247 patients (31.6%).
- After adjustment for age, BMI, FIGO stage, tumour diameter, nodal status, concurrent chemotherapy, and EBRT dose, each one-unit increase in LMR corresponded to an absolute 1.8-percentage-point increase in tumour shrinkage (95% CI 0.1–3.6; p=0.0477).
- Each one-unit increase in LMR was associated with 11.7% higher odds of marked shrinkage (OR 1.117, 95% CI 1.002–1.356; p=0.0241). No significant non-linear relationship was detected (p=0.098).
- Standalone discrimination was poor, with an AUC of 0.579 (95% CI 0.488–0.670), sensitivity of 41.5%, and specificity of 77.5%. In an alternative logit-transformed analysis, the association was no longer statistically significant (p=0.0616).
- LMR was measured only once, HPV status was unavailable for most patients, and response was based on one-dimensional diameter rather than volumetric or pathological assessment. Selection bias and unmeasured treatment differences cannot be excluded.
CLINICAL TAKEAWAY
LMR may contribute modest information to a future multimodal response model, but it cannot identify radioresistant cervical tumours by itself. The weak effect, poor discrimination, and sensitivity to statistical modelling make treatment escalation based on this marker unjustified.