- Email: [email protected]
Yttrium-90 Radioembolization for Liver Malignancies: Prognostic Factors Associated with Survival
Yttrium-90 Radioembolization for Liver Malignancies: Prognostic Factors Associated with Survival Brian L. Dunfee, MD, Ahsun Riaz, MD, Robert J. Lewandowski, MD, Saad Ibrahim, MD, Mary F. Mulcahy, MD, Robert K. Ryu, MD, Bassel Atassi, MD, Kent T. Sato, MD, Steven Newman, MD, Reed A. Omary, MD, Al Benson III, MD, and Riad Salem, MD, MBA
PURPOSE: To identify key prognostic clinical and imaging variables in patients undergoing yttrium-90 radioembolization (90Y) for liver malignancies. MATERIALS AND METHODS: Patients with liver malignancies that progressed despite standard-of-care therapy were treated with 90Y from 2002 to 2006. Baseline functional status, laboratory values, and diagnostic imaging were assessed before therapy. Imaging follow-up was performed 1 month after treatment and subsequently at 3-month intervals. Patients were followed for survival from the time of their first 90Y treatment. RESULTS: Patients with follow-up imaging after radioembolization (N ⴝ 130) were included in this analysis. Primary malignancies included colon, neuroendocrine, and others. The following clinical variables had a significant effect on survival on multivariate analysis: Eastern Cooperative Oncology Group (ECOG) performance status (PS) greater than 0 (hazard ratio [HR], 7.98; 95% CI, 3.98 –16), hepatic tumor burden of 51%–75% (HR, 2.46; 95% CI, 1.01– 6.02), bilirubin level greater than 1.3 mg/dL (HR, 2.60; 95% CI, 1.27–5.34), hepatic metastases from breast cancer (HR, 2.51; 95% CI, 1.13–5.61), response on imaging based on World Health Organization (WHO) criteria (HR, 0.48; 95% CI, 0.24 – 0.94), and lymphocyte depression (HR, 0.56; 95% CI, 0.31– 0.96). Among patients with colorectal cancer metastases to the liver, the HR for survival on univariate analysis for responders compared with nonresponders (per WHO criteria) was 0.26 (95% CI, 0.10 – 0.69). CONCLUSIONS: Cancer-related symptoms (ie, ECOG PS > 0), hepatic tumor burden greater than 50%, increased bilirubin levels, and hepatic metastases from breast cancer were found to be negative prognostic factors. Tumor response to therapy and lymphocyte depression were associated with favorable prognosis. Additionally, WHO response was identified to be a favorable prognostic factor in patients with colorectal cancer metastases. These findings may be useful when counseling patients regarding prognosis of their hepatic disease. J Vasc Interv Radiol 2010; 21:90 –95 Abbreviations:
ECOG ⫽ Eastern Cooperative Oncology Group, HR ⫽ hazard ratio, PS ⫽ performance status, WHO ⫽ World Health Organization
MULTIPLE oncologic factors have been studied in determining the prognostic outcome of patients with metastatic dis-
ease (1– 4). The presence of metastases to the liver predicts a worse outcome in nearly all tumors, with the site of the
From Diagnostic Imaging, Inc., Aria Health Hospitals, Philadelphia, Pennsylvania (B.L.D.); Department of Radiology, Section of Interventional Radiology (A.R., R.J.L., S.I., M.F.M., R.K.R., B.A., K.T.S., R.A.O., R.S.), and Department of Medicine, Division of Hematology and Oncology (M.F.M., S.N., A.B., R.S.), Northwestern Memorial Hospital and Robert H. Lurie Comprehensive Cancer Center, 676 North St. Clair St., Suite 800, Chicago, IL 60611. Received October 10, 2008; final revision received August 19, 2009; accepted September 9,
2009. Address correspondence to R.S.; E-mail: [email protected]
90
From the SIR 2008 Annual Meeting. R.S. is a paid consultant of MDS Nordion (Ottawa, Ontario, Canada). None of the other authors have identified a conflict of interest. © SIR, 2010 DOI: 10.1016/j.jvir.2009.09.011
primary tumor being an additional factor. Performance status (PS) has also been proven useful in predicting the long-term prognosis. The most commonly used system for grading performance was proposed by the Eastern Cooperative Oncology Group (ECOG) (5). Many studies have demonstrated that a higher ECOG PS is related to a higher short-term mortality rate (1). In addition, laboratory values such as bilirubin levels may also assist in predicting the effects of the metastatic deposit on the functional status of the liver (3,4). Many other prognostic factors previously investigated include sex, age, lactic dehy-
Volume 21
Number 1
drogenase level, number of involved segments, and presence of extrahepatic metastases. Tumor markers such as carcinoembryonic antigen and cancer antigen 19-9 have been proposed, but the data are not consistent (4). Additional prognostic indicators can be determined by the response of metastatic lesions after various medical and surgical treatments. Significantly improved survival benefit in patients with metastatic colorectal carcinoma results if the disease does not progress while the patient is receiving neoadjuvant chemotherapy (6). The purpose of the present study was to determine prognostic clinical variables for patients who underwent catheter-directed yttrium-90 therapy in a large cohort of patients with liver metastases. Our goal was determine the prognostic factors that had the most influence on outcome of patients after treatment.
MATERIALS AND METHODS Patient Population An open label cohort study was performed between 2002 and 2006 of 137 patients with liver malignancies refractory to standard-of-care therapy. One hundred thirty patients with follow-up imaging were studied for this analysis. Seven patients were not able to receive follow-up imaging because of early death (⬍2 months after treatment). The data were collected prospectively. The PS of each patient was documented according to the ECOG grading criteria (5). Laboratory values including bilirubin levels were obtained for all patients. Each patient had preprocedural imaging such as computed tomography (CT) and/or magnetic resonance (MR) imaging depending on the primary neoplasm. The study was approved by the investigational review board and was conducted in a manner compliant with the Health Insurance Portability and Accountability Act. The longterm outcomes including safety, toxicity, tumor response, and survival of this patient cohort have been published elsewhere (7). Inclusion and Exclusion Criteria The following inclusion criteria were met for each patient: (i) progressive liver disease despite standard-of-care therapy
Dunfee et al (octreotide for neuroendocrine tumors and chemotherapy for all others); (ii) disease not amenable to surgery; (iii) ECOG PS of 0–2; (iv) noncompromised pulmonary function; (v) ability to undergo angiography and selective visceral catheterization; (vi) adequate hematologic parameters (ie, granulocyte count ⱖ1.5 ⫻ 109/L, platelets ⱖ50 ⫻ 109/L) and renal function (creatinine ⱕ2.0 mg/dL); (vii) adequate liver function (ie, bilirubin ⱕ2.0 mg/dL); and (viii) limited extrahepatic disease (eg, solitary lung metastasis, portal lymphadenopathy ⬎2 cm, adrenal metastases, and limited peritoneal carcinomatosis, ie, deemed as having a minor contribution to long-term morbidity and mortality). Exclusion criteria included (i) evidence of any uncorrectable flow to the gastrointestinal tract observed on angiography or Technetium-99m macroaggregated albumin (99mTc-MAA) scan; (ii) radiographic/pathologic confirmation of hepatocellular carcinoma; and (iii) estimated radiation dose greater than 30 Gy (16.5 mCi) delivered to the lungs in a single administration or 50 Gy on multiple administrations. Technique Patient evaluation.—Patients were referred to the interventional radiology unit for 90Y radioembolization. Baseline functional status of each patient was determined according to the ECOG criteria. Preprocedural angiography was performed to evaluate anatomy and identify vessels that required intervention such as embolization. At this time, 99m Tc-MAA was then administered to estimate the shunted fraction of 90Y to the lungs (8). Finally, baseline imaging was evaluated for the distribution of disease, hepatic tumor burden, and the presence/absence of extrahepatic metastases. Device.—TheraSphere (MDS Nordion, Ottawa, Ontario, Canada) consists of insoluble glass microspheres with 90Y as an integral constituent of the glass. The mean sphere diameter ranges from 20 to 30 m. Yttrium-90 is a pure -emitter with a 64.1-hour physical halflife. The specific activity per microsphere is approximately 2,500 Bq at the time of calibration. The method of activity requirement for treatment has been previously published (9). The 90Y radioisotope is produced by the neutron bombardment of 90Y, yielding a pure
