PET-CT Fusion in Radiation Management of Patients with Anorectal Tumors

Int. J. Radiation Oncology Biol. Phys., Vol. 69, No. 1, pp. 155–162, 2007Copyright � 2007 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/07/$–see front matter

doi:10.1016/j.ijrobp.2007.02.055

CLINICAL INVESTIGATION Rectum

PET-CT FUSION IN RADIATION MANAGEMENT OF PATIENTS WITHANORECTAL TUMORS

CYNTHIA ANDERSON, M.D.,* MARY KOSHY, M.D.,* CHARLES STALEY, M.D.,y NATIA ESIASHVILI, M.D.,*

SHARAM GHAVIDEL, M.S.,* ZACH FOWLER, M.D.,* TIM FOX, PH.D.,* FABIO ESTEVES, M.D.,z

JEROME LANDRY, M.D.,* AND KAREN GODETTE, M.D.*

Departments of *Radiation Oncology and ySurgery and zDivision of Nuclear Medicine, Department of Radiology,Emory University School of Medicine, Atlanta, GA

Purpose: To compare computed tomography (CT) with positron emission tomography–CT (PET-CT) scans withrespect to anorectal tumor volumes, correlation in overlap, and influence on radiation treatment fields and patientcare.Patients and Methods: From March to November 2003, 20 patients with rectal cancer and 3 patients with anal can-cer were treated with preoperative or definitive chemoradiation, respectively. Computed tomography simulationdata generated a CT gross tumor volume (CT-GTV) and CT planning target volume (CT-PTV) and 18F-fluoro-2-deoxy-glucose PET (FDG-PET) created a PET-GTV and PET-PTV. The PET-CT and CT images were fusedusing manual coregistration. Patients were treated with three-dimensional conformal therapy to traditional doses.The PET, CT, and overlap volumes (OVs) were measured in cubic centimeters.Results: Mean PET-GTV was smaller than the mean CT-GTV (91.7 vs. 99.6 cm3). The mean OV was 46.7%. Astumor volume increased, PET and CT OV correlated significantly ( p < 0.001). In 17% of patients PET-CT alteredthe PTV, and in 26% it changed the radiation treatment plan. For 25% of patients with rectal cancer, PET detecteddistant metastases and changed overall management. Ten rectal cancer patients underwent surgery. When thepretreatment PET standardized uptake value was >10 and the posttreatment PET standardized uptake valuewas <6, 100% achieved pathologic downstaging ( p = 0.047).Conclusions: Variation in volume was significant, with 17% and 26% of patients requiring a change in treatmentfields and patient management, respectively. Positron emission tomography can change the management foranorectal tumors by early detection of metastatic disease or disease outside standard radiation fields. � 2007Elsevier Inc.

Anorectal cancer, PET-CT, Radiotherapy.

INTRODUCTION

Positron emission tomography–computed tomography

(PET-CT) fusion is gaining acceptance for use in radiation

management of patients with primary tumors at several sites,

including head and neck, lung, and lymphoid neoplasms

(1–6). For patients with carcinoma of the rectum, PET-CT

is frequently used to diagnose and stage tumors, evaluate

for treatment response, and assess for tumor recurrence.

The utility of PET-CT over PET alone includes better ana-

tomic correlation of areas that show high avidity for 18F-flu-

oro-2-deoxy-glucose (FDG). Cohade et al. (7) determined in

a retrospective study of 45 patients with colorectal cancer that

PET-CT improved the staging and restaging accuracy from

78% to 89% compared with PET alone. In another study,

155

PET-CT was found to alter the staging in 20% of patients

and locate findings not seen in 17% of patients when evalu-

ated by PET alone (8).

Two recent articles have described the benefits of PET-CT

in the treatment of anal carcinoma. In the first, PET detected

nodal metastases in 24% of patients who were considered

node negative on CT scan (9). Cotter et al. (10) reported

the ability of PET imaging to detect correctly 91% of primary

lesions, compared with 59% by CT scan. Twenty-three

patients with clinically negative lymph nodes and 17% of

patients with normal CT scan results were considered node

positive on PET scan.

Although several studies have demonstrated the utility of

PET scans for the detection of colorectal primary tumors,

data on its impact on radiation treatment fields for rectal and

Reprint requests to: Jerome Landry, M.D., Department of Radia-tion Oncology, Emory University Hospital, 1365 Clifton Road NE,Atlanta, GA 30322. Tel: (404) 778-3473; Fax: (404) 778-4139;E-mail: [email protected]

Presented orally at the Annual Meeting of the Radiological Societyof North America, November 28–December 3, 2004, Chicago, IL.

Conflict of interest: none.Received May 11, 2006, and in revised form Feb 9, 2007.

