Share this article on:

FDG PET/CT as a Prognostic Test After 90Y Radioembolization in Patients With Metastatic Hepatic Disease

Zalom, Martina MD*†; Yu, Run MD, PhD; Friedman, Marc MD§; Bresee, Catherine MS; Waxman, Alan MD§

doi: 10.1097/RLU.0b013e318262af7f
Original Articles

The purpose of this study was to determine FDG PET/CT utility in predicting patient outcome undergoing 90Y RE for metastatic liver tumors.

Thirty-one patients with metastatic tumors to the liver underwent 90Y RE between March 15, 2007, and May 5, 2011, at our institution. FDG PET/CT imaging was performed on each patient within 3 months before and after undergoing 90Y RE. Pretreatment and posttreatment FDG PET/CT were evaluated for SUV, number of liver lesions, and presence of disease outside the liver. The Kaplan-Meier method and Cox proportional hazard modeling was used to evaluate if SUV was a predictor of overall survival.

Of the 31 patients, 12 were alive at the end of the study; median survival was 9 months (95% confidence interval, 7–18 months). The 24-month survival rate was 0.28 (95% confidence interval, 0.12–0.48). Kaplan-Meier analysis of preprocedure FDG PET/CT imaging showed no difference in rates of survival by number of lesions observed in the liver (P = 0.114) or presence of disease observed outside the liver (P = 0.719). Cases with new lesions outside the liver after treatment had significantly shorter survival times than cases without new lesions outside the liver (P = 0.002). Cox proportional hazard model showed that SUV levels before and after treatment were not significant predictors of overall survival.

The appearance of new lesions outside the liver on FDG PET/CT within 3 months after 90Y RE was the only statistically significant variable in predicting poor outcome. Absence of new lesions outside the liver on postprocedure FDG PET/CT imaging helps identify patients achieving long-term survival.

From the *Department of Hematology/Oncology, Olive View – UCLA Medical Center, Sylmar; †Samuel Oschin Comprehensive Cancer Center and Departments of ‡Endocrinology, §Imaging, and ∥Biostatistics and Bioinformatics, Cedars-Sinai Medical Center, Los Angeles, CA.

Received for publication March 10, 2012; and revision accepted April 23, 2012.

Conflicts of interest and sources of funding: none declared.

Reprints: Alan D. Waxman, MD, Nuclear Medicine, Cedars-Sinai Medical Center, 8700 Beverly Blvd, Room 1251, Los Angeles, CA 90048. E-mail:

90Y radioembolization (RE) is an increasingly popular minimally invasive procedure used to treat primary and metastatic liver tumors. Microspheres labeled with 90Y isotope are infused through intrahepatic arterial catheters. The microspheres, with a diameter of 30 to 40 μm, are trapped in the tumor’s microvasculature and release beta radiation, which, in the tissue, has an average penetration of 2.5 mm. Radiation is thus applied to the tumor while limiting radiation to normal liver.1–3 90Y RE has been shown to significantly prolong survival in patients with chemotherapy refractory colorectal liver metastases. A phase 2 clinical trial of 50 patients with metastatic colorectal cancer receiving 90Y RE demonstrated a median overall survival of 12.6 months and 2-year survival of 19.6%.4 90Y RE has also been combined with chemotherapy to produce improved progression-free survivals, as shown in a phase 3 trial of 46 patients with metastatic colorectal cancer received either 5-fluorouracil alone or 5-fluorouracil plus 90Y RE. Median time to progression was 2.1 versus 4.5 months.5

Because 90Y RE treatment is palliative in nature and requires interventional procedures, patient selection is essential to maximize efficacy and avoid toxicities. Appropriate candidates for 90Y RE have life expectancy more than 12 weeks, Eastern Cooperative Oncology Group performance status (ECOG) 0 to 2, and reasonable liver reserve. Patients with ECOG performance status of 2 are unable to work; however, they are ambulatory, capable of all self-care, and up and about more than 50% of waking hours. Patients generally harbor unresectable, liver-dominant tumor burden.4 In a retrospective study of 51 patients with metastatic colorectal cancer, overall median survival after 90Y RE was 10.2 months. Imaging response obtained from CT or CT/PET to 90Y RE was evaluated using RECIST criteria. Absence of extrahepatic disease at the time of treatment was associated with improved survival, with a median survival of 17 months, compared with patients with extrahepatic disease at the time of treatment, with a median survival of 6.7 months.6