•
91
-emitter with a high-energy emission of 0.936 keV. Tissue penetration of the activity is 11 mm, with a mean penetration of 2.5 mm. Given the hypervascularity of tumors, the microspheres are preferentially distributed to the tumor and produce localized radiation (10). Clinicians are directed to published Food and Drug Administration guidelines for the use of humanitarian device exemptions in disease conditions other than the approved indication, ie, hepatocellular carcinoma (11). The patients received bilobar or unilobar treatment depending on the extent of disease within the liver at the time of presentation. For bilobar treatment, the first lobe treatment was followed by the second lobe treatment approximately 30 – 40 days later. No whole-liver infusions were performed. Lobar treatment was targeted to deliver 120 Gy according to previously published dosimetry techniques (9). All patients were discharged from the hospital within 6 hours after completion of the procedure. Tumor Response Imaging follow-up was performed 1 month after treatment and subsequently every 3 months. Response was assessed according to World Health Organization (WHO) guidelines (12,13). Tumor response on CT or MR imaging was determined for a maximum of four measurable lesions (⬎1 cm) in terms of the product of the greatest cross-sectional tumor length and the length of the corresponding orthogonal projection (ie, “cross-product”). Corresponding lesions from baseline and posttreatment scans were compared for size changes. Complete response was defined by a change in the sum of the cross-products to zero, partial response as a decrease in the sum of cross-products by at least 50%, stable disease as a decrease in the sum of cross-products by less than 50% or an increase less than 25%, and progression as an increase in the sum of cross-products by at least 25% (12,13). Posttreatment studies were compared and assessed simultaneously with pretreatment images. Laboratory Analyses For this study, the only laboratory parameter used to assess baseline liver function was baseline bilirubin level. Bilirubin levels greater than 1.3 mg/dL
92
•
Prognostic Factors for Survival after Liver
(our institutional normal level) were classified as being increased. As 90Y leads to posttreatment decrease in lymphocyte counts, lymphocyte depression was also studied (14). Lymphocyte depression was defined as a decrease in absolute lymphocyte count from normal (⬎1,000/L) to less than normal at any time after 90Y administration. Statistical Analyses Exploratory analyses to identify variables that may have an effect on survival were performed. For these analyses, patients were classified into two groups: those who were alive at the time of study completion and those who were dead at the time of study completion. The characteristics between these two groups were compared with use of the Fisher exact or 2 test. Survival for the cohort was analyzed from the date of initial intraarterial 90Y microsphere administration. The univariate analyses were performed with use of the log-rank test and the multivariate analyses were performed with use of the Cox proportional-hazards model (ie, stepwise) to determine the prognostic value of each variable analyzed (15). A P value less than .05 was deemed to indicate a significant difference. As the backward stepwise method of Cox proportional hazards eliminates nonsignificant variables sequentially, all variables were entered into the initial multivariate analysis. A substratification analysis was also performed on tumor response in patients with colorectal cancer.
RESULTS
90
January 2010
Y Radioembolization
JVIR
Table 1 Exploratory Analysis of Prognostic Benefit of Variables At Study Completion
Variable
Total (N ⫽ 130)
Alive (n ⫽ 67)
Dead (n ⫽ 63)
54 76
31 (57) 36 (47)
23 (43) 40 (53)
.34
48 82
8 (17) 59 (72)
40 (83) 23 (18)
⬍.001
105 25
51 (49) 16 (89)
54 (51) 9 (20)
.24
0 6 18 106
0 0 (0) 7 (39) 60 (57)
0 6 (100) 11 (61) 46 (43)
.03
66 64
27 (41) 40 (63)
39 (59) 24 (37)
.02
11 119
0 (0) 67 (56)
11 (100) 52 (44)
⬍.001
88 42
46 (53) 21 (50)
42 (47) 21 (50)
.95
48 82
32 (67) 35 (43)
16 (33) 47 (57)
.01
20 17 50 7 36
12 (60) 13 (76) 26 (52) 4 (57) 12 (33)
8 (40) 4 (24) 24 (48) 3 (43) 24 (67)
.046
67 63
32 (48) 31 (49)
35 (52) 32 (51)
Age (y) ⬍65 ⱖ65 ECOG PS ⬎0 0 Lobar involvement Bilobar Unilobar Hepatic tumor burden (%) ⬎75 51–75 26–50 0–25 Extrahepatic metastases Yes No Increased bilirubin (⬎1.3 mg/dL) Yes No Higher dose utilization (ⱖ104 Gy) Yes No Tumor response (WHO Criteria) Yes No Tumor Breast carcinoma Neuroendocrine Colon carcinoma Cholangiocarcinoma Other carcinomas Lymphocyte depression Yes No
P Value
1.00
Note.—Values in parentheses are percentages.
Patient Sample In the study, 130 underwent 218 intraarterial administrations of 90Y microspheres. Seven patients did not have follow-up imaging and hence were not included in the regression analyses. The entire analysis was therefore based on 130 patients. There were 69 women and 61 men, with a median age of 61 years. Primary malignancies included colon (n ⫽ 50), breast (n ⫽ 20), neuroendocrine (n ⫽ 17), cholangiocarcinoma (n ⫽ 7), and others (n ⫽ 36; pancreas, lung, melanoma, renal, esophageal, and adenocarcinoma of unknown primary lesion). All patients were considered to have unresectable disease. A total of 63 patients (48%) had died
and 67 (52%) were alive at the time of study completion (Table 1). Table 2 presents the univariate and multivariate analysis for the study.
younger than 65 years of age was 1.45 (95% CI, 0.87–2.43) on univariate analysis.
Age
Our study population was distributed into two groups before treatment: those with normal functional status (ie, ECOG PS of 0) and those with some level of physical impairment and cancer-related symptoms (ie, ECOG PS ⬎ 0). Of the patients who had an ECOG PS greater than 0 and those who had an ECOG PS of 0, 40 (83%) and 23 (18%) had died at the time of study completion, respectively. The HRs for survival in pa-
Our study population included a spectrum of age ranges, with the majority of patients greater than 65 years of age (58%). Of the patients who were younger than 65 years of age and 65 years of age or older, 23 (43%) and 40 (53%) had died at the time of study completion, respectively. The hazard ratios (HRs) for survival in patients 65 years of age or older compared with those
ECOG PS
Volume 21
Number 1
Dunfee et al
Univariate Analyses
Age (y) ⬍65 ⱖ65 Sex Female Male ECOG PS 0 ⬎0 Lobar involvement Unilobar Bilobar Hepatic tumor burden (%) 51–75 26–50 0–25 Extrahepatic metastases Yes No Increased bilirubin (mg/dL) ⬎1.3 ⱕ1.3 Dose utilization (Gy) ⬎104 ⱕ104 Positive tumor response to treatment Yes No Tumor Breast carcinoma Neuroendocrine Colon carcinoma Cholangiocarcinoma Other carcinomas Lymphocyte depression Yes No
HR (95% CI)
P Value
Multivariate Analysis HR (95% CI)
P Value
Reference 1.45 (0.87–2.43)
.1552
— —
—
Reference 1.16 (0.70–1.90)
.56
— —
—
Reference 10.0 (5.47–18.5)
⬍.0001
Reference 7.98 (3.98–16.0)
⬍.0001
Reference 1.87 (0.92–3.80)
.0860
— —
—
4.50 (1.91–10.6) 1.35 (0.70–2.59) Reference
.0006 .3680
2.46 (1.01–6.02) Reference
.048
2.30 (1.38–3.83) Reference
.0015
— —
2.91 (1.51–5.61) Reference
.0014
2.60 (1.27–5.34) Reference
1.10 (0.65–1.86) Reference
.7286
— —
—
0.30 (0.17–0.55) Reference
.0001
0.48 (0.24–0.94) Reference
.034
1.59 (0.75–3.38) 0.33 (0.12–0.92) 0.84 (0.50–1.39) 0.84 (0.26–2.70) Reference 0.62 (0.38–1.02) Reference
tients with an ECOG PS greater than 0 versus those with an ECOG PS of 0 were 10 (95% CI, 5.47–18.5) and 7.98 (95% CI, 3.98 –16.0) on univariate and multivariate analyses, respectively. Therefore, patients with compromised functional status (ie, ECOG PS of 1/2) are at high risk for worse outcomes after 90Y compared with asymptomatic patients (ie, ECOG PS of 0). Lobar Involvement The majority of patients in our study population had bilobar involvement of metastatic disease. Of the patients who had bilobar and unilobar disease in-
93
(95% CI, 1.01– 6.02) on univariate analysis.
