Detection of Interval Distant MetastasesClinical Utility of Integrated CT-PET Imaging in Patients WithEsophageal Carcinoma After Neoadjuvant Therapy

John F. Bruzzi, FFRRCSI1

Stephen G. Swisher, MD2

Mylene T. Truong, MD1

Reginald F. Munden, MD1

Wayne L. Hofstetter, MD2

Homer A. Macapinlac, MD1

Arlene M. Correa, PhD2

Osama Mawlawi, PhD3

Jaffer A. Ajani, MD4

Ritsuko R. Komaki, MD5

Norio Fukami, MD6

Jeremy J. Erasmus, MD1

1 Division of Diagnostic Imaging, M. D. AndersonCancer Center, Houston, Texas.

2 Department of Thoracic and CardiovascularSurgery, M. D. Anderson Cancer Center, Houston,Texas.

3 Department of Imaging Physics, M. D. Ander-son Cancer Center, Houston, Texas.

4 Department of Gastrointestinal Medical Oncology,M. D. Anderson Cancer Center, Houston, Texas.

5 Division of Radiation Oncology, M. D. AndersonCancer Center, Houston, Texas.

6 Department of Gastrointestinal Medicine andNutrition, M. D. Anderson Cancer Center, Hous-ton, Texas.

BACKGROUND. The objective of the study was to determine the utility of inte-

grated computed tomography / positron emission tomography (CT-PET) imaging

for detecting interval distant metastases and assessing therapeutic response in

patients with locally advanced, potentially resectable esophageal carcinoma after

neoadjuvant therapy.

METHODS. A retrospective study was performed of 88 patients with potentially re-

sectable esophageal carcinoma who received neoadjuvant therapy before planned

surgical resection. CT-PET before and after completion of neoadjuvant was used

for evaluating therapeutic response; response criteria were based on qualitative

and semiquantitative analyses.

RESULTS. Neoadjuvant therapy comprised chemoradiotherapy in 85 patients,

with prior induction chemotherapy in 39 patients. Fifty-five patients proceeded

to esophagectomy. Repeat CT-PET was performed after induction chemotherapy

(n ¼ 23) and after completing chemoradiotherapy (n ¼ 85). CT-PET identified the

interval appearance of metastatic disease in 7 (8%) patients. For assessment of

locoregional therapeutic response, CT-PET was unable to predict pathological

response to neoadjuvant therapy in the primary tumor or locoregional lymph

nodes. CT-PET had sensitivity, specificity, and positive and negative predictive

values of 57%, 46%, 39%, and 64%, respectively, for detection of residual macro-

scopic malignancy within the primary tumor; and sensitivity, specificity, and

positive and negative predictive values of 0%, 90%, 0%, and 69% for detection of

residual malignancy within resected lymph nodes.

CONCLUSION. CT-PET performed after neoadjuvant therapy in patients with

potentially resectable esophageal carcinoma is important for detecting interval

metastases that preclude surgical resection, but is of limited utility for assessing

locoregional therapeutic response. Cancer 2007;109:125–34.

� 2006 American Cancer Society.

KEYWORDS: esophageal cancer, staging, CT-PET, neoadjuvant therapy.

E sophageal cancer is increasing in incidence in the US and an

estimated 14,550 new cases are expected to be diagnosed in

2006.1 Esophagectomy is the treatment of choice in patients with

early-stage disease. In patients with more advanced stage disease,

the role of neoadjuvant therapy before esophagectomy is evolving,2–7

with most studies showing no benefit over surgery alone.6,7 How-

ever, there is evidence to suggest that the subset of patients who

have a pathological response to neoadjuvant therapy will have

improved survival after surgery compared with those patients who

undergo esophagectomy alone.5,6,8–11 Conversely, patients who do

not have a pathological response to neoadjuvant therapy have a

Address for reprints: John F. Bruzzi, Division ofDiagnostic Imaging, M. D. Anderson Cancer Cen-ter, 1515 Holcombe Blvd., Unit 0371, Houston,TX 77030-4009; Fax: (713) 563-0638; E-mail:jbruzzi@mdanderson.org

Received July 24, 2006; revision receivedOctober 12, 2006; accepted October 16, 2006.

ª 2006 American Cancer SocietyDOI 10.1002/cncr.22397Published online 4 December 2006 in Wiley InterScience (www.interscience.wiley.com).

125

worse prognosis after subsequent esophagectomy

than those patients who are resected de novo.6,7,11

In the context of making clinical decisions with

regard to those patients who are most likely to benefit

from esophagectomy after multimodality therapy, the

determination of the pathological response to neoadju-

vant therapy is both clinically difficult and important.

Recently, it has been proposed that 2-fluoro-2-deoxy-d-

glucose (FDG) positron emission tomography (FDG-

PET) imaging can identify those patients who have had

a response to neoadjuvant therapy and who will benefit

from subsequent esophagectomy.12–22 Conversely, stu-

dies have shown that persistent FDG uptake at the site

of the primary tumor after neoadjuvant therapy corre-

lates both with a lack of pathological response to che-

moradiotherapy and with a worse clinical outcome,

implying that nonsurgical management is indicated for

these patients.12–16,19,20 However, there have only been a

limited number of such studies and uncertainty remains

over the clinical utility of computed tomography PET

(CT-PET), particularly with respect to the optimal tim-

ing of performing CT-PET, its specificity for differentiat-

ing between residual tumor and therapy-induced

inflammation, and the detection of distant metastases

after neoadjuvant therapy.23,24 The aim of this study was

to report our clinical experience with CT-PET for detect-

ing distant metastases and in evaluating the therapeutic

response of the primary tumor to neoadjuvant therapy

in patients with locally advanced potentially resectable

esophageal carcinoma.

