Real-Time Monitoring of Hepatocellular Carcinoma Radiofrequency Ablation by Quantitative Temperature MRI

• CASE REPORT
• Preoperative Imaging Findings
• Radiofrequency Thermal Ablation Procedure
• RF ABLATION DEVICE
• MRI TECHNIQUE AND RF NEEDLE POSITIONING
• MAGNETIC RESONANCE TEMPERATURE IMAGING PROTOCOL
• IMAGING FOLLOW-UP
• DISCUSSION
• ABBREVIATIONS
• REFERENCES

Early detection of hepatocellular carcinoma (HCC) through surveillance programs allows potentially curative therapies such as resection, liver transplantation, and percutaneous ablation to be applied in 40% of the patients. In nonsurgical candidates, percutaneous treatments are the best therapeutic approach and may improve survival in patients with well-preserved liver function and small tumors who achieve initial complete response. Applying these criteria, 5-year survival rates above 50% have been achieved.[1]

Minimally invasive, percutaneous ablative therapies under image guidance with thermal energy sources such as radiofrequency (RF) are increasingly used for tumor ablation because of their efficacy and low cost. Radiofrequency thermal ablation (RFTA) provides better objective response rates than ethanol injection, and a survival advantage has been reported.[2]

Sonography or computed tomography (CT) guidance is most often used for RF needle electrode positioning. However, precise and quantitative on-line control of the thermal ablation is not possible with these imaging modalities. The imaging method used for clinical guidance should ideally have a good spatial, temporal, and thermal resolution and should be easily implemented. In addition, the thermal and spatial resolution should be sufficient to detect temperature rise in the tissue around the lesion to: (1) enable real-time control of the ablation of the complete tumor and of a safety margin of 5 to 10 mm around the tumor and (2) avoid unwanted thermal damage in healthy tissue.

Magnetic resonance imaging (MRI) can monitor temperature changes quantitatively using the temperature-dependent shifts of the proton resonance frequency (PRF).[3] Therefore temperature mapping with MRI can be used to predict the treatment outcome during the heating procedure.

CASE REPORT

A 71-year-old man, born in Algeria from Algerian parents and living in France since the age of 27, was admitted to the hospital in April 2004 for the diagnosis and treatment of a liver nodule, which was discovered in 2003. He was regularly followed by his physician for several health problems, which included thyroid insufficiency treated with Levothyrox^® (Merck Lipha, Lyon, France), mild adrenal insufficiency treated with Symbiocort^® (AstraZeneca, Rueil Malmaison, France), a chronic obstructive bronchopneumopathy treated with Singulair^® (Merck-Sharp, Paris, France) and requiring occasional oxygen therapy, and mild type 2 diabetes. The nodule was discovered during an ultrasound examination performed for abdominal pain. Abdominal examination was normal. Liver function tests including albumin, prothrombin time, platelets, bilirubin, transaminases, and alkaline phosphatase were normal. There was a fivefold rise in gamma glutamyl transferase activity. Blood glucose was at the upper limit of normal and his body mass index was 29.4. There was no evidence for portal hypertension by imaging studies. Viral serologies for hepatitis C virus and hepatitis B virus were negative. He used to drink ~80 g of alcohol per day but had stopped for the past 6 years. He smoked less than 2 packs a month.

The diagnosis of HCC was made on MRI (size 30 mm). The patient initially refused to have a liver biopsy and further refused to have his tumor resected. RFTA was performed in May 2004. Prior to RFTA a liver biopsy was performed but missed the tumor. The nontumoral liver was considered as precirrhotic (F3/F4 METAVIR score). Tumor recurrence was noticed 5 months after the RFTA.

Preoperative Imaging Findings

MRI examination was performed 1 month prior to the procedure to treat the recurrent tumor and showed typical imaging for small recurrent HCC with a nodular high signal zone on T2-weighted sequence and early enhancement after gadolinium injection (Fig. [1]). This 26-mm lesion was located above the left portal vein in the internal and lower part of the old RF ablation area. It was not possible to detect precisely the recurrence zone with sonography, contrast-enhanced sonography, and multiphasic CT examination due to heterogeneity of the ablative zone and lack of enhancement on CT examination. Thus, MRI guidance was the only way to perform another RFTA.

