Yale University
Yale University
EliScholar – A Digital Platform for Scholarly Publishing at Yale
EliScholar – A Digital Platform for Scholarly Publishing at Yale
Yale Medicine Thesis Digital Library
School of Medicine
1-1-2019
Identifying Quantitative Enhancement-Based Imaging Biomarkers
Identifying Quantitative Enhancement-Based Imaging Biomarkers
In Patients With Colorectal Cancer Liver Metastases Undergoing
In Patients With Colorectal Cancer Liver Metastases Undergoing
Loco-Regional Tumor Therapy
Loco-Regional Tumor Therapy
Mansur Abdul Ghani
Follow this and additional works at: https://elischolar.library.yale.edu/ymtdl
Part of the Medicine and Health Sciences Commons
Recommended Citation
Recommended Citation
Ghani, Mansur Abdul, “Identifying Quantitative Enhancement-Based Imaging Biomarkers In Patients With
Colorectal Cancer Liver Metastases Undergoing Loco-Regional Tumor Therapy” (2019). Yale Medicine
Thesis Digital Library. 3497.
https://elischolar.library.yale.edu/ymtdl/3497
This Open Access Thesis is brought to you for free and open access by the School of Medicine at EliScholar – A
Digital Platform for Scholarly Publishing at Yale. It has been accepted for inclusion in Yale Medicine Thesis Digital
Library by an authorized administrator of EliScholar – A Digital Platform for Scholarly Publishing at Yale. For more
information, please contact elischolar@yale.edu.
Identifying Quantitative Enhancement-based Imaging Biomarkers in Patients with
Colorectal Cancer Liver Metastases undergoing Loco-regional Tumor Therapy
A Thesis Submitted to the Yale University School of Medicine
in Partial Fulfillment of the Requirements for
the Degree of Doctor of Medicine
by
Mansur Abdul Ghani
2019
Identifying Quantitative Enhancement-based Imaging Biomarkers in Patients with
Colorectal Cancer Liver Metastases undergoing Loco-regional Tumor Therapy
Mansur A. Ghani, Julius Chapiro, and Todd Schlachter. Section of Interventional
Radiology, Department of Radiology and Biomedical Imaging, Yale School of Medicine,
New Haven, CT
The purpose of this study was to test and compare the ability of radiologic
measurements of lesion diameter, volume, and enhancement on baseline magnetic
resonance (MR) images to be predictors of overall survival (OS) and markers of
treatment response in patients with liver-dominant colorectal cancer metastases
undergoing loco-regional tumor therapies.
This retrospective study included 88 patients with colorectal cancer (CRC) liver
metastases, treated with transarterial chemoembolization (TACE) or Y90 transarterial
radioembolization (TARE) between 2001 and 2014. All patients received contrast-
enhanced MRI prior to therapy. Semi-automated whole liver and tumor segmentations of
three dominant lesions were performed on baseline MRI to calculate total tumor volume
(TTV) and total liver volumes (TLV). Quantitative 3D analysis was performed to
calculate enhancing tumor volume (ETV), enhancing tumor burden (ETB, calculated as
ETV/TLV), enhancing liver volume (ELV), and enhancing liver burden (ELB, calculated
as ELV/TLV). Overall and enhancing tumor diameters were also measured. Response
assessment was analyzed in a subset of 63 patients who received 1-month MRI follow-up
imaging using RECIST, mRECIST, change in ELV (DELV), vRECIST and qEASL. A
modified Kaplan-Meier method was used to determine appropriate cutoff values to
stratify patients based on these metrics. The predictive value of each parameter was
assessed by Kaplan-Meier survival curves as well as univariate and multivariate cox
proportional hazard models (statistical significance defined as p < .05).
In baseline imaging analysis, all methods except ELB achieved statistically
significant separation of survival curves. Multivariate analysis showed a HR of 2.1 (95%
CI 1.3-3.4, p=0.004) for enhancing tumor diameter, HR 1.7 (95% CI 1.1-2.8, p=0.04) for
TTV, HR 2.3 (95% CI 1.4-3.9, p<0.001) for ETV, and HR 2.4 (95% CI 1.4-4.0, p=0.001)
for ETB. Among treatment response assessment methods, only vRECIST achieved
statistically significant separation of survival curves and gave a HR of 2.1 (95% CI 1.1-
4.0, p=0.02).