MATERIALS AND METHODSPatient PopulationA retrospective review was performed of consecutive

patients with primary esophageal cancer who were

referred to the Department of Thoracic and Cardiovas-

cular Surgery between January 2003 and July 2005 for

consideration of operativemanagement of their cancers.

Patients were eligible for inclusion if they had previously

untreated, histologically confirmed esophageal malig-

nancy; had locally advanced, potentially resectable car-

cinomas and were scheduled to receive neoadjuvant

therapy before planned resection; and had undergone

PETor CT-PET both for initial staging and for evaluation

of therapeutic response within 10 weeks of completion

of neoadjuvant treatment, and within 10 weeks of eso-

phagectomy in patients undergoing surgery. This study

was approved by the Institutional Review Board of M. D.

Anderson Cancer Center, who waived the requirement

for individual signed consent from the study patients.

TreatmentPatients were treated either with induction chemo-

therapy followed by concurrent chemoradiotherapy

or with concurrent chemoradiotherapy only. Induc-

tion chemotherapy comprised up to 2 6-week cycles

of cisplatin (45 mg/m2), taxotere (33 mg/m2), and 5-

fluorouracil (2 g/m2 as a 24-hour infusion). Concur-

rent chemoradiotherapy comprised 3D conformal

radiotherapy to the primary tumor site and regional

lymph nodes to a planned 50.4 Gy in 28 fractions, to-

gether with chemotherapy consisting of a combina-

tion of a platinum agent (cisplatin or carboplatin)

and 5-fluorouracil or taxotere.

Patients who proceeded to surgical resection

underwent esophagectomy either by a transthoracic

approach (employing a chest anastomosis [Ivor-Lewis

procedure] or a cervical anastomosis), by a transhiatal

approach, or using a minimally invasive technique.

Assessment of Therapeutic ResponseInitial staging included a contrast-enhanced CT scan

of the thorax and abdomen, upper gastrointestinal

endoscopy and endoscopic ultrasound (EGD/EUS)

(including EUS-guided biopsy of suspicious lymph

nodes), and whole-body PET or CT-PET. Assessment

of therapeutic response was performed after induc-

tion chemotherapy and/or after the completion of

concurrent chemoradiotherapy by both EGD/EUS

and CT-PET in all patients.

All CT-PET scans were performed on an integrated

scanner (Discovery ST-8, GE Medical Systems, Milwau-

kee, WI). Patients were injected with a mean 15–

20 mCi of FDG. PET images were acquired during

shallow breathing in the 2D mode for 3 minutes per

bed position 60 to 90 minutes. PET images were recon-

structed using standard vendor-provided reconstruc-

tion algorithms that incorporated ordered subset

expectation maximization. Attenuation correction of

PET images was performed using attenuation data

from the CT component of the examination; emission

data were corrected for scatter, random events, and

dead-time losses using the manufacturer’s software.

PET and fused CT-PET images were analyzed both

qualitatively and semiquantitatively. On qualitative

analysis, foci of abnormal FDG uptake greater than

that of background activity in the adjacent mediasti-

num were identified that corresponded to the known

location of the primary esophageal tumor, locoregio-

nal lymph nodes, and suspected metastases. On re-

evaluation CT-PET the patient was considered to have

had a complete response (no residual FDG uptake in

the primary tumor greater than that of background ac-

tivity), a partial response (persistent focal increased

FDG uptake in the region of the primary esophageal

tumor but with decreased intensity compared with

the baseline scan). or no response (persistent abnor-

mal FDG uptake in the primary tumor equivalent to or

126 CANCER January 1, 2007 / Volume 109 / Number 1

greater than that at baseline imaging). A finding of dif-

fuse increased FDG uptake throughout the esophagus

in a linear pattern was considered to be consistent

with posttreatment esophagitis.

On semiquantitative analysis the intensity of met-

abolic activity within foci of increased FDG uptake (in

the primary tumor, locoregional lymph nodes or dis-

tant metastases) was analyzed on the PET images;

using a semiautomated vendor-provided tool the max-

imum standard uptake value (SUVmax) within the vol-

ume of interest was calculated according to the

formula: SUVmax ¼ maximum measured activity within

the volume of interest (mCi/mL)/injected dose of FDG

(mCi)/body weight (g). The volume of interest was

determined using a user-specified cubic volume drawn

in 3 planes (axial, coronal, and sagittal) to include the

site of abnormal FDG accumulation.

According to previous data based on expected

long-term survival,8,25 a pathological complete re-

sponse (CR) to neoadjuvant therapy was defined by

the absence of any residual viable tumor cells on his-

tological examination of the esophagectomy specimen;

a partial response was defined by residual microscopic

malignancy (�10% viable cells in the primary malig-

nancy); no response to therapy was defined by resid-

ual macroscopic malignancy (>10% viable cells). On

semiquantitative analysis of PET images, patients were

considered to have had a response to induction

chemotherapy if the SUVmax in their primary tumor

decreased by 30% or greater compared with baseline.

Based on previously published studies12,13 and on our

own clinical practice, SUVmax > 4 in the region of the

known esophageal carcinoma after completion of

neoadjuvant chemoradiotherapy was considered to

represent residual viable macroscopic tumor. SUVmax

>2.5 in the locoregional lymph nodes were considered

to represent residual disease.

