Purpose

Hepatocellular carcinoma is associated with significant morbidity and mortality. Ablation therapies are an established treatment, but outcomes remain inconsistent and complete tumor eradication is uncommon. Incomplete ablation of liver tumors can provoke an aggressive and detrimental tumor response (1). Therefore, accurate mapping of the ablative agent’s distribution is critical to determine the delivered dose and confirm efficacy. Additionally, real-time monitoring of ablative agent delivery could ensure safe administration and minimize treatment to undesired areas. Trifluoroacetic acid (TFA) has previously been reported as a possible theranostic chemoablation agent (2). This work investigated the utility of TFA as an ablative agent that permits simultaneous treatment and imaging.

Materials and Methods

Imaging with fluorine-19 magnetic resonance imaging (¹⁹F-MRI) was optimized at 7 T using a custom-built volume coil. A phantom was constructed containing five NMR tubes (5 mm diameter) of varying concentrations of TFA (31.25, 62.50. 125.0, 250.0, and 500.0mM).We acquired fluorine images in triplicate with both rapid acquisition with relaxation enhancement (RARE) and balanced steady-state free precession sequences (bSSFP) with varying parameters (repetition time [TR], echo train length [ETL; RARE only], and flip angle [bSSFP only]) to determine the optimal sequences for the relaxation properties of TFA. Voxel size was fixed at 0.63 × 0.63 × 5.00 mm. Proton images were also acquired with a RARE sequence and superimposed. The signal-to-noise ratios (SNR) of the varying TFA concentrations (from the optimal sequence) were fit to a straight line to determine the sensitivity limit of the system, assuming a conservative minimum necessary SNR of 5 (3). The possibility of real-time imaging of TFA injection was evaluated by injecting 2M TFA into a flexible tube (3.2 mm inner diameter) while imaging with cine ¹⁹F-MRI using the optimal bSSFP sequence. To evaluate the effects of tissue on imaging, we injected 100 μL of TFA (0.25, 0.5, or 1.0M) into ex vivo porcine liver sections and imaged with ¹⁹F- and ¹H-MRI. We then evaluated tissue damage with gross examination, microscopic histology (hematoxylin and eosin), and fluorescence microscopy (phalloidin and 4′,6-diamidino-2-phenylindole to highlight cytoskeletal and nuclear structures, respectively). Average values are reported as the mean value with standard deviation. Differences were determined using Welch’s t-test and considered significant if p < 0.05. All fits used least-squares weighted values.

Results

The TFA in the phantom was successfully imaged at all concentrations, even at short scan times (e.g. 30 s). The optimal RARE TR/ETL were determined to be 3 s/32, and the optimal TR/flip angle for bSSFP were determined to be 2.5 ms/70°. The optimal bSSFP sequence yielded 1.45 times the SNR efficiency (SNR divided by square root of imaging time) of the optimal RARE sequence. The minimum imageable TFA concentration was determined to be 6.7 ± 0.5mM with one minute of scan time (Fig. 1). The injection of 2M TFA was successfully visualized with a temporal resolution of 10 frames/s (Fig. 2). TFA distribution in ex vivo tissue was successfully imaged at all concentrations with adequate signal (Fig. 3). Treatment with TFA successfully coagulated tissue, and the damage was extensive but locally confined. In addition to hepatic lobular architecture and cord disruption, hepatocyte cytoskeletal collapse and chromatin clumping were observed in severely damaged areas.

Discussion

¹⁹F-MRI has been used for more than 40 years because of its 100% natural abundance and high sensitivity (4), but most current applications (e.g. imaging fluorinated drugs or cells labeled with fluorinated compounds) involve very low ¹⁹F concentrations and, hence, very low MR signal intensities. This therapeutic application of TFA involves much higher ¹⁹F concentrations, resulting in significantly increased MR signal. In the phantom, image SNR was found to have a linear relationship with concentration, as expected. Both RARE and bSSFP sequences were successfully optimized and we found that bSSFP yielded higher SNR efficiency than did RARE, in agreement with previous studies (5). Cine ¹⁹F-MRI confirmed that real-time image guidance of TFA ablation therapy is feasible. In the liver sections, TFA distribution could not be visualized with standard proton MRI but was successfully imaged with ¹⁹F-MRI at all concentrations. Results of histologic examination of liver sections treated with TFA ex vivo were consistent with changes that would result in irreversible coagulative necrosis in vivo for the 0.5 and 1.0M concentrations. Future studies will focus on translation to clinical 3 T scanners.

Conclusions

Our results demonstrated that tissue ablation with TFA was both efficacious and imageable, even at low concentrations. We have shown TFA to be a promising theranostic agent for chemical ablation of solid tissue.

Acknowledgements

This work was supported in part by the National Institutes of Health (R01 CA201127-01A1 and P30 CA016672). The authors would like to thank Bryan Tutt, scientific editor.