Genetic Alterations in Colorectal Cancer Have Different Patterns on 18F-FDG PET/CT.

Genetic Alterations in Colorectal Cancer Have

Different Patterns on

18

F-FDG PET/CT

Shang-Wen Chen, MD,*†‡ Chien-Yu Lin, MS,§|| Cheng-Man Ho, MD,

PhD,||¶ Ya-Sian Chang, PhD,||**

Shu-Fen Yang, BS,|| Chia-Hung Kao, MD,†††‡‡ and Jan-Gowth Chang,

MD†||**

C

Over the last 2 decades, multimodality treatment has resulted in great improvements for the treatment of this disease. Recently, new molecular insights and technologies have indicated the initiation, early detection, and prognostic markers of effective drug therapy for CRC. For example, accumulation of molecular alterations, including mutations in KRAS, TP53, and adenomatous polyposis coli (APC), contributes to colorectal carcinogenesis.2,3 In addition, the mutants of these

genomes have been indicated as prognostic factors or therapeutic responder in CRC.3–7

18F-FDG PET imaging is widely used for diagnosis, monitoring

treatment response, and prognostication for CRC. Despite imaging

techniques being critical in the preoperative workup for treatment decisions, a paucity of correlation studies exists between pretreatment image

findings and genetic lesion in this patient setting. A previous

study indicated that SUVmax for the primary tumor and the tumor-toliver

ratio was higher in CRC with KRAS mutations.8 To optimize the

predictive values, however, 2 questions remain unanswered. First, the individual predictive performance across various approaches of PET/ CT-related parameters, such as metabolic tumor volume (MTV) or total lesion glycolysis (TLG), remains unknown. Second, there is a research gap in matching comprehensive PET/CT parameters with various genomic alterations. By comparing various threshold methods, we conducted this study to determine the most effective approach to differentiate mutant and wild-type genomes in CRC. The results can intensify the

optimal therapeutic strategies for CRC patients by combined genetic analysis and functional image.

PATIENTS AND METHODS

Patient Population

One hundred three newly diagnosed CRC patients scheduled to

undergo curative surgical procedures at China Medical University Hospital between January 2009 and December 2012 were included in this

retrospective study (certificate number of local institutional review board, CMUH102-REC2-74). Tumor locations were the colon or sigmoid colon (59 patients) and rectum or rectosigmoid junction (44 patients). The median age was 59 years (range, 26–86 years). Sixty-six

patients were men, and 37 were women. All patients received PET/ CT for pretreatment staging and underwent primary tumor resection

thereafter (median, 7 days; range, 1–28 days). No patient received preoperative chemotherapy or had a history of diabetes. All patients had a

normal serumglucose level before obtaining PET/CTimages. The characteristics of the 103 patients are shown in Table 1.

PET/CT Image Acquisition

All patients fasted for at least 4 hours before 18F-FDG PET/CT

imaging. The images were captured using a PET/CT scanner (16-slice PET/CT, Discovery STE; GE Medical System, Milwaukee, Wis) approximately 60 minutes after administering 370 MBq of 18F-FDG. Patients

were requested to rest during the uptake period. The FDG PET

data were inputted into the workstation, and the images were reviewed to localize the target lesions, as confirmed by 2 nuclear medicine physicians. The physicians were unaware of the information of the preoperative

images. The PET/CT workstation provided a quantification

of FDG uptake for SUV. Nuclear medicine physicians identified the locations of SUVmax and

the values for the primary tumors. This procedure was detailed in our previous report.9

Measurement of MTV and TLG

The PET-based MTV were measured from attenuation-corrected FDG PET images, using an SUV-based automated contouring program (Advantage Workstation Volume Share version 2; GE Health). The MTVs were measured from attenuation-corrected FDG PET images. TheMTV was defined as the sum of metabolic volumes of the primary

tumors. The volume boundaries were drawn sufficiently widely to incorporate each target lesion in the axial, coronal, and sagittal FDG