Statistical AnalysisDifferences in CT-PET findings between responders

and nonresponders were assessed using paired and

unpaired Student t-tests (for continuous variables)

and both univariate and multivariate analyses

employing Fisher exact test and the chi-square test

(for noncontinuous variables). For all analyses, 2-

tailed P-values �.05 were considered significant. In

addition, the sensitivity, specificity, and positive and

negative predictive values of CT-PET and EGD/EUS

were calculated for the detection of residual macro-

scopic malignancy after neoadjuvant therapy. Statisti-

cal analyses were performed (by J.F.B. and A.M.C.)

using GraphPad Instat (Graph Pad Software, San

Diego, CA) and SPSS (SPSS, Chicago, IL) software.

RESULTSPatients, Treatment, and Pathological ResponseBetween January 2003 and July 2005, a total of 88

patients with potentially resectable esophageal can-

cer who underwent combined modality chemora-

diotherapy before planned esophagectomy at M. D.

Anderson Cancer Center were considered eligible for

inclusion in the study. There were 79 men and 9

women (mean age at initial staging, 62 years; age

range, 23–81 years). All patients underwent baseline

staging with either whole-body FDG-PET (n ¼ 2) or

CT-PET (n ¼ 83) imaging (Table 1).

Patients were treated either with induction

chemotherapy followed by concurrent chemora-

diotherapy (n ¼ 39) or with concurrent chemora-

diotherapy only (n ¼ 46); 3 additional patients

received induction chemotherapy but because of dis-

ease progression (detailed below) they did not re-

ceive additional chemoradiotherapy. For those

patients receiving radiotherapy, the median radiation

dose used was 50 Gy (range, 37.8–66 Gy). The me-

dian interval between baseline staging and the com-

pletion of neoadjuvant therapy was 82 days (range,

43–222 days), with 83 (98%) of patients completing

their treatment within 6 months.

After completion of neoadjuvant therapy and af-

ter evaluation of tumor response, 55 patients un-

derwent esophagectomy, either by a transthoracic

approach (Ivor-Lewis procedure [n ¼ 35] or a cervi-

cal anastomosis [n ¼ 8]), by a transhiatal approach

(n ¼ 7), or using a minimally invasive technique

TABLE 1Patient Characteristics

Characteristics

No. patients 88

Age 62 y (23–81 y)

Sex 79 men; 9 women

Esophageal tumor histology

Adenocarcinoma 75

Squamous cell carcinoma 13

Tumor location in esophagus

Upper one-third 1

Middle one-third 9

Distal one-third 58

Gastroesophageal junction 20

AJCC stage of esophageal carcinoma at baseline*

I (T1, N0, M0) 0

IIA (T2 or T3, N0, M0) 30

IIB (T1 or T2, N1, M0) 3

III (T3 or T4, N0 or N1, M0) 55

IV (Any T or N, M1a or M1b) 0

* AJCC: American Joint Committee on Cancer staging system.30 Clinical stage decided by endoscopic

ultrasound (EUS), computed tomography (CT), and CT-PET (positron emission tomography) prior to

neoadjuvant treatment.

Interval Esophageal Cancer Metastases/Bruzzi et al. 127

(n ¼ 5). In 2 additional patients esophagectomy was

abandoned because of tumor invasion into the aorta

(n ¼ 1) or unexpected peritoneal metastases (n ¼ 1),

which were discovered at the time of surgery but

which had not been detected by imaging before the

procedure. A mean of 21 lymph nodes was resected

per patient (range, 1–47). The median time interval

between the completion of chemoradiotherapy and

esophagectomy was 55 days (range, 17–110 days).

The remaining patients did not undergo esopha-

gectomy because of comorbid medical conditions

(n ¼ 18), patient choice (n ¼ 5), failure of local con-

trol (n ¼ 2), or interval appearance of new metastatic

disease (n ¼ 7).

Therapeutic Response Assessment by CT-PET: AfterInduction ChemotherapyIn 23 of the patients who were treated initially with

induction chemotherapy, evaluation of the therapeu-

tic response was performed by CT-PET after 1 (n

¼ 8) or 2 (n ¼ 15) cycles of chemotherapy. CT-PET

was performed in these patients to assess therapeutic

response before planned radiotherapy according to a

clinical trial schedule, either on-protocol (n ¼ 6) or

off-protocol (n ¼ 17); progressive malignancy was

not clinically suspected in any of these patients. The

median time interval between the baseline CT-PET

scan and this repeat CT-PET scan was 83 days (range,

47–180 days).

On baseline CT-PET, all 23 patients had FDG-

avid esophageal tumors (mean SUVmax, 12.7; 95%

confidence interval [CI], 10.3–15). On reevaluation

CT-PET after induction chemotherapy the mean

SUVmax decreased to 6.4 (95% CI, 4.3–8.4) (P < .0001).

By semiquantitative analysis, 10 (43%) of these

patients had a decrease in the SUVmax of their primary

tumor by greater than 50% (mean decrease of 80%,

range 60%–94%). Of the other patients, 7 had a

decrease in their SUVmax of 30%–49%, whereas in

6 patients there was either no appreciable change

(<30% difference) or an increase in SUVmax of the pri-

mary tumor. None of these patients had CT-PET find-

ings suggestive of esophagitis.

