Assessment Of Tumor Response On Functional MR Imaging After Chemoembolization Of Liver Cancer

BackgroundAssessment of tumor response after therapies is important in determining treatment success and in guiding future therapy.Several monitors of tumor response have been used,including histology,tumor markers,and imaging.However,histologic evaluation using tissue biopsy can only be conclusive when it shows viable malignancy.Therefore, repeated negative biopsies do not exclude the presence of residual tumor.Tumor markers sorely are of limited use in assessing tumor response.Various imaging modalities,including Doppler ultrasonography,angiography,computed tomography (CT),and magnetic resonance imaging(MRI) have been used to evaluate therapeutic response.In clinic,standardized criteria for measuring therapeutic response have been established in 1981 by the World Health Organization(WHO).In 2000,the Response Evaluation Criteria in Solid Tumors(RECIST) was introduced to unify response assessment criteria.However,these methods rely on size change of lesions to assess response.Therefore,the early effects of some therapies would not be detected and occasionally patients may not be considered to have exhibited a response despite the presence of tumor necrosis.This realization that anatomy may not change after therapies moved the focus toward new evaluation methods.These include assessment of tumor vascular and cellular integrity,motion of water molecules,and biochemical concentration.In this article we discuss the role of functional MRI in assessing treatment response after transarterial chemoembolization in patients with liver cancer.Objective 1.To assess the diffusion-weighted MR image quality with different b values in the follow-up after transarterial chemoembolization for liver cancer.2.To evaluate the feasibility and capability of diffusion-weighted MR imaging technique in follow-up after transarterial chemoembolization for liver cancer.3.To determine whether the pretreatment apparent diffusion coefficients(ADCs) of liver cancer are predictive of response to transarterial chemoembolization and to compare the ADCs of tumors before and after transarterial chemoembolization.4.To investigate the value of ³¹p MR spectroscopy in monitoring the early response of therapy for hepatocellular carcinoma.Methods1.Diffusion-weighted MR imaging and routine non-enhanced MR imaging were performed in patients with liver cancer after transarterial chemoembolization at 1.5T MRI.DWI were performed with the same SS-SE-EPI sequence and different b value, b＝0 and b≠0(b＝150,350,500 and 800s/mm²,respectively).The signal-noise-ration (SNR) of the lesions after transarterial chemoembolization and the contrast noise ratio (CNR) between the lesion and liver tissue on diffusion-weighted images with different b values and T2-weighted images were measured.The lesion apparent diffusion coefficient(ADC) values were measured on functional diffusion maps.Qualitative evaluations the capability of identification the viable and necrotic tissue in tumor were performed on diffusion-weighted imaging with different diffusion coefficients comparing with DSA and enhanced-CT.2.Diffusion-weighted MR imaging and routine non-enhanced MR imaging were performed in patients with liver cancer after transarterial chemoembolization with the same SS-SE-EPI sequence(b=500s/mm² and 0) at 1.5T MRI.Contrast noise ratio(CNR) between new lesions and liver for diffusion-weighted MR image and T₂WI were measured.The characteristics of the treated lesion on diffusion-weighted image were evaluated.And the capability of discriminating the remaining or recurrent viable tumor and necrotic tissue were evaluated and compared with DSA and enhanced-CT.3.Liver cancers were prospectively evaluated with diffusion-weighted imaging at two b values before and after transarterial chemoemblization.Quantitative ADC maps were calculated with images with b values of 0,and 500s/mm².The mean ADC values of lesions before and after transarterial chemoemblization were compared according to response defined by the effect of transarterial chemoembolization.4.Phosphorus-31 MR spectra of hepatocellular carcinoma obtained before and at 24-48h after transarterial chemoemblization in 15 patients with 17 hepatocellular carcinomas.Alterations of phosphorus metabolism between before and after transarterial chemoembolization were compared with tumor responses evaluated by means of long-term follow-up.Results1.The SNR of the treated lesions were 69.81±18.99,59.33±32.66,59.23±32.94, 54.25±19.71,39.43±11.67 on T2WI,DWI(b＝150s/mm²),DWI(b=350s/mm²), DWI(b=500s/mm²) and