Keywords radioembolization - locoregional therapy - liver-directed therapy
Introduction
Colorectal cancer (CRC) is the third most common cancer worldwide.[1 ] Over one-half of patients with CRC develop metastases (mCRC), most commonly involving
the liver.[2 ]
[3 ] Liver metastases are the principal cause of morbidity and mortality in patients
with mCRC.[4 ]
[5 ] Though systemic chemotherapy is the standard treatment for stage IV colorectal cancer,
the durability of this option is limited in the salvage setting.
Yttrium-90 (90 Y) radioembolization has garnered increasing interest in controlling the growth of
liver CRC metastases in patients who are not candidates for surgical resection or
ablation. It is most frequently used in the salvage setting where randomized prospective
trials have shown a 3-month survival advantage compared with standard chemotherapy.[6 ] In a retrospective matched-pair cohort comparison, a 5-month survival advantage
was demonstrated compared with best supportive care.[7 ] Despite encouraging data, there remains a wide disparity in response for CRC patients
undergoing radioembolization. The reasons for response differences have not been fully
elucidated. Whereas absorbed radiation dose may impact tumor response and lead to
improved outcomes,[8 ] the assessment of baseline characteristics to better stratify patients is needed.
Some early data with diffusion-weighted imaging have shown promise; however, magnetic
resonance is not routinely obtained in patients with colorectal cancer at most centers.[9 ] The presence of nonresponders in the mCRC population is particularly problematic
for several reasons. The use of radioembolization in the early stage may delay the
initiation of a patient's systemic chemotherapy. Furthermore, radioembolization may
result in short- and long-term toxicity that may hinder further systemic therapy and
survival. Therefore, effective stratification of patients as high- or low-likelihood
responders remains an unmet need with potential to optimize clinical outcomes in patients
with CRC liver metastases.
In a minimally embolic therapy such as radioembolization, increased tumor vascularity
should theoretically correlate with better response due to more effective delivery
of 90 Y microspheres into target tumors. The purpose of this study was to determine whether
tumor vascularity predicted imaging response or survival in this setting.
Methods
Patient Selection
A retrospective study was conducted at a single institution by accessing the hospital's
patient information system to identify all patients who underwent 90 Y radioembolization for hepatic mCRC between March 2008 and March 2015. This was collected
into a database and approved by the hospital's institutional review board with waiver
of informed consent.
All patients were discussed in a multidisciplinary liver tumor board. This board comprises
medical oncologists, hepatobiliary surgeons, pathologists, radiation oncologists,
diagnostic radiologists, and interventional radiologists. Radioembolization was not
offered to patients who were candidates for surgical resection or thermal ablation.
Patients were considered for radioembolization if they had liver-only or liver-dominant
mCRC. Patients had either failed at least one line of chemotherapy or radioembolization
was performed as an adjunct to first-line chemotherapy. A total of 89 patients were
identified. Fourteen patients were excluded for incomplete records or lack of follow-up,
leaving a total of 75 patients.
Prior to the procedure, all patients underwent baseline staging with a contrast-enhanced
computed tomography (CT) of the chest, abdomen, and pelvis. Timing of abdominal imaging
was performed in the portal venous phase (70–110 seconds after intravenous contrast
administration). Imaging was obtained within 1 month of the mapping angiogram.
Radioembolization
Radioembolization was performed using standard methods with either glass (TheraSphere,
BTG International) or resin (SirSphere, Sirtex) microspheres. This included a mapping
angiogram with lung shunt fraction determination, 90 Y microsphere infusion, and posttherapy Bremsstrahlung single-photon emission computed
tomography (SPECT)/CT. Device selection (glass vs. resin) was based on multiple factors,
including operator preference, device availability at the study institution at the
time of the procedure, and patients’ medical insurance guidelines. If necessary, prophylactic
coil embolization of extrahepatic arteries was performed to prevent nontarget radioembolization
to extrahepatic structures (e.g., bowel). In all treatments, selective catheterization
and infusion of the tumor-feeding arteries was performed whenever technically feasible.
In cases of multifocal disease in which selective radioembolization was not feasible,
a sequential lobar approach was taken.