Accepted for publication Feb 14, 2007.

mailto:[email protected]

156 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 1, 2007

anal tumors continue to emerge. Ciernik et al. (11) evaluated

the value of PET-CT on radiation planning for patients with

tumors at several sites, including carcinoma of the rectum

(6 patients) and carcinoma of the anus (7 patients). In more

than half of all patients (56%), the gross tumor volume

(GTV) delineation was significantly altered. Gross tumor vol-

ume increases were seen in 3 of 6 patients with rectal prima-

ries, with a mean GTV increase of 50% and a planning target

volume (PTV) increase of 20%. In patients with tumors of the

anal canal, PET-CT influenced the boost volumes but did not

alter the initial standard fields. The mean change of GTV was

34%, with 3 patients experiencing an increase and 1 patient

a decrease in volumes. The mean change to PTV was 18%.

In a study by Thomas et al. (12), PET imaging changed the

overall management of 51% of patients with advanced and

metastatic colorectal cancer. Positron emission tomography

provided additional tumor information in 70% of patients, up-

staging disease in 66% of patients, and downstaging disease in

30%.

We analyzed the clinical impact of PET-CT fusion on ra-

diotherapy planning and management of patients with rectal

and anal primaries. Specifically, we compared the size of tu-

mor volumes on CT with those on PET scans, the correlation

in overlap between both sets of images, and the influence of

PET-CT on the radiation treatment fields and overall man-

agement of patient care.

PATIENTS AND METHODS

The Human Investigations Committee (HIC) at our institution ap-

proved a protocol in 2001 that allowed for prospective and retro-

spective review of all gastrointestinal malignancies in which

radiation was used as part of the overall management. The cohort

of patients in this study was covered as part of the HIC approval.

In addition, all patients provided consent for treatment with chemo-

therapy, radiotherapy, and surgery.

We collected data on all patients with biopsy-proven, intact pri-

mary anorectal tumors treated at a county and private hospital

from March to November 2003. Twenty-three patients underwent

PET-CT fusion scans as part of treatment planning. Twenty patients

(87%) were diagnosed with rectal cancer and were treated with pre-

operative chemoradiotherapy, whereas 3 patients (13%) received

definitive chemoradiation for tumors of the anus.

PET-CT scan and CT simulation proceduresPatients abstained from caloric intake for a minimum of 4 h before

administration of FDG and were required to have a serum glucose

level <200 mg/dL. Ten to fifteen millicuries (370–555 MBq) of

FDG were injected i.v. with 45–60 min allowed for the uptake

phase. Patients with rectal cancer were scanned and treated in the

prone position on a belly board to allow for anterior displacement

of small bowel from the treatment field. Patients with anal cancer

were scanned and treated supine.

Imaging and data acquisition were performed on either an inte-

grated big-bore PET-CT system (Discovery LS; GE Medical Sys-

tems, GE Healthcare, Chalfont St. Giles, United Kingdom) at the

county hospital or a separate PET-CT (DST; GE Medical Systems)

and CT simulation system (Discovery LS; GE Medical Systems)

at the private hospital. On the integrated system, patients were in

the treatment position for simulation on the dedicated integrated

PET-CT system, with automatic fusion of the PET-CT images.

Thirty minutes before scanning, patients with rectal cancer drank

450 mL of barium sulfate suspension (Readi-Cat2; E-Z-EM, West-

bury, NY) for small bowel opacification. Rectal contrast was not used

because of its ability to cause image artifact during interpretation

of the PET-CT images. Computed tomography data were acquired

first during shallow breathing, whereas PET data were acquired

during free breathing and over multiple breathing cycles. Computed

tomography scans were performed without contrast, and parameters

included tube rotation time of 0.5 s per revolution, 140 kV, 180 mA,

with a reconstructed slice thickness of 5 mm. Scanning length was

867 mm, and acquisition time was less than 1 min for the entire axial

field of view.

After CT data acquisition, the table top—with the patient in the

same position—was advanced into the PET scanner, and acquisition

of PET data began at the head. Data were acquired from four to five

table positions for 5 min each. For attenuation correction of the PET

images, CT sections were reduced to PET resolution by smoothing

with a Gaussian filter (8-mm full width, half maximum), and atten-

uation correction slices were reconstructed using an iterative recon-

struction. Images were viewed on a Xeleris workstation (GE

Medical Systems). The PET and CT data sets were then transferred

to a viewing station using Digital Imaging and Communications in

Medicine software.

At the private hospital, all patients received a planning CT simu-

lation scan followed by a PET-CT scan in the same position used dur-

ing the CT simulation. Patients with rectal cancer drank 450 mL of

barium sulfate suspension 20–30 min before both PET-CT scan

and CT simulation. Rectal contrast was not used for CT simulation

to mimic the conditions of the PET-CT scan. The PET-CT was fused

to CT simulation scan images by manual coregistration. A minimum

of three matched anatomic landmarks were selected, with an average

acceptable error of 2 mm. The final data set was sent to an ECLIPSE

treatment-planning system for radiation planning.

