- Email: [email protected]
Fluorodeoxyglucose Positron Emission Tomography Response and Normal Tissue Regeneration After Stereotactic Body Radiotherapy to Liver Metastases
International Journal of
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Gastrointestinal Cancer
Fluorodeoxyglucose Positron Emission Tomography Response and Normal Tissue Regeneration After Stereotactic Body Radiotherapy to Liver Metastases Michelle A. Stinauer, MD, Quentin Diot, PhD, David C. Westerly, PhD, Tracey E. Schefter, MD, and Brian D. Kavanagh, MD, MPH Department of Radiation Oncology, University of Colorado Denver, Aurora, Colorado Received Nov 9, 2011, and in revised form Feb 3, 2012. Accepted for publication Feb 3, 2012
Summary To characterize changes in standard uptake value (SUV) in positron emission tomography (PET) scans and determine the pace of normal tissue regeneration after stereotactic body radiation therapy (SBRT), we reviewed 35 liver metastases. The SUVmax decreased the first months after SBRT to plateau at 3.1. To score local failure by PET criteria, we propose a cutoff SUVmax 6, twice the baseline. PostSBRT values between 4 and 6 are suspicious for local failure. The volume of normal liver reached nadir 6 months after SBRT and regenerated the next 6 months.
Purpose: To characterize changes in standardized uptake value (SUV) in positron emission tomography (PET) scans and determine the pace of normal tissue regeneration after stereotactic body radiation therapy (SBRT) for solid tumor liver metastases. Methods and Materials: We reviewed records of patients with liver metastases treated with SBRT to 40 Gy in 3-5 fractions. Evaluable patients had pretreatment PET and 1 post-treatment PET. Each PET/CT scan was fused to the planning computed tomography (CT) scan. The maximum SUV (SUVmax) for each lesion and the total liver volume were measured on each PET/CT scan. Maximum SUV levels before and after SBRT were recorded. Results: Twenty-seven patients with 35 treated liver lesions were studied. The median follow-up was 15.7 months (range, 1.5-38.4 mo), with 5 PET scans per patient (range, 2-14). Exponential decay curve fitting (rZ0.97) showed that SUVmax declined to a plateau of 3.1 for controlled lesions at 5 months after SBRT. The estimated SUVmax decay half-time was 2.0 months. The SUVmax in controlled lesions fluctuated up to 4.2 during follow-up and later declined; this level is close to 2 standard deviations above the mean normal liver SUVmax (4.01). A failure cutoff of SUVmax 6 is twice the calculated plateau SUVmax of controlled lesions. Parenchymal liver volume decreased by 20% at 3-6 months and regenerated to a new baseline level approximately 10% below the pretreatment level at 12 months. Conclusions: Maximum SUV decreases over the first months after SBRT to plateau at 3.1, similar to the median SUVmax of normal livers. Transient moderate increases in SUVmax may be observed after SBRT. We propose a cutoff SUVmax 6, twice the baseline normal liver SUVmax, to score local failure by PET criteria. Post-SBRT values between 4 and 6 would be suspicious for local tumor persistence or recurrence. The volume of normal liver reached nadir 3-6 months after SBRT and regenerated within the next 6 months. Ó 2012 Elsevier Inc. Keywords: PET imaging, SBRT, Liver metastasis
Reprint requests to: Michelle A. Stinauer, MD, University of Colorado Denver, Department of Radiation Oncology, 1665 N. Aurora Ct, Campus
Int J Radiation Oncol Biol Phys, Vol. 83, No. 5, pp. e613ee618, 2012 0360-3016/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ijrobp.2012.02.008
Mail Stop F- 706, Aurora, CO 80045. Tel: (720) 848-0263; Fax: (720) 8480222; E-mail: [email protected] Conflicts of interest: none.
e614 Stinauer et al.
International Journal of Radiation Oncology Biology Physics
Introduction
were excluded from review if they did not have a pretreatment PET/CT scan and at least 1 follow-up PET/CT scan. Patients with non-FDG avid tumors or having prior treatment with selective internal radiotherapy were also excluded from analysis.
Stereotactic body radiotherapy (SBRT) has become an increasingly common component in the treatment of cancers, whether for primary cancers such as non-small cell lung cancer and hepatocellular carcinoma or in the metastatic setting. Magnetic resonance imaging (MRI) is the best imaging technique for detection and characterization of liver lesions because of its high liver-tolesion contrast and its high spatial resolution (1). MRI has superior anatomic delineation and is especially beneficial for determination of benign liver lesions. However, for detection of liver metastases, MRI and positron emission tomography (PET) using 2- [18F] fluoro-2-deoxyglucose (18FDG) are thought to be comparable (2). In the detection of liver metastases from gastrointestinal primary cancers, PET was found to be more sensitive than MRI in metaanalyses with similar specificity (3, 4). Although MRI has improved spatial delineation of lesions, PET can provide metabolic information not only on hepatic lesions but on other systemic metastases as well. Positron emission tomography uses 18FDG as a surrogate for glucose transportation into the cell and as a substrate for hexokinase during glycolysis. However, 18FDG is unable to undergo metabolism past the point of phosphorylation, which traps 18FDG inside the cell (5). The result is positron emission from cells of high glucose metabolism. Standardized uptake value (SUV) is a semiquantitative differential absorption ratio of the amount of radiotracer in tissue. The maximum SUV (SUVmax) is obtained from a single voxel in a given region of interest (ROI) and has been shown to be quite reproducible (6, 7). In addition to initial diagnosis and staging, PET is commonly used to monitor for response to therapy. Imaging with PET after conventionally fractionated radiotherapy has been shown to be a clinical predictor of not only local control (LC) but also overall survival in a variety of cancer sites. However, the data after SBRT have been limited to lung and pancreatic cancer. The pre-SBRT SUVmax has been shown to be a predictor of progression-free and overall survival in patients undergoing a single 25 Gy treatment for pancreatic cancer (8). However, pretreatment SUVmax in lung cancer patients undergoing SBRT did not predict for LC (9, 10). After SBRT to primary lung cancers, the treated lesions can show continued mild hypermetabolism without treatment failure upon further follow-up (11). Currently there are no such studies of FDG PET changes after SBRT to the liver. The primary objective of this study was to characterize the temporal change in SUV in PET/CT scans after SBRT for solid tumor liver metastases. To improve the efficacy of treatment monitoring with PET/CT we determined quantitative values in SUVmax rather than qualitative change. As a secondary objective, we also set out to describe normal tissue regeneration in the liver after SBRT and any corresponding metabolic changes in normal liver.