For the mapping angiogram, a 4 or 5F diagnostic catheter was used to select the visceral
artery of interest (superior mesenteric and celiac arteries). Selective catheterization
of the lobar hepatic arteries was performed with a 2.8F coaxial microcatheter (Progreat,
Terumo Medical Corporation). Digital subtraction angiography (DSA) was performed at
three frames per second (Philips) with power injection of iodinated contrast (diluted
to 50–70%, depending on patient body habitus). Contrast injection was optimized to
maximize tissue uptake without significant arterial reflux from the microcatheter
tip position. Additional three-dimensional (3D) imaging of the abdomen was performed
using C-arm CT with contrast injection via the microcatheter positioned within the
lobar hepatic artery. Imaging was acquired over a 5-second timeframe with contrast
injection occurring both prior to and during image acquisition. This allowed for enhancement
to be identified both in the arterial and parenchymal phases. For glass microsphere
dosimetry, an intended dose of 120 gray to the target perfused tissue was delivered.
For resin microsphere dosimetry, activity was calculated based on the body-surface
area formula, according to the package insert.
Evaluation of Tumor Imaging Characteristics
Radiologists blinded to patient information and outcome data reviewed imaging and
categorized each tumor as hypo- or hypervascular according to[1 ] baseline contrast-enhanced CT (portal venous phase),[2 ] C-arm CT obtained during mapping angiography, and[3 ] DSA during the mapping procedure. On baseline CT and C-arm CT, if the tumor enhanced
to a greater degree than the surrounding hepatic parenchyma, it was categorized as
hypervascular. Peripheral hypervascularity (with central hypoenhancement) was categorized
as hypervascular. Conversely, if the tumor enhanced less than the surrounding parenchyma,
it was categorized as hypovascular ([Fig. 1 ]). Likewise, for DSA, tumors were categorized as hypo- or hypervascular based on
the degree of vascularity relative to the surrounding hepatic parenchyma ([Fig. 2 ]).
Fig. 1 C-arm CT during mapping angiogram with contrast injection from the left hepatic artery
demonstrating (a ) tumor hypovascularity and (b ) tumor hypervascularity.
Fig. 2 Digital subtraction angiogram during mapping angiogram with contrast injection from
the right hepatic artery demonstrating (a ) tumor hypovascularity and (b ) tumor hypervascularity.
Patient Follow-up and Response Evaluation
Follow-up visits and imaging (portal venous phase contrast-enhanced CT) took place
1 month following radioembolization and every 3 months thereafter. Toxicity was assessed
according to Common Terminology Criteria for Adverse Events (CTCAE), version 4.0.[10 ] Tumor response was categorized as complete response, partial response, stable disease,
or progressive disease, according to Response Evaluation Criteria in Solid Tumors
(RECIST) criteria version 1.1.[11 ]
Statistical Analysis
Categorical variables were summarized as number (percentage) and continuous variables
as mean (interquartile range [IQR]) or median (range). Categorical variables were
compared between vascularity groups using Fisher's exact test and continuous variables
were compared between groups using the Mann-Whitney test. Overall survival (OS) and
progression-free survival (PFS) were assessed using Kaplan-Meier curves. Index tumor
progression was assessed using the cumulative incidence curve to account for the competing
risk of death.[12 ] Median OS, PFS, and time-to-progression were estimated from these curves for each
vascularity group. Event rates were compared between vascularity group using Cox proportional-hazard
models (survival and progression-free survival) and the fine-gray proportion subdistribution
hazard model (index tumor progression).[13 ] Hazard ratios (HRs) were used to summarize differences in event rates between the
vascularity groups; 95% confidence intervals (CIs) were used to assess the extent
of the potential differences between vascularity groups. All statistical calculations
were conducted with the statistical computing language R (version 3.1.1; R Foundation
for Statistical Computing). Throughout, two-sided tests and CIs were used, with statistical
significance defined as p < 0.05.
Results
Seventy-five patients with mCRC treated with radioembolization were included. Patients’
age ranged from 29 to 81 years (median, 57), and 55% of the patient population was
female ([Table 1 ]). The colon was the primary site of malignancy in 79% of patients and rectum in
21% of patients; metastatic disease was diagnosed at the same time as the primary
tumor in 73%. KRAS mutation was identified in 63%. The median follow-up time after
treatment was 375 days (IQR: 186–614 days).