Radiofrequency Thermal Ablation Procedure

RFTA was performed under general anesthesia. After induction of anesthesia, the patient was transferred to the MRI unit and placed in dorsal decubitus position. The abdomen was surrounded by a flexible phased array coil dedicated to abdominal imaging. Vital data, including cardiac and respiratory parameters and blood pressure, were recorded.

RF ABLATION DEVICE

We used a commercially available monopolar RF device (Tyco Healthcare/Radionics, Burlington, MA) operating at a 480-kHz frequency with 200 W of maximum output power. According to the depth of the lesion, magnetic resonance-compatible RF electrode of 15-cm length and 17-gauge diameter was used and internally cooled with sterile water maintained at 4°C using a peristaltic pump (Radionics). The electrically active part of the electrode tip was 3 cm in length, and a magnetic resonance-compatible grounding pad was attached to the skin of the patient's thigh. The RF generator and peristaltic pump were placed outside the magnetic resonance room. The output signal of the generator was filtered with a passive band filter constructed in-house tuned to the magnetic resonance proton frequency, so that the output signal of the generator did not interfere with the RF pulses of the MRI system. Total duration of the RF energy deposition was 12 minutes.

MRI TECHNIQUE AND RF NEEDLE POSITIONING

The procedure was performed in the MRI suite with a 1.5-Tesla MRI system (Philips, Cleveland, OH). The following sequences were performed before RFTA to localize the target lesion: axial in-phase and out-of-phase chemical shift GRE T1-weighted images, axial T2-weighted with half-Fourier acquisition of single-shot turbo spin echo images, with and without spectral fat saturation. In this case, these sequences permitted precise location of the tumor and gadolinium enhancement was unnecessary before treatment. To have an external magnetic resonance-visible marker, two circular water tubes (5-mm diameter) were intersected under the abdominal coil. This marker appeared as hyperintense on T2-weighted sequences, helping to define the entry point on the abdomen and to plan the needle trajectory.

Successive axial T1-weighted sequences were used for needle guidance because they provide a good compromise between satisfactory visibility of the lesion and speed of acquisition. During needle positioning, oblique axial and sagittal sequences were necessary along with axial images to check needle trajectory. Sequences were repeated to correct the location of the needle with its active part transfixing the lesion (Fig. [2]). After completion of the heating procedure and removal of the electrode, transverse T1-weighted, T2-weighted, and gadolinium-enhanced T1-weighted sequences were performed to evaluate response to the RF therapy.

MAGNETIC RESONANCE TEMPERATURE IMAGING PROTOCOL

Magnetic resonance temperature images were performed with a RF-spoiled, segmented GE echo planar imaging sequence, with seven echoes per TR period following a 25-degree slice water-selective spectral-spatial binomial 1-3-3-1 excitation pulse (5-mm thick; TE/TR/FA: 16 ms/43 ms/25 degrees; field of view: 143 × 160 mm; matrix: 90 × 112 [zero filled to 128 × 128]; left-right phase-encode direction; IPR: 1.6 × 1.4 mm). Three adjacent axial slices were acquired. Because of the respiratory gating, the acquisition duration of one slice (340 ms) may be divided over two respiratory cycles with a gating width of ~200 ms at the end of expiration.

Before the RF ablation procedure, a sequence of 50 dynamics was acquired to estimate the phase (and therefore the temperature measurement) stability in the liver tissue surrounding the needle. Then, an identical acquisition was repeated with 200 dynamics and RF power was applied immediately following the 15th dynamic. Temperature maps and evaluation of thermal dose were dynamically calculated in real time with a homemade software, after magnitude and phase images were acquired and reconstructed with standard MRI instrument hardware and software. Temperature and thermal dose maps were superimposed to improve images to visualize the spatial extension of the thermal treatment during the ablation (Fig. [3A]). According to thermal dose map analyses, the necrotic area obtained after RF deposition was 42 × 31 mm and included the whole of tumor volume.