Methods and Materials Patients We reviewed the records of all patients undergoing liver SBRT from October 2003 through December 2009 within an institutional review board-approved retrospective analysis. All patient charts were reviewed for clinical and treatment information. Patients
SBRT technique Stereotactic body radiotherapy was given in 3-5 fractions to a minimum total dose of 40 Gy using a stereotactic technique as previously described (12). Briefly, all patients were immobilized during CT simulation and during treatment with a customized external vacuum-type or synthetic body mold. Abdominal compression was used to limit respiratory motion. The gross tumor volume (GTV) was considered equal to clinical tumor volume, and the planning target volume (PTV) was generated by adding 5 mm radially and 5-10 mm in the superior-inferior direction. More recently, after the adoption of 4-dimensional CT simulation, an internal target volume (ITV) was delineated and a uniform 5-mm margin expansion in all directions was used to generate the PTV. The dose was prescribed to cover 95% of the PTV, normalized to the 60%-80% isodose line. Localization was performed with kilovoltage orthogonal imaging or cone-beam CT.
Evaluation Local failure was scored when there was evidence of tumor viability by either the SUVmax 6 or by the Response Evaluation Criteria in Solid Tumors criterion of expansion of a solid mass with discrete borders within the treated PTV by 20% in the longest dimension relative to the most recent prior CT or MRI. Overall survival was determined from date of treatment completion to date of death or last known follow-up. The PET scans were performed using a hybrid PET/CT scanner (Discovery ST 16 slice; GE Healthcare, Fairfield, CT). Patients were instructed not to exercise 24 hours in advance of imaging and to fast 4 hours before radionuclide injection. Glucose level was obtained at the time of radionuclide injection and had to be <150 mg/dL for nondiabetics and <200 mg/dL for diabetics. Oral diabetic medications could be taken up to 3 hours before study, whereas insulin could be taken up to 2 hours before study. Approximately 18 mCi (range, 12-20 mCi) were injected 60 minutes before scanning. Each PET/CT scan, both pretreatment and follow-up, was fused to the planning CT scan using dedicated medical image analysis software (MIMvista; MIM Software, Cleveland, OH). Fusion was based on rigid alignment, with particular attention given to the treated area. The image analysis software calculated the SUV according to the radionuclide dose, time from radionuclide injection to scan, patient weight, and glucose level at the time of injection. The SUVmax from the GTV or ITV for each treated lesion was measured on each PET/CT scan. The SUVmax levels before and after SBRT were recorded, and an exponential decay curve was fit to the data using the formula: SUVmax t ZSUVmax ekt þ c to describe the change in SUVmax as a function of time after receiving SBRT, where k is the decay constant, and c is the baseline tissue value. Additionally, the liver was contoured on each follow-up PET/ CT scan and compared with the initial liver volume of the planning CT scan. Contouring and volume calculations were also
Volume 83 Number 5 2012 performed with MIMvista. All liver contours were drawn by the same physician to decrease interobserver variation.
Analysis Statistical analysis was performed with Prism software (GraphPad Software, La Jolla, CA) or MATLAB (MathWorks, Natwick, MA). Local control and overall survival were estimated using the KaplanMeier method. Evidence of statistically significant differences was assessed with the log-rank test statistic. All statistical tests were 2-sided, with statistical significance set at the level P.05.
Results Patient population Twenty-seven patients with 35 treated lesions met the study inclusion criteria. All patients were treated for various solid tumor metastases rather than primary hepatocellular cancer. The most common histology was colorectal with 13 patients, followed by head-and-neck cancer, lung, esophageal, and melanoma with 2 patients each. Seventeen patients were considered to have oligometastatic disease with 3 or fewer metastatic sites. Ten patients were considered to have extensive metastatic disease for which treatment was given for symptoms or to control dominant sites of disease. The median number of lesions treated was 1, with range 1-3. See Table for patient characteristics. The SBRT regimen was 40-60 Gy in 5 fractions for 15 lesions and 45-60 Gy in 3 fractions for 20 lesions. The regimen used was at the discretion of the physician and based on clinical objectives
Table Patient and treatment characteristics Patient characteristics Patients 27 Male 14 Female 13 Age (y) 59 (37-79) Lesions 35 Lesions per patient 1 (1-3) Histology Colorectal 13 Lung 2 Head-and-neck 2 Melanoma 2 Esophageal 2 Other 6 Treatment characteristics Gross tumor volume (cm3) 18 (1.9-109) Fractionation schedule 60 Gy in 3 fractions 13 60 Gy in 5 fractions 2 54 Gy in 3 fractions 3 50 Gy in 5 fractions 7 45 Gy in 3 fractions 4 45 Gy in 5 fractions 2 40 Gy in 5 fractions 4 Imaging follow-up (mo) 17.8 (1.5-65.4) Values are number or median (range).
FDG response to liver SBRT e615 and normal tissue dose considerations. The most common regimen was 60 Gy in 3 fractions (nZ13), followed by 50 Gy in 5 fractions (nZ7) and then 40 Gy in 5 fractions and 45 Gy in 3 fractions (both nZ4). Median GTV was 17.9 cm3 (range, 1.9-109.5 cm3). Median imaging follow-up for all lesions was 17.8 months (range, 1.5-65.4 months). See Table for treatment characteristics.
Toxicity Stereotactic body radiotherapy was well tolerated in all patients, with minimal acute toxicity. No long-term toxicity occurred. There were 9 grade 1 events, most commonly abdominal pain (nZ5), followed by nausea (nZ4). One patient had grade 3 nausea and vomiting during definitive chemoradiation to their primary gastric cancer. This toxicity was not thought to be related to SBRT.
Local control The population analyzed represents a subset of a larger cohort of patients treated with liver SBRT at the same institution. The clinical decision to obtain a follow-up PET scan was influenced by numerous factors. In some cases the scans were routine surveillance images; however, in other cases, a PET scan was ordered to evaluate suspicious findings on a post-SBRT CT or MRI scan. In view of the latter consideration, the actuarial rate of LC is possibly skewed by a selection bias toward lower LC. Additionally, whereas some patients suffered early deaths, it is possible that the actuarial LC would have declined with longer follow-up. Nevertheless, the actuarial rate of LC was 90% at 1 year and 85% at 18 months. Within this selected population, log-rank comparison did not reveal the number of fractions, total dose, nor histology to be significant predictors of LC. Although GTV size was also not a significant predictor for LC by log-rank comparison, only 1 failure occurred in lesions with GTV less than the median of 18 cm3. Kaplan-Meier actuarial analysis of pre-SBRT SUVmax greater than or equal to the median (6.9) did not reach a statistical difference in LC (49% vs 87%) at 2 years, although given the large difference this may have been because of the small number of failures. Similar analysis of other SUV cutoffs, including SUVmax 6.0, 7.0, 8.0, 9.0, and 10.0, also did not uncover any statistically significant differences in LC.