Table 2 Univariate and Multivariate Analyses
Variables
•
.23 .03 .49 .78
.06
— .0092
Extrahepatic Metastases Sixty-six patients had extrahepatic metastases in the study population. Of the patients with and without extrahepatic metastases, 39 (59%) and 24 (37%) had died at the time of study completion, respectively. The HR for survival in patients with extrahepatic metastases versus patients without extrahepatic metastases was 2.30 (95% CI, 1.38 –3.83) on univariate analysis. Baseline Bilirubin Levels Increased bilirubin levels were defined as those with any value greater than 1.3 mg/dL during the initial clinical evaluation. Eleven patients had an increased bilirubin level before 90Y microsphere therapy, and all had died before study completion. The HRs for survival in patients with high baseline bilirubin levels versus those with normal baseline bilirubin levels were 2.91 (95% CI, 1.51–5.61) and 2.60 (95% CI, 1.27– 6.34) on univariate and multivariate analyses, respectively. Dose Utilization
2.52 (1.13–5.61) Reference
.26
0.55 (0.31–0.96) Reference
.036
volvement, 54 (51%) and nine (29%) had died at the time of study completion, respectively. The HR for survival in patients with bilobar disease versus patients with unilobar disease was 1.87 (95% CI, 0.92–3.8) on univariate analysis. Hepatic Tumor Burden Of the patients who had 51%–75%, 26%–50% and 0%–25% hepatic tumor burden, six (100%), 11 (61%), and 46 (43%) had died at the time of study completion, respectively. The HRs for survival in patients with hepatic tumor burdens of 50%–75% versus 0%–25% were 4.50 (95% CI, 1.91–10.6) and 2.46
A higher dose for treatment was defined as dose delivery of greater than 104 Gy (16). A total of 88 patients in the study population required a dose greater than 104 Gy and 42 patients received lower treatment doses. Of the patients who received doses greater than 104 Gy and those who received doses of 104 Gy or less, 42 (47%) and 21 (50%) had died at the time of study completion, respectively. The HR for survival in patients who received high radiation doses compared with patients who received lower doses was 1.10 (95% CI, 0.65–1.86) on univariate analysis. Tumor Response Maximum tumor response to treatment was determined by imaging characteristics for all patients and all 332 lesions measured. At the lesional level, response rate was 42.8% (142 of 332). At the patient level, response rate was 37% (48 of 130). Of the patients who did and did not show a response per WHO criteria, 16 (33%) and 47 (57%) had died at the time of study completion, respec-
94
•
Prognostic Factors for Survival after Liver
tively. The HRs for survival in patients who showed a response per WHO criteria compared with those who showed no response were 0.30 (95% CI, 0.17– 0.55) and 0.48 (95% CI, 0.24 – 0.94) on univariate and multivariate analyses, respectively. Primary Tumor A variety of primary tumors were present in the studied population, with the majority being colon, neuroendocrine, breast, cholangiocarcinoma, and others. The tumors classified as “other” were categorized together and used as a reference when calculating the HRs (Tables 1 and 2). Patients with breast cancer metastases to the liver had poor prognosis on multivariate analysis compared with patients with metastases from other sources (HR, 2.52; 95% CI, 1.13–5.61). Lymphocyte Depression One of the known sequelae of 90Y radioembolization is lymphocyte depression (14,17). The lymphocyte levels of each patient were documented before and after 90Y microsphere therapy. Of the 67 patients who showed lymphocyte depression, 32 (51%) died before time of study completion. The HRs for survival among patients with lymphocyte depression versus those without lymphocyte depression were 0.62 (95% CI, 0.38 – 1.02) and 0.55 (95% CI, 0.31– 0.96) on univariate and multivariate analyses, respectively. Subsequently, the Fisher exact test was used to analyze the association of dose (ie, ⬎104 Gy vs ⱕ104 Gy) with the likelihood of lymphocyte depression. No significant association was identified (P ⫽ .7082). Of the 43 patients treated with 104 Gy or less, 23 (54.8%) had lymphocyte depression levels after baseline measurement. A total of 88 patients were treated with more than 104 Gy, and half these patients exhibited lymphocyte depression after treatment. Substratification Analysis or Tumor Response in Colorectal Cancer Given the large proportion of patients with a primary colon carcinoma (n ⫽ 50), the tumor response in this subgroup of patients was independently analyzed. The median survival times of responders and nonresponders per
90
January 2010
Y Radioembolization
JVIR
Table 3 Survival and Tumor Response among Patients with Colon Carcinoma Tumor Response
Total
Median Survival, Months (95% CI)
HR (95% CI)*
P Value
Yes No
20 30
21.9 (13.4–NR) 8.5 (7–19.7)
0.26 (0.10–0.69) Reference
.0071
Note.—NR ⫽ not reached. * Not adjusted for any covariates. Table 4 Statistically Significant Prognostic Indicators on Multivariate Analysis Positive Indicators
Negative Indicators
Response to therapy per WHO criteria* Lymphocyte depression
Poor ECOG performance status (⬎0) Increased bilirubin level (⬎1.3 mg/dL) Hepatic tumor burden (⬎50%) Breast cancer as primary tumor
* Defined as a reduction of the sum of cross-products by ⬎50%.
WHO criteria were 21.9 months and 8.5 months, respectively (HR, 0.26; 95% CI, 0.10 – 0.69; Table 3).