Eleven of these patients underwent subsequent

esophagectomy, of which 5 had a pathological response

(residual viable tumor �10%) and 6 did not. Although

the numbers were too few for conclusive analysis, we

were unable to detect any difference between patients

who had or did not have a pathological response at

esophagectomy (using multiple analyses with respect to

age, gender, tumor histology, stage, and location, quali-

tative appearances on CT-PET images, and differences

in the absolute and percentage changes in SUVmax).

Analyses of the findings on CT-PET are summarized in

Table 2. Of note, of the 6 patients who had residual

macroscopic tumor on esophagectomy, 4 (67%) had an

apparent response on CT-PET performed after induc-

tion chemotherapy (decrease in SUVmax greater than

30%), of whom 1 had an apparent complete response

(no residual FDG uptake).

Nevertheless, CT-PET performed during or after

induction chemotherapy was useful in identifying

new organ metastases in 2 (8%) patients, both of

whom had distal stage III (T3, N1) esophageal ade-

nocarcinomas at baseline staging; 1 of these patients

developed new hepatic metastases, and the other

patient developed both hepatic and pulmonary me-

tastases after 2 cycles of induction chemotherapy.

None of these metastases had been present on base-

line imaging studies. These metastases were detected

by CT-PET at 70 and 112 days from their baseline

CT-PET scan, respectively, and were confirmed by

diagnostic CT imaging and subsequent follow-up.

Therapeutic Response Assessment by CT-PET: AfterConcurrent ChemoradiotherapyAll of the patients in the series (except the 2 patients

who developed metastatic disease after induction

chemotherapy as well as a third patient who died of

progressive malignancy during chemoradiotherapy)

underwent a third CT-PET after the completion of

concurrent chemoradiotherapy, before planned sur-

gery. This third CT-PET (n ¼ 85) was performed at a

median interval of 40 days from the end of therapy

(range, 10–67 days).

In 82 (97%) patients, the primary esophageal

tumor was FDG-avid with SUVmax > 4; the mean

SUVmax was 12.9 (95% CI, 11.2–14.5). After the com-

pletion of neoadjuvant chemoradiotherapy the mean

SUVmax decreased to 4.4 (95% CI, 3.8–5.0) (P < .0001)

(Fig. 1). Fifty (61%) of these patients had a decrease

in the SUVmax of their primary tumor by greater than

50% (mean decrease of 79%; range, 53%–97%). Of

the other patients, 16 (20%) had a decrease in their

SUVmax of 30%–49%, whereas in 16 (20%) patients

there was either no appreciable change (<30% de-

crease) or an increase in SUVmax of the primary tumor.

Fifty-five (63%) patients underwent esophagect-

omy. On examination of the esophagectomy speci-

mens a complete pathological response was found in

18 (33%) patients, with a partial response in an addi-

tional 17 (31%) patients (�10% viable tumor cells in

the primary malignancy). Twenty (36%) patients were

nonresponders (>10% residual viable tumor cells).

Response to neoadjuvant therapy was assessed

before surgery by both EGD/EUS and biopsy and by

CT/PET in 53 of these patients. For detection of re-

sidual viable tumor >10% of the original tumor mass

128 CANCER January 1, 2007 / Volume 109 / Number 1

(corresponding to a lack of pathological response to

neoadjuvant therapy), EGD/EUS had a sensitivity

and positive predictive value of 33% and 88%,

respectively, with a specificity and negative predictive

value of 97% and 69%, respectively.

When assessing the utility of CT-PET performed af-

ter concurrent chemoradiotherapy to differentiate be-

tween patients who had or did not have a pathological

response within the primary esophageal tumor, multi-

ple analyses found no significant difference between

the 2 groups (Table 2). Using a cut-off SUVmax thresh-

old of 4 for the detection of residual viable macroscopic

tumor >10% (corresponding to no pathological

response), CT-PET had sensitivity, specificity, and posi-

tive and negative predictive values of 57%, 46%, 39%,

and 64%, respectively. Of note, 4 (7%) patients who

qualitatively had a complete metabolic response to

therapy by CT-PET had residual macroscopic malig-

nancy (>10% viable tumor cells) on analysis of the eso-

phagectomy specimen. Use of higher cutoff SUVmax

values resulted in increased specificity for detection of

residual tumor, but reduced sensitivity.

Sixteen (29%) patients had residual viable tumor

within regional N1 lymph nodes upon pathological

examination; in 4 (7%) of these patients, residual

viable tumor within regional lymph nodes was found

despite a complete pathological response in the pri-

mary esophageal tumor, whereas in another 4 (7%)

of these patients a partial pathological response in

the primary tumor was found; the remaining 8

patients had no pathological response either in their

primary esophageal tumor or within involved lymph

nodes. CT-PET was unable to detect residual ma-

lignancy within any of the involved locoregional

lymph nodes using qualitative analysis. Using a

cut-off SUVmax of 2.5 for the detection of residual

viable tumor within locoregional lymph nodes, CT-

PET had per-patient sensitivity, specificity, and posi-

tive and predictive values of 0%, 90%, 0%, and 69%,

respectively.