SUV kinetics A total of 192 PET scans were analyzed, with a median of 5 PET scans per patient. The median time after SBRT was 15.7 months. Fig. 1 shows the fitted exponential decay curve, including the SUVmax of the pretreatment SBRT and each post-SBRT PET scan for all lesions considered to be controlled at last follow-up. The calculated time for the SUVmax to decrease by half the original value was 2.0 months. The calculated post-treatment baseline SUVmax was 3.1, which was approached at approximately 5 months. A representative example of images of a controlled lesion is shown in Fig. 2. The median post-treatment SUVmax for controlled lesions was 3.1 (range, 2.1-5.8), consistent with the baseline level calculated using the exponential decay. Lesions judged to have recurred by PET imaging manifested quantitatively different patterns in
e616 Stinauer et al.
International Journal of Radiation Oncology Biology Physics Additionally we evaluated the normal liver for SUV changes in all patients. The entire liver excluding PTV had a median SUVmax of 3.9. The volumes of liver with SUV above a cutoff of 2.5 or 3.0 were not significantly different after radiotherapy. Additionally, the pattern of higher metabolism was diffuse throughout the liver and was not limited to the area surrounding the treated lesion.
Liver volumetrics The change in liver volume from the initial planning CT to followup PET/CT was calculated as a percentage and plotted (Fig. 4). The volumes were binned into groups according to post-SBRT intervals of 3 months. The liver volume decreased a mean of 20% at 3-6 months. The liver steadily regenerated after 12 months to return toward baseline size. However, the liver volume continued to be approximately 10% less than the original volume. See Fig. 5 for a representative patient showing liver contours.
Fig. 1. Maximum standardized uptake value (SUVmax) of the gross tumor volume or internal target volume of all controlled lesions from before and after stereotactic body radiotherapy plotted over time. An exponential decay curve was fitted to the data (solid line).
post-SBRT SUVmax values. Fig. 3 shows all SUVmax values for all patients, connected by line segments to indicate individual patients. Patients were judged to have local failure if the postSBRT SUVmax exceeded a value of 6 after a prior post-SBRT value below 6.
Discussion Assessing the status of tumor control after SBRT to the liver can be a challenging task. Herfarth et al (13) first described the imaging changes on CT scans that follow single-fraction SBRT, notably an enlarged region of relative hypodensity within the high-dose region. This portal venous hypodensity occurs within months after liver SBRT, and histologically it has been characterized as a result of central venous occlusion (14). The changes can mimic tumor progression; however, there is no suggestion of mass effect that deforms the liver contour. One possible advantage of PET/CT is that the tumor metabolic
Fig. 2. Representative case of a controlled lesion. The patient was treated with 50 Gy in 5 fractions. (A) Pretreatment positron emission tomography (PET)/computed tomography (CT) fused to the planning CT showing the internal target volume in green. The maximum standardized uptake value (SUVmax) was 5.8. (B) Five-month follow-up PET/CT was fused to planning CT with SUVmax of 3.4. (C) Ninemonth follow-up PET/CT fused to planning CT with SUVmax of 3.2.
Volume 83 Number 5 2012
Fig. 3. Maximum standardized uptake value (SUVmax) of the gross tumor volume or internal target volume of all lesions plotted over time. All failed lesions had SUVmax 6 (shown as the upper grey horizontal line). information might help differentiate Herfarth changes from tumor activity. Previous reports on SUV changes seen after SBRT to lung cancer have been published (9, 11, 15). In these reports, residual moderate to intense hypermetabolic activity after SBRT did not necessarily indicate residual active tumor. In the Hoopes report, 2 patients had SUV >5.0 without evidence of local recurrence. These patients were followed and subsequently found to have locally controlled disease. Likewise we found that the post-SBRT SUVmax can remain modestly elevated above the surrounding normal liver parenchyma. Among lesions considered controlled at last follow-up, fluctuation up to an SUVmax of 4.2 was observed. Residual post-SBRT low levels of elevated metabolic activity may occur as a result of ongoing normal tissue repair processes. Additionally, the liver itself under normal circumstances is continually metabolically active and can be quite heterogeneous in terms of SUVmax. In a report on FDG activity in 339 normal
Fig. 4. Liver volume changes as a function of post-stereotactic body radiotherapy interval.
