Evaluating transcatheter arterial chemoembolization for primary hepatic cancer by magnetic resonance diffusion-weighted imaging

Abstract

Hepatic cancer is associated with very high mortality and morbidity, and the world’s highest morbidity and mortality rates for this malignant tumor are found in China. Each year, the number of people who die of hepatic cancer in China amounts for 53% of the world total.1 Although resection is the first-choice treatment for hepatic cancer, only 20%-30% of patients have the opportunity to undergo resection.2 Transcatheter arterial chemoembolization (TACE) is a typical intervention therapy. Its advantages include minimal trauma, precise efficacy, and ease of administration. Consequently, it has become a widely used first-line treatment for patients who cannot undergo resection. Effective treatment planning and follow-up require an imaging method for evaluation of tumor response and detection of tumor recurrence. The objective of this study was to evaluate the efficacy of diffusion-weighted imaging (DWI) quantitative technique in identifying the composition of tissue after TACE, detecting remnant tumor after TACE, and detecting the sensitivity and specificity for recurrence by means of magnetic resonance imaging (MRI). METHODS Subjects Participating subjects were twenty-five patients who presented to the Cancer Center, Sun Yat-Sen University for surgical treatment of primary hepatic cancer and underwent TACE between March 2007 and March 2008, according to the Standards of Clinical Diagnosis and Staging for Primary Hepatic Cancer reported at the 8th National Conference on Hepatic Cancer in September 2001. The study population consisted of 19 males and 6 females aged 31-77 years, with a median age of 54 years. On the basis of disease condition and after obtaining patients’ written informed consent, unenhanced + diffusion-weighted + dynamic contrast-enhanced MRI was performed before the first TACE procedure and 4-6 weeks after each operation. MRI detection method A GE Excite 1.5T superconductive scanner and an 8-channel controlled-array software coil were used. All of the subjects underwent routine T1-weighted and T2-weighted MRI, followed by DWI and finally cross-section dynamic contrast-enhanced MRI. The specific imaging sequence was spin-echo echoplanar imaging (SE-EPI) sequence diffusion-enhanced imaging, with repetition time (TR) of 1200 ms, echo time (TE) of 59.0 ms, band width of 62.5 kHz/pixel, matrix of 128 × 128, and number of excitations (NEX) of 4. Three different diffusion-sensitive factors with b values of 1000, 500, and 300 s/mm2 were scanned once. The diffusion gradient adopted three directions: X, Y, and Z. Graphic analysis Two senior physicians specialized in imaging diagnosis comprehensively analyzed the unenhanced MRI, dynamic enhanced MRI, and digital subtraction angiography (DSA) to determine the tumor necrosis, remnant, and recurrence, compared with DWI. Neither enhancement in dynamic-enhanced MRI nor tumor staining in DSA was used as criterion for complete tumor necrosis. DWI signal values in the DWI plot and apparent diffusion coefficient (ADC) values in the ADC plot with b values of 1000, 500, and 300 s/mm2 were measured using function tool software of the GE Excite 1.5T MRI machine. Tissues examined included tumor tissue before operation, remnant tumor tissue after operation (region with enhancement and staining), tissue with coagulative necrosis after operation (region without enhancement and staining), and tissue with tumor recurrence after operation (with enlarged or newly emerged enhancement and staining), and normal liver tissue. When measuring regions of interest, bile ducts, blood vessels, and artifacts should be avoided. While staying within the focal region, the scope should be as large as possible; the area should be measured three times and the average taken. The patterns of DWI signal values and ADC values on the DWI plot for tissues examined at different b values were observed. Qualitative analysis was made with DWI signal values and ADC values in tissues after TACE. Statistical analysis SPSS version 13.0 software (SPSS, Inc., Chicago, USA) was used for statistical analysis. Statistical indices included the mean and standard deviation (mean ± SD) of DWI signal values and ADC values under the b values of 1000, 500, and 300 s/mm2 in each tissue. Data in this study were quantitative data with continuous distribution examined by normal distribution. One-way analysis of variance (ANOVA) was used to analyze DWI signal values and ADC values in each tissue. The Bonferroni adjustment was used for multiple comparisons between groups, with the level of statistical significance set at 0.05. RESULTS Regarding specificity of ADC values in liver tissue before and after TACE as well as tumor tissues, ADC values at different b values in the same tissue were