| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Commentary |
1 Department of Human Oncology, University of Wisconsin-Madison, UW Cancer
Center-Riverview, 410 Dewey St., Wisconsin Rapids, WI 54494.
2 Advisory Committee on Medical Uses of Isotopes, United States Nuclear
Regulatory Commission, Washington, DC.
Received January 29, 2007; accepted after revision January 30, 2007.
Address correspondence to J. S. Welsh
(welsh{at}humonc.wisc.edu).
Keywords: abdominal imaging • CT • interventional radiology • liver • oncologic imaging
This issue of AJR hosts two important studies on a rapidly evolving approach to the treatment of liver malignancies. Miller et al. [1] explore the treatment of hepatic metastases from colorectal and other primary malignancies while Keppke et al. [2] focus on hepatocellular carcinoma (HCC). Both studies used yttrium-90-impregnated microspheres. Currently there are two commercially available 90Y microsphere products: TheraSphere (MDS Nordion Inc.), which are glass microspheres, and SIR-Spheres (SIR-Tex Medical, Inc.), which are resin. The two products followed very different pathways to Federal Drug Administration (FDA) approval resulting in artificial dividing lines regarding their general use, clinical trials, and reported results. TheraSphere was approved as monotherapy (i.e., without concurrent chemotherapy) for HCC under the provisions of a humanitarian device exemption (HDE), whereas SIR-Spheres was granted premarket approval (PMA) for metastatic, unresectable liver tumors from primary colorectal cancer in combination with intraarterial 5-floxuridine chemotherapy.
One of the conditions for approval of an HDE is that there be an institutional review board (IRB) initial review and approval before a humanitarian-use device is used at a facility, as well as continuing review of its use. This has the disadvantage of the sometimes arduous process of designing a clinical protocol and seeking IRB approval before routine use but, on the other hand, this process is more likely to result in publication of an institution's experience. The resin SIR-Spheres do not require an IRB-approved protocol before routine use at an institution but the downside is a lower likelihood of publication of data. The different FDA pathways thus have fostered the use of glass microspheres for HCC and resin microspheres for liver metastases, and have resulted in a disproportionately larger number of presentations and publications on HCC treated with the glass microspheres with a relative paucity of papers on the far more common (at least in the United States) liver metastases. The article by Miller et al. [1] breaks this tradition as it focuses on liver metastases treated with TheraSphere glass 90Y microspheres for liver metastases.
Despite the artificial lines drawn by FDA approval pathways, there are more similarities than differences between the glass and resin microspheres. To start, both use 90Y as their radiation dose engines. Yttrium-90 is frequently called a "pure" beta emitter, which for practical purposes is true as it decays via a 2.28 MeV maximal energy beta particle and accompanying antineutrino 99.98% of the time. While the maximum energy of the emitted beta particles is 2.28 MeV, the average energy is 0.94 MeV. This corresponds to a maximum range of 1.1 cm in tissue, a mean path of 2.5 mm, and an X90 (radius of a sphere in which 90% of energy is deposited) of 5.3 mm. 90Y has a half-life of 64.1 hours (2.67 days) and decays to stable 90Zr. As a rule of thumb, in one kilogram of tissue, 1 GBq of uniformly dispersed 90Y delivers an absorbed dose of approximately 50 Gy. With both the glass and resin microspheres, the radioisotope is embedded within the sphere matrix and does not leach out. After implantation, the 90Y microspheres remain permanently lodged in place.