DISCUSSION Liver metastases are the most common form of hepatic malignancies (18). Various methods for the treatment of hepatic metastases include resection, systemic chemotherapy, radiofrequency ablation, and external-beam radiation. Recently, 90Y microspheres have been extensively used for the treatment of primary and metastatic hepatic disease (14,19–21). Although the standard treatment for metastatic disease to the liver is chemotherapy and surgical resection, there is currently no reported survival benefit for incomplete tumor debulking compared with conservative therapy (22). Many clinical and tumor features have been previously examined to determine which patients benefit most from resection. Such factors include perihepatic lymph node involvement and locally recurrent disease or extrahepatic metastases. Other determinants which represent relative contraindications for resection but vary among studies; some have included large tumor size and multiplicity, increased tumor markers, presence of nondiploid tumors, and margins of resection less than 1 cm (23). The patients in our study population were not deemed candidates for surgical resection given one or more factors previously cited. Therefore, the goal of
brachytherapy was to prevent progression of liver tumors while maintaining residual liver function. We chose the most common variable factors that might influence prognosis. Of the various factors analyzed, poor ECOG PS, increased bilirubin level (ie, ⬎1.3 mg/ dL), hepatic tumor burden greater than 50%, and metastases from breast cancer were variables that demonstrated statistical significance on the multivariate analysis. Patients with a positive tumor response on imaging and lymphocyte depression after 90Y treatment were found to have a favorable prognosis (Table 4). However, unlike the other prognostic indicators, WHO response and lymphocyte suppression may be used only in the decision of whether to continue with additional intraarterial therapies. The presence of extrahepatic metastases was found to be a significant prognostic factor on univariate analysis. However, it did not show significance on multivariate analyses. Patients with secondary liver tumors differ from patients with hepatocellular carcinoma. Patients with hepatocellular carcinoma present with compromised liver function as a result of cirrhosis and the tumor (24). This leads to clinical aberrations such as decrease in PS and biochemical aberrations such as an increase in bilirubin. The decrease in liver function in patients with secondary lesions is most often caused by only the presence of tumor. Hence, the PS and baseline bilirubin values present important and
Volume 21
Number 1
clinically relevant prognostic factors in patients with secondary liver tumors. WHO response has been thought to correlate to survival (25). The data presented herein show that it is a significant prognostic factor for survival on univariate and multivariate analyses. Colon carcinoma represented the largest number of primary tumors in our study population. These patients were further analyzed with respect to their tumor response after treatment. A significantly improved survival was observed in patients in whom WHO response was demonstrated. Multiple types of primary tumors were present in our study population. The largest proportion of tumors included colon, breast, neuroendocrine, and cholangiocarcinoma. The other tumors, as previously listed, were used as a reference to calculate the HR. Despite the positive response in colon and breast carcinomas that 90Y therapy showed, only neuroendocrine primary tumors were associated with a statistically significant HR on univariate analyses. This favorable prognosis of neuroendocrine tumors has been previously described (26,27). Patients with liver metastases from breast cancer had poor prognosis after correcting for other variables, as seen in the multivariate analysis. In conclusion, various prognostic clinical and imaging factors are important for patients referred for catheterdirected intraarterial 90Y therapy for hepatic malignancies. Poor ECOG PS, hepatic tumor burden greater than 50%, and increased bilirubin level (ie, ⬎1.3 mg/dL) were determined to be negative prognostic indicators on univariate and multivariate analyses. Compromised PS was the most powerful negative prognosticator in this population with liver malignancies. Tumor response was a favorable prognostic factor for survival. References 1. Rougier P, Milan C, Lazorthes F, et al. Prospective study of prognostic factors in patients with unresected hepatic metastases from colorectal cancer. Fondation Francaise de Cancerologie Digestive. Br J Surg 1995; 82:1397–1400. 2. Chang AE, Steinberg SM, Culnane M, White DE. Determinants of survival in patients with unresectable colorectal liver metastases. J Surg Oncol 1989; 40: 245–251. 3. Kemeny N, Niedzwiecki D, Shurgot B, Oderman P. Prognostic variables in pa-
Dunfee et al
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
tients with hepatic metastases from colorectal cancer. Importance of medical assessment of liver involvement. Cancer 1989; 63:742–747. Yuste AL, Aparicio J, Segura A, et al. Analysis of clinical prognostic factors for survival and time to progression in patients with metastatic colorectal cancer treated with 5-fluorouracil-based chemotherapy. Clin Colorectal Cancer 2003; 2: 231–234. Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the eastern cooperative oncology group. Am J Clin Oncol 1982; 5:649 – 655. Allen PJ, Kemeny N, Jarnagin W, DeMatteo R, Blumgart L, Fong Y. Importance of response to neoadjuvant chemotherapy in patients undergoing resection of synchronous colorectal liver metastases. J Gastrointest Surg 2003; 7:109–115. Sato KT, Lewandowski RJ, Mulcahy MF, et al. Unresectable chemorefractory liver metastases: radioembolization with 90 Y microspheres—safety, efficacy, and survival. Radiology 2008; 247:507–515. Gray BN, Burton MA, Kelleher DK, Anderson J, Klemp P. Selective internal radiation (SIR) therapy for treatment of liver metastases: measurement of response rate. J Surg Oncol 1989; 42:192– 196. Salem R, Thurston KG. Radioembolization with 90Yttrium microspheres: a state-of-the-art brachytherapy treatment for primary and secondary liver malignancies. Part 1: technical and methodologic considerations. J Vasc Interv Radiol 2006; 17:1251–1278. Meade VM, Burton MA, Gray BN, Self GW. Distribution of different sized microspheres in experimental hepatic tumours. Eur J Cancer Clin Oncol 1987; 23:37– 41. Rhodes S. Humanitarian device exemptions (HDE) regulations: questions and answers; final guidance for industry. Available at http://www.fda. gov/downloads/medicaldevices/device regulationandguidance/guidance documents/ucm071474.pdf. Accessed August 10, 2009. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. European organization for research and treatment of cancer, national cancer institute of the united states, national cancer institute of canada. J Natl Cancer Inst 2000; 92:205–216. World Health Organization. WHO handbook for reporting results of cancer treatment. WHO offset publication no. 48. Geneva: WHO, 1979. Carr BI. Hepatic arterial 90Yttrium glass microspheres (TheraSphere) for unresectable hepatocellular carcinoma: interim
•
95
safety and survival data on 65 patients. Liver Transpl 2004; 10(suppl):S107–S110. 15. Cox DR, Oaks D. Analysis of survival data. London: Chapman & Hall, 1984. 16. Goin JE, Salem R, Carr BI, et al. Treatment of unresectable hepatocellular carcinoma with intrahepatic yttrium 90 microspheres: factors associated with liver toxicities. J Vasc Interv Radiol 2005; 16:205–213. 17. Salem R, Lewandowski RJ, Atassi B, et al. Treatment of unresectable hepatocellular carcinoma with use of 90Y microspheres (TheraSphere): safety, tumor response, and survival. J Vasc Interv Radiol 2005; 16:1627–1639. 18. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin 2005; 55:10 –30. 19. Lewandowski RJ, Thurston KG, Goin JE, et al. 90Y microsphere (TheraSphere) treatment for unresectable colorectal cancer metastases of the liver: response to treatment at targeted doses of 135–150 Gy as measured by [18F]fluorodeoxyglucose positron emission tomography and computed tomographic imaging. J Vasc Interv Radiol 2005; 16:1641–1651. 20. Herba MJ, Thirlwell MP. Radioembolization for hepatic metastases. Semin Oncol 2002; 29:152–159. 21. Gray BN, Anderson JE, Burton MA, et al. Regression of liver metastases following treatment with yttrium-90 microspheres. Aust N Z J Surg 1992; 62:105–110. 22. Scheele J, Stang R, Altendorf-Hofmann A, Paul M. Resection of colorectal liver metastases. World J Surg 1995; 19:59 –71. 23. Doci R, Gennari L, Bignami P, Montalto F, Morabito A, Bozzetti F. One hundred patients with hepatic metastases from colorectal cancer treated by resection: analysis of prognostic determinants. Br J Surg 1991; 78:797– 801. 24. Lewandowski RJ, Kulik LM, Riaz A, et al. A comparative analysis of transarterial downstaging for hepatocellular carcinoma: chemoembolization versus radioembolization. Am J Transplant 2009; 9:1920 –1928. 25. Johnson KR, Ringland C, Stokes BJ, et al. Response rate or time to progression as predictors of survival in trials of metastatic colorectal cancer or non-small-cell lung cancer: a meta-analysis. Lancet Oncol 2006; 7:741–746. 26. Sarmiento JM, Heywood G, Rubin J, Ilstrup DM, Nagorney DM, Que FG. Surgical treatment of neuroendocrine metastases to the liver: a plea for resection to increase survival. J Am Coll Surg 2003; 197:29 –37. 27. Chamberlain RS, Canes D, Brown KT, et al. Hepatic neuroendocrine metastases: does intervention alter outcomes? J Am Coll Surg 2000; 190:432– 445.