PET scan interpretationThe standardized uptake value (SUV) was used as a semiquantita-

tive tool to supplement visual interpretation. Standardized uptake

values were determined by manually drawing a region of interest

over the suspected lesion in the transaxial plane. The SUV in each

pixel was calculated according to the following formula: SUV =

activity concentration/(injected dose/body weight). The maximum

SUV within a region of interest was used to minimize finite resolu-

tion effects. Any suspected lesion with a maximum SUV $2.5

was considered malignant. Histograms of FDG uptake intensity

were not generated.

Radiation and chemotherapy treatment and impact ofPET-CT on management

All patients were treated with concurrent chemotherapy and ra-

diotherapy using three-dimensional conformal therapy. All grossly

visible tumors were contoured prospectively as the gross tumor vol-

ume (GTV). Results of initial diagnostic scans and colonoscopy

were used at the time of contouring. The CT-GTV included the en-

tire circumference of the rectum at the level of the visible abnormal-

ity. The PET-GTV was contoured initially using the coregistered

PET-CT images and was not visible when the CT-GTV was con-

toured so as to minimize bias. The PET-GTV was defined on the

basis of the visual interpretation of the radiation oncologist, and

all contours were analyzed by one radiation oncologist (J.L.).

The PET-GTV and CT-GTV were measured in cubic centimeters.

The overlap volume (OV) was contoured as a separate volume and

PET-CT fusion for anorectal tumors d C. ANDERSON et al. 157

similarly measured. Percentage of overlap was calculated by divid-

ing the OV by the CT-GTV. The CT volume was chosen to deter-

mine the overlap percentage because it is ultimately the data set

that is used for all radiation treatment planning. The clinical target

volume (CTV) was created to account for microscopic disease and

typically was defined as the GTV plus a 2-cm margin. No additional

margin was added to generate the planning target volume.

Initial radiation fields for patients with rectal cancer were de-

signed for a four-field box technique with adjustments made to field

size when necessary according to PET data. On the anterior and pos-

terior fields, the superior border was L5/S1, the inferior border was

the bottom of the obturator foramen with a minimum of a 2-cm mar-

gin on the most distal aspect of the CTV, and the lateral borders were

placed 2 cm distal to the pelvic inlet. For the lateral fields, the ante-

rior margin was placed posterior to the pubic symphysis for patients

with T3 disease and anterior to the symphysis for T4 disease. The

initial field was treated to 45 Gy with a reduced field boost that en-

compassed the CTV alone to a mean total dose of 50.4 Gy (range,

50.4–59.4 Gy). Chemotherapy was delivered concurrently and con-

sisted of continuous infusion 5-fluorouracil at 225 mg/m2 during

radiation. One month after surgery the patients received adjuvant

5-fluorouracil and leucovorin for 6 cycles. Patients were scheduled

for follow-up PET-CT scanning to assess tumor response within

2 weeks of completing chemoradiation. Surgery was planned for

1 to 2 weeks thereafter.

Patients with cancer of the anal canal were treated in an anterior–

posterior fashion, with the initial field being wider anteriorly than

posteriorly. The initial superior border was placed at the junction

of L5/S1, and the inferior border was 3 cm below the anal tumor.

Lateral borders covered the lateral inguinal lymph nodes anteriorly

and the pelvic inlet posteriorly. Adjustments to the borders were

made according to PET data. At 30.6 Gy, the superior border

was lowered to the bottom of the sacroiliac joints for both fields,

and the reduced field received a boost dose of 14.4 Gy. Finally,

the GTV plus a 2-cm margin received a 5.4–14.4-Gy boost to

give a total dose of 50.4–59.4 Gy according to the stage of disease.

The medial and lateral inguinal nodes were boosted with electron

fields to deliver 45 Gy in clinically N0 patients and 55.8–59.4

Gy to patients with involved lymph nodes. Patients received 5-fluo-

rouracil at 1000 mg/m2 � 96 h along with mitomycin-C at 10 mg/

m2 � 1 during the first and fourth week of radiation. Follow-up

scans were obtained between 1 and 3 months after completion of

treatment.

Data analysisMean GTV and OVs from the CT and PET were calculated. The

impact of PET on formation of GTV and PTV and its ability to

correspond to OV were analyzed. Additional analyses included

changes in treatment strategy based on PET data and correlation

of posttreatment PET imaging with pathologic response when

appropriate.

Data were analyzed with JMP Statistical Discovery Software

5.0.1a (SAS Institute, Cary, NC). Results were calculated by analy-

sis of variance and root mean square methods.

RESULTS

Patient demographicsTwenty patients (87%) were diagnosed with rectal cancer,

and 3 patients (13%) were treated for tumors of the anus.

Table 1 shows demographic information on all patients.

