| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Northwestern Memorial Hospital, Northwestern
University Feinberg School of Medicine, 676 N St. Clair, Ste. 800, Chicago, IL
60611.
2 Department of Preventive Medicine, Northwestern Memorial Hospital,
Northwestern University Feinberg School of Medicine, Chicago, IL 60611.
Received June 27, 2006;
accepted after revision August 18, 2006.
Address correspondence to F. H. Miller.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. CT images of 42 patients with 76 90Y-treated HCC lesions were analyzed. We used four response criteria: WHO size, RECIST size, necrosis, and combined criteria (RECIST and necrosis). Imaging features of treated lesions included both nodular and peripheral rim enhancement. Survival was assessed with the Kaplan-Meier method.
RESULTS. The response rate was 23% according to RECIST criteria, 26% according to WHO criteria, 57% according to necrosis criteria, and 59% according to combined criteria. Response according to necrosis and combined criteria was detected earlier than response according to size criteria alone. Ten responding lesions initially increased in size. After therapy, enhancing peripheral nodules increased in size in 10 lesions, decreased in size in two lesions, and disappeared in two lesions. Twenty-one of 25 lesions with thin rim enhancement after 90Y administration responded to treatment. The median survival times were 660 and 236 days for Okuda stage I and Okuda stage II disease, respectively.
CONCLUSION. Use of combined size and necrosis criteria may lead to more accurate assessment of response to 90Y therapy than use of size criteria alone. Imaging features after 90Y treatment, including size, necrosis, peripheral enhancing nodules, and thin rim enhancement, are described.
Keywords: abdominal imaging • brachytherapy • CT • hepatocellular carcinoma • interventional radiology • liver • oncologic imaging • radioembolization • yttrium-90
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Accurate evaluation of response to 90Y treatment on imaging studies is important for adequate clinical management. Guidelines from the World Health Organization (WHO) and the Response Evaluation Criteria in Solid Tumors (RECIST) group represent the oncologic criteria for response [11, 12]. These criteria are based on changes in tumor size after therapy. Although numerous studies have been conducted to evaluate therapeutic response of HCC according to size criteria, in only a few studies has necrosis been used as a criterion of response. Moreover, the role of lesion necrosis in the therapeutic response to 90Y has not been satisfactorily investigated.
The purpose of our study was to describe the imaging features of HCC after 90Y treatment and to compare size criteria (WHO and RECIST) with necrosis criteria and combined criteria (RECIST and necrosis) for assessment of response. To our knowledge, this study is the largest to date of the cross-sectional imaging features of HCC after 90Y therapy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-fetoprotein level (one
[2%] of the patients), or imaging findings (22 [52%] of the patients).
Unilobar or bilobar 90Y treatments were administered depending on
the number and location of the liver lesions. Only patients with favorable
response or stable disease in the first treated lobe received treatment to the
other lobe. In patients with bilobar disease, each hepatic lobe was treated
separately at 30- to 60-day intervals. A total of 52 lobes were treated, and 76 HCC lesions were analyzed. One to four index lesions per lobe were analyzed. Physical examination, medical history interview, and laboratory tests were performed in all cases before 90Y treatment and repeated as clinically indicated. CT and MRI were performed immediately before treatment, a mean of 30 days after treatment, and at approximately 60- to 90-day intervals subsequently. Two to six CT examinations were performed per patient. The length of imaging follow-up after treatment ranged from 21 to 334 days (mean, 125 days). Follow-up imaging studies were categorized into intervals of 50 days for analytic purposes. Institutional review board approval was obtained. This retrospective study was compliant with the Health Insurance Portability and Accountability Act. All patients provided signed informed consent for 90Y treatment.
90Y Treatment
Hepatic arteriography and 99mTc-macroaggregated albumin
scintigraphy for detection of extrahepatic shunting of blood were performed
before 90Y treatment. Shunting to the lungs with a cumulative dose
of less than 30 Gy was acceptable. Significant flow to extrahepatic organs was
corrected by arterial coil embolization or catheter positioning. The most
commonly embolized vessels were the gastroduodenal and the right gastric
arteries.
Administration of 90Y has been described previously [7]. In short, 90Y glass microspheres were delivered into a lobar branch of the hepatic artery. The 90Y dose was based on liver volume (not tumor burden), which was calculated with CT or MRI. The corresponding liver mass was determined with the conversion factor 1.03 g/cm3 [13]. The activity required to deliver the desired dose to the liver was the product of the intended radiation absorbed dose and liver volume divided by 50. The desired 90Y dose was 100-120 Gy.