Table 1
Patient characteristics by vascularity
Vascularity by DSAa
Vascularity by C-arm CTa
Variable
All
(n = 75)
Hypervascular
(n = 37)
Hypovascular
(n = 38)
p Valueb
Hypervascular
(n = 22)
Hypovascular
(n = 15)
p Valueb
Abbreviations: CT, computed tomography; DSA, digital subtraction angiography; ECOG,
Eastern Cooperative Oncology Group.
a Values are no. (%) or median (range).
b Fisher's exact test or the Mann-Whitney test.
c Patients with missing values for primary tumor site (n = 3), timing of diagnosis (n = 2), KRAS status (n = 26), or carcinoembryonic antigen (n = 23) were excluded.
Sex
Male
34 (45.3)
15 (40.5)
19 (50.0)
0.49
7 (31.8)
9 (60.0)
0.11
Female
41 (54.7)
22 (59.5)
19 (50.0)
15 (68.2)
6 (40.0)
Age (y)
57 (29–81)
57 (30–81)
58 (29–75)
0.71
57 (35–72)
55 (30–75)
> 0.99
ECOG score
0
39 (52.0)
20 (54.0)
19 (50.0)
> 0.99
9 (40.9)
10 (66.7)
0.39
1
31 (41.3)
15 (40.5)
16 (42.1)
10 (45.5)
4 (26.7)
2
5 (6.7)
2 (5.4)
3 (7.9)
3 (13.6)
1 (6.7)
Primary tumor sitec
Colon
57 (79.2)
31 (86.1)
26 (72.2)
0.25
20 (90.9)
5 (38.5)
0.002
Rectum
15 (20.8)
5 (13.9)
10 (27.8)
2 (9.1)
8 (61.5)
Timing of hepatic metastatis diagnosisc
Metachronous with primary
20 (27.4)
7 (19.4)
13 (35.1)
0.19
4 (18.2)
5 (35.7)
0.27
Synchronous with primary
53 (72.6)
29 (80.6)
24 (64.9)
18 (81.8)
9 (64.3)
KRAS mutation statusc
Wild type
18 (36.7)
9 (37.5)
9 (36.0)
> 0.99
8 (44.4)
4 (36.4)
0.72
Mutant
31 (63.3)
15 (62.5)
16 (64.0)
10 (55.6)
7 (63.6)
Carcinoembryonic antigenc
32.6 (1.2–1,658)
32.6 (1.2–1,658)
28.2 (1.5–993)
0.48
28.9 (1.8–694)
81.8 (2.4–993)
0.41
Bilirubin
0.7 (0.3–1.9)
0.7 (0.3–1.9)
0.6 (0.3–1.5)
0.058
0.7 (0.3–1.5)
0.7 (0.4–1.3)
0.73
Albumin
3.7 (1.1–4.5)
3.8 (1.1–4.5)
3.6 (2.0–4.3)
0.53
3.8 (1.1–4.5)
3.5 (2.7–4.3)
0.36
Of the 75 patients treated, 37 (49%) had hypervascular tumors determined by DSA. Thirty-seven
patients had C-arm CT imaging during mapping angiography, of which 22 (59%) were classified
as hypervascular and 15 (41%) hypovascular. Patients who did not have C-arm CT were
either due to lack of this technology in one of the angiography suites, or due to
significant respiratory motion in which C-arm CT could not be performed. Although
they were not available in all patients, C-arm CT images were available in a similar
fraction of patients with hypervascular tumors per DSA as patients with hypovascular
tumors (49% vs. 50%, p > 0.99). Of the 37 patients with both DSA and C-arm CT images, there was concordance
of imaging hypervascularity in 29 (78%).
Tumor vascularity was also examined on baseline contrast-enhanced CT. By CT, only
5 of 75 (7%) patients’ tumors were classified as hypervascular, all of which were
concordant with DSA and C-arm CT. Because of the limited sample size of tumors classified
as hypervascular by CT, these classifications by baseline CT were not analyzed further.