Thirteen patients (56.5%) were male, and mean age was 58

years (range, 30–81 years). Patients were staged clinically ac-

cording to the American Joint Committee on Cancer Staging

Manual, 6th edition, on the basis of physical examination, co-

lonoscopy report, pathologic reports from the biopsy speci-

men, and pretreatment imaging, including CT scan of the

chest, abdomen, and pelvis, PET-CT scan, rectal ultrasound,

and chest X-ray. Staging CT scans and serum carcinoem-

bryonic antigen levels were measured before PET-CT imag-

ing. Lymph nodes were considered positive if they were

palpable on clinical examination, measured greater than 1.5

cm on CT scan, were enlarged on transrectal ultrasound, or

had an SUV >2.5 on PET-CT.

Of the patients with rectal primaries, 12 (60%) were staged

T3N0 (IIA); 2 (10%) were T4N0 (IIB); 3 (15%) were

T3NxM1 (IV), and T3N1 (IIIB), T2N2 (IIIC), and TxNxM1

(IV) were each represented with 1 patient (5% each). The 4

patients with metastatic disease had suspicious lesions de-

tected on PET-CT in the liver, lung, para-aortic lymph nodes,

and descending colon, respectively. Each patient subse-

quently underwent a dedicated spiral high-resolution CT

scan of the abdomen or lung to further evaluate the abnormal

areas. Radiographically suspicious lesions were presumed to

represent metastases and were not biopsied. Two patients

(67%) with cancer of the anus were staged as T3N0 (II),

and the third (33%) was staged T3N2 (IIIB). Seventeen pa-

tients (74%) were treated at a county hospital and 6 (26%)

at a private institution.

Observed volumesPositron emission tomography volumes, on average, were

smaller than CT volumes (Table 2). The mean PET-GTV was

91.7 cm3 (median, 37.1 cm3; range, 2.9–859 cm3), and the

mean CT-GTV was 99.6 cm3 (median, 66 cm3; range,

>17–570 cm3). The mean OV was 46.7% and ranged from

11% to 99% (Fig. 1).

Table 1. Patient demographics

GenderMale 13 (56.5)Female 10 (43.5)

Age (y)Range 30–81Mean (y) 58

Tumor locationRectum 20 (87)Anus 3 (13)

Stage by locationRectum

T3N0 (IIA) 12 (60)T4N0 (IIB) 2 (10)T3N1 (IIIB) 1 (5)T2N2 (IIIC) 1 (5)T3NxM1 (IV) 3 (15)TxNxM1 (IV) 1 (5)

AnusT3N0 (II) 2 (67)T3N2 (III) 1 (13)

Values are number (percentage) unless otherwise noted.


Table 2. PET volume, CT volume, overlap volume, and overlap percentage

Patient no. Site Stage CT volume (cm3) PET volume (cm3) Overlap volume (cm3) Overlap (% of CT volume)

1 Rectum TxNxM1 62 48 36 582 Rectum T3NxM1 105.8 31.2 25.7 243 Rectum T3N0Mx 29.6 18.3 10 344 Rectum T3N0Mx 82 23 16 205 Rectum T3N0M1 47 50 26 556 Rectum T3N0Mx 62 49 32 527 Rectum T3N0Mx 121 141 84 698 Rectum T3N0Mx 22.4 2.9 15.1 679 Rectum T3N0Mx 56.9 37.1 21.9 38

10 Rectum T3N0Mx 24 25 3 1311 Rectum T3N0Mx 119 139 84 7112 Rectum T3N0Mx 117 77 60 5113 Rectum T3N1Mx 32 4 4 1114 Rectum T2N2Mx 68 27 23 3415 Rectum T3N0Mx 66 34.7 27.6 4216 Rectum T3N0Mx 170 169 49 2917 Rectum T4NxMx 156 119 70 4518 Rectum T3N0Mx 37 32 21 5719 Rectum T4N0Mx 197.1 127.7 92 4720 Rectum T3N0M1 50 26 23 4621 Anal canal T3N0Mx 570 859 562 9922 Anal canal T3N0Mx 17 16 10 5923 Anal canal T3N2Mx 79 52 43 55

Abbreviations: PET = positron emission tomography; CT = computed tomography.

Correlation between CT and PET volumes andoverlap volumes

The majority of patients had PET or CT volumes between

0 and 200 cm3, but initial analysis revealed the presence of

a single patient with rectal cancer whose PET volume was

859 cm3 and CT volume was 570 cm3. To account for the in-

fluence of this patient on the results, statistical analyses were

performed both including and excluding his data.

As tumor volume increased on CT or PET imaging, so too

did OV in a statistically significant fashion ( p < 0.001). This

result was observed and remained significant if the outlier pa-

tient was excluded ( p < 0.001). In addition, as the size of the

tumor increased, there was a significant correlation between

the PET and CT volumes (R2 = 0.94, p < 0.0001) that was still

present without the outlier (R2 = 0.75, p < 0.0001).

The overlap percentage was calculated as a percentage of

OV to CT tumor volume and was intended to determine

how much of the OV corresponded to the PET vs. CT volume

(Fig. 1). As both PET and CT volumes increased, there was

Overlap Percentage of CT Volume

0%

20%

40%

60%

80%

100%

120%

1 3 5 7 9 11 13 15 17 19 21 23Patient

% O

verla

p

Fig. 1. Overlap percentage of computed tomography volume.

a statistically significant correlation between the PET volume

and overlap percentage or CT volume and overlap percentage.