Thus, the similarity between the two commercial products is greater than the dissimilarities. For instance, there is less difference than that between the two commercially available anti-CD20 non-Hodgkin's lymphoma radioimmunotherapy products, Zevalin (ibritumomab tiuxetan, Biogen Idec) and Bexxar (tositumomab, Corixa), which use 90Y and 131I, respectively, and different immunoglobulin carrier molecules. There are some important distinctions, however. The resin microspheres have an average diameter of 32 µm ± 10 µm and the estimated initial activity per microsphere is 50 Bq. A typical whole-liver treatment might consist of an infusion of 2.0 GBq, approximately 40 million microspheres on average. The actual number infused depends on the time interval between calibration and infusion, thus between 20-80 million spheres might be infused. These figures contrast sharply with the glass microsphere indices, which have a median diameter of about 25 µm. Because the activity per sphere is much higher with the glass microspheres (2,500 Bq per sphere) far fewer are administered: A typical whole-liver treatment of 5 GBq consists of about 2 million microspheres on average (1.2-8 million spheres based on decay time)—less than 10% of the number of resin microspheres for similar cases. The vastly different number of infused microspheres leads to possible differences in embolic physiology. The specific gravity of the two commercially available microspheres also differs substantially, with the glass microspheres having twice the density of the resin microspheres (3.2 g/mL vs 1.6 g/mL), leading to some speculation about deposition characteristics. However, a detailed analysis of four resected liver specimens posttreatment showed no obvious difference between glass or resin microspheres regarding embolic location [3]. Both are small enough to penetrate deeply into tumors but large enough to become permanently wedged in the end arterioles, which have an 8 µm mean diameter. There have been no formal clinical investigations that actually compare the two types of microspheres.
The dual blood supply to the liver provides a natural selectivity for tumor therapy since the majority of blood supply to tumors derives from the hepatic artery and the majority of blood to normal liver parenchyma comes from the portal vein (thus the alternative name "selective internal radiation therapy" or SIRT for 90Y microsphere therapy). After infusion via a branch of the hepatic arterial system, the microspheres preferentially lodge in the periphery around the tumors. Campbell et al. [4] found 50-70-fold more microspheres in this peripheral zone than in normal liver after administration, confirming the proposed selectivity. They reported an average dose to tumors of 200-600 Gy with minimal tumor doses between 70 and 190 Gy and less than 1% of normal liver receiving more than 30 Gy. In an elegant study, Kennedy et al. [3] examined four explanted livers after 90Y microsphere therapy with glass microspheres for HCC or resin microspheres for colorectal liver metastases. The microspheres surrounded tumors, collecting in small groups of one to four, but occasionally up to 20 for the resin microspheres, with tumor to normal tissue ratios of up to 200:1 and radiation doses between 100 Gy and 3,000 Gy. The metastatic lesions showed > 90% necrosis in all examined tumor nodules and there was no pathological evidence of venoocclusive disease or widespread radiation hepatitis.
Yttrium-90 can be derived as a decay product of 90Sr, a nuclear
fission byproduct or it can be produced through neutron bombardment of
89Y in an (n,
) neutron capture reaction. As such, its
distribution and handling are under the regulation of the Nuclear Regulatory
Commission (NRC) (or a particular state in agreement states). Interventional
radiologists often call this type of treatment "radioembolization"
due to some technical similarities to chemoembolization and bland
embolization. From a regulatory perspective, however, radioactive microsphere
therapy is quite different. Initially, 90Y microsphere therapy was
considered a form of "permanent implant" manual brachytherapy and
authorized users had to meet the training and experience requirements outlined
in 10 CFR 35.490 (Radiation Oncology; manual brachytherapy). Radiation
oncologists tend to refer to this treatment as microsphere brachytherapy or
microbrachytherapy to distinguish it from other forms of brachytherapy. The
differences between microsphere brachytherapy and temporary, high-dose-rate
(HDR) brachytherapy are obvious but there are significant differences between
microsphere brachytherapy and other permanent brachytherapy procedures as
well. For example, permanent low-dose-rate prostate (LDR) brachytherapy with
125I, 131Cs, or 103Pd uses seeds that are
clearly visible, individually assayable, and easily counted whereas the
microspheres cannot be counted or even seen with the naked eye. From the FDA's
perspective, the 90Y microspheres are considered a radiotherapeutic