CT and MR Technique
CT examinations were performed with MDCT scanners (LightSpeed QX/i, GE
Healthcare; and Sensation 16 and Sensation 64, Siemens Medical Solutions).
According to our liver protocol, unenhanced and arterial and portal venous
phase images were acquired. With the LightSpeed scanner, unenhanced 5-mm
contiguous axial CT images of the abdomen were obtained in the high-quality
mode with a rotation speed of 0.8, table speed of 15 mm/rotation, 120 kV, and
220 mAs. A rotation speed of 0.5 and effective tube current of 250 mAs were
used for the Sensation scanners. Contrast-enhanced images were obtained with a
delay of approximately 30 seconds (arterial phase) and 70 seconds (venous
phase) after injection of 125 mL of iohexol (Omnipaque 350, Amersham Health)
at a rate of 3-5 mL/s with a mechanical power injector (EHU 700, Medrad). For
three patients, hepatic CT angiograms were obtained for better delineation of
the vascular distribution and lesions.
MRI was performed with a 1.5-T system (Magnetom Sonata, Siemens Medical
Solutions) with a highperformance gradient system (40 mT/m amplitude, 200
mT/m/ms slew rate) and phased-array body coil. Our MRI liver protocol
consisted of axial and coronal multislice thin-section T2-weighted HASTE,
axial unenhanced fat-suppressed T1-weighted spoiled gradientecho, and dynamic
gadolinium-enhanced axial T1-weighted spoiled gradient-echo sequences with fat
suppression in the arterial (based on fluoroscopy-preparation timing sequence)
and venous phases (45-60 and 90 seconds) and delayed phases 2-5 minutes after
contrast administration. Gadopentetate dimeglumine (Magnevist, Berlex
Pharmaceuticals) was administered with a power injector (Spectris, Medrad) at
a dose of 0.1 mmol/kg (
20 mL), followed by 20 mL of saline flush.
T1-weighted fat-suppressed spoiled gradient-echo images with shared prepulses
were obtained with the following parameters: TR/TE, 120-160/1.9; flip angle,
70°; slice thickness, 6 mm; gap, 1.8 mm; matrix size, 125 x 256;
rectangular field of view, 30-40 cm; 23 slices acquired in breath-hold of 20
seconds.
Image Evaluation
CT and MR studies were analyzed retrospectively on a PACS workstation by
two radiologists with 5 and 7 years of experience. Both radiologists in
consensus evaluated all studies. When disagreement occurred, a third
radiologist with more than 15 years of experience was invited to review the CT
and MR images, needed in only four of the CT cases, and his judgment
prevailed. Unenhanced and arterial and venous phase images were analyzed.
Treatment response was evaluated according to size criteria (WHO and RECIST),
necrosis criteria, and combined criteria (RECIST and necrosis)
(Table 1). Complete response,
partial response, stable disease, or progressive disease was determined by
comparison of each set of follow-up images with the baseline images.
Measurements of the lesions were obtained on baseline images and on each set
of follow-up images. The percentage change in the sum of the product of the
cross-sectional diameters of the index lesions was calculated according to WHO
criteria [1]. The percentage
change in the sum of the longest diameter of the index lesions was calculated
according to RECIST criteria.
|
After therapy, lesions may not change in size or may become smaller or larger. Lesions also may become more heterogeneous in density or signal intensity. On arterial phase images, residual tumor tends to be hyperdense compared with the surrounding liver and necrotic portion. On portal venous phase images, tumor can be hypodense relative to the liver but become more enhanced than the necrotic portion. The tumor also can be isodense or hyperdense relative to the liver. We examined the size of the lesions and the necrosis and enhancement patterns.
Necrosis was defined as no enhancing tissue. A maximum increase of 10 H on CT after contrast administration was accepted for necrotic tissue because it was considered insignificant. We determined the change in the average percentage of necrosis of the index lesions by comparing the extent of necrosis before therapy with each posttherapy scan based on volume. Complete necrosis of a lesion was considered complete response. Partial response required an at least 30% increase in the percentage of lesion necrosis. This threshold for response according to necrosis criteria was based on percentages similar to those in the RECIST criteria, which require an at least 30% decrease in size for partial response. A third category consisted of insufficient change in lesion necrosis to be classified as complete or partial response. Necrosis criteria were not used to differentiate stable and progressive disease.
|
|
|
In addition to response by lobe, response by patient was evaluated. When both hepatic lobes were treated, the best response by lobe represented the response of the patient. For example, if a patient had stable disease in one lobe and partial response in the other lobe, the patient was considered to have a partial response. In addition to changes in lesion size and necrosis after 90Y treatment, imaging findings such as thin rim enhancement and enhancing peripheral nodules in treated lesions were analyzed and correlated with response to treatment.
