Role of MRI with diffusion-weighted imaging (DWI) in patients with uterine cervical cancer for intracavitary brachytherapy treatment alignment
Role of MRI with diffusion-weighted imaging (DWI) in patients with uterine cervical cancer for intracavitary brachytherapy treatment alignment
Chiara Solmi, Federica Fiocchi, Anna Scarpa, Maria Chiara Gibertini, Giuseppina Demarco, Annarita Pecchi, Pietro Torricelli
Cervical cancer (CC) is the fourth most common cancer among women worldwide despite prevention programs [1]. Intracavitary brachytherapy (BRT) is a crucial component for achieving radical treatment of locally advanced cervical cancer (LACC) (2018 FIGO stages IB3-IVA) and MRI has a key role for its planning, allowing dose escalation to the target volumes (HR-CTV and GTV) while sparing organ at risk (OAR) [2]. Our study aims to demonstrate the dimensional reduction of the lesion after external beam radiotherapy (EBRT) and intracavitary BRT and to evaluate tumour response using DWI sequences with quantification of the apparent diffusion coefficient (ADC) mean value.
1. Patient Selection This ongoing retrospective and prospective study, approved by the ethics committee, includes 17 patients who completed CC intracavitary BRT from July 2021 until June 2023 at the Policlinico of Modena. 2. MRI Protocol MRI examinations on 1.5T scanner (MR SIGNA™ Voyager, GE healthcare) with a 5-channel external phased array torso coil were performed before and after EBRT, prior to each BRT session (one to four) and at the end of treatment. The exams acquired after applicator (tandem and ovoids) implantation included a panoramic view axial T2 weighted image (WI) (slice thickness 3 mm, interslice gap 1 mm, FOV 34 × 34 mm, matrix 292 × 192, number of excitations 2), sagittal T2WI (slice thickness 4 mm, interslice gap 1 mm, FOV 34 × 34 mm, matrix 292 × 292, number of excitations 2), axial oblique T2WI (perpendicular to the cervix; slice thickness 3 mm, interslice gap 0 mm, FOV 34 × 34 mm, matrix 352 × 352, number of excitations 2) and axial oblique DWI (b-values 50, 400, 1000 s/mm2). 3. Image Analysis Three radiologists with varying experience levels independently analysed the MR images by calculating tumour volume multiplying the three main diameters (measured using axial oblique DWI-MRI and sagittal T2WI) by a constant (0.523) and by estimating lowest ADC value drawing a freehand region of interest (ROI) on the lesion (figure 1). After analysing the images, a final consensus was reached. Pre-BRT session axial oblique T2WI were used by radiation oncologists as a reference for contouring the high risk – clinical target volume (HR-CTV) and Gross Tumour Volume (GTV) (figure 2) as per the GEC ESTRO recommendations [3]. 4. Statistical Analysis First, to test the repeatability of image analysis the correlations in pre-BRT session MR images between volumes measured by radiologists, HR-CTV and GTV were studied using multiple Pearson correlations. Furthermore, to check how the volumes and ADC values evolved over time during therapy, two different repeated measures ANOVAs (rm-ANOVA) were performed using time as an independent variable and volumes / ADC values as dependent variables. Regarding “time” as a variable, it was differentiated in three different stages: at primary staging (T1), after EBRT (T2) and at follow-up after BRT (T3). Paired t-tests were used for post-hoc comparisons. Additionally, ADC values and volumes between responders and relapsing patients were compared using independent samples t-tests both at the beginning and at the end of therapy. Finally, using the same t-test, the comparison between the difference both in volume and ADC after therapy between the two studied groups was evaluated.
Figure 1. 53-year-old woman with FIGO stage III CC. First staging MRI: the lesion has 7,17 cm of cranio-caudal extension measured on the sagittal T2WI slice (A), axial diameters of 5,25 cm (latero-lateral) and 5,04 cm (antero-posterior) on axial oblique DWI (B) and ADC mean value of 0,70 × 10−3 mm2 (C).
Figure 2. Brachytherapy planning with MRI of the same patient after EBRT.