Despite its limited ability in determining thera-

peutic response within the primary esophageal mass

or regional lymph nodes, CT-PET after concurrent

chemoradiotherapy and before planned surgery was

useful in identifying new metastatic disease in 5 (6%)

patients. In all of these patients the histology of the

primary esophageal carcinoma was adenocarcinoma;

the stage of the esophageal malignancy at initial eva-

TABLE 2CT-PET Evaluation of Therapeutic Response in Primary Esophageal Tumor

CT-PET findings*

Pathological response

(����10% residual viable tumor)

No pathological response

(>10% residual viable tumor) P

Mean SUVmax

CT-PET 1 (n ¼ 68) 11.3 (9.1–13.5)y 10.0 (7.7–12.3) .43

CT-PET 2 (n ¼ 11) 7.2 (1.8–12.6) 8.6 (1.9–15.4) .68

CT-PET 3 (n ¼ 68) 4.0 (3.2–4.9) 4.3 (3.1–5.5) .70

Qualitative analysis: no. patients with a complete or partial response with respect to CT-PET 1

CT-PET 2 4 5 1

CT-PET 3 29 16 1

Mean decrease in SUVmax with respect to CT-PET 1 (absolute, %)

CT-PET 2 (n ¼ 11) 6.6 (37.5%) 4.0 (35%) .49 (.92)

CT-PET 3 (n ¼ 68) 7.3 (41.6%) 5.7 (46.8%) .38 (.81)

No. patients with >30% decrease in SUVmax with respect to CT-PET 1

CT-PET 2 3 4 1

CT-PET 3 29 14 .32

No. patients with >50% decrease in SUVmax with respect to CT-PET 1

CT-PET 2 2 1 .55

CT-PET 3 21 10 .58

No. patients with residual SUVmax

<4 16 9

�4 19 11 1

<5 21 13

�5 14 7 .78

<6 27 15

�6 8 5 1

CT-PET indicates computed tomography/positron emission tomography; SUVmax, maximum standard uptake value.

* CT-PET 1: baseline scan; CT-PET 2: scan performed during or after induction chemotherapy; CT-PET 3: scan performed after completion of concurrent chemoradiotherapy.y Numbers in parentheses refer to 6 95% confidence intervals.

Interval Esophageal Cancer Metastases/Bruzzi et al. 129

FIGURE 1. A 65-year-old man with stage III adenocarcinoma of the distal esophagus. (A) Baseline staging CT-PET demonstrates intensely FDG-avid primary tumor(arrow), but no evidence of metastatic disease. (B) CT-PET 20 days after completion of neoadjuvant chemoradiotherapy demonstrates new foci of increased FDG

uptake (arrowhead, short arrow) suspicious for new metastases. Diffuse linear FDG uptake in the esophagus is consistent with postradiation esophagitis (arrow). No

definite abnormality is evident on the CT images (C); however, axial fused CT-PET image confirms an abnormal focus of increased FDG uptake in the right gluteus

medius muscle (D) (short arrow). An intramuscular metastasis from the patient’s esophageal carcinoma was confirmed by subsequent biopsy.

luation was IIB (n ¼ 1) or III (n ¼ 4). These metasta-

ses were detected at a median interval of 145 days

from baseline staging. These new interval metastases

comprised distant organ metastases (M1b) alone in 3

patients, nonregional lymph node metastases (M1a)

alone in 1 patient, and both distant organ and nonre-

gional lymph node metastases in 1 patient. The

organ metastases were located in the liver (n ¼ 2),

bones (n ¼ 3), and skeletal muscle (n ¼ 1). Two of

these patients had metastases that were radiologi-

130 CANCER January 1, 2007 / Volume 109 / Number 1

FIGURE 2. 71-year-old man with stage IIB adenocarcinoma of the distal esophagus, status post neoadjuvant chemoradiotherapy. (A) Axial CT image fromrepeat CT-PET scan demonstrates a subtle osteolytic lesion of the right acetabulum (arrow). (B) Fused axial CT-PET image confirms the presence of an FDG-

avid metastasis (arrow).

cally occult or were located outside the range of rou-

tine surveillance imaging studies, and in whom there

was no evidence of metastatic disease elsewhere (1

patient with skeletal muscle metastases within the

left infraspinatus and right gluteal muscles, and 1

patient with a bone marrow metastasis in the right

acetabulum) (Figs. 1, 2). These metastases were con-

firmed by percutaneous biopsy (n ¼ 1) and by fol-

low-up CT-PET imaging (n ¼ 1).

Therapeutic Response Assessment by CT-PET: Detectionof Interval MetastasesIn total, CT-PET performed after induction chemo-

therapy or neoadjuvant chemoradiotherapy detected

unexpected metastases in 7 (8%) patients. The me-

dian interval between baseline staging and detection

of the new interval metastases was 159 days. These

metastases were confirmed either by biopsy (n ¼ 1)

or by additional follow-up imaging (n ¼ 5); 1 patient

developed disseminated metastases of the liver and

lung and did not undergo biopsy or follow-up ima-

ging. Six (7%) of these patients were still being

actively considered for esophagectomy at the time of

their repeat CT-PET scan, but the appearance of new

metastases precluded further consideration of sur-

gery. They instead received further nonsurgical man-

agement, comprising palliative radiotherapy directed

to the metastases (n ¼ 2), palliative chemotherapy

(n ¼ 3), or supportive care (n ¼ 1). Details concern-

ing the patient and tumor characteristics in these

patients are summarized in Table 3.

DISCUSSIONOur results show that CT-PET is useful in detecting

distant metastases in patients with locally advanced,

potentially resectable esophageal carcinoma after

neoadjuvant therapy. In this regard, unexpected in-

terval metastases were detected in 8% of patients.

Additionally, the results of our study suggest that CT-

PET has limited clinical utility in assessing therapeu-

tic response within the primary esophageal tumor

and within locoregional lymph nodes.