FDG response to liver SBRT e617
Fig. 5. Representative case of liver volume change over time. The planning liver volume is shown in blue. The liver contoured on a 4-month follow-up scan and fused to the planning scan is shown in red. At 4 months it is 18.3% smaller than the original treatment volume. The liver contoured on a 13-month follow-up scan fused to the original plan is shown in green. It regenerated to 9.6% less than the original planning volume. healthy adult livers, the median SUVmax was 2.89 (16). This correlates well with our posttreatment lesion baseline SUVmax of 3.1. Our pretreatment liver (minus GTV) mean SUVmax was 3.94, which is higher than in normal controls. This may indicate that some patients had small foci of metastatic disease other than the treated lesion(s) because the SUVmax is for a single voxel. Alternatively, the presence of a metastatic lesion might stimulate inflammation in the surrounding parenchyma that is associated with elevated metabolic activity. In the Lin et al report (16), the mean voxel SUV plus 2 standard deviations summed to a level of 4.01. In the present series, as noted above, patients considered controlled at last follow-up had a post-SBRT SUVmax ranging up to approximately this same level. We have applied a proposed cutoff SUVmax 6, which is twice the baseline FDG activity and more than 5 standard deviations above mean normal liver voxel SUV, to score local failure by PET criteria. Fig. 3 demonstrates an apparent sharp contrast in the pattern of post-SBRT SUVmax between patients considered to have recurred by this criterion and those considered to remain controlled. The LC rate of 90% at 1 year is within the range of other reports of SBRT to liver metastases (17, 18). Given the aforementioned caveats about selection bias, whereby PET was often performed selectively for patients with suspicious findings on post-SBRT CT or MRI scans, it is not possible from this experience to draw firm conclusions about risk factors for local failure. The high doses delivered during SBRT not only ablate the tumor cells but also cause temporary injury to the surrounding normal tissue. The mean volume loss seen maximally at 3-6 months was 20%, consistent with a prior report from our institution by Olsen et al (14), in which the maximum percentage volume loss was further characterized as a function of mean normal liver dose. The normal liver begins to regenerate and hypertrophy in response, and returns to a baseline volume at approximately 12 months, which is similar to surgical resection data (19). Interestingly, in the Olsen et al study (16), the linear regression model relating mean liver dose to maximum percentage volume reduction did not intercept the y-axis at a value of 0. Likewise, in this study with longer intervals of follow-up and
e618 Stinauer et al. observation of regeneration from 6-12 months after SBRT, we observed a recovery plateau approximately 10% below the pretreatment level. We hypothesize that these two observations, taken together, reveal what is likely an initial tumor-related expansion of the liver before treatment. In other words, the tumor likely contributes volume to the liver as a combination of the malignant cells themselves along with local tissue edema and inflammation. Ablation of the tumor with SBRT seems to remove this stimulus for an artifactually increased liver volume, providing an explanation for these observations. There was no perceived correlation between liver regrowth and increased hypermetabolism in the liver after SBRT treatment, either focally around the treated lesion or generally throughout the liver. The absence of increased metabolism during the time of regeneration, most notably in the 6- to 12-month range, is likely because liver regeneration is a subtle, gradual process over months and not easily detected during a single scan. The limitations of this study include the small patient numbers, short follow-up, and previously discussed selection biases. Additionally, our observations might not apply to patients with primary liver cancer, who tend to have worse baseline liver function than patients with metastases. Nevertheless, the current observations are informative to clinicians following patients after liver SBRT and provide a framework for interpreting post-SBRT PET scans.
International Journal of Radiation Oncology Biology Physics
4.
5. 6.
7.
8.
9.
10.
11.
Conclusion Maximum SUV decreases over the first few months after SBRT to a plateau of 3.1 that is roughly equal to the median SUVmax of normal healthy livers. Transient moderate SUV increases, with SUVmax up to approximately 4, may be observed and do not correspond to subsequent local failure. We propose a cutoff SUVmax 6, twice the baseline FDG activity, to score local failure by PET criteria. Values from 4-6 would be suspicious for local recurrence. Finally, the average normal liver tissue volume reduction after SBRT was 20% and reached nadir at 3-6 months after SBRT. The liver regenerated to a new baseline level approximately 10% less than the pretreatment volume at 1 year.
References 1. Sahani DV, Kalva SP. Imaging the liver. Oncologist 2004;9:385e397. 2. Yang M, Martin DR, Karabulut N, et al. Comparison of MR and PET imaging for the evaluation of liver metastases. J Magn Reson Imaging 2003;17:343e349. 3. Kinkel K, Lu Y, Both M, et al. Detection of hepatic metastases from cancers of the gastrointestinal tract by using noninvasive imaging
12.
13.
14.
15.
16. 17.
18.
19.
methods (US, CT, MR imaging, PET): a meta-analysis. Radiology 2002;224:748e756. Bipat S, van Leeuwen MS, Comans EF, et al. Colorectal liver metastases: CT, MR imaging, and PET for diagnosisdmeta-analysis. Radiology 2005;237:123e131. Sols A, Crane RK. The inhibition of brain hexokinase by adenosinediphosphate and sulfhydryl reagents. J Biol Chem 1954;206:925e936. Velasquez LM, Boellaard R, Kollia G, et al. Repeatability of 18F-FDG PET in a multicenter phase I study of patients with advanced gastrointestinal malignancies. J Nucl Med 2009;50:1646e1654. Nahmias C, Wahl LM. Reproducibility of standardized uptake value measurements determined by 18F-FDG PET in malignant tumors. J Nucl Med 2008;49:1804e1808. Schellenberg D, Quon A, Minn AY, et al. 18Fluorodeoxyglucose PET is prognostic of progression-free and overall survival in locally advanced pancreas cancer treated with stereotactic radiotherapy. Int J Radiat Oncol Biol Phys 2010;77:1420e1425. Hoopes DJ, Tann M, Fletcher JW, et al. FDG-PET and stereotactic body radiotherapy (SBRT) for stage I non-small-cell lung cancer. Lung Cancer 2007;56:229e234. Burdick MJ, Stephans KL, Reddy CA, et al. Maximum standardized uptake value from staging FDG-PET/CT does not predict treatment outcome for early-stage non-small-cell lung cancer treated with stereotactic body radiotherapy. Int J Radiat Oncol Biol Phys 2010;78: 1033e1039. Henderson MA, Hoopes DJ, Fletcher JW, et al. A pilot trial of serial 18F-fluorodeoxyglucose positron emission tomography in patients with medically inoperable stage I non-small-cell lung cancer treated with hypofractionated stereotactic body radiotherapy. Int J Radiat Oncol Biol Phys 2010;76:789e795. Schefter TE, Kavanagh BD, Timmerman RD, et al. A phase I trial of stereotactic body radiation therapy (SBRT) for liver metastases. Int J Radiat Oncol Biol Phys 2005;62:1371e1378. Herfarth KK, Hof H, Bahner ML, et al. Assessment of focal liver reaction by multiphasic CT after stereotactic single-dose radiotherapy of liver tumors. Int J Radiat Oncol Biol Phys 2003;57:444e451. Olsen CC, Welsh J, Kavanagh BD, et al. Microscopic and macroscopic tumor and parenchymal effects of liver stereotactic body radiotherapy. Int J Radiat Oncol Biol Phys 2009;73:1414e1424. Matsuo Y, Nakamoto Y, Nagata Y, et al. Characterization of FDG-PET images after stereotactic body radiation therapy for lung cancer. Radiother Oncol 2010;97:200e204. Lin CY, Ding HJ, Lin CC, et al. Impact of age on FDG uptake in the liver on PET scan. Clin Imaging 2010;34:348e350. Rusthoven KE, Kavanagh BD, Cardenes H, et al. Multi-institutional phase I/II trial of stereotactic body radiation therapy for liver metastases. J Clin Oncol 2009;27:1572e1578. Herfarth KK, Debus J, Lohr F, et al. Stereotactic single-dose radiation therapy of liver tumors: results of a phase I/II trial. J Clin Oncol 2001; 19:164e170. Chen MF, Hwang TL, Hung CF. Human liver regeneration after major hepatectomy. A study of liver volume by computed tomography. Ann Surg 1991;213:227e229.