However, the statistical significance of these findings was not

maintained once the outlier patient was removed from analy-

sis. This suggests that when tumor volumes are small on either

imaging modality (between 0 and 200 cm3), there may be

wide variation between the locations of the primary tumor

on PET as compared with CT, but as volumes become larger,

the degree of correlation becomes more pronounced.

Figure 2 illustrates the radiation portals of a patient with

T3N0 adenocarcinoma of the rectum whose planning fields

underwent significant change as a result of PET data.

Colonoscopy revealed a mass that originated 4 cm from

the anal verge; the colonoscope could only be advanced

to 7 cm, thus the superior aspect was not visualized. The

initial PET scan demonstrated an SUV uptake of 16 within

the lesion. The posttreatment PET-CT scan showed a small

focus in the inferior rectum with an SUV of 3.3. Unfortu-

nately, the patient did not present for surgery to allow for

pathologic correlation, but the dramatic change in SUV

supports tumor instead of inflammation in the area of

FDG avidity.

Impact on radiation treatment planningIn 4 of 23 patients (17%; one anal canal tumor and three

rectal tumors), integration of the PET volume with the plan-

ning volumes resulted in a change in the PTV. Twenty-six

percent of patients (6 of 23) experienced a change in the

radiation treatment-planning process. Changes included

increasing field sizes because traditional fields would have

cut through a contoured PET tumor volume or changing


Fig. 2. Radiation field change as a result of positron emission tomography (PET) data for a patient with T3N0 adenocar-cinoma of the rectum. Colonoscopy described a mass 4 cm from the anal verge. Note the location of the anterior border ofthe lateral field on the image on the right based on computed tomography gross tumor volume (CT GTV) (blue) comparedwith PET GTV.

a treatment course from definitive to palliative because of the

detection of distant metastases.

Impact on overall managementFor the 20 patients with rectal cancer, PET imaging

resulted in a change in overall management of 5 patients

(25%), predominantly by avoiding unnecessary surgery and

early detection of distant metastases. Positron emission

tomography findings did not impact the overall treatment

for any of the patients with anal carcinoma.

Postradiation preoperative PET-CT and pathologicfindings in patients with rectal primaries

Fourteen of 20 patients (70%) with rectal primaries

received a PET-CT scan 2 weeks after the completion of ra-

diotherapy to assess tumor response. In all patients, the SUV

decreased in the location of the tumor from a range of 10–25

to 5–7, representing a mean change of 66%.

Of the 15 patients who remained candidates for surgery,

10 underwent surgical resection. Subset analysis was

performed to determine whether the change in degree of

PET avidity from pretreatment to posttreatment imaging

could be predictive of pathologic downstaging and, thus,

save some patients from an unnecessary surgery (Table

3). Results indicated that 100% of patients (5 of 5) who

had a pretreatment PET SUV >10 and a posttreatment

PET SUV <6 achieved pathologic downstaging after neoad-

juvant therapy with negative lymph nodes and #T3 tumors

( p = 0.047, Fisher’s exact test). The mean SUV change in

these patients was 68%, which is not sufficiently different

from the overall mean change to assist in clinical decision

making.

When pathologic reports of patients who underwent sur-

gery were compared with the results of posttreatment PET-

CT scans, discrepancies were noted in 3 of 9 patients

(33%) with regard to nodal staging. All 3 patients’ posttreat-

ment scans indicated there had been metabolic improvement

of the rectal primary mass and no hypermetabolic regional

adenopathy. However, pathologic specimens in each case

demonstrated positive regional lymph nodes and residual vi-

able tumor. Although this represents discordance in the nodal

disease in one third of the patients, the small number of


Table 3. PET changes after chemoradiation for rectal cancer and pathologic findings

Patient no. Pretreatment SUV Posttreatment SUV % SUV changePretreatment SUV >10 and

posttreatment SUV <6 Pathologic stage

2 14.7 7.4 50 No T3N13 10 5.8 42 Yes T3N04 29.2 17.8 39 No T3N06 15.9 5.5 65 Yes T1micN08 23 7 70 No T3N19 6 0 100 No T3N1

12 10.6 5 53 Yes T3N015 14.9 0 100 Yes T1micN016 16 3.3 79 Yes T0N019 19.1 8 58 No T4N0

Abbreviation: PET = positron emission tomography; SUV = standardized uptake value.

patients in this subset analysis precludes calculation of sensi-

tivity and specificity.

All 4 patients who demonstrated distant metastatic disease

at the time of diagnosis received follow-up imaging. The 2 pa-

tients with presumed lung or para-aortic metastases showed

resolution at the primary site but eventual progression of dis-

ease in the suspected areas. The patient with an additional

colon lesion and the patient with liver lesions demonstrated

worsening disease at these locations at first follow-up.