In conclusion, tumor enhancement of CRC liver metastases on baseline MR
imaging is strongly associated with patient survival after loco-regional tumor therapy,
suggesting that ETV and ETB are better prognostic indicators than non-enhancement
based and one-dimensional based markers. However, while volumetric-based methods
are superior to 1D methods, enhancement-based methods of treatment response
assessment were not successful in predicting survival. A potential implication of these
findings as novel staging markers warrants prospective validation.
Acknowledgements
I greatly appreciate the Yale School of Medicine Office of Student Research and
the Radiological Society of North America for their generous funding of this research.
I am incredibly grateful to Dr. Todd Schlachter and Dr. Julius Chapiro for making
this thesis possible, and going above and beyond the obligations of good mentors. Under
their tutelage and guidance, the Yale Interventional Oncology Research Lab has become
another family to me that has fostered all areas of my professional and personal growth.
Thank you to my parents, Shahid Ghani and Arshia Rahman for their
unconditional love and support. I owe any and every success I achieve in my life to your
sacrifices. To my brother, Yusuf Ghani, thank you for always reminding me what really
matters in life. Finally, to my best friend and beautiful wife, Naureen Rashid, thank you
for being the greatest companion I could ask for, and I can’t wait to experience all of
life’s adventures with you.
Table of Contents
INTRODUCTION .................................................................................................................................. 1
METHODS
............................................................................................................................................. 5
RESULTS PART I: BASELINE MR IMAGING ANALYSIS
.............................................................15
RESULTS PART II: TREATMENT RESPONSE ASSESSMENT .....................................................25
DISCUSSION ........................................................................................................................................37
REFERENCES
......................................................................................................................................42
1
Introduction
Colorectal cancer (CRC) is the third most common cancer in the world and the
second leading cause of cancer-related deaths worldwide, resulting in 700,000 deaths per
year (1). The mortality rate from CRC has dropped over the last several decades due to
increased screening and prevention as well as more effective treatment options; the 1-
year and 5-year survival of patients with CRC have improved to 83.2 and 64.3%
respectively. However, the occurrence of CRC metastases to distant organs drops the 5-
year survival to 11.7% (2). The liver is the most common site of metastatic disease,
present in approximately 25% of patients at diagnosis with a prevalence of nearly 65%
during the course of disease (3). Although surgical resection of the primary tumor and
liver metastases is currently the most effective treatment option, this is generally feasible
if there are £5 metastases per liver lobe, at least two adjacent tumor-free segments, and a
liver remnant after surgery >20% (4). Only 10 to 25% of patients with hepatic metastases
from CRC are candidates for hepatic resection at diagnosis (5). The remainder are treated
with systemic chemotherapy with the goal of improving survival and, in some,
downsizing to allow for liver resection (6). However, this is unable to prevent the
development of progressive disease in the majority of patients (7). As a result, liver-
directed loco-regional treatments for patients with unresectable hepatic metastases, in the
form of image-guided intra-arterial therapies (IAT), including yttrium-90 (90Y)
transarterial radioembolization (TARE), conventional transarterial chemoembolization
(cTACE), or drug-eluting bead TACE (DEB-TACE), are often indicated for palliative
therapy (8–10).
2
Current guidelines by the National Comprehensive Cancer Network (NCCN)
recommend that IATs may be used in patients with liver dominant metastatic disease
(³80% of tumor burden located in the liver) and when the level of hepatic involvement is
not greater than 60% (11). Compared with systemic therapies, IATs can result in
significantly higher concentrations of drugs within the tumor as well as a lower incidence
of systemic toxicities and adverse events (12). In general, IATs mitigate drug toxicity and
yield more robust local tumor control by targeting the mostly arterially supplied tumor
tissue while sparing non-tumoral liver parenchyma, which is mainly fed through the
portal vein (13). Three common IATs include cTACE, DEB-TACE and TARE.