compared. Statistically significant differences of ADC values were found in liver tissue, tumor tissue before operation, remnant tumor tissue after operation, and in tissue with tumor recurrence after operation. As the b value decreased, the ADC value in tissues increased. Using one-way ANOVA, data at three different b values showed statistically significant differences (F=23.92-43.25, P <0.05). ADC values in coagulative necrosis tissue showed no statistically significant differences (F=0.83, P >0.05). In comparing ADC values at the same b value in different tissues, the ADC values arranged from the lowest to the highest were those of tumor tissue before operation, remnant tumor tissue after operation, liver tissue, tissue with tumor recurrence after operation, and tissue with coagulative necrosis after operation. The b value of 1000 s/mm2 was most suitable because the extent of overlapping and variation of ADC values in tissues were small and could lead to better identification of results (Table 1).Table 1: ADC values (×10-3 mm2/s) in each tissue before and after TACE by b valueWith b=1000 s/mm2, comparison of ADC values in different tissues revealed that the value was the highest in tissue with coagulative necrosis after operation, which was statistically different from that in other tissues (F=23.25, P <0.05). The value was the lowest in tumor tissue before operation, and the value in tissue with tumor recurrence after operation was significantly higher than the former (P <0.05). The ADC value in remnant tumor tissue after operation was intermediate between the previous two values, but with no statistically significant differences (Table 2 and Figures 1-3). Regarding specificity of DWI signal intensity for hepatic tissue before and after TACE as well as tissue in the tumor region, statistically significant differences were found between DWI signals of the same tissue at different b values. As b values decreased, DWI signal values in each tissue increased. One-way ANOVA showed statistically significant differences in this group with three b values (F=13.46-188.87, P <0.05) (Table 3).Table 2: Bonferroni analysis of ADC values in each tissue at b=1000 s/mm2Figure 1.: 43-year-old male patient with differentiated hepatic carcinoma in right lobe. Resection was performed before operation. 1A. Tumor staining for right lobe in digital subtraction angiography (DSA). 1B. When the b value was 1000 s/mm2, a high signal existed for diffusion-weighted imaging (DWI), with a signal value of 73.09 and apparent diffusion coefficient (ADC) value of 1.15 ×10-3 mm2/s.Figure 2. Same patient as in Figure 1 after one TACE procedure. 2A.The tumor was stained as shown by DSA. 2B. When the b value was 1000 s/mm2, a high signal existed for DWI, with a signal value of 70.10, lower than that before the operation. The ADC value was 1.20 × 10-3 mm2/s.Figure 3. Same patient as in Figures 1 and 2, after two TACE procedures. 3A. DSA showed no tumor staining. The tumor with coagulative necrosis indicated by staining on right of image. 3B. When the b value was 1000 s/mm2, a high signal existed for DWI, with a signal value of 58.30, lower than that before the operation. The ADC value was 1.32 ×10-3 mm2/s, significantly higher than that before the operation.Table 3: DWI signal intensity in each tissue before and after TACE by b valueComparing DWI signals in each tissue with b=1000 s/mm2 revealed that signal in tissue with coagulative necrosis after operation was the lowest, followed by that in liver tissue. No statistically significant difference was found between these two values, both of which were significantly lower than those in tumor tissue before operation, remnant tumor tissue after operation, and tissue with tumor recurrence after operation (F=23.75, P <0.05). The signal was the highest in tumor tissue before operation, followed by that in remnant tumor tissue after operation, and then that in tissue with tumor recurrence after operation. No statistically significant differences were found between these three values. DISCUSSION Magnetic resonance DWI reflects the diffusion movement sequence in voxels of water molecules in tissue, with the capacity to detect the movement condition of water molecules in biological tissue. This sequence can reveal morphological changes by means of different signal intensities of tissues and facilitate qualitative analysis of tissue by means of ADC values pertaining to dual imaging for morphology and function.3,4 It may allow accurate qualitative analysis of tissues after TACE. DWI diffusion sensitivity (also called the diffusion-sensitive factor) is expressed as the b value, with larger b values indicating greater sensitivity.5 A high b value represents a longer TE for imaging. Regarding the liver with a shorter T2 value, the signal-to-noise ratio (SNR) of the image will obviously decrease. To increase the stability of DWI in