PET images. To define the contouring margins around the tumor, we used SUVmax of 2.5 (MTV2.5) and SUVmax of 3.0 (MTV3.0), as previously

reported. 9 Furthermore, volumes above 20% of SUVmax

(MTV20%), 30% of SUVmax (MTV30%), 40% of SUVmax (MTV40%),

and 50% of SUVmax (MTV50%) were analyzed. The contour around

the target lesions within the boundaries was automatically produced, and the voxels presenting an SUV intensity of SUVmax 2.5 or 3.0 or

greater and 20%, 30%, 40%, or 50% or greater of SUVmax within the

contouring margin were incorporated to define the tumor volumes. TheMTVs for the primary tumors included adjacent lymph nodes with small volumes. The small lymph nodes adjacent to the primary tumor cannot be segmented from the primary tumor by PET/CT when they appear similar to a part of the primary lesion. However, large lymph

nodes neighboring the primary tumor can be segmented using an automatic volume of interest tool on PET/CT, even if they are partially contiguous to the primary tumor.

The TLGs were also measured from attenuation-corrected FDG

PET images, using an SUV-based automated contouring program (Advantage Workstation Volume Share version 2; GE Health). The TLG

was calculated according to the following formula: TLG = SUVmean _

MTV.10,11 We used the same threshold levels as the MTVs, namely,

TLG2.5, TLG3.0, TLG20%, TLG30%, TLG40%, and TLG50%.

Measurement of PET-Based Maximal Tumor Width

The PET-based maximal tumor width (TW) was also measured

from attenuation-corrected FDG PET images, using an SUV-based automated contouring program. To define the maximal TWaround the tumor,

we used the measured distance above 30% of SUVmax (TW30%)

and 40% of SUVmax (TW40%). Using this tool, the calculated unit for

the analyses was millimeter. The details were described previously.12

Genetic Mutation Analysis

A routine pathology preparation was performed after tumor resection, and the tumor and normal tissues were preserved in tissue bank at China Medical University Hospital.

We used high-resolution melting (HRM) methods for genetic

mutation analysis. A good amplicon design is essential for obtaining robust and reproducible HRManalysis. The difference between wild-type

and heterozygote curves becomes smaller and more difficult to differentiate when the product length increases. All the amplicons have been

suggested to be designed smaller than 300 base pairs.13 In the present

study, the primers for HRM analysis were selected using Primer3 software. Supplemental Table 1, http://links.lww.com/CNM/A7, summarizes

primer sequences and general information for each gene. Polymerase chain reaction (PCR) was carried out in duplicate in a volume of 10 μL with LightCycler 480 High-Resolution Melting

Master (reference 04909631001, Roche Diagnostics), 1_ buffer containing Taq polymerase, nucleotides, the dye ResoLight 10 ng DNA,

0.3 μM primers, and 2.5 mM MgCl2, for amplification of APC gene.

In additional, the analyses of TP53, PIK3CA, KRAS, and BRAF genes were carried out in a volume of 15 μL with Type-it HRM PCR Kit (Qiagen, Hilden, Germany), 1_ HRM PCR master mix containing HotStar Taq Plus DNA polymerase, Type-it HRM PCR buffer, Qsolution, dNTP and EVA green dye, and 15 ng DNA, 0.66 μM of each

TP53 and BRAF genes primer, 0.67 μM of each PIK3CA and KRAS genes primer. High-resolution melting assays were conducted with commercial software (LightCycler 480 Gene Scanning Software Version 1.5; Roche Diagnostics). With SYBR Green I filter (533 nm),

PCR program consisted of an initial denaturation-activation step at 95°C for 10 minutes, followed by a 45-cycle program for detecting APC gene (denaturation at 95°C for 10 seconds, annealing at 60°C

10 seconds, and elongation at 72°C for 25 seconds for reading of fluorescence; acquisition mode, single). The 40-cycle program for detecting