device but in practice, 90Y microspheres are in a liquid suspension
and handled similarly to radiopharmaceuticals. Unlike unsealed radioisotope
radiopharmaceutical agents, however, if 90Y microspheres spill and
dry up they can roll around and get into cracks and crevices or aerosolize and
be inhaled, posing different radiation safety concerns. Microsphere
brachytherapy thus does not cleanly fit into 10 CFR 35.300 (unsealed isotopes
requiring a written directive) either. Because of the difficulties in
classifying 90Y microsphere therapy, the NRC created a new
category, 10 CFR 35.1000 (emerging technologies) to address the specific
issues and concerns surrounding this therapy. There is currently a rule change
petition under consideration that may allow interventional radiologists to
become authorized users if they have received the requisite training and
experience, but given the present political atmosphere surrounding the
availability, handling, and usage of radioactive materials, it may be a while
before the NRC guidance formally changes. Partly because of these strict
regulations, microsphere brachytherapy requires a well-integrated team
approach, with participants from many disciplines such as interventional
radiology, nuclear medicine, radiation oncology, surgery, medical physics, and
medical oncology. In an ideal world, this should enhance patient care thanks
to the expert input from various participants, but in the real world, it poses
several logistical challenges and may explain why so few patients who
potentially could benefit are instead receiving care under the management of
one discipline alone.
Both articles in this issue of AJR discuss the importance of necrosis in assessing response to treatment. Most malignant tumors have a fair amount of necrotic tissue naturally, as the rapidly proliferating cells outpace their blood supply. Tumors thus actively recruit neovasculature to meet their oxygen and nutrient needs. This phenomenon has led to the clinical use of anti-angiogenesis agents such as bevacizumab (Avastin, Genentech) in the treatment of liver metastases and other malignancies. The degree of gross necrosis after treatment is sometimes an invaluable reflection of response and is of important prognostic significance. For example, after chemotherapy for osteosarcoma, necrosis in the subsequently resected specimen is of significant importance with a low percentage of necrosis portending a poorer outcome [5-7].
Similar findings have been reported with unresectable HCC [8]. Thus, the findings reported here by Keppke et al. [2] are likely to be more than mere refinements of radiographic response criteria; they will probably prove to be of prognostic significance in disease-free and overall survival in this disease. HCC is the most common primary hepatic malignancy worldwide and the 3rd most common cause of cancer mortality worldwide. In parts of Asia and Africa, the rates are as high as 500 cases per 100,000 people. Although relatively rare in the United States, the incidence is predicted to increase as the infection rates of hepatitis B and C continue rising.
Phase II studies have shown that pathologic complete response after preoperative therapy that converted unresectable tumors into resectable ones appears to improve survival [9]. Interestingly, in this particular study of 50 patients there were no radiographic complete responses and only 13 (26%) partial responses. Nine of the partial responders underwent resection and four of these had no viable tumor cells on histological review, confirming that radiographic response using the size-based RECIST or WHO criteria does indeed underestimate the degree of response to treatments.
The work of Keppke et al. [2] suggests that adding necrosis criteria (a 30% increase in the percentage of tumor necrosis posttherapy was used as their threshold) substantially changes the response rates to 90Y microsphere brachytherapy. Their new combined size and necrosis criteria increased the response rates from 23% (RECIST) and 26% (WHO) to 59%. Understanding that all patients had progressive disease before treatment and that a halting of progression was likely to indicate clinical benefit, 90Y microsphere brachytherapy showed an impressive 78% response or stable disease rate using RECIST/WHO criteria, which was improved to 88% according to their combined criteria. The authors speculate that these figures may be a more accurate reflection of clinical benefit than the traditional definitions. Unfortunately, this retrospective study could not stratify clinical outcome on the specific definition of response nor definitively document a survival benefit for those responding versus those not. The reported median survivals of 660 days for Okuda stage I patients and 236 days for Okuda stage II do compare favorably with no-treatment figures of 249 days and 60 days, respectively [10]. We await prospective data that hopefully will confirm these findings and also show correlation of the new CT combined response criteria with clinical outcome.