PURPOSE: To identify key prognostic clinical and imaging variables in patients undergoing yttrium-90 radioembolization (90Y) for liver malignancies. MATERIALS AND METHODS: Patients with liver malignancies that progressed despite standard-of-care therapy were treated with 90Y from 2002 to 2006. Baseline functional status, laboratory values, and diagnostic imaging were assessed before therapy. Imaging follow-up was performed 1 month after treatment and subsequently at 3-month intervals. Patients were followed for survival from the time of their first 90Y treatment. RESULTS: Patients with follow-up imaging after radioembolization (N ⴝ 130) were included in this analysis. Primary malignancies included colon, neuroendocrine, and others. The following clinical variables had a significant effect on survival on multivariate analysis: Eastern Cooperative Oncology Group (ECOG) performance status (PS) greater than 0 (hazard ratio [HR], 7.98; 95% CI, 3.98 –16), hepatic tumor burden of 51%–75% (HR, 2.46; 95% CI, 1.01– 6.02), bilirubin level greater than 1.3 mg/dL (HR, 2.60; 95% CI, 1.27–5.34), hepatic metastases from breast cancer (HR, 2.51; 95% CI, 1.13–5.61), response on imaging based on World Health Organization (WHO) criteria (HR, 0.48; 95% CI, 0.24 – 0.94), and lymphocyte depression (HR, 0.56; 95% CI, 0.31– 0.96). Among patients with colorectal cancer metastases to the liver, the HR for survival on univariate analysis for responders compared with nonresponders (per WHO criteria) was 0.26 (95% CI, 0.10 – 0.69). CONCLUSIONS: Cancer-related symptoms (ie, ECOG PS > 0), hepatic tumor burden greater than 50%, increased bilirubin levels, and hepatic metastases from breast cancer were found to be negative prognostic factors. Tumor response to therapy and lymphocyte depression were associated with favorable prognosis. Additionally, WHO response was identified to be a favorable prognostic factor in patients with colorectal cancer metastases. These findings may be useful when counseling patients regarding prognosis of their hepatic disease. J Vasc Interv Radiol 2010; 21:90 –95 Abbreviations:
ECOG ⫽ Eastern Cooperative Oncology Group, HR ⫽ hazard ratio, PS ⫽ performance status, WHO ⫽ World Health Organization
MULTIPLE oncologic factors have been studied in determining the prognostic outcome of patients with metastatic dis-
ease (1– 4). The presence of metastases to the liver predicts a worse outcome in nearly all tumors, with the site of the
From Diagnostic Imaging, Inc., Aria Health Hospitals, Philadelphia, Pennsylvania (B.L.D.); Department of Radiology, Section of Interventional Radiology (A.R., R.J.L., S.I., M.F.M., R.K.R., B.A., K.T.S., R.A.O., R.S.), and Department of Medicine, Division of Hematology and Oncology (M.F.M., S.N., A.B., R.S.), Northwestern Memorial Hospital and Robert H. Lurie Comprehensive Cancer Center, 676 North St. Clair St., Suite 800, Chicago, IL 60611. Received October 10, 2008; final revision received August 19, 2009; accepted September 9,
2009. Address correspondence to R.S.; E-mail: [email protected]
90
From the SIR 2008 Annual Meeting. R.S. is a paid consultant of MDS Nordion (Ottawa, Ontario, Canada). None of the other authors have identified a conflict of interest. © SIR, 2010 DOI: 10.1016/j.jvir.2009.09.011
primary tumor being an additional factor. Performance status (PS) has also been proven useful in predicting the long-term prognosis. The most commonly used system for grading performance was proposed by the Eastern Cooperative Oncology Group (ECOG) (5). Many studies have demonstrated that a higher ECOG PS is related to a higher short-term mortality rate (1). In addition, laboratory values such as bilirubin levels may also assist in predicting the effects of the metastatic deposit on the functional status of the liver (3,4). Many other prognostic factors previously investigated include sex, age, lactic dehy-
Volume 21
Number 1
drogenase level, number of involved segments, and presence of extrahepatic metastases. Tumor markers such as carcinoembryonic antigen and cancer antigen 19-9 have been proposed, but the data are not consistent (4). Additional prognostic indicators can be determined by the response of metastatic lesions after various medical and surgical treatments. Significantly improved survival benefit in patients with metastatic colorectal carcinoma results if the disease does not progress while the patient is receiving neoadjuvant chemotherapy (6). The purpose of the present study was to determine prognostic clinical variables for patients who underwent catheter-directed yttrium-90 therapy in a large cohort of patients with liver metastases. Our goal was determine the prognostic factors that had the most influence on outcome of patients after treatment.
MATERIALS AND METHODS Patient Population An open label cohort study was performed between 2002 and 2006 of 137 patients with liver malignancies refractory to standard-of-care therapy. One hundred thirty patients with follow-up imaging were studied for this analysis. Seven patients were not able to receive follow-up imaging because of early death (⬍2 months after treatment). The data were collected prospectively. The PS of each patient was documented according to the ECOG grading criteria (5). Laboratory values including bilirubin levels were obtained for all patients. Each patient had preprocedural imaging such as computed tomography (CT) and/or magnetic resonance (MR) imaging depending on the primary neoplasm. The study was approved by the investigational review board and was conducted in a manner compliant with the Health Insurance Portability and Accountability Act. The longterm outcomes including safety, toxicity, tumor response, and survival of this patient cohort have been published elsewhere (7). Inclusion and Exclusion Criteria The following inclusion criteria were met for each patient: (i) progressive liver disease despite standard-of-care therapy
Dunfee et al (octreotide for neuroendocrine tumors and chemotherapy for all others); (ii) disease not amenable to surgery; (iii) ECOG PS of 0–2; (iv) noncompromised pulmonary function; (v) ability to undergo angiography and selective visceral catheterization; (vi) adequate hematologic parameters (ie, granulocyte count ⱖ1.5 ⫻ 109/L, platelets ⱖ50 ⫻ 109/L) and renal function (creatinine ⱕ2.0 mg/dL); (vii) adequate liver function (ie, bilirubin ⱕ2.0 mg/dL); and (viii) limited extrahepatic disease (eg, solitary lung metastasis, portal lymphadenopathy ⬎2 cm, adrenal metastases, and limited peritoneal carcinomatosis, ie, deemed as having a minor contribution to long-term morbidity and mortality). Exclusion criteria included (i) evidence of any uncorrectable flow to the gastrointestinal tract observed on angiography or Technetium-99m macroaggregated albumin (99mTc-MAA) scan; (ii) radiographic/pathologic confirmation of hepatocellular carcinoma; and (iii) estimated radiation dose greater than 30 Gy (16.5 mCi) delivered to the lungs in a single administration or 50 Gy on multiple administrations. Technique Patient evaluation.—Patients were referred to the interventional radiology unit for 90Y radioembolization. Baseline functional status of each patient was determined according to the ECOG criteria. Preprocedural angiography was performed to evaluate anatomy and identify vessels that required intervention such as embolization. At this time, 99m Tc-MAA was then administered to estimate the shunted fraction of 90Y to the lungs (8). Finally, baseline imaging was evaluated for the distribution of disease, hepatic tumor burden, and the presence/absence of extrahepatic metastases. Device.—TheraSphere (MDS Nordion, Ottawa, Ontario, Canada) consists of insoluble glass microspheres with 90Y as an integral constituent of the glass. The mean sphere diameter ranges from 20 to 30 m. Yttrium-90 is a pure -emitter with a 64.1-hour physical halflife. The specific activity per microsphere is approximately 2,500 Bq at the time of calibration. The method of activity requirement for treatment has been previously published (9). The 90Y radioisotope is produced by the neutron bombardment of 90Y, yielding a pure