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Gastrointestinal Cancer
Fluorodeoxyglucose Positron Emission Tomography Response and Normal Tissue Regeneration After Stereotactic Body Radiotherapy to Liver Metastases Michelle A. Stinauer, MD, Quentin Diot, PhD, David C. Westerly, PhD, Tracey E. Schefter, MD, and Brian D. Kavanagh, MD, MPH Department of Radiation Oncology, University of Colorado Denver, Aurora, Colorado Received Nov 9, 2011, and in revised form Feb 3, 2012. Accepted for publication Feb 3, 2012
Summary To characterize changes in standard uptake value (SUV) in positron emission tomography (PET) scans and determine the pace of normal tissue regeneration after stereotactic body radiation therapy (SBRT), we reviewed 35 liver metastases. The SUVmax decreased the first months after SBRT to plateau at 3.1. To score local failure by PET criteria, we propose a cutoff SUVmax 6, twice the baseline. PostSBRT values between 4 and 6 are suspicious for local failure. The volume of normal liver reached nadir 6 months after SBRT and regenerated the next 6 months.
Purpose: To characterize changes in standardized uptake value (SUV) in positron emission tomography (PET) scans and determine the pace of normal tissue regeneration after stereotactic body radiation therapy (SBRT) for solid tumor liver metastases. Methods and Materials: We reviewed records of patients with liver metastases treated with SBRT to 40 Gy in 3-5 fractions. Evaluable patients had pretreatment PET and 1 post-treatment PET. Each PET/CT scan was fused to the planning computed tomography (CT) scan. The maximum SUV (SUVmax) for each lesion and the total liver volume were measured on each PET/CT scan. Maximum SUV levels before and after SBRT were recorded. Results: Twenty-seven patients with 35 treated liver lesions were studied. The median follow-up was 15.7 months (range, 1.5-38.4 mo), with 5 PET scans per patient (range, 2-14). Exponential decay curve fitting (rZ0.97) showed that SUVmax declined to a plateau of 3.1 for controlled lesions at 5 months after SBRT. The estimated SUVmax decay half-time was 2.0 months. The SUVmax in controlled lesions fluctuated up to 4.2 during follow-up and later declined; this level is close to 2 standard deviations above the mean normal liver SUVmax (4.01). A failure cutoff of SUVmax 6 is twice the calculated plateau SUVmax of controlled lesions. Parenchymal liver volume decreased by 20% at 3-6 months and regenerated to a new baseline level approximately 10% below the pretreatment level at 12 months. Conclusions: Maximum SUV decreases over the first months after SBRT to plateau at 3.1, similar to the median SUVmax of normal livers. Transient moderate increases in SUVmax may be observed after SBRT. We propose a cutoff SUVmax 6, twice the baseline normal liver SUVmax, to score local failure by PET criteria. Post-SBRT values between 4 and 6 would be suspicious for local tumor persistence or recurrence. The volume of normal liver reached nadir 3-6 months after SBRT and regenerated within the next 6 months. Ó 2012 Elsevier Inc. Keywords: PET imaging, SBRT, Liver metastasis
Reprint requests to: Michelle A. Stinauer, MD, University of Colorado Denver, Department of Radiation Oncology, 1665 N. Aurora Ct, Campus
Int J Radiation Oncol Biol Phys, Vol. 83, No. 5, pp. e613ee618, 2012 0360-3016/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ijrobp.2012.02.008
Mail Stop F- 706, Aurora, CO 80045. Tel: (720) 848-0263; Fax: (720) 8480222; E-mail: [email protected] Conflicts of interest: none.
e614 Stinauer et al.
International Journal of Radiation Oncology Biology Physics
Introduction
were excluded from review if they did not have a pretreatment PET/CT scan and at least 1 follow-up PET/CT scan. Patients with non-FDG avid tumors or having prior treatment with selective internal radiotherapy were also excluded from analysis.
Stereotactic body radiotherapy (SBRT) has become an increasingly common component in the treatment of cancers, whether for primary cancers such as non-small cell lung cancer and hepatocellular carcinoma or in the metastatic setting. Magnetic resonance imaging (MRI) is the best imaging technique for detection and characterization of liver lesions because of its high liver-tolesion contrast and its high spatial resolution (1). MRI has superior anatomic delineation and is especially beneficial for determination of benign liver lesions. However, for detection of liver metastases, MRI and positron emission tomography (PET) using 2- [18F] fluoro-2-deoxyglucose (18FDG) are thought to be comparable (2). In the detection of liver metastases from gastrointestinal primary cancers, PET was found to be more sensitive than MRI in metaanalyses with similar specificity (3, 4). Although MRI has improved spatial delineation of lesions, PET can provide metabolic information not only on hepatic lesions but on other systemic metastases as well. Positron emission tomography uses 18FDG as a surrogate for glucose transportation into the cell and as a substrate for hexokinase during glycolysis. However, 18FDG is unable to undergo metabolism past the point of phosphorylation, which traps 18FDG inside the cell (5). The result is positron emission from cells of high glucose metabolism. Standardized uptake value (SUV) is a semiquantitative differential absorption ratio of the amount of radiotracer in tissue. The maximum SUV (SUVmax) is obtained from a single voxel in a given region of interest (ROI) and has been shown to be quite reproducible (6, 7). In addition to initial diagnosis and staging, PET is commonly used to monitor for response to therapy. Imaging with PET after conventionally fractionated radiotherapy has been shown to be a clinical predictor of not only local control (LC) but also overall survival in a variety of cancer sites. However, the data after SBRT have been limited to lung and pancreatic cancer. The pre-SBRT SUVmax has been shown to be a predictor of progression-free and overall survival in patients undergoing a single 25 Gy treatment for pancreatic cancer (8). However, pretreatment SUVmax in lung cancer patients undergoing SBRT did not predict for LC (9, 10). After SBRT to primary lung cancers, the treated lesions can show continued mild hypermetabolism without treatment failure upon further follow-up (11). Currently there are no such studies of FDG PET changes after SBRT to the liver. The primary objective of this study was to characterize the temporal change in SUV in PET/CT scans after SBRT for solid tumor liver metastases. To improve the efficacy of treatment monitoring with PET/CT we determined quantitative values in SUVmax rather than qualitative change. As a secondary objective, we also set out to describe normal tissue regeneration in the liver after SBRT and any corresponding metabolic changes in normal liver.