Grouped according to vascularity, baseline patient characteristics are shown in [Table 1 ], and baseline tumor characteristics are shown in [Table 2 ]. There were no significant differences in baseline patient characteristics between
patients who had hyper- or hypovascular tumors ([Table 1 ]), with the exception of the site of the primary tumor. The only tumor or treatment
characteristic that differed significantly between hypo- and hypervascular tumors
was the 90 Y device used ([Table 2 ]), where hypervascular tumors were more likely to be treated using glass microspheres
than hypovascular tumors based on DSA and C-arm CT.
Table 2
Tumor and treatment characteristics by vascularity
Vascularity by angiographya
Vascularity by C-arm CTa
Variable
All
(n = 75)
Hypervascular
(n = 37)
Hypovascular
(n = 38)
p Valueb
Hypervascular
(n = 22)
Hypovascular
(n = 15)
p Valueb
Abbreviation: CT, computed tomography.
a Values are no. (%) or median (range).
b Fisher's exact test or the Mann-Whitney test.
c Patients with missing values for prior chemotherapy (n = 1), chemotherapy after 90 Y (n = 14), or antiangiogenic agent after 90 Y (n = 16) were excluded.
No. of tumors
1
5 (6.7)
2 (5.4)
3 (7.9)
0.13
1 (4.5)
1 (6.7)
0.37
2–3
11 (14.7)
9 (24.3)
2 (5.3)
6 (27.3)
1 (6.7)
4–5
10 (13.3)
5 (13.5)
5 (13.2)
4 (18.2)
2 (13.3)
6+
49 (65.3)
21 (56.8)
28 (73.7)
11 (50.0)
11 (73.3)
Tumor distribution
Bilobar
62 (82.7)
29 (78.4)
33 (86.8)
0.38
18 (81.8)
13 (86.7)
> 0.99
Unilobar
13 (17.3)
8 (21.6)
5 (13.2)
4 (18.2)
2 (13.3)
Long axis tumor (largest) dimension (mm)
44 (14–140)
48 (15–130)
40 (14–140)
0.39
49 (23–130)
34 (14–119)
0.16
Infiltrative tumor
8 (10.7)
5 (13.5)
3 (7.9)
0.48
3 (13.6)
2 (13.3)
> 0.99
Tumor burden > 50%
7 (9.3)
2 (5.4)
5 (13.2)
0.43
2 (9.1)
2 (13.3)
> 0.99
Portal vein thrombus
4 (5.3)
2 (5.4)
2 (5.3)
> 0.99
2 (9.1)
1 (6.7)
> 0.99
Definite extrahepatic spread
13 (17.3)
5 (13.5)
8 (21.1)
0.54
5 (22.7)
5 (33.3)
0.71
Prior liver treatment
Resection
19 (25.3)
11 (29.7)
8 (21.1)
0.44
6 (27.3)
2 (13.3)
0.43
Ablation
16 (21.3)
9 (24.3)
7 (18.4)
0.58
1 (4.5)
4 (26.7)
0.14
Chemoembolization
1 (1.3)
1 (2.7)
0 (0.0)
0.49
0 (0.0)
0 (0.0)
> 0.99
External beam radiation
2 (2.7)
1 (2.7)
1 (2.6)
> 0.99
1 (4.5)
1 (6.7)
> 0.99
Prior lines of chemotherapyb
0
1 (1.4)
0 (0.0)
1 (2.6)
0.71
0 (0.0)
1 (6.7)
0.54
1
25 (33.8)
13 (36.1)
12 (31.6)
10 (45.5)
4 (26.7)
2
28 (37.8)
12 (33.3)
16 (42.1)
8 (36.4)
7 (46.7)
3–4
20 (27.0)
11 (30.6)
9 (23.7)
4 (18.2)
3 (20.0)
Lung shunt fraction (%)
3.1 (1.0–12.4)
3.0 (1.0–12.4)
3.2 (1.2–9.8)
0.71
3.3 (1.0–12.4)
3.2 (2.0–9.8)
0.99
90 Y device infused
Resin
43 (57.3)
15 (40.5)
28 (73.7)
0.005
4 (18.2)
8 (53.3)
0.036
Glass
32 (42.7)
22 (59.5)
10 (26.3)
18 (81.8)
7 (46.7)
Location of 90 Y infusions
Right lobe only
7 (9.3)
1 (2.7)
6 (15.8)
0.12
3 (13.6)
3 (20.0)
0.50
Left lobe only
31 (41.3)
18 (48.6)
13 (34.2)
12 (54.5)
5 (33.3)
Both lobes
37 (49.3)
18 (48.6)
19 (50.0)
7 (31.8)
7 (46.7)
After treatment, most clinical complications and biochemical toxicities were uncommon,
and not significantly different between the different groups of patients ([Table 3 ]). Of the 68 patients with imaging follow-up, 5.9% had a complete response by RECIST
criteria. Though rates of complete response were higher in the hypovascular groups
by DSA and C-arm CT, these differences did not reach statistical significance ([Table 4 ]). There were no significant differences in OS (median: 439 vs. 342 days, p = 0.96), PFS (median: 111 versus 128 days, p = 0.41), or hepatic time-to-progression (median: 313 vs. 244 days, p = 0.83) between DSA hyper- and hypovascular tumors ([Fig. 3 ]). The associated HRs are summarized in [Table 5 ]. Results were not meaningfully changed after adjusting for the 90 Y device used.
