KEY POINTS
- This retrospective feasibility study evaluated 40 MRI datasets from 20 patients treated with definitive chemoradiotherapy: 19 had cervical cancer and one vault cancer, 95% had locally advanced disease, 50% had pelvic or para-aortic nodal involvement and 85% had squamous-cell carcinoma.
- Treatment followed an EMBRACE II-based approach with pelvic external-beam RT to 45 Gy in 25 fractions, involved-node boosting to 55–57.5 Gy, weekly concurrent chemotherapy and image-guided brachytherapy delivered as 7 Gy × 4 fractions over two applications. Thirteen of 20 patients (65%) required combined intracavitary/interstitial brachytherapy.
- Gross tumor volume was independently generated using conventional T2-weighted MRI alone and T2 combined with qualitative diffusion-weighted imaging and ADC information. DWI-positive tumor required diffusion hyperintensity with corresponding ADC hypointensity; contours were generated by consensus between a radiologist and radiation oncologist.
- Before external-beam treatment, median GTV was 35.25 cm³ on T2 versus 23.75 cm³ with T2/DWI, but the paired difference was not statistically significant (p=0.117) and volumes correlated strongly (ρ=0.943; p<0.0001). Fourteen of 20 patients had smaller DWI-informed volumes, five larger volumes and one equal volume.
- Geometry nevertheless differed even at baseline. Median maximum right-sided lateral tumor extent was 2.63 cm with T2 versus 2.05 cm with T2/DWI (p<0.0001). Overall spatial agreement remained moderate, with median Dice 0.75, Jaccard 0.61 and volume ratio 0.90; 75% of paired lateral-extent measurements differed by ≤5 mm.
- The largest discrepancy appeared before brachytherapy, after chemoradiation. Median residual GTV was 1.50 versus 1.05 cm³ (p=0.017), while spatial agreement fell sharply to Dice 0.33, Jaccard 0.19 and volume ratio 0.79 despite a still-moderate correlation between absolute volumes (ρ=0.707; p=0.0005). Approximately 65% of cases had smaller residual volumes with DWI/ADC.
- Maximum right lateral extent before brachytherapy was also larger on T2 alone at 0.95 versus 0.81 cm (p=0.033). The authors propose that post-RT inflammation and fibrosis may contribute to abnormal T2 signal without representing viable tumor, but the study did not establish pathology, recurrence mapping or brachytherapy dose-response as ground truth.
CLINICAL TAKEAWAY
The important finding is not that DWI simply makes cervical GTVs smaller: after chemoradiotherapy, T2 and DWI/ADC could describe quite different residual tumor shapes despite similar absolute volumes. That could matter in brachytherapy with steep dose gradients, but 20 retrospective patients are insufficient to change current MRI-based IGABT contouring standards.
SOURCE
Technical Innovations & Patient Support in Radiation Oncology