IMAGING FOLLOW-UP

After RFTA, the liver of the patient was imaged with fat-suppressed T1-weighted, fat-suppressed T2-weighted, and dynamic gadolinium-enhanced T1-weighted sequences. The size of the ablated zone was respectively estimated as 30 × 20 mm, 32 × 16 mm, and 34 × 25 mm with these sequences. The discrepancy between thermal dose maps data and gadolinium-enhanced T1-weighted images can be explained by contrast uptake in a peripheral hyperemic zone, which was included in the necrosis volume on thermal dose maps, according to the lethal thermal dose concept (Figs. [3B] and [3C]).[4]

Control MRI examinations were performed at 6 weeks and 3, 6, 12, and 18 months. No local recurrence was found on these examinations.

DISCUSSION

RFTA is widely applied to the treatment of unresectable malignant liver lesions such as HCC and tumor metastasis. Usually, ultrasound or CT guidance and monitoring are used for RFTA. Unfortunately, these techniques have been shown to be less sensitive than MRI for depicting small lesions, HCC in particular.[5] Therefore, MRI guidance is some times the only way for RFTA to be performed, especially when the lesion is not detected with ultrasonography or multiphasic CT procedure.

Estimation of the real size of the ablated zone during the procedure is another important limitation of this mini-invasive technique, when ultrasound or CT monitoring is used. With ultrasound monitoring, the hyperechoic zone due to gas emission during the procedure is only a rough approximation of the area of induced tissue necrosis.[6] As for CT, assessment of the extent of the ablative zone is only possible at the end of the procedure because of the impracticality of repeated injections. Furthermore, enhancement due to the hyperemic area and inflammatory process around the ablative zone does not permit any correct analysis of the margins immediately after RFTA.

A discrepancy between histopathologic data and imaging results from CT and MRI has been demonstrated in two recent studies,[7] [8] with overestimation of the RF response as compared with liver explant analysis. In fact, techniques such as ultrasonography, CT, and MRI fail to provide accurate RFTA monitoring and immediate posttreatment evaluation.[6] [7] [8] [9] Precise evaluation of the extent of thermal ablation is especially critical during treatment of tumors located next to large intrahepatic vessels such as the portal and hepatic vein. The cooling effect of blood flow (heat-sink effect) provides indeed a high risk for incomplete necrosis or local recurrence for these lesions.[10] Thus, having another method beyond posttreatment imaging to precisely determine the extent of tissue damage is necessary for optimal thermal treatment.

Magnetic resonance thermometry seems to be a very promising technique because of its ability to provide a near real-time quantitative measurement of the tissue temperature distribution.[11] [12] [13] [14] Moreover, because magnetic resonance thermometry provides an evaluation of spatial and temporal temperature evolution, calculation of thermal dose in every voxel of the ablative zone becomes accessible.[4] [15] Up to now, RFTA performed with magnetic resonance thermometry has been considered as a challenging procedure[16] because of susceptibility artifacts induced by metallic electrodes and the need for efficient filters to avoid interferences between the RF generator and MRI. Feasibility and effectiveness of the technique has been demonstrated in a recent in vivo study performed on rabbit livers.[17] In the same study, thermal dose maps were shown to be more predictive and precise than other magnetic resonance images to evaluate the final RFTA zone size. The major advantage of the thermal dose maps on the gadolinium-enhanced sequences is that it can be repeated as many times as necessary throughout the procedure. It is thus possible with this technique to check during the course of treatment whether or not the lethal thermal dose has been delivered to the entire tumor. In the event of incomplete ablation after the first RF application, analysis of the thermal dose maps permits accurate guidance for needle repositioning. Up to now we have successfully performed 10 RFTAs with quantitative temperature MRI. Fig. [4] illustrates such a beneficial effect overcoming the cooling effect of blood flow.

For our patient, magnetic resonance temperature imaging monitoring had two advantages: first, magnetic resonance guidance was the only way to ensure precise positioning of the needle electrode because the lesion was undetectable by sonography or CT; second, temperature maps were very useful to ensure that the entire lesion was ablated despite the “heat-sink effect of flow” due to its proximity to the left portal vein branch.