Fig. 3 Kaplan-Meier curves showing overall survival (left panel) and progression-free survival
(right panel) for patients with hyper- and hypovascular lesions by DSA. Tick marks
indicate censoring times. VA, vascularity by angiography.
Table 3
Short-term clinical and biochemical toxicity by vascularity
Vascularity by DSAb
Vascularity by C-arm CTb
Clinical toxicitya
All
(n = 74)
Hyper
(n = 37)
Hypo
(n = 37)
p Valuec
Hyper
(n = 22)
Hypo
(n = 14)
p Valuec
Abbreviations: AST, aspartate aminotransferase; ALT, alanine transaminase; CT, computed
tomography; CTCAE, Common Terminology Criteria for Adverse Events; DSA, digital subtraction
angiography.
a Patients with insufficient follow-up to assess clinical toxicity (n = 1), hospital readmission (n = 3), or biochemical toxicity (n = 5) were excluded.
b Values are no. (%).
c Fisher's exact test.
Fatigue
45 (60.8)
24 (64.9)
21 (56.8)
0.63
14 (63.6)
9 (64.3)
> 0.99
Pain
12 (16.2)
6 (16.2)
6 (16.2)
> 0.99
3 (13.6)
4 (28.6)
0.39
Postembolization syndrome
10 (13.5)
5 (13.5)
5 (13.5)
> 0.99
2 (9.1)
1 (7.1)
> 0.99
Hospital readmission within 30 d
6 (8.3)
3 (8.3)
3 (8.3)
> 0.99
3 (14.3)
2 (14.3)
> 0.99
Ulcer
5 (6.8)
1 (2.7)
4 (10.8)
0.36
0 (0.0)
0 (0.0)
> 0.99
Ascites
4 (5.4)
3 (8.1)
1 (2.7)
0.61
2 (9.1)
0 (0.0)
0.51
Liver failure
3 (4.0)
3 (8.1)
0 (0.0)
0.24
2 (9.1)
0 (0.0)
0.51
Death
1 (1.4)
1 (2.7)
0 (0.0)
> 0.99
1 (4.5)
0 (0.0)
> 0.99
Encephalopathy
0 (0.0)
0 (0.0)
0 (0.0)
> 0.99
0 (0.0)
0 (0.0)
> 0.99
Biloma
0 (0.0)
0 (0.0)
0 (0.0)
> 0.99
0 (0.0)
0 (0.0)
> 0.99
Abscess
0 (0.0)
0 (0.0)
0 (0.0)
> 0.99
0 (0.0)
0 (0.0)
> 0.99
Biochemical toxicity
All
DSA Hyper
DSA Hypo
CT Hyper
CT Hypo
(CTCAE grade ≥ 3)
a
(n = 70)
(n = 35)
(n = 35)
p Value
c
(n = 21)
(n = 14)
p Value
c
Any biochemical toxicity
5 (7.2)
3 (8.6)
2 (5.9)
> 0.99
2 (10.0)
0 (0.0)
0.50
Hypoalbuminemia
4 (5.7)
2 (5.7)
2 (5.7)
> 0.99
2 (9.5)
0 (0.0)
0.51
Increased total bilirubin
2 (2.9)
2 (5.7)
0 (0.0)
0.49
2 (10.0)
0 (0.0)
0.50
Leukopenia
1 (1.4)
1 (2.9)
0 (0.0)
> 0.99
0 (0.0)
0 (0.0)
> 0.99
Increased AST
1 (1.4)
1 (2.9)
0 (0.0)
> 0.99
1 (4.8)
0 (0.0)
> 0.99
Increased ALT
0 (0.0)
0 (0.0)
0 (0.0)
> 0.99
0 (0.0)
0 (0.0)
> 0.99
Table 4
Best response measured by RECIST criteria
Vascularity by angiographyb
Vascularity by C-arm CTb
Variable
Alla
(n = 68)
Hyper
(n = 37)
Hypo
(n = 31)
p Valuec
Hyper
(n = 22)
Hypo
(n = 12)
p Valuec
Abbreviations: CR, complete response; CT, computed tomography; PR, partial response;
PD, progression of disease; RECIST, Response Evaluation Criteria in Solid Tumors;
SD, stable disease.