Conventional TACE delivers an emulsion of conventional chemotherapeutic agents
carried by Lipiodol to the tumor-feeding artery. Lipiodol is an iodinated poppy seed oil-
based medium that works as an effective drug carrier, partial embolic agent and contrast
agent which is easily visualized under fluoroscopy and computed tomography (CT),
helping to confirm targeting and complete tumor coverage (14). Polymer-based drug-
eluting beads (DEBs) were developed with the hopes of delivering higher concentrations
of chemotherapy to the tumor while improving systemic toxicities caused by cTACE
(15). DEB-TACE results in a controlled release of chemotherapeutics over several hours
to days after injection (16). TARE involves delivery to the tumor of radioactive
microspheres that emit b-radiation into the surrounding tissue. It is also a safe and
effective treatment for unresectable, chemorefractory colorectal cancer metastases to the
liver (17).
The success of IATs in clinical trials has firmly established these interventional
techniques as mainstays in palliative therapy for advanced hepatic metastatic disease, and
3
research efforts to improve the efficacy of these modalities continue to grow. However,
currently there is no agreed upon prognostic staging system that can give accurate
prognostic information regarding patients with advanced CRC (18). A number of studies
have proposed classification and staging systems based on a variety of variables
including the number of metastatic nodules, size of metastases, unilobar versus bilobar
involvement, the extent of liver involvement, performance status, and serum alkaline
phosphatase, but none of these systems have gained universal acceptance (19–23).
Instead, much current work has centered on the accurate assessment of treatment
response. The primary clinical purpose of follow-up imaging is to be able to determine
responders and non-responders with the purpose of informing therapeutic decisions.
Several standard guidelines have been established to evaluate tumor morphology for this
purpose. The two most common protocols are Response Evaluation Criteria in Solid
Tumors (RECIST) and World Health Organization (WHO) criteria, which measure tumor
diameters in one and two dimensions, respectively (24). However, these measures are
poor indicators of response following IATs, as these procedures usually rely on
embolization of tumor-feeding arteries resulting in necrosis of the tumor without
immediate effects on overall size (25).
Due to this shortcoming, modified RECIST (mRECIST) and European
Association for the Study of the Liver (EASL) criteria, which measure enhancing tumor
diameter on contrast-enhanced MRI in one dimension or two dimensions, respectively,
were developed. However, these 1D and 2D image assessment techniques are susceptible
to inherent inaccuracies, including limited reproducibility and inability to quantify
heterogeneous tumors (26). As a result, three-dimensional quantitative image analysis
4
techniques including volumetric RECIST (vRECIST) and quantitative EASL (qEASL)
have been developed to more accurately quantify tumor response via volumetric tumor
measurements and enhancing tumor volume (27). Preliminary studies have demonstrated
the superiority of these techniques in predicting survival after intra-arterial therapy (28–
30). Two recent studies determined that baseline 3D enhancement-based tumor burden
analysis in hepatocellular carcinoma (HCC) patients better predicted survival than
diameter- and non-enhancement-based measurements (31,32).
While assessment of treatment response is certainly beneficial in helping guide
therapeutic decision-making, it may take anywhere from one to six months after the first
IAT session to determine response depending on what assessment guidelines are used. A
prognostic staging system is advantageous in its ability to inform clinical decision-
making at the time of diagnosis. Tumor enhancement on imaging may be an important
component of such a staging system. However, to date, no studies have investigated 3D
enhancement-based analysis in CRC liver metastases prior to TACE or TARE.
Additionally, there is a desirability to utilize a whole-liver approach to quantitative
response. Currently available means of quantitating tumor enhancement requires
segmentation of the tumor to delineate tumor borders from normal liver parenchyma.
This can be a time-intensive process and the accuracy varies with the expertise of the
operator. A whole-liver approach, on the other hand, quantitates the enhancement in the
entire liver volume. This method only requires segmentation of the whole liver, which is
much faster to generate, eliminates the subjectivity associated with lesion-based analysis,
and accounts for tumor heterogeneity (29). It is important to address these gaps in
knowledge to validate the use of 3D quantitative imaging techniques as indicators of
5
therapeutic efficacy in an increasing number of clinical trials and to inform clinical
treatment recommendations for patients with hepatic metastases of CRC.