There are, however, several limitations to the use of this technique in clinical practice: MRI guidance is not practical for needle electrode positioning because of the narrowness of the tunnel, and it takes much more time than under sonographic guidance (almost 2 hours versus 40 minutes). The use of whole body MRI open magnet or closed magnet with larger diameter would certainly make it possible to improve the accessibility of the patients to the puncture.[18] Nevertheless, the current open magnet systems are limited to a 1-Tesla field. We point out that the technique of thermometry by MRI from PRF must preferably be used on intense magnets (equal to or higher than 1 Tesla) to draw an advantage from the strong signal/noise ratio. Until high field open systems become available, a pragmatic solution would consist of reserving the use of MRI only to the time of thermometry and to perform the puncture under echographic guidance with a patient lying on a removable bed of MRI, because the very great majority of the liver lesions are perfectly visible with this method. Another limitation is dependent on the artifact of the magnetic resonance-compatible electrode that is quite large, with an increase of almost 800% of the apparent diameter of the electrode. This made the measurement of the temperature at the vicinity of the electrode tip impossible. Nevertheless, this artifact did not prevent us from obtaining information on the reliable thermal dose at the periphery of the tumors. In clinical practice, the risk of incomplete treatment is related to insufficient heating at the periphery of the tumor and therefore incomplete necrosis, resulting in local recurrences at the margin of the ablated lesion.

In conclusion, this was a 71-year-old man who underwent RF ablation of a 26-mm recurrent HCC under magnetic resonance thermometry guidance. This technique permitted accurate positioning of the needle electrode through the lesion, which was undetectable by sonography and CT examination. Furthermore, it was possible to ensure ablation of the entire tumor volume during the procedure by real-time thermal dose maps analyses. Eighteen-month follow-up of the patient revealed no local recurrence on repeated MRI examinations.