FDG PET detects glucose uptake and usage in tumor cells. More intense FDG uptake reflects a more metabolically active tumor.7–9 FDG PET may be more useful than CT in detecting early response to 90Y RE. In a retrospective study of 42 patients with unresectable liver metastases treated with 90Y RE, results were evaluated with CT using WHO, RECIST, and necrosis criteria and compared with response on FDG PET. FDG PET detected significantly more responses to treatment than CT, with 63% of patients showing response to treatment compared with CT by RECIST criteria (6%) or combined criteria (24%).10 Other studies also demonstrated that FDG PET consistently detects more response to 90Y RE than anatomic imaging.11–13 It is not clear, however, whether responsiveness in FDG PET activity would translate into clinical benefits.

The goal of this study was to determine whether metabolic imaging with FDG PET/CT provides important information on clinical outcomes in patients having undergone 90Y RE. Our study may help identifying patients most likely to benefit from this procedure.

Back to Top | Article Outline


Medical records of patients who underwent 90Y RE with resin micospheres at our institution between March 15, 2007, and May 5, 2011 were retrospectively reviewed. Of 107 patients evaluated for 90Y RE, 73 underwent the procedure. Of 73 patients, 31 underwent serial FDG PET/CT imaging. This study was reviewed and approved by our institution’s review board. All patients had received treatment for their metastatic cancers before their participation in this study. Prior treatments included chemotherapy, chemoembolization, and/or surgery. Selection criteria to undergo 90Y RE were as follows: liver-dominant metastatic burden, ECOG performance status of 2 or lower, life expectancy greater than 3 months, and tumor refractory to standard chemotherapy. Patients were excluded from receiving 90Y RE because of previous external beam radiation to the liver, total bilirubin greater than 1.5 mg/dL, albumin less than 3 g/dL, transaminases greater than 5 times the upper limit of normal, portal vein thrombosis, greater than 50% hepatic parenchymal involvement, greater than 20% liver/lung blood flow shunting, inability to isolate hepatic artery from splanchnic branches during mapping angiogram, disseminated extrahepatic disease, or avastin within 1 month. Disseminated disease was defined as organ involvement, other than liver, including that of bone, brain, and lung, but not including lymph node involvement.

FDG PET/CT imaging was performed within 3 months before 90Y RE and 3 months after 90Y RE. Before FDG PET/CT, patients were required to fast for a minimum of 6 hours before injection of 10 to 15 mCi of FDG. Finger stick blood glucose levels ranged between 65 and 120 mg/dL. Patients were imaged 90 minutes after injection with images obtained from the base of the brain through the midthigh region. CT was used for attenuation correction and localization. SUV normalized to body weight was used for all calculations.

FDG PET/CT images were interpreted by experienced nuclear medicine physicians. Preprocedure and postprocedure FDG PET/CT were evaluated for SUV of the most active liver lesion, change in SUV of the most active liver lesion between preprocedure and postprocedure scans, more than 3 liver lesions, presence of disease outside the liver, and change in size of liver lesions measured as larger, smaller, or stable.

Survival time was measured by time from preprocedure FDG PET/CT imaging to either date of last observation or death. Median observational period was 7 months (range, 2–35 months). Patient symptoms were assessed at 2 months after 90Y RE by chart review of clinic visits.

Survival times were analyzed by Kaplan-Meier method and tested across strata with the log-rank test. Cox proportional hazard modeling was used to evaluate if SUV was a predictor of overall survival, and tests of significance were performed with a Wald test. . Pretreatment versus posttreatment SUV data and change in SUV were compared with paired t tests, whereas ordered categorical data of symptom improvement were analyzed across strata with the nonparametric Wilcoxon rank sum test. All tests were 2-sided, and P < 0.05 was considered statistically significant. Analysis was performed with SAS v9.3 statistical software (SAS Institute, Cary, NC).