this study, three b values, 1000 s/mm2, 500 s/mm2, and 300 s/mm2, were used to generate a better imaging SNR. We found no statistically significant differences in ADC value of tissue with coagulative necrosis at different b values (P >0.05). For ADC values of all other tissues and DWI signal values in all tissues, significant differences were found (P <0.05), with the value increasing as the b value decreased. The degree of increase in the liver and active tumor tissue was significantly greater than that in tissue with coagulative necrosis. As the ADC values of the liver and active tumor tissue increased, the difference from that in tissue with coagulative necrosis decreased. This phenomenon was most obvious when the b value was 300 s/mm2. Because the blood supply was greater in the liver and active tumor tissue than in tissue with coagulative necrosis, the smaller diffusion gradient when b was 300 s/mm2 and the increased effect of blood infusion resulted in an inability to accurately reflect the diffusion movement of water molecules. When b was 1000 s/mm2, the effect of blood infusion decreased significantly. At this time, DWI accurately reflected the diffusion movement of water molecules, making the difference in ADC values between tissue with coagulative necrosis and the liver and active tumor tissue more apparent, better distinguishing the tissues. This study used ADC values to perform quantitative analysis of tissues of hepatic cancer after TACE. The ADC values in tumor tissue before operation were basically the same as the ADC values in malignant hepatic tumor.5-9 Its ADC value is lower, possibly because of the higher cell density of malignant tumor tissue, increased ratio of nucleus to plasma, and decreased extracellular water, causing restricted diffusion of water molecules. The ADC values of tissue with coagulative necrosis were significantly higher than those in remnant liver tumor tissue. Geschwind et al9-12 found that after chemoembolization intervention surgery, the combination of embolization of the end tumor blood vessel by iodized oil — causing anemia and hypoxia — and cytotoxicity from local high-concentration chemotherapy drugs damaged the focal tumor tissue, with increased cell membrane permeability, rupture, and lysis; a large amount of intracellular plasma leakage; and increased space between cells and tissues. This resulted in increased free diffusion of water molecules outside the cell and significant apoptosis in addition to necrosis in tumor tissue after TACE.11,13 The ADC values of tissue with tumor recurrence were lower than those of coagulative tissue but higher than those of tumor tissue before operation. Because primary hepatic cancer has abundant side-chain circuitry and a multicentric tumor source, it is difficult to completely inactivate tumor with TACE, leaving part of the fiber gap and remnant tumor cell under involucrum. At this time, they are barely detectable with routine and enhanced MRI, even with DSA, becoming the source of subsequent recurrence. The ADC values decrease, however. The change is earlier than that of T2WI and even histological changes. 12 Statistically significant differences were found between these three types of pathological changes. In the past, DWI signal was used to characterize the anatomic structure of an image and was rarely used in quantitative analysis. This is mainly because DWI scans include the combined effects of T2, proton, and ADC value changes, giving DWI signal a larger variation. The results of the current study showed that using DWI signal for quantitative analysis of tissues in hepatic cancer after TACE, despite larger standard deviations, can still identify the specificity of tissues, achieving a result similar to that of ADC quantitative analysis. When b was 1000 s/mm2, DWI signal of tissue with coagulative necrosis after operation was the lowest, while that of liver tissue was higher, although the difference was not statistically significant. Both of these values were significantly lower than those of tumor tissue before operation, remnant tumor tissue after operation, and tissue with tumor recurrence after operation (P <0.05). The decrease from a high signal before operation to a lower signal showed that the tumor had significant necrosis with specificity, with no need for further treatment. Thus DWI signal is an important reference value for evaluating the efficacy of TACE for the treatment of hepatic cancer. Follow-up studies of TACE for primary hepatic cancer have been reported, but generally only within 1 month after single TACE.14 Long-term follow-up studies have not yet been reported. DWI can clearly show the morphology of the tumor focal area, while ADC value and DWI signal can qualify the tissue composition after TACE. This can help achieve the goal of timely detection of remnant tumor and recurrence in hepatic cancer through MRI.