OBJECTIVETo investigate the clinical application of diffusion weighted imaging (DWI) in uterine cervical cancer and the apparent diffusion coefficient (ADC) value in diagnosis and predicting treatment response.

METHODSTwenty-eight patients with advanced primary cervical cancer confirmed by pathology and 10 cases of normal uterine cervix as control were recruited in this prospective clinical trial. To analyze the correlation between tumor volume measured in DWI and tumor maximum diameter measured according to the RECIST criteria. To compare the ADC value differences among the uterine cervical cancer, uterine myometrium, and normal uterine cervix. To compare the ADC values in 17 cancer patients before and after treatment.

RESULTSThe illustration of tumor boundary in DWI was superior to conventional T2WI and post-enhancement T1WI. The DWI with higher b value (2000 s/mm(2)) had a better signal-to-noise ratio. The tumor volume measured in DWI has good correlation with tumor maximum diameter according to RECIST criteria (r = 0.759, P < 0.01). When b = 800 s/mm(2), the ADC values of the uterine cervical cancer, uterine myometrium, and normal uterine cervix were (9.85 ± 1.55)×10(-3) mm(2)/s, (14.20 ± 2.80)×10(-3) mm(2)/s, and (14.14 ± 0.45) ×10(-3) mm(2)/s. When b = 2000 s/mm(2), the ADC values of the uterine cervical cancer, uterine myometrium and normal uterine cervix were (7.38 ± 0.98)×10(-3) mm(2)/s, (8.52 ± 2.38)×10(-3) mm(2)/s, and (8.60 ± 0.63)×10(-3) mm(2)/s, respectively. There were significant differences between the cervical cancer and normal cervix or uterine myometrium (P < 0.001 for both). When b = 800 s/mm(2), the ADC value was (9.85 ± 1.55)×110(-3) mm(2)/s before and (13.41 ± 2.93)×10(-3) mm(2)/s after treatment (P < 0.001). When b = 2000 s/mm(2), the ADC value was (7.38 ± 0.98)×10(-3) mm(2)/s before and (8.93 ± 1.92)×10(-3) mm(2)/s after treatment (P = 0.008). Univariate logistic regression analysis showed that 25% ADC, 50%ADC, and 75%ADC in the tumor ADC value histogram before treatment were significantly correlated to the treatment outcome of cervical cancer (P < 0.05 for all). Multivariate regression analysis showed that 25%ADC, 50%ADC, and 75%ADC in the tumor ADC value histogram before treatment were not significantly correlated to the treatment outcome of cervical cancer (P > 0.05 for all). The values of ROC curves were 25%ADC = 0.818, 50%ADC = 0.775, and 75%ADC = 0.716 (P > 0.05), however, the 25% ADC showed a relatively stronger statistical power.

CONCLUSIONSDWI helps to confirm the morphology and exact target zone of the tumor for radiotherapy. DWI volume measurement is well correlated with RECIST criteria, particularly in volume measurement of irregular tumors. ADC value has a potential in quantitatively monitoring treatment response and predicting outcome of cervical cancers.