ABBREVIATIONS

• CT computed tomography

• FA flip angle

• GE gradient echo

• GRE gradient recalled echo

• HCC hepatocellular carcinoma

• IPR in-plane resolution

• MRI magnetic resonance imaging

• PRF proton resonance frequency

• RF radiofrequency

• RFTA radiofrequency thermal ablation

• TE time echo

• TR time recovery

REFERENCES

• 1 Llovet J M, Sala M. Non-surgical therapies of hepatocellular carcinoma.  Eur J Gastroenterol Hepatol. 2005;  17 505-513
  CrossrefPubMedSearch in Google Scholar
• 2 Shiina S, Teratani T, Obi S et al.. A randomized controlled trial of radiofrequency ablation with ethanol injection for small hepatocellular carcinoma.  Gastroenterology. 2005;  129 122-130
  CrossrefPubMedSearch in Google Scholar
• 3 Quesson B, de Zwart J A, Moonen C T. Magnetic resonance temperature imaging for guidance of thermotherapy.  J Magn Reson Imaging. 2000;  12 525-533
  CrossrefPubMedSearch in Google Scholar
• 4 Sapareto S A, Dewey W C. Thermal dose determination in cancer therapy.  Int J Radiat Oncol Biol Phys. 1984;  10 787-800
  CrossrefPubMedSearch in Google Scholar
• 5 Giannini E G. Review: MRI is more sensitive but less specific than ultrasonography or spiral CT for diagnosis of hepatocellular carcinoma.  ACP J Club. 2006;  145 21
  PubMedSearch in Google Scholar
• 6 Leyendecker J R, Dodd III G D, Halff G A et al.. Sonographically observed echogenic response during intraoperative radiofrequency ablation of cirrhotic livers: pathologic correlation.  AJR Am J Roentgenol. 2002;  178 1147-1151
  CrossrefPubMedSearch in Google Scholar
• 7 Mazzaferro V, Battiston C, Perrone S et al.. Radiofrequency ablation of small hepatocellular carcinoma in cirrhotic patients awaiting liver transplantation: a prospective study.  Ann Surg. 2004;  240 900-909
  CrossrefPubMedSearch in Google Scholar
• 8 Lu D S, Yu N C, Raman S S et al.. Radiofrequency ablation of hepatocellular carcinoma: treatment success as defined by histologic examination of the explanted liver.  Radiology. 2005;  234 954-960
  CrossrefPubMedSearch in Google Scholar
• 9 Raman S S, Lu D S, Vodopich D J, Sayre J, Lassman C. Creation of radiofrequency lesions in a porcine model: correlation with sonography, CT, and histopathology.  AJR Am J Roentgenol. 2000;  175 1253-1258
  PubMedSearch in Google Scholar
• 10 Lu D S, Raman S S, Limanond P et al.. Influence of large peritumoral vessels on outcome of radiofrequency ablation of liver tumors.  J Vasc Interv Radiol. 2003;  14 1267-1274
  CrossrefPubMedSearch in Google Scholar
• 11 Peters R D, Hinks R S, Henkelman R M. Ex vivo tissue-type independence in proton-resonance frequency shift MR thermometry.  Magn Reson Med. 1998;  40 454-459
  PubMedSearch in Google Scholar
• 12 Wlodarczyk W, Boroschewski R, Hentschel M et al.. Three-dimensional monitoring of small temperature changes for therapeutic hyperthermia using MR.  J Magn Reson Imaging. 1998;  8 165-174
  PubMedSearch in Google Scholar
• 13 Moriarty J A, Chen J C, Purcell C M et al.. MRI monitoring of interstitial microwave-induced heating and thermal lesions in rabbit brain in vivo.  J Magn Reson Imaging. 1998;  8 128-135
  PubMedSearch in Google Scholar
• 14 Chen J C, Moriarty J A, Derbyshire J A et al.. Prostate cancer: MR imaging and thermometry during microwave thermal ablation-initial experience.  Radiology. 2000;  214 290-297
  PubMedSearch in Google Scholar
• 15 McDannold N, King R L, Jolesz F A, Hynynen K. The use of quantitative temperature images to predict the optimal power for focused ultrasound surgery: in vivo verification in rabbit muscle and brain.  Med Phys. 2002;  29 356-365
  CrossrefPubMedSearch in Google Scholar
• 16 Rhim H, Goldberg S N, Dodd III G D et al.. Essential techniques for successful radio-frequency thermal ablation of malignant hepatic tumors.  Radiographics. 2001;  21 S17-S39
  PubMedSearch in Google Scholar
• 17 Lepetit-Coiffe M, Quesson B, Seror O et al.. Real-time monitoring of radiofrequency ablation of rabbit liver by respiratory-gated quantitative temperature MRI.  J Magn Reson Imaging. 2006;  24 152-159
  CrossrefPubMedSearch in Google Scholar
• 18 Huppert P E, Trubenbach J, Schick F et al.. MRI-guided percutaneous radiofrequency ablation of hepatic neoplasms: first technical and clinical experiences.  Rofo. 2000;  172 692-700
  PubMedSearch in Google Scholar

Hervé LaumonierM.D. 