Most previous studies of FDG-PET imaging in

patients with esophageal carcinoma have evaluated

its utility in assessing therapeutic response of the

primary esophageal tumor to neoadjuvant therapy

and there has been little emphasis on its role in

detecting interval distant metastases. Specifically,

there are only 3 recent PET studies that report the

rate of detection of interval distant metastases after

neoadjuvant therapy. In these studies, distant metas-

tases were detected in 8%–17% of patients after

neoadjuvant therapy and before planned esophagect-

omy.15,16,21 However, the small number of patients in

these studies together with a limited explanation of

the clinical context and the method of detection of

the interval metastases precludes assessment of the

clinical utility of PET imaging in restaging patients

with esophageal cancer after neoadjuvant therapy.

In the study by Flamen et al.15 there were only

36 patients and 6 (16.6%) of these patients had dis-

tant metastases detected by imaging. The location of

these metastases and the role of PET imaging in their

detection are not indicated. Specifically, it is unclear

whether PET imaging contributed to the detection of

metastases when compared with CT imaging of the

chest and abdomen. Additionally, inclusion in the

study required a clinical diagnosis of locally ad-

vanced nonresectable primary tumor (cT4), an inclu-

sion criteria that potentially limits applicability to

most surgical series that have a higher preponder-

ance of earlier stage disease. Although the rate of

detection of M1b metastases by imaging after neoad-

Interval Esophageal Cancer Metastases/Bruzzi et al. 131

juvant therapy in the study by Cerfolio et al.21 is sim-

ilar to our study (4 of 48 [8%] patients), the addi-

tional benefit of PET imaging compared with CT was

small, ie, 3 of the 4 metastases were in the liver and

were detected by both studies. Finally, the study by

Weber et al.16 examined early therapeutic response

evaluation after 14 days of chemotherapy without

radiation therapy. Subsequent imaging performed later

in the course of treatment detected metastases in 4

(11%) patients, but further details were not specified.

In our study, CT-PET performed after neoadju-

vant therapy detected interval metastases in 7 (8%)

of the 88 patients in the study group. Because our

study group comprised consecutive patients under-

going evaluation for planned surgical resection after

neoadjuvant therapy, the distribution of disease was

predominantly stage II and III. This distribution of

stage is representative of most surgical series and

consequently we believe our study provides a valid

indication as to the accuracy of PET imaging in the

detection of interval metastases after neoadjuvant

therapy in patients with esophageal carcinoma. Of

particular clinical relevance, the use of PET imaging

allowed detection of metastases in 2 of the 7 patients

that were not detected on conventional staging.

These metastases were located outside the range of

routine restaging CT imaging in these patients. Addi-

tionally, in 1 patient the organ metastases were in an

unusual site (skeletal muscle). We have previously

reported the proclivity of esophageal carcinoma to

manifest with metastases in unusual and uncommon

locations after induction therapy26 and this experi-

ence, together with the results of our present study,

indicate that CT-PET may be the optimal imaging

modality for their detection. Of additional interest,

metastases were detected in 2 patients after only

2 cycles of induction chemotherapy, and in 2

patients there were CT-PET findings consistent with

a complete response of the primary tumor to therapy

despite the appearance of new distant organ metasta-

ses. The detection of new metastases in these

patients precluded any further consideration of eso-

phagectomy, but instead guided palliative nonsurgi-

cal management of the new metastases.

The second major finding of clinical importance

in our study pertains to the evaluation of locoregio-

nal therapeutic response in the primary esophageal

tumor. Although CT-PET was useful in detecting

interval metastases after neoadjuvant therapy, we

found that CT-PET was less useful for assessing ther-

apeutic response in the primary esophageal carci-

noma and locoregional lymph nodes. In our study,

we found no significant difference between respon-

ders and nonresponders with respect to CT-PET find-

ings at completion of neoadjuvant therapy and

before planned surgery. These findings are at var-

iance with reports from previous studies, in which

investigators have shown that pathologic responders

can be differentiated from nonresponders based on

the absolute and percentage decrease in FDG uptake

between baseline and completion of neoadjuvant

therapy (with ‘‘optimal’’ thresholds ranging between

35% and 60%)14–16,18–20 or on the degree of residual

metabolic activity within the primary tumor after

completion of treatment.12,13,17 It is important to

note that in all of these previous reports, criteria for

determining therapeutic response based on FDG-PET

findings were calculated retrospectively based on re-

TABLE 3Demographic and Clinical Details of Patients Who Were Diagnosed With New Metastatic Disease After Neoadjuvant Chemotherapyor Chemoradiotherapy

Patient no.

Age

(y) Sex

Histology ofprimary esophageal

malignancy

Stage ofesophageal cancer

at baseline evaluation

Anatomical locationof primary esophageal

malignancy

Neoadjuvant

therapy

Anatomical location

of new metastases

1 46 Men Adenocarcinoma III Distal 1/3 Induction CT Liver, upper abdominal lymph nodes

2 62 Men Adenocarcinoma III GEJ Induction CT Liver, lungs

3 65 Men Adenocarcinoma III Distal 1/3 Induction CT þconcurrent CRT

Right gluteal muscle þleft infraspinatus muscle

4 62 Men Adenocarcinoma III Distal 1/3 Induction CT þconcurrent CRT

Liver

5 71 Men Adenocarcinoma III GEJ Induction CT þconcurrent CRT

Left supraclavicular lymph nodes

6 71 Men Adenocarcinoma IIB Distal 1/3 Concurrent CRT Right acetabulum

7 65 Men Adenocarcinoma III Distal 1/3 Concurrent CRT Liver, multiple osseous sites,

supraclavicular lymph nodes

CT indicates chemotherapy; CRT, chemoradiotherapy; GEJ, gastroesophageal junction.