Methods and Materials Patients We reviewed the records of all patients undergoing liver SBRT from October 2003 through December 2009 within an institutional review board-approved retrospective analysis. All patient charts were reviewed for clinical and treatment information. Patients
SBRT technique Stereotactic body radiotherapy was given in 3-5 fractions to a minimum total dose of 40 Gy using a stereotactic technique as previously described (12). Briefly, all patients were immobilized during CT simulation and during treatment with a customized external vacuum-type or synthetic body mold. Abdominal compression was used to limit respiratory motion. The gross tumor volume (GTV) was considered equal to clinical tumor volume, and the planning target volume (PTV) was generated by adding 5 mm radially and 5-10 mm in the superior-inferior direction. More recently, after the adoption of 4-dimensional CT simulation, an internal target volume (ITV) was delineated and a uniform 5-mm margin expansion in all directions was used to generate the PTV. The dose was prescribed to cover 95% of the PTV, normalized to the 60%-80% isodose line. Localization was performed with kilovoltage orthogonal imaging or cone-beam CT.
Evaluation Local failure was scored when there was evidence of tumor viability by either the SUVmax 6 or by the Response Evaluation Criteria in Solid Tumors criterion of expansion of a solid mass with discrete borders within the treated PTV by 20% in the longest dimension relative to the most recent prior CT or MRI. Overall survival was determined from date of treatment completion to date of death or last known follow-up. The PET scans were performed using a hybrid PET/CT scanner (Discovery ST 16 slice; GE Healthcare, Fairfield, CT). Patients were instructed not to exercise 24 hours in advance of imaging and to fast 4 hours before radionuclide injection. Glucose level was obtained at the time of radionuclide injection and had to be <150 mg/dL for nondiabetics and <200 mg/dL for diabetics. Oral diabetic medications could be taken up to 3 hours before study, whereas insulin could be taken up to 2 hours before study. Approximately 18 mCi (range, 12-20 mCi) were injected 60 minutes before scanning. Each PET/CT scan, both pretreatment and follow-up, was fused to the planning CT scan using dedicated medical image analysis software (MIMvista; MIM Software, Cleveland, OH). Fusion was based on rigid alignment, with particular attention given to the treated area. The image analysis software calculated the SUV according to the radionuclide dose, time from radionuclide injection to scan, patient weight, and glucose level at the time of injection. The SUVmax from the GTV or ITV for each treated lesion was measured on each PET/CT scan. The SUVmax levels before and after SBRT were recorded, and an exponential decay curve was fit to the data using the formula: SUVmax t ZSUVmax ekt þ c to describe the change in SUVmax as a function of time after receiving SBRT, where k is the decay constant, and c is the baseline tissue value. Additionally, the liver was contoured on each follow-up PET/ CT scan and compared with the initial liver volume of the planning CT scan. Contouring and volume calculations were also
Volume 83 Number 5 2012 performed with MIMvista. All liver contours were drawn by the same physician to decrease interobserver variation.
Analysis Statistical analysis was performed with Prism software (GraphPad Software, La Jolla, CA) or MATLAB (MathWorks, Natwick, MA). Local control and overall survival were estimated using the KaplanMeier method. Evidence of statistically significant differences was assessed with the log-rank test statistic. All statistical tests were 2-sided, with statistical significance set at the level P.05.
Results Patient population Twenty-seven patients with 35 treated lesions met the study inclusion criteria. All patients were treated for various solid tumor metastases rather than primary hepatocellular cancer. The most common histology was colorectal with 13 patients, followed by head-and-neck cancer, lung, esophageal, and melanoma with 2 patients each. Seventeen patients were considered to have oligometastatic disease with 3 or fewer metastatic sites. Ten patients were considered to have extensive metastatic disease for which treatment was given for symptoms or to control dominant sites of disease. The median number of lesions treated was 1, with range 1-3. See Table for patient characteristics. The SBRT regimen was 40-60 Gy in 5 fractions for 15 lesions and 45-60 Gy in 3 fractions for 20 lesions. The regimen used was at the discretion of the physician and based on clinical objectives
Table Patient and treatment characteristics Patient characteristics Patients 27 Male 14 Female 13 Age (y) 59 (37-79) Lesions 35 Lesions per patient 1 (1-3) Histology Colorectal 13 Lung 2 Head-and-neck 2 Melanoma 2 Esophageal 2 Other 6 Treatment characteristics Gross tumor volume (cm3) 18 (1.9-109) Fractionation schedule 60 Gy in 3 fractions 13 60 Gy in 5 fractions 2 54 Gy in 3 fractions 3 50 Gy in 5 fractions 7 45 Gy in 3 fractions 4 45 Gy in 5 fractions 2 40 Gy in 5 fractions 4 Imaging follow-up (mo) 17.8 (1.5-65.4) Values are number or median (range).
FDG response to liver SBRT e615 and normal tissue dose considerations. The most common regimen was 60 Gy in 3 fractions (nZ13), followed by 50 Gy in 5 fractions (nZ7) and then 40 Gy in 5 fractions and 45 Gy in 3 fractions (both nZ4). Median GTV was 17.9 cm3 (range, 1.9-109.5 cm3). Median imaging follow-up for all lesions was 17.8 months (range, 1.5-65.4 months). See Table for treatment characteristics.
Toxicity Stereotactic body radiotherapy was well tolerated in all patients, with minimal acute toxicity. No long-term toxicity occurred. There were 9 grade 1 events, most commonly abdominal pain (nZ5), followed by nausea (nZ4). One patient had grade 3 nausea and vomiting during definitive chemoradiation to their primary gastric cancer. This toxicity was not thought to be related to SBRT.
Local control The population analyzed represents a subset of a larger cohort of patients treated with liver SBRT at the same institution. The clinical decision to obtain a follow-up PET scan was influenced by numerous factors. In some cases the scans were routine surveillance images; however, in other cases, a PET scan was ordered to evaluate suspicious findings on a post-SBRT CT or MRI scan. In view of the latter consideration, the actuarial rate of LC is possibly skewed by a selection bias toward lower LC. Additionally, whereas some patients suffered early deaths, it is possible that the actuarial LC would have declined with longer follow-up. Nevertheless, the actuarial rate of LC was 90% at 1 year and 85% at 18 months. Within this selected population, log-rank comparison did not reveal the number of fractions, total dose, nor histology to be significant predictors of LC. Although GTV size was also not a significant predictor for LC by log-rank comparison, only 1 failure occurred in lesions with GTV less than the median of 18 cm3. Kaplan-Meier actuarial analysis of pre-SBRT SUVmax greater than or equal to the median (6.9) did not reach a statistical difference in LC (49% vs 87%) at 2 years, although given the large difference this may have been because of the small number of failures. Similar analysis of other SUV cutoffs, including SUVmax 6.0, 7.0, 8.0, 9.0, and 10.0, also did not uncover any statistically significant differences in LC.