We compared the response rate of patients with portal vein thrombosis (PVT)
with that of patients without PVT according to our combined criteria. Imaging
response was correlated with -fetoprotein level in patients with a
baseline -fetoprotein level greater than 400 ng/mL. Time to response
was measured from the treatment date until the criteria for response were
first met. Survival was calculated from the date of first 90Y
treatment until death or the patient's being alive on May 13, 2006, whichever
came first. Twenty-seven patients died before May 13, 2006, owing to tumor
progression or cirrhosis.
Statistical Analysis
Categoric measurements were summarized with count and percentage. The
response rates based on the combined criteria were compared by use of
two-sided Fisher's exact test in two groups: first between subjects with and
those without PVT and then between subjects with stable or increased
-fetoprotein levels and those with at least a 50% decrease in
-fetoprotein level after treatment. Student's t test was used
to compare the time to response between the group with and that without PVT.
Logistic regression analysis was used to examine whether the baseline tumor
size was associated with the outcome of interest: partial or complete response
(yes vs no). The significance level was set at p = 0.05. The
Kaplan-Meier method was used to produce survival curves and to estimate median
survival time and corresponding 95% CI for patients with disease in each Okuda
stage and for all patients.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
Treatment Response by Lobe
Response rate by lobe, including complete and partial response, was 19%
(10/52) according to RECIST criteria, 21% (11/52) according to WHO criteria,
51% (27/52) according to necrosis criteria, and 53% (28/52) according to
combined criteria. According to our combined criteria, 9% (5/52) of lobes had
complete response, 44% (23/52) had partial response, 34% (18/52) had stable
disease, and 11% (6/52) had progressive disease. In the lobes with complete
response, the lesions became completely necrotic and decreased in size over
time. Of the 23 lobes with partial response, 14 (60%) met necrosis criteria
only, one (4%) met size criteria (RECIST) only, and eight (34%) met both
necrosis and size criteria for response. Among the patients who did not
respond to therapy (n =6) and in whom disease immediately progressed,
time to progressive disease ranged from 22 to 127 days (median, 33 days; mean,
48 days). Three lobes with complete response and seven lobes with partial
response according to our combined criteria had an increase in size of the
lesions on the first follow-up examination (19-75 days after treatment; mean,
31 days) (Fig. 1A,
1B,
1C).
|
PVT
PVT was found before treatment in 19 patients. The thrombus appeared to be
tumorous in 15 patients (in whom 10 thrombi appeared attached to parenchymal
masses and five were not clearly attached), and four patients appeared to have
bland thrombus. According to our combined criteria, 10 (52%) of the 19
patients with PVT and 15 (65%) of the 23 patients without PVT responded to
treatment (Fig. 2A,
2B). The difference in the
response rates in these two groups of patients was not statistically
significant (p = 0.67, Fisher's exact test). The time to first
response in patients with PVT was not statistically different from that in
patients without PVT (p = 0.4612).
-Fetoprotein
In five (41%) of 12 patients with a baseline -fetoprotein level
greater than 400 ng/mL, -fetoprotein level decreased at least 50% after
90Y therapy. Four (80%) of these five patients responded to
treatment, and one (20%) had progressive disease according to our combined
criteria. There was no significant correlation between -fetoprotein
level and treatment response on CT (p = 0.11, Fisher's exact test).
One patient with complete necrosis of the lesions after treatment had a
transient increase in -fetoprotein level from 232 to 568 ng/mL (Fig.
3A,
3B).
Survival
The median survival time was 660 days (95% CI, 447-upper limit is beyond
the last observed time in the study group) for patients with Okuda stage I
(n = 15) and 236 days (95% CI, 145-upper limit is beyond the last
observed time in the study group) for patients with Okuda stage II (n
= 27) disease. The median overall survival time was 431 days (95% CI,
192-upper limit is beyond the last observed time in the study group)
(Fig. 4).