Department of Radiology, Hôpital Saint André

CHU Bordeaux, 1, rue jean Burguet, 33075, Bordeaux Cedex, France

REFERENCES

• 1 Llovet J M, Sala M. Non-surgical therapies of hepatocellular carcinoma.  Eur J Gastroenterol Hepatol. 2005;  17 505-513
  CrossrefPubMedSearch in Google Scholar
• 2 Shiina S, Teratani T, Obi S et al.. A randomized controlled trial of radiofrequency ablation with ethanol injection for small hepatocellular carcinoma.  Gastroenterology. 2005;  129 122-130
  CrossrefPubMedSearch in Google Scholar
• 3 Quesson B, de Zwart J A, Moonen C T. Magnetic resonance temperature imaging for guidance of thermotherapy.  J Magn Reson Imaging. 2000;  12 525-533
  CrossrefPubMedSearch in Google Scholar
• 4 Sapareto S A, Dewey W C. Thermal dose determination in cancer therapy.  Int J Radiat Oncol Biol Phys. 1984;  10 787-800
  CrossrefPubMedSearch in Google Scholar
• 5 Giannini E G. Review: MRI is more sensitive but less specific than ultrasonography or spiral CT for diagnosis of hepatocellular carcinoma.  ACP J Club. 2006;  145 21
  PubMedSearch in Google Scholar
• 6 Leyendecker J R, Dodd III G D, Halff G A et al.. Sonographically observed echogenic response during intraoperative radiofrequency ablation of cirrhotic livers: pathologic correlation.  AJR Am J Roentgenol. 2002;  178 1147-1151
  CrossrefPubMedSearch in Google Scholar
• 7 Mazzaferro V, Battiston C, Perrone S et al.. Radiofrequency ablation of small hepatocellular carcinoma in cirrhotic patients awaiting liver transplantation: a prospective study.  Ann Surg. 2004;  240 900-909
  CrossrefPubMedSearch in Google Scholar
• 8 Lu D S, Yu N C, Raman S S et al.. Radiofrequency ablation of hepatocellular carcinoma: treatment success as defined by histologic examination of the explanted liver.  Radiology. 2005;  234 954-960
  CrossrefPubMedSearch in Google Scholar
• 9 Raman S S, Lu D S, Vodopich D J, Sayre J, Lassman C. Creation of radiofrequency lesions in a porcine model: correlation with sonography, CT, and histopathology.  AJR Am J Roentgenol. 2000;  175 1253-1258
  PubMedSearch in Google Scholar
• 10 Lu D S, Raman S S, Limanond P et al.. Influence of large peritumoral vessels on outcome of radiofrequency ablation of liver tumors.  J Vasc Interv Radiol. 2003;  14 1267-1274
  CrossrefPubMedSearch in Google Scholar
• 11 Peters R D, Hinks R S, Henkelman R M. Ex vivo tissue-type independence in proton-resonance frequency shift MR thermometry.  Magn Reson Med. 1998;  40 454-459
  PubMedSearch in Google Scholar
• 12 Wlodarczyk W, Boroschewski R, Hentschel M et al.. Three-dimensional monitoring of small temperature changes for therapeutic hyperthermia using MR.  J Magn Reson Imaging. 1998;  8 165-174
  PubMedSearch in Google Scholar
• 13 Moriarty J A, Chen J C, Purcell C M et al.. MRI monitoring of interstitial microwave-induced heating and thermal lesions in rabbit brain in vivo.  J Magn Reson Imaging. 1998;  8 128-135
  PubMedSearch in Google Scholar
• 14 Chen J C, Moriarty J A, Derbyshire J A et al.. Prostate cancer: MR imaging and thermometry during microwave thermal ablation-initial experience.  Radiology. 2000;  214 290-297
  PubMedSearch in Google Scholar
• 15 McDannold N, King R L, Jolesz F A, Hynynen K. The use of quantitative temperature images to predict the optimal power for focused ultrasound surgery: in vivo verification in rabbit muscle and brain.  Med Phys. 2002;  29 356-365
  CrossrefPubMedSearch in Google Scholar
• 16 Rhim H, Goldberg S N, Dodd III G D et al.. Essential techniques for successful radio-frequency thermal ablation of malignant hepatic tumors.  Radiographics. 2001;  21 S17-S39
  PubMedSearch in Google Scholar
• 17 Lepetit-Coiffe M, Quesson B, Seror O et al.. Real-time monitoring of radiofrequency ablation of rabbit liver by respiratory-gated quantitative temperature MRI.  J Magn Reson Imaging. 2006;  24 152-159
  CrossrefPubMedSearch in Google Scholar
• 18 Huppert P E, Trubenbach J, Schick F et al.. MRI-guided percutaneous radiofrequency ablation of hepatic neoplasms: first technical and clinical experiences.  Rofo. 2000;  172 692-700
  PubMedSearch in Google Scholar

Hervé LaumonierM.D. 

Department of Radiology, Hôpital Saint André

CHU Bordeaux, 1, rue jean Burguet, 33075, Bordeaux Cedex, France