Session Item

Despite numerous experiences published in the literature, few clinical trials today use radiomics to personalize cancer treatment. This is because radiomic models generally use complex parameters and advanced techniques, that do not allow a clear interpretation of the mechanism behind the prediction: the low interpretability leads clinicians to not trust these models. As a consequence, new biomarkers are emerging that are simple to calculate and based on robust radiobiological theories: Early Regression Index (ERI), is an image-based biomarker that has recently reported interesting results in predicting pathological complete response (pCR) after neoadjuvant chemoradiotherapy (nCRT) in case of rectal cancer. Such parameter is generally calculated on T2 MR images and consists in modelling the early tumour regression combining the GTV volume measured during simulation (V_pre) and at mid therapy (V_mid).

ERI=-ln(1-(Vmid/Vpre)^Vpre))
This study aims to evaluate the feasibility of using ERI in Locally Advanced Cervical Cancer (LACC), evaluating its ability in predicting pCR not only considering T2-w MR images, but also DWI

A total of 88 patients affected by LACC (FIGO IB2-IVA) were enrolled. All the patients underwent nCRT, combining weekly 40 mg/m² of cisplatin with concurrent RT, prescribing 50.6 Gy to PTV1 (CTV1+5 mm) and 39.6 Gy to PTV2 (CTV2+7mm) in 22 fractions. An MRI protocol consisting in two acquisitions (T2-w and DWI) in two times (before treatment and at mid therapy) was applied. GTV was delineated and ERI was calculated for both imaging modalities. A radical hysterectomy was performed for each patient within 8 weeks after nCRT: pCR was considered in case of absence of any residual tumour cells at any site (pR0). The ERI performance in identifying pCR patients was quantified calculating the area (AUC) under the Receiver Operating Characteristic (ROC) curve. The best threshold value was obtained maximizing the Youden Index and the values of sensitivity and specificity were calculated at this threshold.

The ROC curves obtained for ERI calculated starting from the volumes measured on T2-w MR images (ERI_T2) and on DWI images (ERI_DWI) are reported in figure 1.

The performance of ERI_DWI (AUC=0.81 with 95%CI of 0.70-0.91) are superior to those reported by ERI_T2 (AUC=0.76 with 0.65-0.87 as 95% CI). At the best cut-off threshold, ERI_T2 shows high specificity (97,4%) with low sensitivity (43%), while ERI_DWI high sensitivity (86.5%) and limited specificity (64.1%).

This study confirmed ERI as a good biomarker also in case of LACC, with higher performance in DW images. In particular, the high value of specificity of ERI_DWI demonstrates high reliability in early identifying patients who will not go through a complete response, so allowing to the clinicians to modify the treatment approach in time. An external validation study is required before to implement such biomarker in clinical practice