Imaging Findings
Thin rim enhancement—After treatment, 25 (32%) of 76 lesions
had thin peripheral rim enhancement less than 5 mm thick (Fig.
5A,
5B). Eighty-four percent
(21/25) of these lesions responded to treatment, 12% (3/25) became stable, and
4% (1/25) progressed according to our combined criteria. Rim enhancement was
first seen 22-181 days (mean, 52 days) after treatment and lasted 25-282 days
(mean, 131 days). This finding eventually disappeared in seven (28%) of the
lesions. The length of follow-up for this specific group with thin rim
enhancement ranged from 22 to 334 days (mean, 150), a rate similar to that in
the overall study group.
|
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Imaging studies play an integral role in monitoring response to 90Y therapy [7, 8, 17, 18]. Traditionally, a decrease in tumor size according to the WHO and RECIST criteria has been considered CT evidence of treatment response [11, 12]. Although they represent the oncologic reference standard, these criteria are often unreliable, particularly after regional liver therapies [7, 19, 20]. The imaging response of HCC to 90Y therapy has been evaluated in only a few studies. In most of these studies, traditional size criteria were used for determining response to treatment.
Acknowledging the limitations of size criteria, the panel of experts on HCC of the European Association for the Study of the Liver proposed considering lesion necrosis in the evaluation of therapeutic response of malignant tumors [21, 22]. To our knowledge, a study of the role of lesion necrosis relative to changes in lesion size in the therapeutic response of HCC to 90Y has not been performed. Also, to our knowledge, the efficacy of the combined use of size and necrosis criteria to evaluate response has not been investigated. We conceived a method for response evaluation that combines size and necrosis criteria and allows comparison of responses according to each set of criteria separately. Our method allows identification of responses evidenced solely or predominantly by lesion necrosis. We used 30% necrosis, not any amount of necrosis, as the minimal threshold for response.
In our study, the response rates according to size criteria (WHO and RECIST) were underestimated compared with necrosis and combined criteria. In addition, use of necrosis and combined criteria led to detection of response much earlier than use of WHO and RECIST criteria. This finding shows the advantage of the necrosis criteria, because early identification of likelihood of response and lack of response is valuable in the care of patients with advanced HCC.
Necrosis, cystic degeneration, hemorrhage, and edema can increase the size of responding tumors [17]. Therefore increase in lesion size is not sufficient evidence for a diagnosis of progressive disease. In three lobes in our study, the treated lesions became completely necrotic but increased in size and were erroneously categorized as progressive disease on the basis of traditional size criteria. These responding lesions lacked enhancement, unlike tumors that increased in size but remained solid, indicating progression of disease. The addition of necrosis criteria to size criteria is important for detecting favorable response despite an increase in size after 90Y treatment.
Although the response rates according to size criteria were low, in most patients the size of the lesions became stable after 90Y therapy. Because all patients had progressive disease before treatment, and the disease likely would have progressed over time, stabilization of disease was an indication of clinical benefit. The response rates, including stable disease, were 78% (33/42) according to RECIST and WHO criteria and 88% (37/42) according to combined criteria. These response rates may be an accurate reflection of benefit from 90Y therapy.
A median survival period of 660 days was attained by patients with Okuda stage I disease and of 236 days by patients with Okuda stage II disease. The median overall survival period was 431 days. These survival times are consistent with those previously reported for patients with unresectable HCC undergoing 90Y microsphere treatment and show that 90Y therapy has a positive effect on the survival of these patients [5, 6, 9, 10]. In a study by Okuda et al. [1], the median survival time among 229 patients who received no specific treatment was 1.6 months, 0.7 month for patients with stage III disease, 2.0 months for patients with stage II disease, and 8.3 months for patients with stage I disease.
In patients with -fetoprotein values greater than 400 ng/mL before
treatment, we did not find a statistically significant association between
-fetoprotein level and imaging response according to our combined
criteria. A significant decrease in -fetoprotein level after
90Y treatment likely reflects response of hepatic lesions
[7,
8]. An increase in
-fetoprotein level after treatment, however, may be due to extrahepatic
progression of disease rather than worsening liver disease. A transient
increase in -fetoprotein level may also result from tumor lysis or
necrosis, as occurred in one of our patients (Fig.
3A,
3B). Therefore
-fetoprotein value cannot be used without imaging in the assessment of
treatment response.