Among the 17 patients (mean age 63 years-old, SD = 16 years) who completed BRT treatment, seven patients underwent all four MRI alignments, while ten patients also had some BRT alignments made with Computed Tomography interposed. In only one case there was a complication during the applicator implantation (figure 3), consisting in cervical perforation by the uterine tandem, subsequently resolved without any serious consequences at the next MRI. Concordance was found between the visible lesion volume measured by radiologists and by radiation oncologists (GTV), with a significant positive correlation (r(26) = 0.69, p < 0.001). The HR-CTV contains the demonstrable GTV plus a margin for subclinical tumour spread [3]: consistently with this, we found a significant positive correlation between GTV and HR-CTV (r(26) = 0.61, p = 0.001). Additionally, a significant positive correlation was found between tumour volume and HR-CTV (r(26) = 0.51, p = 0.006). To study the evolution of volume and ADC during the treatment, a subgroup of the thirteen patients who underwent follow-up MRI (T3) was considered. The rm-ANOVA performed with volume as a dependent variable, showed a significant main effect of time (F(2) = 20.78, p < 0.001). The t-test revealed that the studied variable is significantly different at T1 and at the T2 (t(12) = 4.259, p = 0.001), at T1 and T3 (t(12) = 4.906, p = < 0.001), and at T2 and T3 (t(12) = 3.099, p = 0.009), showing a decreasing trend overtime (figure 4). Therefore, the therapy seems to be effective in reducing tumour volume both at T2 (after EBRT) and T3 (after BRT). As it may be observed in the graphic (figure 4), the decrease in volume after EBRT seems steeper than after BRT, consistently with the idea that EBRT is primarily directed at massive reduction of tumour volume [4] while BRT to the radicality concerning parametrial invasion and reduction of recurrence incidence [2]. The second rm-ANOVA performed with ADC as a dependent variable showed the same significant main effect of time (F(2) = 13.13, p < 0.001). However, in this case the pattern of results is different: the t-test revealed that the ADC values were significantly different between T1 (M = 0.71 × 10−3 mm2/s, SD = 0.06 × 10−3 mm2/s) and T2 (M = 1.00 × 10−3 mm2/s, SD = 0.13 × 10−3 mm2/s) (t(12) = -6.374, p < 0.001) and T1 and T3 (M = 1.29 × 10−3 mm2/s, SD = 0.50 × 10−3 mm2/s) (t(12) = -4.330, p = 0.001), while the difference between T2 and T3 was not significant (t(12) = -2.146, p = 0.053), possibly due to the small sample size. In this instance, as expected from previous findings [5], the trend is opposite compared to the volume: the value of ADC increases somehow linearly over time (figure 4). In this case the therapy seems to have a significant increasing effect over the ADC. Three out of thirteen patients experienced recurrence. The independent samples t-test performed showed no significant difference in initial volume between patients with complete clinical response (M = 105.06 cm3, SD = 89.16 cm3) and relapsing patients (M = 122.17 cm3, SD = 42.32 cm3) (t(11) = -0.315, p = 0.759). Initial ADC was not significantly different in the two analysed groups (responders: M = 0.72 × 10−3 mm2/s, SD 0.05 × 10−3 mm2/s; relapsing: M = 0.68 × 10−3 mm2/s, SD 0.10 × 10−3 mm2/s; t(11) = 0.975, p = 0.351), showing that both initial ADC and volume were not predictive of recurrence. The difference between initial and post-treatment volume in the two groups was also not significant (t(11) = -0.328, p = 0.759), while there was a significant difference in ADC mean values (t(11) = -3.477, p = 0.005) (figure 5), resulting smaller in relapsing patients.
Figure 3. 65-year-old woman with FIGO stage IVA CC. Coronal and axial T2WI (A,B) show cervical perforation by the uterine tandem. Coronal and axial T2WI (C,D), performed at the next brachytherapy session, show a proper placement of the tandem and a small amount of fluid in the cul-de-sac as the only complication.
Figure 4. Mean volumes / ADC values at primary staging (T1), after EBRT (T2) and after BRT (T3). The first graphic shows a significant lesion volume reduction after treatment, while the second graphic shows a linear ADC value increase during treatment.
Figure 5. Difference between pre-treatment and post-treatment ADC (ADC 0 – ADC 1) in patients with complete response to treatment (PR) and in patients who had a relapse (PD).
Intracavitary BRT appears to have a smaller effect on volume compared to EBRT, as it serves as a boost to achieve local radicality on LACC. Detailed tumour identification after EBRT is essential to correctly identify the residual lesion before BRT sessions and MRI proved to be an accurate tool both for radiologists and radiation oncologists. ADC may be a valuable parameter for assessing response to EBRT / BRT treatment and, hopefully, predicting recurrence. References: [1] Arbyn M, Weiderpass E, Bruni L, et al. Estimates of incidence and mortality of cervical cancer in 2018: a worldwide analysis [published correction appears in Lancet Glob Health. 2022 Jan;10(1):e41]. Lancet Glob Health. 2020;8(2):e191-e203. [2] Chino J, Annunziata CM, Beriwal S, et al. Radiation Therapy for Cervical Cancer: Executive Summary of an ASTRO Clinical Practice Guideline. Pract Radiat Oncol. 2020;10(4):220-234. [3] Haie-Meder C, Potter R, Limbergen EV, et al. Recommendations from Gynaecological (GYN) GEC-ESTRO Working Group (I): Concepts and terms in 3D image based 3D treatment planning in cervix cancer brachytherapy with emphasis on MRI assessment of GTV and CTV. Radiother Oncol 2005;74:235–245. [4] Jastaniyah N, Yoshida K, Tanderup K, et al. A volumetric analysis of GTVD and CTVHR as defined by the GEC ESTRO recommendations in FIGO stage IIB and IIIB cervical cancer patients treated with IGABT in a prospective multicentric trial (EMBRACE). Radiother Oncol 2016;120:404-411. [5] McVeigh PZ, Syed AM, Milosevic M, Fyles A, Haider MA. Diffusion-weighted MRI in cervical cancer. Eur Radiol. 2008;18:1058–64.