Back to Top | Article Outline


Origins of primary tumor of patients who received 90Y RE and underwent serial FDG PET/CT imaging were as follows: 17 colorectal, 3 neuroendocrine, 3 breast, 8 other origin. Of all patients, 14 (45%) men and 17 (55%) were women. There was no significant difference in the type of primary diagnosis in patients who received the procedure and those who did not (χ2 test, P = 0.3058). Among just the 73 patients who received the procedure, there was a significant difference in primary diagnosis between patients who received FDG PET/CT imaging and those who did not. Colorectal adenocarcinoma was more heavily represented in the group who received PET-CT imaging, with 17 patients, compared with 4 patients in the group who did not receive FDG PET/CT imaging (P < 0.0001). Neuroendocrine tumors were more heavily represented in the group who did not receive imaging, with 31 patients, compared with 3 patients in the group who did receive FDG PET/CT imaging. Not surprisingly, Kaplan-Meier analysis of the 73 patients who received the procedure showed significantly poorer prognosis in cases who underwent imaging compared with those who did not (log-rank test, P < 0.001), likely because of the overrepresentation of colorectal carcinoma (worse prognosis) and underrepresentation of neuroendocrine tumors (better prognosis) in the former group.

At baseline, 19 patients (61%) had 3 or more liver lesions and 14 patients (45%) had lesions outside the liver. Mean SUV level of the most active hepatic lesion decreased from before [7.1 (4.7)] to after treatment [5.4 (3.9)], although this difference was not statistically significant (Table 1). Two months after treatment, 9 cases were rated as better, 14 cases as same, and 8 cases as worse.



Twelve of 31 patients were alive at the end of the study, with a median survival of 9 months [95% confidence interval (CI), 7–18]. The 24-month survival rate was 0.28 (95% CI, 0.12–0.48).

Kaplan-Meier analysis of preprocedure FDG PET/CT imaging showed no difference in rates of survival by the number of lesions observed in the liver (P = 0.114) or disease observed outside the liver (P = 0.719). Postprocedure FDG PET/CT results with more lesions in the liver also were not a significant predictor of survival (P = 0.121). On the basis of the categorical change in lesion size from before to after treatment, the 12-month survival rate for those patients with tumor shrinkage was 0.45 (95% CI, 0.13–0.74) and 0.25 (95% CI, 0.04–0.56) for cases with tumor enlargement, although cases with no change had a 12-month survival rate of 0.86 (95% CI, 0.33–0.98), although this difference was only a statistical trend (P = 0.076; Fig. 1). Of 18 patients with liver lesions either smaller or larger, 11 had new lesions noted outside the liver on postprocedure FDG PET/CT. Of patients with liver lesions that were unchanged in size, fewer had new lesions noted outside the liver, 3 of 8 patients. Data for the change in the size of liver lesions were not available for 5 of the patients in our study; in 2 patients, because they died before postprocedure scans could be performed; and in 3 patients, because either preprocedure or postprocedure scans were not performed at our institution and hence were not available for direct comparison.



Patients with new lesions outside the liver after treatment had 12-month survival rates of 0.21 (95% CI, 0.05–0.44), which was significantly shorter than the 12-month rate for patients without any new lesions at 0.77 (95% CI, 0.34–0.94; P = 0.002; Fig. 2). Cox proportional hazard modeling showed SUV levels of hepatic lesions, including change in SUV, were not significant predictors of overall survival. Pretreatment SUV mean was 7.1 [hazard ratio (HR), 0.97; 95% CI, 0.89–1.07; P = 0.58], posttreatment SUV mean was 5.4 (HR, 1.08; 95% CI, 0.96–1.22; P = 0.18), and change was −1.8 (HR, 1.06; 95% CI, 0.96–1.16; P = 0.25).



Patients with colorectal cancer were evaluated as a subgroup. At the end of the study, 6 of 17 patients with colorectal cancer were alive, with a median survival time of 9 months (95% CI, 7–34 months). As previously mentioned, none of the criteria used to evaluate FDG PET/CT were significantly correlated with survival except for new lesions outside the liver observed on posttreatment FDG PET/CT (P = 0.027; Fig. 3). Cox proportional hazard modeling showed that neither SUV levels before or after treatment nor changes in SUV were significant predictors of overall survival.