132 CANCER January 1, 2007 / Volume 109 / Number 1

ceiver operating characteristics (ROC) acquired from

the study itself, and have varied from investigator to

investigator. Furthermore, these previous studies

have varied widely with respect to patient numbers

(10–103 patients), image acquisition techniques, FDG

uptake quantification, types of neoadjuvant therapy

(chemotherapy vs concurrent chemoradiotherapy),

and the timing of FDG-PET imaging.

In our study the assessment of therapeutic re-

sponse of the primary esophageal carcinoma was

limited by the high rate of false-positive findings for

residual viable tumor (specificity and positive predic-

tive values of 46% and 39%, respectively). The persis-

tent increased FDG uptake in the region of the

esophageal tumor in patients who did have a patho-

logical response was presumably a result of treat-

ment-induced inflammation and ulceration. In a

previous study from our group, we have shown how

inflammation and ulceration in the esophagus result-

ing from chemoradiotherapy are an important cause

of false-positive findings on FDG-PET images per-

formed before planned surgical resection, and ad-

versely affect the diagnostic accuracy of FDG-PET

imaging in discriminating between responders and

nonresponders.24 In the present study, 55% of pa-

tients who received chemoradiotherapy (to a mean

radiation dose of 50 Gy) before esophagectomy had

ulceration detected at EGD.

A recent meta-analysis concluded that FDG-PET

imaging is useful in evaluating therapeutic response

of the primary esophageal malignancy after com-

bined modality therapy.27 As discussed above, this

differs from our recent experience with integrated

CT-PET. While it is important to realize that our

study is limited by the retrospective nature and the

variable periods of time between completion of

neoadjuvant therapy and CT-PET evaluation, 1 im-

portant factor that may account for this difference

between our study and most of the earlier reports is

the use of integrated CT-PET rather than dedicated

PET imaging. In integrated CT-PET, the CT images

used for attenuation correction are acquired in sus-

pended respiration, whereas the PET images are

acquired while the patient is breathing, leading to a

potential mismatch in registration between the CT

images and the PET images. This is in contrast to

dedicated PET imaging, where both the emission and

transmission PET images are acquired while the

patient is breathing. The partial volume averaging

that results from the misregistration of CT and PET

images can account for errors in SUVmax calculations

of up to 30% to 50%,28 particularly in tumors that are

located close to the diaphragm and that are therefore

subjected to pronounced movement during breath-

ing (greater than 2 cm). However, continued innova-

tions in CT-PET—specifically, the ability to correct

for respiratory motion artifact by the use of respira-

tory gated 4D CT-PET (the fourth dimension being

time)—will undoubtedly improve the reproducibility

and accuracy of CT-PET for initial evaluation and

assessment of therapeutic response of patients with

esophageal cancer, particularly distal tumors.28,29

ConclusionCT-PET performed after neoadjuvant therapy can

detect distant metastases in up to 8% of patients

with potentially resectable, locally advanced esopha-

geal carcinoma. Because these metastases can be

clinically occult and may occur in unusual and

uncommon locations after induction therapy, whole-

body CT-PET is the best imaging method for their

detection. Whereas CT-PET can detect unsuspected

metastases that preclude curative resection and

which can influence palliative therapy, it has limited

clinical utility in the assessment of therapeutic

response of the primary esophageal tumor.

REFERENCES1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2006. CA

Cancer J Clin. 2006;56:106–130.

2. Walsh TN, Noonan N, Hollywood D, Kelly A, Keeling N,

Hennessy TP. A comparison of multimodal therapy and

surgery for esophageal adenocarcinoma. N Engl J Med.

1996;335:462–467.

3. Medical Research Council Oesophageal Cancer Working

Group. Surgical resection with or without preoperative

chemotherapy in oesophageal cancer: a randomised con-

trolled trial. Lancet. 2002;359:1727–1733.

4. Bosset JF, Gignoux M, Triboulet JP, et al. Chemoradiother-

apy followed by surgery compared with surgery alone in

squamous-cell cancer of the esophagus. N Engl J Med.

1997;337:161–167.

5. Urba SG, Orringer MB, Turrisi A, Iannettoni M, Forastiere

A, Strawderman M. Randomized trial of preoperative che-

moradiation versus surgery alone in patients with locore-

gional esophageal carcinoma. J Clin Oncol. 2001;19:305–

313.

6. Kelsen DP, Ginsberg R, Pajak TF, et al. Chemotherapy fol-

lowed by surgery compared with surgery alone for loca-

lized esophageal cancer. N Engl J Med. 1998;339:1979–1984.

7. Law S, Fok M, Chow S, Chu KM, Wong J. Preoperative

chemotherapy versus surgical therapy alone for squamous

cell carcinoma of the esophagus: a prospective randomized

trial. J Thorac Cardiovasc Surg. 1997;114:210–217.

8. Roth JA, Pass HI, Flanagan MM, Graeber GM, Rosenberg

JC, Steinberg S. Randomized clinical trial of preoperative

and postoperative adjuvant chemotherapy with cisplatin,

vindesine, and bleomycin for carcinoma of the esophagus.

J Thorac Cardiovasc Surg. 1988;96:242–248.

9. Forastiere AA, Orringer MB, Perez-Tamayo C, Urba SG,

Zahurak M. Preoperative chemoradiation followed by

transhiatal esophagectomy for carcinoma of the esopha-

gus: final report. J Clin Oncol. 1993;11:1118–1123.

