KEY POINTS
- Five cervical cancer cell lines—CaSki, HeLa, MS-751, SiHa, and SW-756—were exposed to 1% versus 21% oxygen for 24 hours across three passages. Investigators identified 402 differentially expressed genes, narrowed these to 61 consistently regulated candidates, and mapped 55 genes into the final classifier.
- The signature included established hypoxia-response genes such as CA9, VEGFA, BNIP3, EGLN1, LDHA, and HK2. Cellular response to hypoxia showed 45.2-fold enrichment (p=1.87×10⁻⁹), alongside glycolytic and metabolic pathways.
- The classifier was trained using 71 TCGA patients and evaluated in an internal validation subset. Hypoxic tumours had worse overall survival (p=0.0049) and progression-free survival (p=0.0063); hypoxia remained independently associated with mortality after stage adjustment (HR 2.69, 95% CI 1.29–5.61; p=0.009).
- The Manchester validation cohort included 153 pretreatment FFPE biopsies, of which 57 were classified as hypoxic and 96 as normoxic. Hypoxia was associated with advanced stage, tumour size ≥4 cm, pelvic nodal involvement, and hydronephrosis after correction for multiple comparisons.
- In Manchester, hypoxia predicted worse progression-free survival (p=0.042) and overall survival (p=0.017). It remained independently prognostic for mortality after adjustment for clinical factors (HR 1.95, 95% CI 1.08–3.51; p=0.026), including within the 129 patients treated with primary radiotherapy.
- Adding the signature to clinical covariates increased the C-statistic from 0.716 to 0.740 (likelihood-ratio p=0.01). External validation also showed worse outcomes among hypoxic tumours in Seoul (n=300, disease-free survival p=0.0025) and Norway (n=283, progression-free survival p=0.0062; overall survival p=0.0033).
- In the Norwegian cohort, classification agreed with an independently developed six-gene hypoxia signature in 202 of 283 tumours (71%). The study used inferred rather than directly measured hypoxia, bulk transcriptomics, retrospective cohorts, and non-locked assay thresholds; it did not show that the signature predicts benefit from hypoxia-modifying therapy.
CLINICAL TAKEAWAY
The signature may eventually identify cervical cancers with a biologically aggressive, hypoxia-associated phenotype using routine pretreatment biopsy material. It is currently a prognostic research classifier—not a basis for intensifying chemoradiotherapy, adding hypoxia-directed treatment, or changing brachytherapy strategy.