SUV kinetics A total of 192 PET scans were analyzed, with a median of 5 PET scans per patient. The median time after SBRT was 15.7 months. Fig. 1 shows the fitted exponential decay curve, including the SUVmax of the pretreatment SBRT and each post-SBRT PET scan for all lesions considered to be controlled at last follow-up. The calculated time for the SUVmax to decrease by half the original value was 2.0 months. The calculated post-treatment baseline SUVmax was 3.1, which was approached at approximately 5 months. A representative example of images of a controlled lesion is shown in Fig. 2. The median post-treatment SUVmax for controlled lesions was 3.1 (range, 2.1-5.8), consistent with the baseline level calculated using the exponential decay. Lesions judged to have recurred by PET imaging manifested quantitatively different patterns in
e616 Stinauer et al.
International Journal of Radiation Oncology Biology Physics Additionally we evaluated the normal liver for SUV changes in all patients. The entire liver excluding PTV had a median SUVmax of 3.9. The volumes of liver with SUV above a cutoff of 2.5 or 3.0 were not significantly different after radiotherapy. Additionally, the pattern of higher metabolism was diffuse throughout the liver and was not limited to the area surrounding the treated lesion.
Liver volumetrics The change in liver volume from the initial planning CT to followup PET/CT was calculated as a percentage and plotted (Fig. 4). The volumes were binned into groups according to post-SBRT intervals of 3 months. The liver volume decreased a mean of 20% at 3-6 months. The liver steadily regenerated after 12 months to return toward baseline size. However, the liver volume continued to be approximately 10% less than the original volume. See Fig. 5 for a representative patient showing liver contours.
Fig. 1. Maximum standardized uptake value (SUVmax) of the gross tumor volume or internal target volume of all controlled lesions from before and after stereotactic body radiotherapy plotted over time. An exponential decay curve was fitted to the data (solid line).
post-SBRT SUVmax values. Fig. 3 shows all SUVmax values for all patients, connected by line segments to indicate individual patients. Patients were judged to have local failure if the postSBRT SUVmax exceeded a value of 6 after a prior post-SBRT value below 6.
Discussion Assessing the status of tumor control after SBRT to the liver can be a challenging task. Herfarth et al (13) first described the imaging changes on CT scans that follow single-fraction SBRT, notably an enlarged region of relative hypodensity within the high-dose region. This portal venous hypodensity occurs within months after liver SBRT, and histologically it has been characterized as a result of central venous occlusion (14). The changes can mimic tumor progression; however, there is no suggestion of mass effect that deforms the liver contour. One possible advantage of PET/CT is that the tumor metabolic
Fig. 2. Representative case of a controlled lesion. The patient was treated with 50 Gy in 5 fractions. (A) Pretreatment positron emission tomography (PET)/computed tomography (CT) fused to the planning CT showing the internal target volume in green. The maximum standardized uptake value (SUVmax) was 5.8. (B) Five-month follow-up PET/CT was fused to planning CT with SUVmax of 3.4. (C) Ninemonth follow-up PET/CT fused to planning CT with SUVmax of 3.2.
Volume 83 Number 5 2012
Fig. 3. Maximum standardized uptake value (SUVmax) of the gross tumor volume or internal target volume of all lesions plotted over time. All failed lesions had SUVmax 6 (shown as the upper grey horizontal line). information might help differentiate Herfarth changes from tumor activity. Previous reports on SUV changes seen after SBRT to lung cancer have been published (9, 11, 15). In these reports, residual moderate to intense hypermetabolic activity after SBRT did not necessarily indicate residual active tumor. In the Hoopes report, 2 patients had SUV >5.0 without evidence of local recurrence. These patients were followed and subsequently found to have locally controlled disease. Likewise we found that the post-SBRT SUVmax can remain modestly elevated above the surrounding normal liver parenchyma. Among lesions considered controlled at last follow-up, fluctuation up to an SUVmax of 4.2 was observed. Residual post-SBRT low levels of elevated metabolic activity may occur as a result of ongoing normal tissue repair processes. Additionally, the liver itself under normal circumstances is continually metabolically active and can be quite heterogeneous in terms of SUVmax. In a report on FDG activity in 339 normal
Fig. 4. Liver volume changes as a function of post-stereotactic body radiotherapy interval.
FDG response to liver SBRT e617
Fig. 5. Representative case of liver volume change over time. The planning liver volume is shown in blue. The liver contoured on a 4-month follow-up scan and fused to the planning scan is shown in red. At 4 months it is 18.3% smaller than the original treatment volume. The liver contoured on a 13-month follow-up scan fused to the original plan is shown in green. It regenerated to 9.6% less than the original planning volume. healthy adult livers, the median SUVmax was 2.89 (16). This correlates well with our posttreatment lesion baseline SUVmax of 3.1. Our pretreatment liver (minus GTV) mean SUVmax was 3.94, which is higher than in normal controls. This may indicate that some patients had small foci of metastatic disease other than the treated lesion(s) because the SUVmax is for a single voxel. Alternatively, the presence of a metastatic lesion might stimulate inflammation in the surrounding parenchyma that is associated with elevated metabolic activity. In the Lin et al report (16), the mean voxel SUV plus 2 standard deviations summed to a level of 4.01. In the present series, as noted above, patients considered controlled at last follow-up had a post-SBRT SUVmax ranging up to approximately this same level. We have applied a proposed cutoff SUVmax 6, which is twice the baseline FDG activity and more than 5 standard deviations above mean normal liver voxel SUV, to score local failure by PET criteria. Fig. 3 demonstrates an apparent sharp contrast in the pattern of post-SBRT SUVmax between patients considered to have recurred by this criterion and those considered to remain controlled. The LC rate of 90% at 1 year is within the range of other reports of SBRT to liver metastases (17, 18). Given the aforementioned caveats about selection bias, whereby PET was often performed selectively for patients with suspicious findings on post-SBRT CT or MRI scans, it is not possible from this experience to draw firm conclusions about risk factors for local failure. The high doses delivered during SBRT not only ablate the tumor cells but also cause temporary injury to the surrounding normal tissue. The mean volume loss seen maximally at 3-6 months was 20%, consistent with a prior report from our institution by Olsen et al (14), in which the maximum percentage volume loss was further characterized as a function of mean normal liver dose. The normal liver begins to regenerate and hypertrophy in response, and returns to a baseline volume at approximately 12 months, which is similar to surgical resection data (19). Interestingly, in the Olsen et al study (16), the linear regression model relating mean liver dose to maximum percentage volume reduction did not intercept the y-axis at a value of 0. Likewise, in this study with longer intervals of follow-up and
e618 Stinauer et al. observation of regeneration from 6-12 months after SBRT, we observed a recovery plateau approximately 10% below the pretreatment level. We hypothesize that these two observations, taken together, reveal what is likely an initial tumor-related expansion of the liver before treatment. In other words, the tumor likely contributes volume to the liver as a combination of the malignant cells themselves along with local tissue edema and inflammation. Ablation of the tumor with SBRT seems to remove this stimulus for an artifactually increased liver volume, providing an explanation for these observations. There was no perceived correlation between liver regrowth and increased hypermetabolism in the liver after SBRT treatment, either focally around the treated lesion or generally throughout the liver. The absence of increased metabolism during the time of regeneration, most notably in the 6- to 12-month range, is likely because liver regeneration is a subtle, gradual process over months and not easily detected during a single scan. The limitations of this study include the small patient numbers, short follow-up, and previously discussed selection biases. Additionally, our observations might not apply to patients with primary liver cancer, who tend to have worse baseline liver function than patients with metastases. Nevertheless, the current observations are informative to clinicians following patients after liver SBRT and provide a framework for interpreting post-SBRT PET scans.