TP53 gene was with the following settings: denaturation at 95°C for 10 seconds, annealing at 63°C for 35 seconds, and elongation at 72°C for 10 seconds for reading of fluorescence; acquisition mode, single. The 40-cycle program for detecting PIK3CA and KRAS genes was with the following settings: denaturation at 95°C for 10 seconds, annealing at 60°C for 30 seconds, and elongation at 72°C for 10 seconds for reading of fluorescence; acquisition mode, single. The 40-cycle program

for detecting BRAF gene was with the following settings: denaturation at 95°C for 10 seconds, annealing at 56°C for 30 seconds, and elongation

at 72°C for 20 seconds for reading of fluorescence; acquisition mode, single. The melting program included denaturing at 95°C for

1 minute, annealing at 40°C for 1 minute, and subsequent melting that consists of a continuous fluorescent reading of fluorescence from 55°C to 90°C at the rate of 25 acquisitions per degree celsius. Shapes of difference plot curves of each DNA duplicate sample must be reproducible

both in shape and peak height. To confirm results of HRM

analysis, a sequencing analysis was also performed in all the amplicons containing abnormal melting curve, and some represent amplicons with normal melting curve.

Statistical Analysis

All values are expressed as mean (SD). Correlations between genomic alterations or PET/CT-related parameters were examined using

Pearson correlation, with the α level set at 0.01. Differences in SUVmax

or various thresholds of PET/CT-related parameters between mutated and wild-type genome were tested using theMann-Whitney U test. The analyseswere tested by receiver operating characteristic (ROC) curve analysis

to compare the predictive ability. In addition, the predictive values of PET-related parameters for the genomic expression were examined using multivariate logistic regression analysis. All analyses were 2-sided, with P values less than 0.05 considered statistically significant. Statistical analyses were performed using SPSS, version 13.0 (SPSS Inc, Chicago, Ill).

RESULTS

Frequency of Genetic Alterations

As shown in Table 1, genetic alterations in TP53 (exons 2–11), KRAS (codons 12 and 13), and APC were identified in 41 (40%), 34 (33%), and 27 (26%) of the patients, respectively. Tumors in 29 patients had mutations in KRAS codon 12 whereas 5 mutants in KRAS codon 13. The mutants in PIK3CA and BRAF were 5 (5%) and 4 (4%), respectively. Thirteen patients (13%) had both TP53 and KRAS mutations.

Our data showed that there was no obvious correlation between

the 2 genetic alterations (P = 0.82, γ = 0.23). Similarly, therewas no association between TP53 and APC (P = 0.91, γ = 0.011), or KRAS and

APC (P = 0.33, γ = 0.098). Because of the rarity of genetic alterations in PIK3CA and BRAF, these genes were excluded in this analysis.

Predictive Value of PET/CT-Related Parameters

for the TP53 Mutation

The Mann-Whitney U test showed that CRC tumors with

mutated TP53 were associated with a trend of increased SUVmax and

serum levels of carcinoembryonic antigen (CEA) compared with wildtype (Supplemental Table 2, http://links.lww.com/CNM/A8). The ROC

curves showed that the SUVmax, MTV3.0, TLG40%, and TW40% predicted

the TP53 mutation most accurately. Thus, they were chosen as

regression analysis disclosed that SUVmax was the only predictor

for TP53 mutations with the odds ratio (OR) of 1.28 (P = 0.04; 95% confidence interval [CI], 1.01-1.61). The quantitative difference of the SUVmax between the 2 groups is illustrated in Figure 1. Themean

(SD) SUVmax for patient with and without TP53 mutation was 11.0

(5.9) and 9.4 (5.1), respectively.