•
91
-emitter with a high-energy emission of 0.936 keV. Tissue penetration of the activity is 11 mm, with a mean penetration of 2.5 mm. Given the hypervascularity of tumors, the microspheres are preferentially distributed to the tumor and produce localized radiation (10). Clinicians are directed to published Food and Drug Administration guidelines for the use of humanitarian device exemptions in disease conditions other than the approved indication, ie, hepatocellular carcinoma (11). The patients received bilobar or unilobar treatment depending on the extent of disease within the liver at the time of presentation. For bilobar treatment, the first lobe treatment was followed by the second lobe treatment approximately 30 – 40 days later. No whole-liver infusions were performed. Lobar treatment was targeted to deliver 120 Gy according to previously published dosimetry techniques (9). All patients were discharged from the hospital within 6 hours after completion of the procedure. Tumor Response Imaging follow-up was performed 1 month after treatment and subsequently every 3 months. Response was assessed according to World Health Organization (WHO) guidelines (12,13). Tumor response on CT or MR imaging was determined for a maximum of four measurable lesions (⬎1 cm) in terms of the product of the greatest cross-sectional tumor length and the length of the corresponding orthogonal projection (ie, “cross-product”). Corresponding lesions from baseline and posttreatment scans were compared for size changes. Complete response was defined by a change in the sum of the cross-products to zero, partial response as a decrease in the sum of cross-products by at least 50%, stable disease as a decrease in the sum of cross-products by less than 50% or an increase less than 25%, and progression as an increase in the sum of cross-products by at least 25% (12,13). Posttreatment studies were compared and assessed simultaneously with pretreatment images. Laboratory Analyses For this study, the only laboratory parameter used to assess baseline liver function was baseline bilirubin level. Bilirubin levels greater than 1.3 mg/dL
92
•
Prognostic Factors for Survival after Liver
(our institutional normal level) were classified as being increased. As 90Y leads to posttreatment decrease in lymphocyte counts, lymphocyte depression was also studied (14). Lymphocyte depression was defined as a decrease in absolute lymphocyte count from normal (⬎1,000/L) to less than normal at any time after 90Y administration. Statistical Analyses Exploratory analyses to identify variables that may have an effect on survival were performed. For these analyses, patients were classified into two groups: those who were alive at the time of study completion and those who were dead at the time of study completion. The characteristics between these two groups were compared with use of the Fisher exact or 2 test. Survival for the cohort was analyzed from the date of initial intraarterial 90Y microsphere administration. The univariate analyses were performed with use of the log-rank test and the multivariate analyses were performed with use of the Cox proportional-hazards model (ie, stepwise) to determine the prognostic value of each variable analyzed (15). A P value less than .05 was deemed to indicate a significant difference. As the backward stepwise method of Cox proportional hazards eliminates nonsignificant variables sequentially, all variables were entered into the initial multivariate analysis. A substratification analysis was also performed on tumor response in patients with colorectal cancer.
RESULTS
90
January 2010
Y Radioembolization
JVIR
Table 1 Exploratory Analysis of Prognostic Benefit of Variables At Study Completion
Variable
Total (N ⫽ 130)
Alive (n ⫽ 67)
Dead (n ⫽ 63)
54 76
31 (57) 36 (47)
23 (43) 40 (53)
.34
48 82
8 (17) 59 (72)
40 (83) 23 (18)
⬍.001
105 25
51 (49) 16 (89)
54 (51) 9 (20)
.24
0 6 18 106
0 0 (0) 7 (39) 60 (57)
0 6 (100) 11 (61) 46 (43)
.03
66 64
27 (41) 40 (63)
39 (59) 24 (37)
.02
11 119
0 (0) 67 (56)
11 (100) 52 (44)
⬍.001
88 42
46 (53) 21 (50)
42 (47) 21 (50)
.95
48 82
32 (67) 35 (43)
16 (33) 47 (57)
.01
20 17 50 7 36
12 (60) 13 (76) 26 (52) 4 (57) 12 (33)
8 (40) 4 (24) 24 (48) 3 (43) 24 (67)
.046
67 63
32 (48) 31 (49)
35 (52) 32 (51)
Age (y) ⬍65 ⱖ65 ECOG PS ⬎0 0 Lobar involvement Bilobar Unilobar Hepatic tumor burden (%) ⬎75 51–75 26–50 0–25 Extrahepatic metastases Yes No Increased bilirubin (⬎1.3 mg/dL) Yes No Higher dose utilization (ⱖ104 Gy) Yes No Tumor response (WHO Criteria) Yes No Tumor Breast carcinoma Neuroendocrine Colon carcinoma Cholangiocarcinoma Other carcinomas Lymphocyte depression Yes No
P Value
1.00
Note.—Values in parentheses are percentages.
Patient Sample In the study, 130 underwent 218 intraarterial administrations of 90Y microspheres. Seven patients did not have follow-up imaging and hence were not included in the regression analyses. The entire analysis was therefore based on 130 patients. There were 69 women and 61 men, with a median age of 61 years. Primary malignancies included colon (n ⫽ 50), breast (n ⫽ 20), neuroendocrine (n ⫽ 17), cholangiocarcinoma (n ⫽ 7), and others (n ⫽ 36; pancreas, lung, melanoma, renal, esophageal, and adenocarcinoma of unknown primary lesion). All patients were considered to have unresectable disease. A total of 63 patients (48%) had died
and 67 (52%) were alive at the time of study completion (Table 1). Table 2 presents the univariate and multivariate analysis for the study.
younger than 65 years of age was 1.45 (95% CI, 0.87–2.43) on univariate analysis.
Age
Our study population was distributed into two groups before treatment: those with normal functional status (ie, ECOG PS of 0) and those with some level of physical impairment and cancer-related symptoms (ie, ECOG PS ⬎ 0). Of the patients who had an ECOG PS greater than 0 and those who had an ECOG PS of 0, 40 (83%) and 23 (18%) had died at the time of study completion, respectively. The HRs for survival in pa-
Our study population included a spectrum of age ranges, with the majority of patients greater than 65 years of age (58%). Of the patients who were younger than 65 years of age and 65 years of age or older, 23 (43%) and 40 (53%) had died at the time of study completion, respectively. The hazard ratios (HRs) for survival in patients 65 years of age or older compared with those
ECOG PS
Volume 21
Number 1
Dunfee et al
Univariate Analyses
Age (y) ⬍65 ⱖ65 Sex Female Male ECOG PS 0 ⬎0 Lobar involvement Unilobar Bilobar Hepatic tumor burden (%) 51–75 26–50 0–25 Extrahepatic metastases Yes No Increased bilirubin (mg/dL) ⬎1.3 ⱕ1.3 Dose utilization (Gy) ⬎104 ⱕ104 Positive tumor response to treatment Yes No Tumor Breast carcinoma Neuroendocrine Colon carcinoma Cholangiocarcinoma Other carcinomas Lymphocyte depression Yes No
HR (95% CI)
P Value
Multivariate Analysis HR (95% CI)
P Value
Reference 1.45 (0.87–2.43)
.1552
— —
—
Reference 1.16 (0.70–1.90)
.56
— —
—
Reference 10.0 (5.47–18.5)
⬍.0001
Reference 7.98 (3.98–16.0)
⬍.0001
Reference 1.87 (0.92–3.80)
.0860
— —
—
4.50 (1.91–10.6) 1.35 (0.70–2.59) Reference
.0006 .3680
2.46 (1.01–6.02) Reference
.048
2.30 (1.38–3.83) Reference
.0015
— —
2.91 (1.51–5.61) Reference
.0014
2.60 (1.27–5.34) Reference
1.10 (0.65–1.86) Reference
.7286
— —
—
0.30 (0.17–0.55) Reference
.0001
0.48 (0.24–0.94) Reference
.034
1.59 (0.75–3.38) 0.33 (0.12–0.92) 0.84 (0.50–1.39) 0.84 (0.26–2.70) Reference 0.62 (0.38–1.02) Reference
tients with an ECOG PS greater than 0 versus those with an ECOG PS of 0 were 10 (95% CI, 5.47–18.5) and 7.98 (95% CI, 3.98 –16.0) on univariate and multivariate analyses, respectively. Therefore, patients with compromised functional status (ie, ECOG PS of 1/2) are at high risk for worse outcomes after 90Y compared with asymptomatic patients (ie, ECOG PS of 0). Lobar Involvement The majority of patients in our study population had bilobar involvement of metastatic disease. Of the patients who had bilobar and unilobar disease in-
93
(95% CI, 1.01– 6.02) on univariate analysis.
