Apparent diffusion coefficient values predict response to brachytherapy in bulky cervical cancer

Elizabeth E. Dong & Alexander N. Hanania et al. · 2024-03-13

Abstract

Background

Diffusion-weighted magnetic resonance imaging (DWI) provides a measurement of tumor cellularity. We evaluated the potential of apparent diffusion coefficient (ADC) values obtained from post-external beam radiation therapy (EBRT) DWI and prior to brachytherapy (BT) to predict for complete metabolic response (CMR) in bulky cervical cancer.

Methods

Clinical and DWI (b value = 500 s/mm 2 ) data were obtained from patients undergoing interstitial BT with high-risk clinical target volumes (HR-CTVs) > 30 cc. Volumes were contoured on co-registered T2 weighted images and 90th percentile ADC values were calculated. Patients were stratified by CMR (defined by PET-CT at three months post-BT). Relation of CMR with 90th percentile ADC values and other clinical factors (International Federation of Gynecology and Obstetrics (FIGO) stage, histology, tumor and HR-CTV size, pre-treatment hemoglobin, and age) was assessed both in univariate and multivariate logistic regression analyses. Youden’s J statistic was used to identify a threshold value.

Results

Among 45 patients, twenty-eight (62%) achieved a CMR. On univariate analysis for CMR, only 90th percentile ADC value was significant ( p  = 0.029) while other imaging and clinical factors were not. Borderline significant factors were HR-CTV size ( p =  0.054) and number of chemotherapy cycles ( p =  0.078). On multivariate analysis 90th percentile ADC ( p <  0.0001) and HR-CTV size ( p <  0.003) were highly significant. Patients with 90th percentile ADC values above 2.10 × 10 − 3 mm 2 /s were 5.33 (95% CI, 1.35–24.4) times more likely to achieve CMR.

Conclusions

Clinical DWI may serve to risk-stratify patients undergoing interstitial BT for bulky cervical cancer.

Authors
Elizabeth E. Dong, Junqian Xu, Joo-Won Kim, Jason Bryan, Jewel Appleton, Daniel A. Hamstra, Michelle S. Ludwig, Alexander N. Hanania