Predictive Value of PET/CT-Related Parameters

for the KRAS Mutation

Several threshold methods, such as SUVmax, TLGs, and TWs,

showed a significant trend of higher accumulation of FDG uptake in CRC tumors with mutated KRAS compared with wild-type (Supplemental Table 3, http://links.lww.com/CNM/A9). After using ROC

curves to compare similar parameters with each other, SUVmax, MTV3.0,

TLG40%, and TW40% predicted the KRAS mutation most accurately

with the areas under the curve of 0.66 ± 0.06, 0.64 ± 0.06, 0.66 ± 0.06, and 0.68 ± 0.05, respectively. As summarized in Table 3, TW40% was found to be an independent parameter for predicting

KRAS mutations. The OR was 1.15 (P = 0.001; 95% CI, 1.06–1.24). The mean (SD) value of TW40% for patient with and without mutation

was 3.0 (1.0) cm and 2.4 (0.8) cm, respectively. Figure 2 depicts

the quantitative difference of TW40% between the 2 groups. When analyzing

tumors harboring codon 12 mutation only (n = 29), TW40%

remained the predictor for KRAS mutant with the OR of 1.09 (P = 0.02; 95% CI, 1.01–1.21). Our data showed no close association

between TW40% and SUVmax (P = 0.73, γ = 0.34).

Predictive Value of PET/CT-Related Parameters

for the APC Mutation

As listed in Table 4, CRC tumors with APC mutation had a trend with an increased TLG40% in the univariate analysis (P = 0.02), but not

in the multivariate analysis. The mean (SD) TLG40% value in patients

with and without mutation was 227.4 (204.6) g and 144.8 (134.1) g, respectively (Fig. 3). No other parameters were associated APC mutation

(Supplemental Table 4, http://links.lww.com/CNM/A10).

Difference in FDG Accumulation in the Cases of

Combined TP53 and KRAS Mutations

According to our data, patients with combined TP53 and KRAS mutations or nodal metastasis were 2 prognosticators for inferior survival when compared with those without the 2 adverse features.

Thus, further analysis was carried out to examine the association between the 2 combined mutants and PET/CT parameters. Using the

Mann-Whitney U test, SUVmax was the only predictor for tumors with

combined TP53 and KRAS mutations. The mean (SD) SUVmax for patient with and without

this combined mutants was 11.7 (4.3) and 9.8

(5.6) (P = 0.03), respectively. Tumors with pathological stage T3 to T4 were found to be the only factor associated with nodal metastasis (P = 0.001), whereas the PET/CT-related parameters or genetic alterations did not relate to positive nodal status.

Accuracy in Predicting Genetic Alteration

Based on the parameters for TP53 and KRAS mutations mentioned previously, we sought to determine the optimal cutoff to distinguish between the mutants and wild-types. When using the

optimal cutoff value of SUVmax at 10, the sensitivity, specificity,

and accuracy for predicting the TP53 mutation were 49%, 68%, and 60%, respectively. If applying the median value of TW40% (2.6

cm) as a cutoff for KRAS mutation, the sensitivity, specificity, and accuracy were 62%, 61%, and 61%, respectively.

DISCUSSION

Although interest is developing in the role of FDG PET in staging or monitoring response in CRC, FDG PET has rarely been investigated in the correlation of genetic alteration. A unique advantage of FDG PET scanning is the ability to use the quantitative information

of the glucose uptake within the tumor or to automatically create a contour around the tumor. Compared with CTor MRI, this autocontouring

process substantially reduces the interobserver variability in the interpretation of images.14–16 Mechanisms affecting FDG accumulation in

cancer tissues are complex.17,18 In CRC, some studies have suggested

that glucose transporter 1–mediated FDG accumulation is more essential than hexokinase type 2 activity.8,19 Using a large sample size and

various threshold methods, we first demonstrated the differences across various PET/CT-related parameters in predicting 3 major genomic expressions in CRCs.

Oncogenic activation of KRAS can influence several cellular processes that regulate biological course.20 KRAS mutation, which occurs

in approximately 40% of CRCs, is particularly crucial because it can predict a lack of responses to therapies with antibodies targeted to the epidermal growth factor receptor.4,5 Two studies indicated that

anti-epidermal growth factor receptor monoclonal antibodies have significant efficacy in the treatment of metastatic CRC patients with wildtype

KRAS tumors.21,22 In addition, the American Society of Clinical

Oncology suggested that patients with metastatic CRC, having a KRAS mutation in codon 12 or 13, should not receive anti-epidermal growth factor receptor antibody treatment.23 In this study, we found that several

methods measuring FDG accumulation were higher in mutated KRAS. However, the importance of our findings has 2 aspects. First, we disclosed that TW40% method can achieve higher accuracy through the

comprehensive comparison with the other threshold approaches. Second, we first highlighted CRC tumorswith codon 12 rather than codon 13 mutation is the key genetic change correlatedwith an increased FDG uptake.