Thin rim enhancement was seen in 32% (25/76) of the lesions in our study. Twenty-one (84%) of these lesions responded to treatment. This finding likely corresponds to the presence of granulation tissue related to inflammatory treatment reaction [23-25]. Thin rim enhancement has been reported after different ablation techniques, but to our knowledge, it has not been reported after 90Y therapy. According to the literature, thin rim enhancement is a transient finding with a maximal duration of 6 months [26, 27]. It was transient in only 28% (7/25) of the lesions in our study. However, sufficient follow-up to show resolution of this finding may not have been performed in all cases; 11 lesions did not have follow-up data for more than 6 months.
Enhancing peripheral nodules were seen in 18% (14/76) of treated lesions. The size of these nodules increased progressively in 10 (71%) of the 14 lesions and gradually filled out the lesions, indicating that they generally represented residual viable tumor. Residual tumor likely results from irregular distribution of 90Y microspheres inside the lesions [28, 29]. In addition, many cases of nodular enhancement represent incompletely treated tumor because they are in the watershed area between two vascular distributions. For example, the liver dome tumors may have supply from segments VIII and IV. At follow-up examinations, one distribution is treated, and the nodular area that has increased in size represents essentially untreated tumor. Some of the tumors present after treatment were believed to represent stable disease with residual tumor or scar rather than active aggressive tumor. They had decreased progressively in size in two of 14 (14%) lesions and decreased in size and eventually disappeared in two lesions that became completely necrotic. Tsuda et al. [30] found lack of growth and decreased enhancement of peripheral nodules in ablated HCC and attributed these findings to the slow development of coagulation necrosis in residual tumor. Delayed necrosis may explain the decrease in size or disappearance of peripheral nodules in our study. Some of the patients with peripheral nodular enhancement were treated after closure of the study owing to clinical progression. For this study, the statisticians and we considered it imperative to lock the database on a certain date.
HCC is frequently associated with PVT. Yttrium-90 treatment seems to be safe for these patients because, unlike transarterial chemoembolization, it does not produce a significant embolic effect, which can cause hepatic artery occlusion and liver failure in the setting of PVT [31]. We investigated whether PVT had any effect on response to 90Y treatment as evaluated with imaging. According to our results, 90Y treatment is as effective in HCC patients with PVT as it is in those without PVT. Therefore 90Y treatment seems to be an appealing therapy in this subset of patients with HCC and PVT.
We surmise that some patients do not respond to therapy because of radiation sensitivity and the cell cycle of the tumor. Using 90Y treatment, we treat at only one moment in time compared with standarddose fraction radiation therapy.
This retrospective study had limitations. Histopathologic correlation of imaging findings after therapy was not routinely obtained because of the potential for sampling error after therapy, and biopsy was considered too invasive without significant effect on future management. The HCC tumors were unresectable, and 90Y treatment was palliative. Correlation with functional information from PET studies was not obtained because the diagnostic sensitivity of PET for HCC is limited and, as a result, this examination is not routinely reimbursed by insurance.
In conclusion, as evaluated with imaging studies and survival analysis, 90Y treatment seems to be effective for patients with HCC, including patients with PVT. The combination of size (RECIST) and necrosis criteria may be superior to size criteria (WHO and RECIST) alone for determining response of HCC to 90Y treatment. We describe imaging features of HCC after 90Y treatment, including peripheral enhancing nodules and thin rim enhancement. Large, multiple-institution studies are necessary to validate the clinical efficacy, response criteria, and imaging findings among patients with HCC treated with 90Y microspheres.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Related articles in AJR:
This article has been cited by other articles:
|
|
 |
|
  B. Atassi, A. K. Bangash, A. Bahrani, G. Pizzi, R. J. Lewandowski, R. K. Ryu, K. T. Sato, V. L. Gates, M. F. Mulcahy, L. Kulik, et al. Multimodality Imaging Following 90Y Radioembolization: A Comprehensive Review and Pictorial Essay RadioGraphics, January 1, 2008; 28(1): 81 - 99. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. P. Kalva, A. Thabet, and S. Wicky Recent Advances in Transarterial Therapy of Primary and Secondary Liver Malignancies RadioGraphics, January 1, 2008; 28(1): 101 - 117. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Welsh Radiographically Identified Necrosis After 90Y Microsphere Brachytherapy: A New Standard for Oncologic Response Assessment? Am. J. Roentgenol., March 1, 2007; 188(3): 765 - 767. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