The purpose of our study was to (1) determine whether 3D whole-liver and tumor
enhancement features can serve as a staging biomarker in patients with CRC metastases
to the liver and (2) determine if a whole-liver approach can be used to measure treatment
response.
Methods
Study cohort
This single-institution study was conducted in compliance with the Health
Insurance Portability and Accountability Act and approved by the institutional review
board. Between 2001 and December 2014, a total of 126 patients with liver-only or liver-
dominant metastatic colorectal cancer (mCRC) underwent their first session of IAT
within our institution and received contrast-enhanced MR imaging within 60 days
following IAT. Patients were excluded if their baseline imaging was missing from the
database (N=20). Additional patients were excluded if imaging was truncated or poor
quality (e.g., motion artifact) (N=17). One patient was excluded because of failure of
registration between the pre-and post-contrast images. The remaining 88 patients, treated
with cTACE, DEB-TACE, or TARE, were included in the final analysis.
Of these 88 patients, 70 received one month post-procedure follow-up MR
imaging. Seven were missing imaging from our patient database, resulting in 63 patients
included for follow-up analysis (Fig. 1).
6
Fig. 1 Flowchart for patient selection process and exclusion criteria
Evaluation and staging
All included patients underwent a full clinical examination and baseline
laboratory tests (liver function; serum albumin, prothrombin time, total bilirubin,
aspartate transaminase, alanine transaminase). Eastern Cooperative Oncology Group
(ECOG) performance status was recorded in all patients.
Intra-arterial therapy
Experienced interventional radiologists performed all procedures. A consistent
approach according to our standard institutional protocol was used. Initially, all patients
126 Patients
(2001-2014)
88 Patients
Included for Baseline
Analysis
63 Patients
Included for Follow-up
Analysis
– Missing imaging
– Poor image quality
– Failure of registration
– Did not receive
follow-up imaging
7
underwent a diagnostic angiogram to define the hepatic arterial anatomy and to determine
portal venous patency. Patients undergoing cTACE were treated with selective (lobar or
segmental) injections. A solution containing 100 mg of cisplatin, 50 mg of doxorubicin
and 10 mg of mitomycin C in a 1:1 mixture with Lipiodol (Guerbet, France) was infused
and followed by administration of 100- to 300-µm-diameter microspheres (Embospheres,
Merit Medical, USA). Substantial arterial flow reduction to the tumor was defined as the
technical end point of the procedure. Patients undergoing DEB-TACE received
chemoembolization using LC Bead-M1 (70-150 µm), loaded with 100mg irinotecan
(BTG, UK). The beads were mixed with a non-ionic contrast media in the vial
immediately prior to use according to the instructions and delivered into the artery slowly
(in 1 ml aliquots followed by saline over an approximately 3-5 min period). Patients
treated with TARE were subjected to angiographic evaluation and, if required,
embolization of collateral arteries was performed to avoid off-target radiation injury. In
order to evaluate the degree of hepato-pulmonary shunting and to detect gastrointestinal
deposition, 5–6 mCi of 99mTC-labelled macroaggregated albumin was injected into the
hepatic artery. This shunt study preceded the treatment by at least 1 week. Depending on
the extent of the disease within the liver, patients received either unilobar or bilobar (right
and left) treatment in multiple sessions. In order to avoid liver injury, no whole liver
single session infusion was performed. The administration of Y90 microspheres
(TheraSpheres®, MDS Nordion, Ottawa, Canada) was performed in accordance with
institutional radiation safety guidelines. All patients who received cTACE or DEB-TACE
were admitted overnight. Patient who received TARE were discharged the same day of
the procedure after clinical monitoring in the recovery area.
8
MR imaging technique
All patients included in this study underwent a standardized MRI protocol before
treatment. MRI was performed on a 1.5-Tesla scanner (Siemens Magnetom Avanto,
Erlangen, Germany) using a phased array torso coil. The protocol included breath-hold
unenhanced and contrast-enhanced (0.1 mmol/kg intravenous gadopentetate; Magnevist;
Bayer, Wayne, NJ) T1-weighted three-dimensional fat-suppressed spoiled gradient-echo
imaging (repetition time ms/echo time ms, 5.77/2.77; field of view, 320–400 mm; matrix,
192×160; slice thickness, 2.5 mm; receiver bandwidth, 64 kHz; flip angle, 10°) in the
hepatic arterial phase (20 s), portal venous phase (70 s) and delayed phase (3 min).