Across the 3 group ratings of symptom change, there was no statistically significant difference in SUV level, although patients rated as “better” had larger decreases in SUV ratings than cases rated as “worse” or “same” (P = 0.078; Table 2. Cases rated “better” after treatment were statistically more likely to be rated as not having increased disease burden in the liver (P = 0.045). The preprocedure FDG PET/CT findings of the number of liver lesions and disease outside the liver were not significant predictors of posttreatment symptoms. Posttreatment FDG PET/CT findings of increased hepatic tumor burden and worsened disease outside the liver were also not significantly correlated with posttreatment symptoms.



Four patients involved in the study harbored neuroendocrine tumors; one of whom had neuroendocrine tumor of the breast. All had disease outside the liver at the time of 90Y RE. Of the 4 patients, all were alive at the end of the study, with time from treatment to end of study as 8, 11, 33, and 34 months. After 90Y RE, 2 of the patients experienced symptomatic improvement; both had decreased SUV of liver lesions after the procedure, one by 3 units and the other by 1 unit. Two patients experienced no change in their symptoms after 90Y RE; neither had change to SUV of liver lesions after the procedure and one had new lesions outside the liver.

Back to Top | Article Outline


90Y RE is an emerging treatment modality for primary or metastatic liver lesions. 90Y RE is an invasive, expensive, but relatively safe procedure. Patient selection is critical to achieve the maximal benefits to risks-and-benefits-to-cost ratios. Clinical criteria of patient selection help to a certain degree, but predicting outcomes of 90Y RE has been a challenge. FDG PET reflects glucose uptake and metabolic activity of tumor cells.7–9 Prior studies have validated FDG PET as a useful tool in detecting early response to 90Y RE treatment but did not sufficiently correlate PET activity and clinical benefits.10,11,13 It is reasonable to hypothesize that FDG PET uptake before and after 90Y RE would predict clinical outcomes. Our present study reveals interesting and somewhat unexpected results on the utility of FDG PET in predicting clinical outcome of 90Y RE.

The most important result of our study is that posttreatment FDG PET/CT findings are important in identifying patients expecting long-term survival after 90Y RE. Although the numbers are relatively small, no new extrahepatic lesions at 3 months portend a favorable prognosis, indicating that posttreatment FDG PET/CT is a significant tool in determining survival. Our observation is consistent with the clinical experience that extrahepatic metastasis of colorectal cancer is associated with adverse outcome and the results of a previous study showing a small survival advantage of patients with only hepatic disease,14 which supports the use FDG PET at 3 months after 90Y RE. Wong et al,15 using FDG PET/CT, demonstrated correlation between liver tumor burden and the presence of extrahepatic disease before 90Y RE treatment. FDG PET/CT may further help predict development of extrahepatic disease and identify patients unlikely to derive benefit from 90Y RE treatment.

Previous studies have suggested that decrease in FDG uptake is associated with favorable clinical outcome. A retrospective study of 78 patients undergoing 90Y RE showed a mean overall survival time of 25.63 (1.52) months in responders and 20.45 (2.11) months in nonresponders (P = 0.04).14 In our study, however, we have shown that reduction of SUV in liver lesions does not confer prognostic value in predicting survival, although it was associated with symptomatic improvement. The different results of 2 studies may reflect the differences in the studied patient population. In the previous study, a significant number of patients harbored hepatocellular carcinoma; in the present study, most patients harbored metastatic colorectal carcinoma. Our results may thus be more applicable to patients with metastatic colorectal carcinoma and suggest that liver response to 90Y RE as shown by FDG uptake changes is correlated with symptoms control rather than survival in such patients.

Our study was limited by a relatively small number of patients and by the heterogeneity of primary tumor diagnoses. Larger studies may be able to identify a subgroup for which reduction in SUV or the other criteria studied will affect prognosis. The retrospective nature of our study also limits its statistical persuasion. No a priori power analysis was performed because of the exploratory nature of this study. Furthermore, a post hoc power analysis would not be valid because the analysis of nonsignificant findings would result in the conclusion that the study was underpowered.