DWI(b=800s/mm²),respectively(F=4.43,P=0.0024＜0.05). And CNR between the treated lesion and liver were 19.11±11.33,17.69±9.20, 21.38±10.10,19.90±13.75 and 13.24±11.02 on T2WI,DWI(b=150s/mm²), DWI(b=350s/mm²),DWI(b=500s/mm²) and DWI(b=800s/mm²),respectively(F=1.70, P=0.1556＞0.05).The treated lesion ADC values were(2.35±0.80)×10^-3mm²/s,(1.95±0.59)×10^-3mm²/s,(1.78±0.44)×10^-3mm²/s and(1.54±0.37)×10^-3mm²/s on functional diffusion maps with b=150,350,500 and 800s/mm²,respectively(F=21.96, P=0.0001＜0.05).The viable and necrotic tissue in treated lesion can be identified more clearly on diffusion-weighted images with b=500s/mm² and with b=800s/mm² than diffusion-weighted images with b=150s/mm² and with b=350s/mm².2.There was significant difference in CNR between new cancer lesion and liver tissue on diffusion-weighted and T2-weighted images(46.36±19.49 vs.33.24±17.26, t=2.400,P=0.025).Liver cancer after transarterial chemoemblization has variable signal intensity on diffusion-weighted MR image.On diffusion-weighted image,the hypo-intensity regions in the treated lesions correspond to regional no enhancement on enhanced-CT or no tumor stain tissue on angiogram that means necrotic tissue because of decreased cellular density;and the hyper-intensity regions in the treated lesions mean the viable tissue because of higher cellular density and hyper-vascular which enhancement on enhanced-CT and tumor stain on angiogram.The results of receiver operator characteristic(ROC) analysis for differentiation of viable and necrotic tumor tissue with mean apparent diffusion coefficients for images with b values of 0 and 500 s/mm² showed that threshold ADC value of 1.84×10^-3mm²/s had 92.3%sensitivity and 100%specificity for identification of necrotic tumor tissue.3.Nonresponding lesions had a significantly higher pretreatment mean ADC than did responding lesions(1.687×10^-3 mm²/s vs.1.278×10^-3 mm²/s,P＜0.05).The results of receiver operator characteristic(ROC) analysis for differentiation of nonresponding and responding lesions with mean apparent diffusion coefficients for images with b values of 0 and 500 s/mm² showed that threshold ADC value of 1.618×10^-3 mm²/s had 96.0%sensitivity and 77.8%specificity for identification of nonresponding lesions. After transarterial chemoembolization,responding lesions had a significant increase in %ADC values than did nonresponding lesions(32.63%vs.5.24%,P=0.025).The results of receiver operator characteristic(ROC) analysis for differentiation of nonresponding and responding lesions with mean apparent diffusion coefficients changes for images with b values of 0 and 500 s/mm² before and after transarterial chemoembolization showed that threshold%ADC value of 16.21%had 72%sensitivity and 100%specificity for identification of responding lesions.No significant ADC value change was observed in normal liver parenchyma(t=0.174,P=0.862) and spleen (t=1.964,P=0.052) after transarterial chemoembolization.4.For alterations in phosphorus metabolism in HCCs,the median phosphomonoesters(PME) level in 17 HCCs after treatment was 1.38×10⁷(range, 0.665×10⁷-6.21×10⁷),which was significantly decreased than that before treatment (median,2.98×10⁷;range,0.846×10⁷-102.5×10⁷)(P＜0.05).In long-term follow-up studies,the ratio of phosphomonoester to nucleoside triphosphate(PME/NPT)(P＜0.01) and the ratio of phosphodiester to nucleoside triphosphate(PDE/NPT)(P＜0.01) was remarkable decreased before and after treatment in the HCCs(11/17) which showed a response to therapy;in other 6 HCCs classified as no response,each PME/NPT ratio significantly diminished(P＜0.05),while the PDE/NPT ratio rose slightly,but no significant(P＞0.05). Conclusions1.In the follow-up after transarterial chemoembolization for liver cancer with DWI, diffusion coefficient influences the image quality.The diffusion-weighted MR imaging with suitable diffusion coefficient can achieve high image quality and identify viable and necrotic tissue in tumor more clearly.2.Diffusion-weighted MR imaging is a sensitive and useful method for detecting new lesion and the remaining or recurrent viable tumor for liver cancer after transarterial chemoembolization.3.High pretreatment mean ADC of liver cancers was predictive of poor response to chemoembolization.A significant increase in mean ADC was observed in lesions that responded to chemoembolization.These findings may have implications for development of individualized therapy.4.The results demonstrate ³¹P MR spectroscopy is a suitable method for monitoring the early response of HCC to transarterial chemoembolization.