Image Analysis
Two radiology residents with 2-3 years of experience performed tumor
radiological measurements. All measurements made by the two readers were done using
standardized electronic calipers using Digital Imaging in Communications and Medicine
(DICOM) files. Prior to the measurements, images were examined in axial, coronal and
sagittal reconstructions to visually identify the largest tumor dimension (for diameter and
enhancement, respectively). The respective slice with the largest dimension of the tumor
was then used for individual manual measurements. Native T1 images as well as triphasic
contrast-enhanced T1 images were used to visually distinguish tumor enhancement from
false-positive hyperintense T1 signal (e.g. from hemorrhage) and measurements were
performed on the portal venous phase images (33). The portal venous phase was selected
because it is the phase in which hypo-vascular liver metastases such as from lung, breast,
9
stomach, and colorectal cancer are most conspicuous . The three largest lesions were
selected for analysis. The sums for overall tumor diameter and enhancing tumor diameter
of the three largest lesions were determined.
3D quantitative image analysis was performed by research medical student MG
who had 1 year of experience with the software prototype used in the study (Medisys,
Philips Research, Suresnes, France) (27) and was verified by a radiology resident with 2
years of experience. The accuracy and reader-independent reproducibility of the
semiautomatic tumor segmentation as well as the radiological–pathological correlation of
the technique was described and verified in previous papers (34–37). First, portal venous
phase images were registered to the pre-contrast image using an affine transformation
method in the BioImage Suite software (Fig. 2a) (38). Then, whole-livers were
segmented in three-dimensions using the semi-automatic segmentation software (Fig.
2b). The total liver volume (TLV) was calculated on the basis of this segmentation. The
software performed semi-automatic 3D tumor segmentation on the portal venous phase,
contrast-enhanced MRI (Fig. 2c). The total tumor volume (TTV) was directly calculated
on the basis of this segmentation. Enhancing volumes were determined using the qEASL
calculation based on image subtraction (Fig. 2d) (27,39). In brief, the 3D segmentation
mask was transferred onto the subtraction image and a region of interest (ROI) was
placed into extratumoral liver parenchyma as a reference to calculate the relative
enhancement values within the tumor. The patient-specific, average signal intensity
within the ROI was then defined as a threshold to estimate enhancement within the 3D
mask. Subsequently, enhancing regions were expressed as a percentage of the previously
calculated overall tumor volume and visualized using a color map overlay on the portal
10
venous phase MRI scan. qEASL analysis of both the whole-liver and tumor segmentation
mask gives enhancing liver volume (ELV) and enhancing tumor volume (ETV),
respectively. ELV divided by TLV gives enhancing liver burden (ELB). ETV divided by
TLV gives enhancing tumor burden (ETB). Tumor response after IAT was determined by
calculating the change between baseline and one month follow-up imaging in the
measured parameters of the same lesions (lesion diameter for RECIST, enhancing lesion
diameter for mRECIST, enhancing liver volume for DELV, total tumor volume for
vRECIST and enhancing tumor volume for qEASL. Table 1 gives a glossary of terms
used in this study and Fig. 3 gives an overview of all anatomic and enhancement-based
methods.
11
Fig. 2 Image processing workflow. a) Portal venous phase images were registered to the
pre-contrast image using an affine transformation method in the BioImage Suite software.
b) Whole-livers were segmented in three-dimensions using semi-automatic segmentation
software. c) Another software performed semi-automatic 3D tumor segmentation. d) The
3D segmentation mask was transferred onto the subtraction image and a region of interest
(ROI) was placed into extratumoral liver parenchyma as a reference to calculate the
relative enhancement values within the tumor.