a Patients with insufficient imaging follow up (n = 7) were excluded.
b Values are no. (%).
c Fisher's exact test.
Best response
CR
4 (5.9)
1 (2.7)
3 (9.7)
0.55
1 (4.5)
2 (16.7)
0.74
PR
22 (32.4)
14 (37.8)
8 (25.8)
7 (31.8)
4 (33.3)
SD
23 (33.8)
12 (32.4)
11 (35.5)
8 (36.4)
3 (25.0)
PD
19 (27.9)
10 (27.0)
9 (29.0)
6 (27.3)
3 (25.0)
Table 5
Differences in survival and progression between vascularity groups defined by DSA
Event incidencea
Event
No.
Hyper (n = 37)
Hypo (n = 38)
HR
(95% CI)
p Value
Abbreviations: CI, confidence interval; DSA, digital subtraction angiography; HR,
hazard ratio.
HR > 1 indicates hypervascular group is at higher risk than hypovascular group.
a Values are no. (%); patients with insufficient imaging follow-up (n = 7) were excluded.
Death
75
32 (86.5)
30 (78.9)
0.99
(0.60–1.63)
0.96
Death or overall tumor progression
75
35 (94.6)
34 (89.5)
1.22
(0.76–1.96)
0.41
Index tumor progression
68
22 (61.1)
22 (68.8)
0.94
(0.52–1.68)
0.83
Similarly, there were no significant differences in OS (median: 489 vs. 342 days,
p = 0.74), PFS (median: 100 vs. 152 days, p = 0.62), or hepatic time-to-progression (median: 231 vs. 273 days, p = 0.42) between C-arm CT hyper- and hypovascular tumors ([Fig. 4 ]). The associated HRs are summarized in [Table 6 ].
Fig. 4 Kaplan-Meier curves showing overall survival (left panel) and progression-free survival
(right panel) for patients with hyper- and hypovascular lesions by C-arm CT. Tick
marks indicate censoring times. VC, vascularity by c-carm CT.
Table 6
Differences in survival and progression between vascularity groups defined by C-arm
CT
Event incidencea
Event
No.
Hyper
(n = 22)
Hypo
(n = 15)
HR
(95% CI)
p Value
Abbreviations: CI, confidence interval; HR, hazard ratio.
HR > 1 indicates hypervascular group is at higher risk than hypovascular group.
a Values are no. (%); patients with insufficient imaging follow up were excluded.