Interval Esophageal Cancer Metastases/Bruzzi et al. 133

10. Donington JS, Miller DL, Allen MS, Deschamps C, Nichols

FC 3rd, Pairolero PC. Tumor response to induction chemora-

diation: influence on survival after esophagectomy. Eur J

Cardiothorac Surg. 2003;24:631–636; discussion 36–37.

11. Ancona E, Ruol A, Santi S, et al. Only pathologic complete

response to neoadjuvant chemotherapy improves signifi-

cantly the long term survival of patients with resectable

esophageal squamous cell carcinoma: final report of a ran-

domized, controlled trial of preoperative chemotherapy

versus surgery alone. Cancer. 2001;91:2165–2174.

12. Swisher SG, Maish M, Erasmus JJ, et al. Utility of PET, CT,

and EUS to identify pathologic responders in esophageal

cancer. Ann Thorac Surg. 2004;78:1152–1160; discussion

52–60.

13. Swisher SG, Erasmus J, Maish M, et al. 2-Fluoro-2-deoxy-

D-glucose positron emission tomography imaging is pre-

dictive of pathologic response and survival after preopera-

tive chemoradiation in patients with esophageal carcinoma.

Cancer. 2004;101:1776–1785.

14. Wieder HA, Brucher BL, Zimmermann F, et al. Time course

of tumor metabolic activity during chemoradiotherapy of

esophageal squamous cell carcinoma and response to

treatment. J Clin Oncol. 2004;22:900–908.

15. Flamen P, Van Cutsem E, Lerut A, et al. Positron emission

tomography for assessment of the response to induction

radiochemotherapy in locally advanced oesophageal can-

cer. Ann Oncol. 2002;13:361–368.

16. Weber WA, Ott K, Becker K, et al. Prediction of response to

preoperative chemotherapy in adenocarcinomas of the

esophagogastric junction by metabolic imaging. J Clin

Oncol. 2001;19:3058–3065.

17. Kato H, Kuwano H, Nakajima M, et al. Usefulness of posi-

tron emission tomography for assessing the response of

neoadjuvant chemoradiotherapy in patients with esopha-

geal cancer. Am J Surg. 2002;184:279–283.

18. Kroep JR, Van Groeningen CJ, Cuesta MA, et al. Positron

emission tomography using 2-deoxy-2-[18F]-fluoro-D-glu-

cose for response monitoring in locally advanced gastro-

esophageal cancer; a comparison of different analytical

methods. Mol Imaging Biol. 2003;5:337–346.

19. Brucher BL, Weber W, Bauer M, et al. Neoadjuvant therapy

of esophageal squamous cell carcinoma: response evalua-

tion by positron emission tomography. Ann Surg. 2001;233:

300–309.

20. Downey RJ, Akhurst T, Ilson D, et al. Whole body 18FDG-

PET and the response of esophageal cancer to induction

therapy: results of a prospective trial. J Clin Oncol. 2003;

21:428–432.

21. Cerfolio RJ, Bryant AS, Ohja B, Bartolucci AA, Eloubeidi

MA. The accuracy of endoscopic ultrasonography with

fine-needle aspiration, integrated positron emission to-

mography with computed tomography, and computed to-

mography in restaging patients with esophageal cancer

after neoadjuvant chemoradiotherapy. J Thorac Cardiovasc

Surg. 2005;129:1232–1241.

22. Wieder HA, Beer AJ, Lordick F, et al. Comparison of

changes in tumor metabolic activity and tumor size during

chemotherapy of adenocarcinomas of the esophagogastric

junction. J Nucl Med. 2005;46:2029–2034.

23. Brink I, Hentschel M, Bley TA, et al. Effects of neoadjuvant

radio-chemotherapy on 18F-FDG-PET in esophageal carci-

noma. Eur J Surg Oncol. 2004;30:544–550.

24. Erasmus J, Munden R, Truong MT, et al. Pre-operative che-

moradiation-induced ulceration in patients with esopha-

geal cancer: a confounding factor in tumor response assessment

in integrated CT-PET imaging. J Thorac Oncol. 2006;1:478–

486.

25. Swisher SG, Ajani JA, Komaki R, et al. Long-term outcome

of phase II trial evaluating chemotherapy, chemoradiother-

apy, and surgery for locoregionally advanced esophageal

cancer. Int J Radiat Oncol Biol Phys. 2003;57:120–127.

26. Bruzzi J, Swisher S, Truong M, Sabloff B, Munden R, Eras-

mus J. Pattern of distant organ metastases detected by

integrated PET/CT imaging. Curr Prob Diagn Radiol.

In press.

27. Westerterp M, van Westreenen HL, Reitsma JB, et al. Esopha-

geal cancer: CT, endoscopic US, and FDG PET for assessment

of response to neoadjuvant therapy—systematic review. Radi-

ology. 2005;236:841–851.

28. Pan T, Mawlawi O, Nehmeh SA, et al. Attenuation correc-

tion of PET images with respiration-averaged CT images in

PET/CT. J Nucl Med. 2005;46:1481–1487.

29. Nehmeh SA, Erdi YE, Pan T, et al. Quantitation of respira-

tory motion during 4D-PET/CT acquisition. Med Phys.

2004;31:1333–1338.

30. Esophagus. American Joint Committee on Cancer: AJCC

Staging Manual, 5th ed. Philadelphia: Lippincott-Raven; 1997:

65–69.

134 CANCER January 1, 2007 / Volume 109 / Number 1