Postradiation PET-CT findings in patients with analprimaries

For the 2 patients with Stage II anal carcinoma, PET-CT

imaging showed improved metabolic uptake within the pri-

mary lesion at the time of first follow-up scan. Persistent

thickening was noted on a follow-up CT scan for the patient

with Stage IIIB disease.

DISCUSSION

Combined-modality PET-CT is increasingly being applied

in oncology to assist with initial staging of tumors as well as

monitoring response to therapy in several disease sites. This

has translated into increased incorporation of combined-mo-

dality PET-CT into radiation oncology for treatment planning

(13). There are several benefits for using a combined PET-CT

over a PET or CT scan alone. Most notably, a combined scan

provides a noninvasive way to assess a tumor’s biologic activ-

ity and incorporates spatial localization to help improve stag-

ing accuracy. For colorectal cancer, PET-CT has an accuracy

between 83% and 93% for initial staging, a 96% sensitivity

and 97% specificity for local recurrence, and a 95% sensitivity

and 98% specificity for detecting distant metastases (14, 15).

For cancer of the anal canal, the sensitivity and specificity of

PET have not been fully determined.

There are few studies on the impact of PET-CT data on the

radiation management of these lesions. The primary endpoint

of this investigation was to assess the impact of PET-CT fu-

sion on both the radiotherapy planning and overall manage-

ment of patients with anorectal tumors. Specifically, we

compared the overlap between tumor volumes contoured on

the PET and CT images and then calculated the frequency

with which the contours led to a change in radiation field

design.

Mean PET and CT volumes were quite similar, but the

ranges and median volumes differed more dramatically. The

CT-GTVs tended to be larger than what was contoured on

PET. There are several potential explanations for this observa-

tion. If it is difficult to distinguish between malignant and re-

active tissues on CT imaging, an over-definition of the true

GTV can occur. In contrast, areas that display high avidity

for FDG on PET imaging often have sharper borders that

can help in GTV contouring. Conversely, the inability of

a portion of tumor to uptake FDG secondary to hypoxia or

necrosis can falsely underestimate an area at risk. This is par-

ticularly troublesome if the rim of residual tumor cells is less

than 1 cm thick because of the resolution limitations of the

PET scanner.

The distinction between inflammation, physiologic bowel

activity, and tumor can be difficult to interpret on PET-CT as

well. In our study, SUV was used as the method to determine

whether an area represented malignancy when the PET-CT

scans were initially interpreted. During the contouring, we

relied on a visual assessment of the tumor to generate the

PET-GTV. Recently, the National Cancer Institute issued

consensus recommendations for the use of PET imaging to

determine response to therapy for patients with cancer (16).

The investigators discussed the advantages and disadvan-

tages of the different methods for assessing FDG uptake,

including the SUV and visual methods. Visual interpretation

was considered the easiest method but can also be subjected to

reader interpretation. In contrast, SUV assessment provides

a semiquantitative method for analysis but can be defined

numerous ways and requires attenuation correction. No con-

sensus guidelines were set to recommend one method over

another, and both remain acceptable. All of our patients with

rectal primaries who received a posttreatment PET-CT scan

showed a decrease in FDG avidity in their tumors, with a mean

reduction of 66%. This supports the idea that these areas rep-

resented viable tumor instead of inflammation or normal

bowel activity; however, future studies are needed to deter-

mine a more reliable way to make this distinction.


When the outlier patient was both included and excluded

from analysis, the observation that increases in tumor volume

on either CT or PET lead to increases in the overlap volume

remained statistically significant. The clinical implication of

this finding is that larger tumors are more likely to have over-

lapping CT and PET volumes and are less likely to have in-

adequate tumor volume coverage during radiation treatment

planning. It is also possible that tighter tumor margins can

be used to define the PTV when both the PET and CT tumor

volumes are large. However, the same significance was not

maintained when we correlated overlap percentage to either

CT or PET volumes.

Without inclusion of the outlier patient, the observation

that increases in PET or CT volumes correlated with overlap

percentage no longer retained significance. Again, several

potential explanations can be found. As primary lesions of

the rectum or anus increase in size, their adherence to adja-

cent structures can limit the variability of tumor location

between imaging modalities as a function of peristalsis or

respiration. Furthermore, larger tumors may be more easily

visible for correct contouring on either CT or PET. Thus,

when tumor volumes are small (between 0 and 200 cm3 in

our series), there may be wide variation between the locations

of the primary lesion on PET as compared with CT, but as the

volumes become larger, the degree of correlation becomes

more pronounced.

The lack of a strong correspondence between PET and CT

in patients with smaller primaries can also be due to the pit-

falls of manual coregistration. Meltzer et al. (6) indicated that

this technique can be fraught with limitations when used to

delineate treatment volumes in small areas with dense anat-

omy, such as the head and neck. Although shortcomings of

manual coregistration remain in the management of anorectal

tumors, it may not have the same clinical significance as in

the head and neck.