International Journal of Radiation Oncology Biology Physics
4.
5. 6.
7.
8.
9.
10.
11.
Conclusion Maximum SUV decreases over the first few months after SBRT to a plateau of 3.1 that is roughly equal to the median SUVmax of normal healthy livers. Transient moderate SUV increases, with SUVmax up to approximately 4, may be observed and do not correspond to subsequent local failure. We propose a cutoff SUVmax 6, twice the baseline FDG activity, to score local failure by PET criteria. Values from 4-6 would be suspicious for local recurrence. Finally, the average normal liver tissue volume reduction after SBRT was 20% and reached nadir at 3-6 months after SBRT. The liver regenerated to a new baseline level approximately 10% less than the pretreatment volume at 1 year.
References 1. Sahani DV, Kalva SP. Imaging the liver. Oncologist 2004;9:385e397. 2. Yang M, Martin DR, Karabulut N, et al. Comparison of MR and PET imaging for the evaluation of liver metastases. J Magn Reson Imaging 2003;17:343e349. 3. Kinkel K, Lu Y, Both M, et al. Detection of hepatic metastases from cancers of the gastrointestinal tract by using noninvasive imaging
12.
13.
14.
15.
16. 17.
18.
19.
methods (US, CT, MR imaging, PET): a meta-analysis. Radiology 2002;224:748e756. Bipat S, van Leeuwen MS, Comans EF, et al. Colorectal liver metastases: CT, MR imaging, and PET for diagnosisdmeta-analysis. Radiology 2005;237:123e131. Sols A, Crane RK. The inhibition of brain hexokinase by adenosinediphosphate and sulfhydryl reagents. J Biol Chem 1954;206:925e936. Velasquez LM, Boellaard R, Kollia G, et al. Repeatability of 18F-FDG PET in a multicenter phase I study of patients with advanced gastrointestinal malignancies. J Nucl Med 2009;50:1646e1654. Nahmias C, Wahl LM. Reproducibility of standardized uptake value measurements determined by 18F-FDG PET in malignant tumors. J Nucl Med 2008;49:1804e1808. Schellenberg D, Quon A, Minn AY, et al. 18Fluorodeoxyglucose PET is prognostic of progression-free and overall survival in locally advanced pancreas cancer treated with stereotactic radiotherapy. Int J Radiat Oncol Biol Phys 2010;77:1420e1425. Hoopes DJ, Tann M, Fletcher JW, et al. FDG-PET and stereotactic body radiotherapy (SBRT) for stage I non-small-cell lung cancer. Lung Cancer 2007;56:229e234. Burdick MJ, Stephans KL, Reddy CA, et al. Maximum standardized uptake value from staging FDG-PET/CT does not predict treatment outcome for early-stage non-small-cell lung cancer treated with stereotactic body radiotherapy. Int J Radiat Oncol Biol Phys 2010;78: 1033e1039. Henderson MA, Hoopes DJ, Fletcher JW, et al. A pilot trial of serial 18F-fluorodeoxyglucose positron emission tomography in patients with medically inoperable stage I non-small-cell lung cancer treated with hypofractionated stereotactic body radiotherapy. Int J Radiat Oncol Biol Phys 2010;76:789e795. Schefter TE, Kavanagh BD, Timmerman RD, et al. A phase I trial of stereotactic body radiation therapy (SBRT) for liver metastases. Int J Radiat Oncol Biol Phys 2005;62:1371e1378. Herfarth KK, Hof H, Bahner ML, et al. Assessment of focal liver reaction by multiphasic CT after stereotactic single-dose radiotherapy of liver tumors. Int J Radiat Oncol Biol Phys 2003;57:444e451. Olsen CC, Welsh J, Kavanagh BD, et al. Microscopic and macroscopic tumor and parenchymal effects of liver stereotactic body radiotherapy. Int J Radiat Oncol Biol Phys 2009;73:1414e1424. Matsuo Y, Nakamoto Y, Nagata Y, et al. Characterization of FDG-PET images after stereotactic body radiation therapy for lung cancer. Radiother Oncol 2010;97:200e204. Lin CY, Ding HJ, Lin CC, et al. Impact of age on FDG uptake in the liver on PET scan. Clin Imaging 2010;34:348e350. Rusthoven KE, Kavanagh BD, Cardenes H, et al. Multi-institutional phase I/II trial of stereotactic body radiation therapy for liver metastases. J Clin Oncol 2009;27:1572e1578. Herfarth KK, Debus J, Lohr F, et al. Stereotactic single-dose radiation therapy of liver tumors: results of a phase I/II trial. J Clin Oncol 2001; 19:164e170. Chen MF, Hwang TL, Hung CF. Human liver regeneration after major hepatectomy. A study of liver volume by computed tomography. Ann Surg 1991;213:227e229.