In summary, we have demonstrated that posttreatment FDG PET/CT follow-up is an important predictor of survival. Presence or absence of new lesions outside the liver defines which patients will achieve long-term survival. Larger series with longer follow-up may be important in determining if the trend in improvement of hepatic tumor SUV after 90Y-RE will achieve statistical significance in predicting long-term survival.

Back to Top | Article Outline


1. Sangro B, Bilbao JI, Iñarrairaegui M, et al.. Treatment of hepatocellular carcinoma by radioembolization using 90Y microspheres. Dig Dis. 2009; 27: 164–169.
2. Gulec SA, Fong Y. Yttrium 90 microsphere selective internal radiation treatment of hepatic colorectal metastases. Arch Surg. 2007; 142: 675–682.
3. Kennedy AS, Salem R. Radioembolization (yttrium-90 microspheres) for primary and metastatic hepatic malignancies. Cancer J. 2010; 16: 163–175.
4. Cosimelli M, Golfieri R, Cagol PP, et al.. Italian Society of Locoregional Therapies in Oncology (SITILO). Multi-centre phase II clinical trial of yttrium-90 resin microspheres alone in unresectable, chemotherapy refractory colorectal liver metastases. Br J Cancer. 2010; 103: 324–331.
5. Hendlisz A, Van den Eynde M, Peeters M, et al.. Phase III trial comparing protracted intravenous fluorouracil infusion alone or with yttrium-90 resin microspheres radioembolization for liver-limited metastatic colorectal cancer refractory to standard chemotherapy. J Clin Oncol. 2010; 28: 3687–3694.
6. Nace GW, Steel JL, Amesur N, et al.. Yttrium-90 radioembolization for colorectal cancer liver metastases: a single institution experience. Int J Surg Onc. 2011; 2011: 1–9.
7. Delberke D, Martin WH. Update of PET and PET/CT for hepatobiliary and pancreatic malignancies. HPB. 2005; 7: 166–179.
8. Pantaleo MA, Nannini M, Maleddu A, et al.. Conventional and novel PET tracers for imaging in oncology in the era of molecular therapy. Cancer Treat Rev. 2008; 34: 103–121.
9. Berry JD, Cook GJ. Positron emission tomography in oncology. Br Med Bull. 2006; 79–80: 171–186.
10. Miller FH, Keppke AL, Reddy D, et al.. Response of liver metastases after treatment with yttrium-90 microspheres: role of size, necrosis, and PET. AJR Am J Roentgenol. 2007; 188: 776–783.
11. Szyszko T, Al-Nahhas A, Canelo R, et al.. Assessment of response to treatment of unresectable liver tumours with 90Y microspheres: value of FDG PET versus computed tomography. Nucl Med Commun. 2007; 28: 15–20.
12. Wong CY, Salem R, Raman S, et al.. Evaluating 90Y-glass microsphere treatment response of unresectable colorectal liver metastases by [18F]FDG PET: a comparison with CT or MRI. Eur J Nucl Med Mol Imaging. 2002; 29: 815–820.
13. Jiao LR, Szyszko T, Al-Nahhas A, et al.. Clinical and imaging experience with yttrium-90 microspheres in the management of unresectable liver tumours. Eur J Surg Oncol. 2007; 33: 597–602.
14. Kucuk ON, Soydal C, Lacin S, et al.. Selective intraarterial radionuclide therapy with yttrium-90 (Y-90) microspheres for unresectable primary and metastatic liver tumor. World J Surg Oncol. 2011; 9: 86.
15. Wong CY, Gates VL, Tang B, et al.. Fluoro-2-deoxy-D-glucose positron emission tomography/computed tomography predicts extrahepatic metastatic potential of colorectal metastasis: a practical guide for yttrium-90 microsphere liver-directed therapy. Cancer Biother Radiopharm. 2010; 25: 233–236.

90Y radioembolization; FDG PET/CT; prognosis

© 2012 Lippincott Williams & Wilkins, Inc.