Fusion imaging of ultrasound and MRI in the assessment of locally advanced cervical cancer: a prospective study

Abstract

Background: Fusion imaging is a new diagnostic method that integrates MRI and ultrasound. It may improve the detection and staging of locally advanced cervical cancer.

Objective: To evaluate the feasibility and accuracy of fusion imaging in patients with locally advanced cervical cancer.

Methods: Patients with suspicion of locally advanced cervical cancer at clinical examination and/or imaging, who were candidates for neoadjuvant treatment (chemotherapy or chemoradiation) followed by surgery, were prospectively enrolled between March and November 2018. MRI, ultrasound, and fusion images were obtained before and after neoadjuvant treatment. Feasibility, success of the fusion examination, and time needed to perform fusion studies were evaluated. The rates of concordance between MRI and ultrasound before and after performing fusion, using Cohen, Spearman, and McNemar tests were calculated. The agreement between MRI and ultrasound examination, and the agreement between radiologist and gynecologist during the fusion technique in assessing local extension of disease and the presence of residual disease after neoadjuvant therapy, were also analyzed. The rates of concordance between MRI and ultrasound examination before and after performing fusion imaging, using Cohen's kappa and Spearman's rank correlation coefficient were calculated. A McNemar test was used to assess if there were statistical significant differences in the parameters' agreement before and after performing fusion imaging.

Results: 40 patients were selected and of these, 33 were analyzed. A total of 52 fusion examinations were performed: 33 (63.5%) of 52 at the time of diagnosis and 19 (36.5%) of 52 after neoadjuvant treatment. Fusion imaging was feasible in 50 (96%) of 52 studies. The median overall time of fusion execution was 13 min (range 6-30) and the time spent in performing a fusion examination decreased from the first to the last examination (20 vs 6 min). The agreement between MRI and ultrasound parameters increased after performing fusion, particularly for parametrial infiltration (74% vs 86%, p=0.014 for the right posterior parametrium; 66% vs 80%, p=0.008 for the left posterior parametrium, 70% vs 82%, p=0.014 for the right lateral parametrium).

Conclusions: Fusion of MRI and ultrasound is feasible in patients with locally advanced cervical cancer and may increase the diagnostic accuracy of the single imaging methods. Fusion provides multiple diagnostic opportunities in gynecological oncology.

Keywords: MRI; cervical cancer; fusion; ultrasonography.