The potential pitfalls of manual coregistration are avoided

when the CT portion of a PET-CT study is used for radiation

treatment planning, as is the case at our county institution. At

our private facility, the treatment-planning CT is obtained

separately. However, we attempt to minimize the potential

for error by pursuing a PET-to-CT fusion with a 2-mm mar-

gin of error, as is calculated by our fusion software.

It is difficult to assess the impact of not using rectal con-

trast at the time of CT simulation. Because rectal contrast

can cause artifact interference on PET-CT images, we abstain

from using it during the planning CT so that both studies are

performed under nearly identical circumstances. Rectal con-

trast can provide improve delineation of the gross tumor, but

it may not be as helpful in targeting the extraluminal extent of

disease, which is where some field modifications occurred.

The inability to use rectal contrast in both studies, though,

may be a potential limitation of incorporating PET-CT in

treatment planning.

The goal of the overlap percentage was to ascertain

whether the overlap volume tended to correspond more

commonly with the PET or CT volume. The CT volume was

used in the denominator to calculate the overlap percentage

because it is ultimately the CT data set that is used for all

radiation treatment planning. Although including or not

including the outlier patient in this analysis led to mixed sta-

tistical significance, the trend of an increasing overlap per-

centage with increased PET and CT volume was consistent

when all patients were included.

The ultimate clinical significance of this study was to de-

termine the frequency with which the use of PET-CT fusion

altered radiation treatment, overall management, and surgical

outcomes in patients. In our series, 26% of patients experi-

enced a change in the radiation treatment-planning process

either by increasing traditional radiation treatment field sizes

that would have cut through a PET tumor volume or changing

a patient’s overall treatment course from definitive to pallia-

tive. Furthermore, the integration of PET data altered the

PTV in 17% of patients. We believe that this dramatic change

in field design justifies routine use of PET-CT fusion in radio-

therapy planning.

It is important to consider the impact of PET-CT fusion on

the care of patients with rectal and anal primaries separately.

For the 20 patients with rectal carcinoma, the results of PET

imaging altered the management of 25% of patients by de-

tecting distant metastases and avoiding surgery. This sup-

ports the use of PET-CT imaging as part of the routine

initial workup of patients with rectal cancer.

The majority of patients received a posttreatment scan at

2 weeks, and all patients showed a dramatic response to neo-

adjuvant chemoradiation with a decrease in SUV by at least

half in the primary lesions. We sought to determine whether

the degree of response to preoperative therapy noted on

PET-CT imaging could be predictive of pathologic down-

staging. On subset analysis, all patients with advanced tu-

mors whose pretreatment SUV was >10 and posttreatment

SUV was <6 achieved pathologic downstaging (all pathol-

ogy specimens were #T3N0) at the time of surgery. How-

ever, the number of patients in our subset is small, and the

percentage change in SUV between those who experienced

pathologic downstaging and those who did not is not dra-

matic. Thus, we cannot suggest that a certain percent change

exists that may safely eliminate surgery or predict patho-

logic downstaging in patients. Additional investigation in

a prospective fashion with a larger number of patients is

needed.

For patients with rectal primaries who underwent surgical

resection, we compared the results of pathologic reports with

preoperative PET-CT scans. Our results demonstrated the

limitations of PET imaging to define accurately the extent

of involved regional adenopathy. Of 20 patients who under-

went surgery, 3 were thought to have no hypermetabolic

adenopathy after chemoradiation on imaging, but their

pathologic specimens demonstrated positive lymph nodes.

Although our numbers are too small to calculate sensitivity,

specificity, or predictive values, other studies note the limita-

tion of FDG-PET in analyzing lymph nodes, with a sensitivity

of 29% (17). This limitation of PET is understandable be-

cause involved lymph nodes may be too small after treatment

to be adequately characterized by PET. Alternatively, the


contained tumor cells may have become metabolically less

active and are therefore less FDG avid.

Results for patients with anal primaries were less dramatic;

the number of patients in our series is small. However, few

studies have investigated the utility of PET-CT for manage-

ment of carcinoma of the anus. Of the three patients in our se-

ries, PET-CT fusion altered the radiation treatment planning

in 1 patient during the creation of the initial PTV. This result

is similar to that reported by Ciernik et al. (11), who noted

that incorporation of PET-CT data impacted the size of the

fields designed to treat the primary lesion during the final

boost but did not increase the size of the initial large radiation

fields. Although PET-CT may not commonly change initial

fields, its improved ability to detect nodal disease over CT

alone is significant because this may lead to additional radi-

ation dose to the involved groins.

