Session Item

To investigate the effect of brachytherapy (BT) applicator and implant type on morbidity and local control in locally advanced cervix cancer treated on the EMBRACE I study.

1071 patients treated with tandem&ring (T&R) (n=725) and tandem&ovoids (T&O) (n=346) from 19 EMBRACE-I centres were analyzed. After receiving EBRT (45-50Gy in 25-30fx) and MR-guided BT (IGBT) using intracavitary (IC) (n=559) or intracavitary/interstitial (IC/IS) (n=512) implants, patients were prospectively followed (in median 48 months). Local control (LC) and physician-assessed (CTCAE v.3) late morbidity were compared based on (1) applicator type: T&R vs T&O, and (2) implant type: IC vs IC/IS. Moderate-to-severe (G≥2) genito-urinary (GU: cystitis, frequency), gastro-intestinal (GI: proctitis, bleeding, diarrhea) and vaginal (stenosis, mucositis) symptoms were analysed separately. Severe-to-life-threatening events (G≥3) were pooled for evaluation of GU, GI and vaginal morbidity. Comparisons between applicator and implant types were evaluated by Cox proportional hazard multiple regression model, adjusting for: patient (baseline morbidity, age, body mass index), disease (local FIGO stage, organ involvement, tumour dimensions, histology, tumour necrosis), and treatment-related confounders (CTV-HR D90%, EBRT dose) (Table 1), which were included in the multivariable analysis (MVA) if significant in the univariate analysis (p ≤ 0.15).

OAR doses (bladder and rectum D2cm3; ICRU points) for T&O vs T&R were higher by 5-7Gy in IC implants and by 1-5Gy in IC/IS implants, while target dose (CTV-HR D90%) was lower by on average 2.5Gy in T&O compared to T&R applicators. Table 1 shows the hazard-ratios (HR) for individual G≥2 morbidity endpoints and pooled G≥3 GU/GI/vaginal/fistula morbidity in patients treated with T&O IC vs. T&R IC, T&O IC/IS vs. T&R IC/IS and IC vs. IC/IS applicators. In 3/7 and 6/7 evaluated individual symptoms (in IC and IC/IS implants, respectively), the T&O showed a statistically significant (p<0.05) higher risk for G≥2 morbidity compared to T&R. Figure 1 shows examples of adjusted cumulative hazard of individual endpoints (G≥2 cystitis, proctitis and vaginal stenosis). Crude incidence of local failure was 7.3% (25/343) and 6.6% (47/712) in patients treated with T&O and T&R, respectively. In MVA, local control did not statistically differ between the two groups (p>0.1).

In this patient cohort treated between 2008-2015, T&O applicators were associated with higher OAR and lower target doses as compared to T&R. The risk for G≥2 and G≥3 morbidity was higher for patients treated with T&O, while local control was similar. For IC/IS-treated large CTV-HR volumes >35cm3, target doses were higher by 3.5Gy compared to IC applicators. Nevertheless, combined IC/IS applicators did not increase the risk of morbidity compared to IC applicators.