In contrast, Kawada et al8 conducted a pilot predictive study by using

SUVmax and tumor-to-liver ratio in a cohort of 51 CRC patients, and they

showed that SUVmax had an OR of 1.17 with an accuracy of 75% in forecasting

mutated KRAS when using a cutoff of 13.

Our study showed that TP53 and KRAS mutations rarely

coexisted in the same tumors. This agreed that mutations in 2 genes are on separate pathways in CRC tumorigenesis.3 TP53 plays a crucial

role in maintaining genome stability and integrity. When its function is jeopardized, uncontrolled cell proliferation might be facilitated because of genomic instability. Although some studies have reported association of somatic mutations in TP53 with poor survival or lack of response to therapy, the clinical significance of TP53 status still remains controversial.

23,24 To date, there was a paucity of human data reporting the association

of genetic alteration and FDG PET finding. A cell line study

have reported that TP53 mutant cells had 1.5- to 2-fold higher FDG uptake than wild-type TP53 cells in basal condition, and the difference of

FDG uptake was greater after 188Re treatment.25 Nonetheless, our data

showed that mutant TP53 tumors did not show a consistent increased FDG accumulation across several thresholds as the cases of KRAS. The 2 parameters distinguishing thewild-type andmutantwere SUVmax

and serum level of CEA. Despite that there was no significant correlation between TP53 mutation and increased Tor N stage, the effect of tumor burden could not be entirely excluded unless studies with a

controlled tumor size had been done. Given that there was no association between the SUVmax and the TW40%, the 2 methods represent different

single point; theoretically, the TWmethods may reflect the real macroscopic tumor burden more precisely because they stand for a range of

TWby using a fixed threshold. Because of the low sensitivity and accuracy in predicting TP53 mutation, further studies are necessary to validate our findings.

Mutations in the APC gene are responsible for familial adenomatous polyposis and the majority of sporadic CRC.26 The APC gene mutation

rate is lower in Taiwan than in other countries.7 APC mutation

alone was less reported as a poor prognostic factor in CRC compared with KRAS or TP53. Our study first showed that tumors with APC mutation had a trendwith an increased TLG40%; however, there is a need to

conduct further studies to confirm our results because of the lack of consistency across the other PET/CT parameters.

This research should be interpreted with several concerns. First, CRC tumors with mutated KRAS or TP53 showed only 1.15- to 1.28-fold increase in FDG uptakes with low accuracy. Together with the results from the study of Kawada et al,8 currently, FDG PET/CT is not

sufficient to replace the mutational testing. Second, PET/CTmay represent the gross status of the tumors. Heterogeneity of the mutations

within a CRC tumor has been reported.27 As a result, the correlation

study might be biased because dissected specimens for mutational testing may not reflect the exact macroscopic status of the entire tumor. Finally, some bias might existed in patient selection or genomic analysis

because of the retrospective study. Particularly, the difference of tumor biology between rectal and colon tumors should be clarified.28 Nonetheless,

based on the different patterns of FDG accumulation, our study

implied that the specific changes in the glucose uptake pathway might be responsive for special genetic alterations. To optimize the therapeutic effect of PET/CT in CRC, future studies should include more participants prospectively and use standardized protocols for FDG PETacquisition and correction of the partial volume effect or false-positive PET

findings.29 Furthermore, as described in a recent study,30 it is more interesting

to explore the underlying mechanisms by which mutated genomes increased 18F-FDG accumulation.

CONCLUSIONS

PET/CT-related parameters in CRC tumors can be used for

supplementing genetic analysis to correlate genetic alterations. The mutated KRAS tumors are associated with higher TW40%, whereas tumors

with mutated TP53 have an increased SUVmax. However, further studies