LimitationsIt is important to note the limitations of our study. Our in-

vestigation involved a small series of patients, the majority of

whom had carcinoma of the rectum, which makes conclu-

sions about the role of PET-CT fusion in the management

of anal primaries less generalizable. At our private institution,

manual coregistration of the PET-CT to the simulation CT

REFER

differed from the practice at our county institution, in which

the CT data set obtained at the time of PET-CT was used for

radiation treatment planning. This certainly can lead to differ-

ences in contours of the GTV and overlap but may represent

what occurs in some clinical practice applications. The clin-

ical significance of the outlier patient could be more clearly

defined if there were more patients with large, locally ad-

vanced tumors. However, because both imaging modalities

provide useful data on the extent of tumor in patients with

anorectal primaries, we advocate incorporating data from

both modalities in radiation treatment planning.

CONCLUSION

We recommend routine use of PET-CT in the initial stag-

ing and workup for patients with cancer of the rectum or anus.

Data gathered from this imaging modality can lead to invalu-

able data on the true extent of the primary lesion and meta-

static spread and can lead to changes in radiation treatment

planning that may alter clinical outcomes. Future studies in-

clude comparison and long-term follow-up to determine the

ultimate clinical impact of PET-CT fusion on management

of patients with anorectal tumors.

ENCES

1. Paulino AC, Koshy M, Howell R, et al. Comparison of CT- andFDG-PET defined gross tumor volume in intensity modulatedradiotherapy for head-and-neck cancer. Int J Radiat OncolBiol Phys 2005;61:1385–1392.

2. Bradley J, Thorstad WL, Mutic S, et al. Impact of FDG-PET onradiation therapy volume delineation in non-small cell lungcancer. Int J Radiat Oncol Biol Phys 2004;59:78–86.

3. Erdi YE, Rosenweig K, Erdi AK, et al. Radiotherapy treatmentplanning for patients with non-small cell lung cancer using pos-itron emission tomography (PET). Radiother Oncol 2002;62:51–60.

4. Zinzani PL, Magagnoli M, Chierchetti F, et al. The role of pos-itron emission tomography (PET) in the management of lym-phoma patients. Ann Oncol 1999;10:1181–1184.

5. Isasi CR, Lu P, Blaufox MD. A meta-analysis of 18F-fluoro-2-deoxy-glucose positron emission tomography in the staging andrestaging of patients with lymphoma. Cancer 2005;104:1066–1074.

6. Metzer CC, Bransetter BF. Review of role of PET-CT fusion inhead and neck cancer. Oncology 2005;19:250–253.

7. Cohade C, Osman M, Leal J, et al. Direct comparison of 18F-FDG-PET and PET-CT in patients with colorectal carcinoma.J Nucl Med 2003;44:1797–1803.

8. Rosen RS, Patel PV, Wahl RL, et al. The role of FDG-PET inthe initial staging of newly diagnosed and untreated colorectalcancer [abstract 1069]. Presented at the 51st Annual Meetingof the Society of Nuclear Medicine; June 19–23, 2004; Philadel-phia, PA.

9. Trautmann TG, Zuger JH. Positron Emission Tomography forpretreatment staging and posttreatment evaluation in cancer ofthe anal canal. Mol Imaging Biol 2005;7:309–313.

10. Cotter SE, Grigsby PW, Siegel BA, et al. FDG-PET/CT in theevaluation of anal carcinoma. Int J Radiat Oncol Biol Phys2006;65:720–725.

11. Ciernik IF, Dizendorf E, Baumert BG, et al. Radiation treatmentplanning with an integrated positron emission and computer to-mography (PET/CT): A feasibility study. Int J Radiat Oncol BioPhys 2003;57:853–863.

12. Thomas C, Kassam Z, Wong W, et al. The influence of positronemission tomography (PET) in changing treatment options inlocally advanced metastatic colorectal cancer [abstract 206].Presented at the 2005 Gastrointestinal Cancers Symposium;January 27–29, 2005; Hollywood, FL.

13. Paulino AC, Johnstone PAS. FDG-PET in radiotherapy treat-ment planning: Pandora’s box? Int J Radiat Oncol Biol Phys2004;59:4–5.

14. Sachelaire I, Kerr K, Ghesani M, et al. Integrated PET-CT:Evidence-based review of oncology indications. Oncology2005;19:481–490.

15. Delbeke D, Martin WH. PET and PET-CT for evaluation ofcolorectal carcinoma. Semin Nucl Med 2004;34:209–223.

16. Shankar LK, Hoffman JM, Bacharach S, et al. Consensus rec-ommendations for the use of 18F-FDG PET as an indicator oftherapeutic response in patients in National Cancer Institute Tri-als. Consensus recommendations for the use of 18F-FDG PETas an indicator of therapeutic response in patients in NationalCancer Institute Trials. J Nucl Med 2006;47:1059–1066.

17. Kantorova I, Lipska L, Belohlavek O, et al. Routine (18)F-FDGPET preoperative staging of colorectal cancer: Comparison withconventional staging and its impact on treatment decisionmaking. J Nucl Med 2003;44:1784–1788.

PET-CT Fusion in Radiation Management of Patients with Anorectal Tumors

Documents

Transcript of PET-CT Fusion in Radiation Management of Patients with Anorectal Tumors