Table 2 Univariate and Multivariate Analyses
Variables
•
.23 .03 .49 .78
.06
— .0092
Extrahepatic Metastases Sixty-six patients had extrahepatic metastases in the study population. Of the patients with and without extrahepatic metastases, 39 (59%) and 24 (37%) had died at the time of study completion, respectively. The HR for survival in patients with extrahepatic metastases versus patients without extrahepatic metastases was 2.30 (95% CI, 1.38 –3.83) on univariate analysis. Baseline Bilirubin Levels Increased bilirubin levels were defined as those with any value greater than 1.3 mg/dL during the initial clinical evaluation. Eleven patients had an increased bilirubin level before 90Y microsphere therapy, and all had died before study completion. The HRs for survival in patients with high baseline bilirubin levels versus those with normal baseline bilirubin levels were 2.91 (95% CI, 1.51–5.61) and 2.60 (95% CI, 1.27– 6.34) on univariate and multivariate analyses, respectively. Dose Utilization
2.52 (1.13–5.61) Reference
.26
0.55 (0.31–0.96) Reference
.036
volvement, 54 (51%) and nine (29%) had died at the time of study completion, respectively. The HR for survival in patients with bilobar disease versus patients with unilobar disease was 1.87 (95% CI, 0.92–3.8) on univariate analysis. Hepatic Tumor Burden Of the patients who had 51%–75%, 26%–50% and 0%–25% hepatic tumor burden, six (100%), 11 (61%), and 46 (43%) had died at the time of study completion, respectively. The HRs for survival in patients with hepatic tumor burdens of 50%–75% versus 0%–25% were 4.50 (95% CI, 1.91–10.6) and 2.46
A higher dose for treatment was defined as dose delivery of greater than 104 Gy (16). A total of 88 patients in the study population required a dose greater than 104 Gy and 42 patients received lower treatment doses. Of the patients who received doses greater than 104 Gy and those who received doses of 104 Gy or less, 42 (47%) and 21 (50%) had died at the time of study completion, respectively. The HR for survival in patients who received high radiation doses compared with patients who received lower doses was 1.10 (95% CI, 0.65–1.86) on univariate analysis. Tumor Response Maximum tumor response to treatment was determined by imaging characteristics for all patients and all 332 lesions measured. At the lesional level, response rate was 42.8% (142 of 332). At the patient level, response rate was 37% (48 of 130). Of the patients who did and did not show a response per WHO criteria, 16 (33%) and 47 (57%) had died at the time of study completion, respec-
94
•
Prognostic Factors for Survival after Liver
tively. The HRs for survival in patients who showed a response per WHO criteria compared with those who showed no response were 0.30 (95% CI, 0.17– 0.55) and 0.48 (95% CI, 0.24 – 0.94) on univariate and multivariate analyses, respectively. Primary Tumor A variety of primary tumors were present in the studied population, with the majority being colon, neuroendocrine, breast, cholangiocarcinoma, and others. The tumors classified as “other” were categorized together and used as a reference when calculating the HRs (Tables 1 and 2). Patients with breast cancer metastases to the liver had poor prognosis on multivariate analysis compared with patients with metastases from other sources (HR, 2.52; 95% CI, 1.13–5.61). Lymphocyte Depression One of the known sequelae of 90Y radioembolization is lymphocyte depression (14,17). The lymphocyte levels of each patient were documented before and after 90Y microsphere therapy. Of the 67 patients who showed lymphocyte depression, 32 (51%) died before time of study completion. The HRs for survival among patients with lymphocyte depression versus those without lymphocyte depression were 0.62 (95% CI, 0.38 – 1.02) and 0.55 (95% CI, 0.31– 0.96) on univariate and multivariate analyses, respectively. Subsequently, the Fisher exact test was used to analyze the association of dose (ie, ⬎104 Gy vs ⱕ104 Gy) with the likelihood of lymphocyte depression. No significant association was identified (P ⫽ .7082). Of the 43 patients treated with 104 Gy or less, 23 (54.8%) had lymphocyte depression levels after baseline measurement. A total of 88 patients were treated with more than 104 Gy, and half these patients exhibited lymphocyte depression after treatment. Substratification Analysis or Tumor Response in Colorectal Cancer Given the large proportion of patients with a primary colon carcinoma (n ⫽ 50), the tumor response in this subgroup of patients was independently analyzed. The median survival times of responders and nonresponders per
90
January 2010
Y Radioembolization
JVIR
Table 3 Survival and Tumor Response among Patients with Colon Carcinoma Tumor Response
Total
Median Survival, Months (95% CI)
HR (95% CI)*
P Value
Yes No
20 30
21.9 (13.4–NR) 8.5 (7–19.7)
0.26 (0.10–0.69) Reference
.0071
Note.—NR ⫽ not reached. * Not adjusted for any covariates. Table 4 Statistically Significant Prognostic Indicators on Multivariate Analysis Positive Indicators
Negative Indicators
Response to therapy per WHO criteria* Lymphocyte depression
Poor ECOG performance status (⬎0) Increased bilirubin level (⬎1.3 mg/dL) Hepatic tumor burden (⬎50%) Breast cancer as primary tumor
* Defined as a reduction of the sum of cross-products by ⬎50%.
WHO criteria were 21.9 months and 8.5 months, respectively (HR, 0.26; 95% CI, 0.10 – 0.69; Table 3).