a)
b
)
c)
d
)
12
Table 1 Glossary of terms
Term
Abbreviation
Definition
Total Tumor Volume
TTV
Volume of tumor based on tumor
segmentation mask
Total Tumor Burden
TTB
TTV divided by liver volume
Enhancing Liver Volume
ELV
Volume of enhancement on whole-liver
segmentation mask
Enhancing Liver Burden
ELB
ELV divided by liver volume
Enhancing Tumor Volume
ETV
Volume of enhancement on tumor
segmentation mask
Enhancing Tumor Burden
ETB
ETV divided by liver volume
Change in Enhancing Liver
Volume
DELV
Percentage change in ELV between
baseline and follow-up image
Response Evaluation
Criteria In Solid Tumors
RECIST
Percentage change in tumor diameters
Modified RECIST
mRECIST
Percentage change in enhancing tumor
diameters
Volumetric RECIST
vRECIST
Percentage change in TTB between
baseline and follow-up image
Quantitative EASL
qEASL
Percentage change in ETV between
baseline and follow-up image
13
Fig. 3 MRI assessment techniques. a) Measurements of one-dimensional overall
diameter. b) One-dimensional measurement of enhancing tumor diameter. c) Red outline
shows liver segmentation that gives total liver volume (TLV). Subsequent qEASL
analysis gives enhancing liver volume (ELV). ELV/TLV gives enhancing liver burden
(ELB). d) Red outline shows tumor segmentation which gives total tumor volume (TTV).
Subsequent qEASL analysis gives enhancing tumor volume (ETV). ETV/TLV gives
enhancing tumor burden (ETB).
Statistical analysis
All statistical computations were performed using the commercial statistical
software SPSS (IBM, version 23.0, Armonk, NY, USA). The summary of data was
performed using descriptive statistics. Count and frequency were used for categorical
variables. Mean and range were used for continuous variables. A non-Gaussian
distribution was confirmed and a non-parametric Wilcoxon matched-pair test was used.
OS was defined from the date of the IAT session until death or last available follow-up.
14
In order to stratify patients into two groups based on baseline imaging parameters
and 3D response assessment methods, the modified Kaplan-Meier method proposed by
Contal and O’Quigley was used to determine optimal thresholds (40). In brief, this
method tests each unique value that exists for the given variable as a potential cut-off
point. For each potential cut-off point, a Kaplan-Meier analysis and a log-rank test
statistic is performed. The lowest p-value and greatest log-rank test statistic is selected as
the cut-off point.
Survival curves were estimated with the Kaplan–Meier method and plotted for
each stratifying parameter. The median OS and the 95 % confidence interval (CI) for low
tumor burden and high tumor burden were calculated for every method. The predictive
value of each radiological technique was assessed using Cox proportional hazard ratios
(HR). This was followed by a univariate and multivariate analysis, which was performed
in two steps. In the first step, a univariate Cox regression model was used to evaluate the
association of overall survival with clinical factors assessed on baseline: age, race, sex,
number of lesions, treatment type, bilirubin level, existence of extrahepatic metastases,
synchronous disease, previous surgery of primary tumor, and previous hepatic resection.
In the second step, adjusted hazard ratios for all radiological measurements were
estimated from the Cox regression model which simultaneously included the respective
radiological method as well as clinical factors that were found to be significantly
predictive of overall patient survival (p < 0.05) (41).
15
Results Part I: Baseline MR Imaging Analysis
Patient characteristics and clinical outcome
Baseline patient characteristics are summarized in Table 2. The average age of the
cohort at the time of treatment was 59.3 ± 11.4 years. Table 3 gives disease
characteristics and treatment history. A majority of patients (N=68, 77.3%) had
multifocal disease. The majority of patients (96.6%) received previous colorectal
resection, but only one patient received previous hepatic resection. The cohort is
approximately evenly split between those who received TARE (N=47, 53.4%) and those
who received TACE (N=41, 46.6%). All IATs were technically successful and no major
toxicities were reported. The mean interval between baseline imaging and IAT was 19.8
days (range, 1-60 days). Median OS of the cohort was 7.6 months (95% CI 6.1-9.0), and
by the end of the study observation date (December 1st, 2016), a total of 79 patients
(89.8%) were deceased.