Death
37
17 (77.3)
11 (73.3)
0.88
(0.40–1.90)
0.74
Death or overall tumor progression
37
20 (90.9)
14 (93.3)
1.19
(0.6–2.38)
0.62
Index tumor progressiona
34
13 (59.1)
7 (58.3)
1.44
(0.60–3.47)
0.42
Discussion
Growing evidence suggests that 90 Y radioembolization in addition to chemotherapy may be superior to chemotherapy alone
for refractory liver mCRC.[6 ]
[14 ]
[15 ]
[16 ] However, a subset of patients may experience toxicity or may fail to achieve hepatic
tumoral stability or response. According to a recently published phase III trial,
although addition of radioembolization to fluorouracil (5FU) increased the proportion
of patients with stable disease from 35 to 76% and resulted in partial response in
10% of patients (compared with 0% in the 5FU-only group), there were still 10% of
patients with disease progression after radioembolization.[6 ] Similarly, prospective studies have reported that between 14.8 and 37% of patients
have progressive disease despite radioembolization, and retrospective studies have
reported as high as 23% progressive disease based on radiographic response.[17 ] Therefore, it is reasonable to conclude that there is a sizeable portion of patients
who do not benefit from radioembolization, and it remains an open question how this
cohort can be identified a priori.
Several factors have been associated with poor outcomes after radioembolization. Two
groups have reported that more than 3 prior lines of chemotherapy are associated with
poor response to radioembolization.[18 ]
[19 ] It is unclear whether this is a result of the effect of prior chemotherapy on the
susceptibility of tumor tissue to radiotherapy, or if multiple prior lines of chemotherapy
are simply a marker of aggressive tumors. Documented predictors of adverse outcomes
after radioembolization include large hepatic burden of disease, extrahepatic disease,
and lymph node involvement.[17 ]
[18 ]
[20 ]
[21 ] Additional factors associated with adverse outcomes after radioembolization include
elevated baseline tumor markers and rectal primary mCRC.[18 ]
[20 ] Though multiple prognostic factors are likely required to identify appropriate patients
for radioembolization, our study examines the potential utility of radiographic measures
of tumor vascularity to augment patient selection.
This study demonstrates that tumor vascularity, as measured by C-arm CT and DSA, does
not correlate with tumor response or survival. These data suggest that although biomarkers
for high-likelihood radioembolization-responders remain in need, radiographic measures
of tumor vascularity are unlikely to be of use for this purpose. Though these results
are unexpected, they have potentially important consequences for patient selection
for radioembolization.
There are several possible explanations for the results in this study. There is extensive
literature to suggest that tissue oxygen is a radiosensitizer and that hypoxia is
radioprotective.[22 ]
[23 ]
[24 ]
[25 ] First, hypoxia is known to occur in many solid tumors and local hypoxia that may
promote tumor angiogenesis.[26 ] Therefore, there may not necessarily be a direct or predictable relationship between
a tumor's vascularity and the partial pressure of oxygen (PO2 ) in the tumor tissue.[27 ] Indeed, if vascularity varies in concert with tissue metabolism, there may be no
overall relation between vascularity and PO2 as demand may exceed supply. There is great interest in the development of technology
for imaging tumor hypoxia.[28 ] Understanding the PO2 within target tumors may help with patient selection or dose planning for radioembolization.
Whereas radioembolic particles may be more effectively delivered to hypervascular
tumors, such tumors may be inherently more aggressive. Alternatively, additional unidentified
components in the tumoral stroma and micro-environment may affect both vascularity
and susceptibility to radiation.
Limitations to this study exist. This is a retrospective study, and therefore both
patient selection and radioembolization device is nonuniform for the patient cohort.
Second, though the assessment of vascularity was performed by a consensus of physicians
who were blinded to the results, no standardized criteria currently exist for the
scoring of tumor vascularity. Additionally, because CT images are acquired over an
instant, the appearance of the vascularity of the tumor could be affected by the timing
of the acquisition relative to the contrast material injection. Nevertheless, the
CT protocol used reflects a relatively conventional method for portal venous phase
timing, generalizable to other patient populations. Third, it is possible that hypervascular
tumors may exhibit more aggressive behavior compared with hypovascular tumors. In
that scenario, radioembolization could be more effective at treating (aggressive)
hypervascular tumors than at treating (less aggressive) hypovascular tumors, leading
to similar outcomes in two tumor populations that might behave differently without
treatment.
Conclusion
In summary, this study concludes that radiographic measures of tumor vascularity correlate
poorly with outcome after 90 Y radioembolization. Therefore, the degree of vascularity of a tumor should not impact
candidacy for radioembolization given current evidence.
Note
This research did not receive any specific grant from funding agencies in the public,
commercial, or not-for-profit sectors.