Miller et al. [1] focus on the far more common problem of metastatic disease involving the liver. In the United States, according to American Cancer Society statistics, there were an estimated 106,680 new cases of colon cancer and 41,930 cases of rectal adenocarcinoma with the liver being the most common site of metastases. Miller and colleagues point out the pitfalls of RECIST and WHO criteria in assessing response to treatment with 90Y microspheres and, like Keppke et al. [2], use necrosis and combined criteria to assess response (again using a 30% increase in baseline necrosis as the threshold for defining partial response). Unlike the case with HCC, 18FFDG PET is reliable in evaluating metastatic disease to the liver and was used along with CT data. Their findings suggest that necrosis occurs early after treatment, often preceding any observable decrease in tumor size. For patients who responded based on all criteria, median time to response was 116 days by RECIST and 68 days by WHO but only 29 days by necrosis and 34 days by combined criteria. Importantly—but not surprisingly—seven hepatic lobes with response by necrosis actually showed volume increases on early CT and would have been misclassified as having progressive disease. Four of the seven lobes were evaluated with FDG PET, which confirmed response to 90Y microsphere therapy. In this study, PET appeared not only to detect response sooner but also had greater sensitivity for response than CT. In those patients undergoing PET, PET detected more responses (63%) than CT by RECIST (6%) or combined criteria (24%). PET missed four small (8 mm mean diameter) new lesions detected on CT, however, revealing less than perfect sensitivity. In the entire cohort, about 50% showed CT evidence of disease stabilization after 90Y microsphere therapy, which may be clinically relevant since all patients were progressing prior to therapy; this figure incidentally exactly matches the response rate by their combined CT criteria.
Interestingly, the authors also point out that in many lobes classified as progressive disease, the index lesion(s) may have responded but the failure was based on new disease appearance. This illustrates a possible weakness of RECIST and WHO criteria since such patients might possibly benefit from repeat 90Y microsphere treatment despite being categorized as failures.
Miller et al. [1] did not
report changes in carcinoembryonic antigen (CEA) in their study, perhaps
because not all of their 42 patients had liver metastases secondary to
colorectal cancers. In a large, seven-institution retrospective review
involving 208 patients exclusively with unresectable metastatic colorectal
cancer treated with resin 90Y microspheres, Kennedy et al.
[11] reported that CEA
frequently increased during the first 2 weeks posttreatment (up to 120%
baseline) but by 6 weeks an obvious decline was seen in responders to <
80%; by 12 weeks it had fallen to 50%. In their study, PET showed response in
91% (compared with 35% RECIST CT response) and this response appeared earlier
than CT responses, similar to the findings of Miller et al.
[1]. Keppke et al.
[2] report limited
-fetoprotein (AFP) data on only 12 patients with baseline AFP levels
over 400 ng/mL. In five of these 12, AFP declined by at least 50%. They found
no correlation between AFP and treatment response. However, the time course
for AFP was not described, raising the possibility that AFP levels after
90Y microsphere brachytherapy might follow a delayed time course
similar to the CEA changes reported by Kennedy et al.
[11]. In fact, one patient
initially showed increases in both AFP and tumor size on CT but the volumetric
increase appeared to be due to necrosis; this would have been classified as
progressive disease. One month later, AFP dropped by around 50%, suggesting a
transient elevation due to tumor lysis rather than disease progression. A
similar finding is occasionally seen in prostate cancer patients undergoing
external beam radiation therapy wherein prostate specific antigen (PSA) levels
might increase during and shortly after therapy before showing a steady,
gradual decline.
Overall, the studies by Miller et al. [1] and Keppke et al. [2] make a strong case for the use of adding CT evidence of necrosis as a criterion of response after 90Y microsphere brachytherapy for both HCC and liver metastases. For liver metastases, FDG PET also appears to have an important role. These two articles might herald a new standard of oncologic assessment of response after 90Y microsphere brachytherapy.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