DISCUSSION Liver metastases are the most common form of hepatic malignancies (18). Various methods for the treatment of hepatic metastases include resection, systemic chemotherapy, radiofrequency ablation, and external-beam radiation. Recently, 90Y microspheres have been extensively used for the treatment of primary and metastatic hepatic disease (14,19–21). Although the standard treatment for metastatic disease to the liver is chemotherapy and surgical resection, there is currently no reported survival benefit for incomplete tumor debulking compared with conservative therapy (22). Many clinical and tumor features have been previously examined to determine which patients benefit most from resection. Such factors include perihepatic lymph node involvement and locally recurrent disease or extrahepatic metastases. Other determinants which represent relative contraindications for resection but vary among studies; some have included large tumor size and multiplicity, increased tumor markers, presence of nondiploid tumors, and margins of resection less than 1 cm (23). The patients in our study population were not deemed candidates for surgical resection given one or more factors previously cited. Therefore, the goal of
brachytherapy was to prevent progression of liver tumors while maintaining residual liver function. We chose the most common variable factors that might influence prognosis. Of the various factors analyzed, poor ECOG PS, increased bilirubin level (ie, ⬎1.3 mg/ dL), hepatic tumor burden greater than 50%, and metastases from breast cancer were variables that demonstrated statistical significance on the multivariate analysis. Patients with a positive tumor response on imaging and lymphocyte depression after 90Y treatment were found to have a favorable prognosis (Table 4). However, unlike the other prognostic indicators, WHO response and lymphocyte suppression may be used only in the decision of whether to continue with additional intraarterial therapies. The presence of extrahepatic metastases was found to be a significant prognostic factor on univariate analysis. However, it did not show significance on multivariate analyses. Patients with secondary liver tumors differ from patients with hepatocellular carcinoma. Patients with hepatocellular carcinoma present with compromised liver function as a result of cirrhosis and the tumor (24). This leads to clinical aberrations such as decrease in PS and biochemical aberrations such as an increase in bilirubin. The decrease in liver function in patients with secondary lesions is most often caused by only the presence of tumor. Hence, the PS and baseline bilirubin values present important and
Volume 21
Number 1
clinically relevant prognostic factors in patients with secondary liver tumors. WHO response has been thought to correlate to survival (25). The data presented herein show that it is a significant prognostic factor for survival on univariate and multivariate analyses. Colon carcinoma represented the largest number of primary tumors in our study population. These patients were further analyzed with respect to their tumor response after treatment. A significantly improved survival was observed in patients in whom WHO response was demonstrated. Multiple types of primary tumors were present in our study population. The largest proportion of tumors included colon, breast, neuroendocrine, and cholangiocarcinoma. The other tumors, as previously listed, were used as a reference to calculate the HR. Despite the positive response in colon and breast carcinomas that 90Y therapy showed, only neuroendocrine primary tumors were associated with a statistically significant HR on univariate analyses. This favorable prognosis of neuroendocrine tumors has been previously described (26,27). Patients with liver metastases from breast cancer had poor prognosis after correcting for other variables, as seen in the multivariate analysis. In conclusion, various prognostic clinical and imaging factors are important for patients referred for catheterdirected intraarterial 90Y therapy for hepatic malignancies. Poor ECOG PS, hepatic tumor burden greater than 50%, and increased bilirubin level (ie, ⬎1.3 mg/dL) were determined to be negative prognostic indicators on univariate and multivariate analyses. Compromised PS was the most powerful negative prognosticator in this population with liver malignancies. Tumor response was a favorable prognostic factor for survival. References 1. Rougier P, Milan C, Lazorthes F, et al. Prospective study of prognostic factors in patients with unresected hepatic metastases from colorectal cancer. Fondation Francaise de Cancerologie Digestive. Br J Surg 1995; 82:1397–1400. 2. Chang AE, Steinberg SM, Culnane M, White DE. Determinants of survival in patients with unresectable colorectal liver metastases. J Surg Oncol 1989; 40: 245–251. 3. Kemeny N, Niedzwiecki D, Shurgot B, Oderman P. Prognostic variables in pa-
Dunfee et al
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
tients with hepatic metastases from colorectal cancer. Importance of medical assessment of liver involvement. Cancer 1989; 63:742–747. Yuste AL, Aparicio J, Segura A, et al. Analysis of clinical prognostic factors for survival and time to progression in patients with metastatic colorectal cancer treated with 5-fluorouracil-based chemotherapy. Clin Colorectal Cancer 2003; 2: 231–234. Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the eastern cooperative oncology group. Am J Clin Oncol 1982; 5:649 – 655. Allen PJ, Kemeny N, Jarnagin W, DeMatteo R, Blumgart L, Fong Y. Importance of response to neoadjuvant chemotherapy in patients undergoing resection of synchronous colorectal liver metastases. J Gastrointest Surg 2003; 7:109–115. Sato KT, Lewandowski RJ, Mulcahy MF, et al. Unresectable chemorefractory liver metastases: radioembolization with 90 Y microspheres—safety, efficacy, and survival. Radiology 2008; 247:507–515. Gray BN, Burton MA, Kelleher DK, Anderson J, Klemp P. Selective internal radiation (SIR) therapy for treatment of liver metastases: measurement of response rate. J Surg Oncol 1989; 42:192– 196. Salem R, Thurston KG. Radioembolization with 90Yttrium microspheres: a state-of-the-art brachytherapy treatment for primary and secondary liver malignancies. Part 1: technical and methodologic considerations. J Vasc Interv Radiol 2006; 17:1251–1278. Meade VM, Burton MA, Gray BN, Self GW. Distribution of different sized microspheres in experimental hepatic tumours. Eur J Cancer Clin Oncol 1987; 23:37– 41. Rhodes S. Humanitarian device exemptions (HDE) regulations: questions and answers; final guidance for industry. Available at http://www.fda. gov/downloads/medicaldevices/device regulationandguidance/guidance documents/ucm071474.pdf. Accessed August 10, 2009. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. European organization for research and treatment of cancer, national cancer institute of the united states, national cancer institute of canada. J Natl Cancer Inst 2000; 92:205–216. World Health Organization. WHO handbook for reporting results of cancer treatment. WHO offset publication no. 48. Geneva: WHO, 1979. Carr BI. Hepatic arterial 90Yttrium glass microspheres (TheraSphere) for unresectable hepatocellular carcinoma: interim
•
95
safety and survival data on 65 patients. Liver Transpl 2004; 10(suppl):S107–S110. 15. Cox DR, Oaks D. Analysis of survival data. London: Chapman & Hall, 1984. 16. Goin JE, Salem R, Carr BI, et al. Treatment of unresectable hepatocellular carcinoma with intrahepatic yttrium 90 microspheres: factors associated with liver toxicities. J Vasc Interv Radiol 2005; 16:205–213. 17. Salem R, Lewandowski RJ, Atassi B, et al. Treatment of unresectable hepatocellular carcinoma with use of 90Y microspheres (TheraSphere): safety, tumor response, and survival. J Vasc Interv Radiol 2005; 16:1627–1639. 18. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin 2005; 55:10 –30. 19. Lewandowski RJ, Thurston KG, Goin JE, et al. 90Y microsphere (TheraSphere) treatment for unresectable colorectal cancer metastases of the liver: response to treatment at targeted doses of 135–150 Gy as measured by [18F]fluorodeoxyglucose positron emission tomography and computed tomographic imaging. J Vasc Interv Radiol 2005; 16:1641–1651. 20. Herba MJ, Thirlwell MP. Radioembolization for hepatic metastases. Semin Oncol 2002; 29:152–159. 21. Gray BN, Anderson JE, Burton MA, et al. Regression of liver metastases following treatment with yttrium-90 microspheres. Aust N Z J Surg 1992; 62:105–110. 22. Scheele J, Stang R, Altendorf-Hofmann A, Paul M. Resection of colorectal liver metastases. World J Surg 1995; 19:59 –71. 23. Doci R, Gennari L, Bignami P, Montalto F, Morabito A, Bozzetti F. One hundred patients with hepatic metastases from colorectal cancer treated by resection: analysis of prognostic determinants. Br J Surg 1991; 78:797– 801. 24. Lewandowski RJ, Kulik LM, Riaz A, et al. A comparative analysis of transarterial downstaging for hepatocellular carcinoma: chemoembolization versus radioembolization. Am J Transplant 2009; 9:1920 –1928. 25. Johnson KR, Ringland C, Stokes BJ, et al. Response rate or time to progression as predictors of survival in trials of metastatic colorectal cancer or non-small-cell lung cancer: a meta-analysis. Lancet Oncol 2006; 7:741–746. 26. Sarmiento JM, Heywood G, Rubin J, Ilstrup DM, Nagorney DM, Que FG. Surgical treatment of neuroendocrine metastases to the liver: a plea for resection to increase survival. J Am Coll Surg 2003; 197:29 –37. 27. Chamberlain RS, Canes D, Brown KT, et al. Hepatic neuroendocrine metastases: does intervention alter outcomes? J Am Coll Surg 2000; 190:432– 445.