16
Table 2 Baseline Patient and Tumor Characteristics
Parameter
N (%)
Demographics
Age
<65
61 (69)
≥65 years
27 (31)
Sex
Male
60 (68)
Female
28 (32)
Race
White
68 (77)
African-American
14 (16)
Other
6 (7)
ECOG Score
0
59 (67)
1
26 (30)
2
3 (3)
Bilirubin (mg/dL)
£1.2
77 (88)
>1.2
11 (12)
17
Table 3 Disease Characteristics and Treatment History
Parameter
N (%)
Number of lesions/patient
1
20 (23)
2
13 (15)
3
8 (9)
≥4
47 (53)
First IAT Received
TARE
47 (53)
TACE
41 (47)
Synchronous disease
Yes
52 (59)
No
36 (41)
Extrahepatic metastases
Yes
31 (35)
No
57 (65)
Tumor location
Bilobar
68 (77)
Unilobar
20 (23)
Previous Systemic
Chemotherapy
Yes
77 (88)
No
11 (13)
Previous Surgery of Primary
Tumor
Yes
85 (97)
No
3 (3)
Previous Hepatic Resection
Yes
1 (1)
No
87 (99)
18
Image Analysis
Liver and tumor characteristics as well as the results of 1D and 3D measurements
are summarized in Table 4. One-dimensional analysis gave a mean overall tumor
diameter of 15.6±6.8 cm and an enhancing tumor diameter of 8.9±4.1 cm. As for 3D
analysis, mean liver volume was 2165±778 cm3 (range 862-4583 cm3). Whole-liver 3D
assessment gave an ELV of 818±433 cm3 (range 104-2262 cm3) and an ELB of
38.1±16.4% (range 10.1-79.3%). Three-dimensional measurements acquired from the
tumor segmentations gave an ETV of 94.7±163 cm3 (range 0.01- 886 cm3) and an ETB of
3.6±19.4% (range 0.01-24.3%). Table 5 gives the threshold value used to stratify the
cohort into high and low burden groups for each parameter based on the modified
Kaplan-Meier method as already described.
19
Table 4 Tumor/Liver Characteristics and 1D and 3D measurements
Parameter
Liver Volume (cm3)
Mean
2165
Range
862-4583
1D Measurements
Overall Tumor Diameter (cm)
Mean
15.6
SD
6.8
Enhancing Tumor Diameter (cm)
Mean
8.9
SD
4.1
3D Measurements
Enhancing Liver Volume [ELV] (cm3)
Mean
818
SD
433
Enhancing Liver Burden [ELB]
(%)
Mean
38.1
SD
16.4
Total Tumor Volume [TTV]
(cm3)
Mean
499
SD
626
Total Tumor Burden [TTB] (%)
Mean
19
Range
0.2-99
Enhancing Tumor Volume [ETV]
(cm3)
Mean
94.7
SD
163
Enhancing Tumor Burden [ETB]
(%)
Mean
3.6
Range
0.01-24.3
20
Table 5 Optimal cutoff values for high and low tumor burden
Image Parameter
Cutoff
Overall tumor diameter
11.5 cm
Enhancing tumor diameter
8.0 cm
Total tumor volume (TTV)
335 cm3
Total tumor burden (TTB)
15%
Enhancing liver volume (ELV)
1060 cm3
Enhancing liver burden (ELB)
32%
Enhancing tumor volume (ETV)
60 cm3
Enhancing tumor burden (ETB)
3.2%
Survival Analysis
Univariate analysis of baseline clinical parameters identified a significant
correlation between the lobar distribution of disease (bilobar disease, hazard ratio [HR]
2.12 [95 % CI 1.22-3.7], p=0.01), ECOG score (ECOG >0, HR 1.79 [95% CI 1.09-2.9],
p=0.02), bilirubin level (bilirubin >1.2 mg/dL, HR 1.9 [95% CI 1.1-3.6], p=0.05), and
previous systemic chemotherapy (HR 0.48 [95 % CI 0.24-0.97], p=0.04) with OS. The
other baseline characteristics included for univariate analysis (age, race, sex, number of
lesions, treatment type, existence of extrahepatic metastases, synchronous disease,
previous surgery of primary tumor, and previous hepatic resection) did not show
significant correlation with OS.
For the diameter-based thresholds, the log-rank test demonstrated that survival
curves showed good separation when stratified both by overall tumor diameter (p=0.004)
