Int. J. Radiation Oncology Biol. Phys., Vol. 69, No. 4, pp. 1167–1172, 2007Copyright � 2007 Elsevier Inc.

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

doi:10.1016/j.ijrobp.2007.04.047

CLINICAL INVESTIGATION Rectum

CLINICAL PARAMETERS PREDICTING PATHOLOGIC TUMOR RESPONSE AFTERPREOPERATIVE CHEMORADIOTHERAPY FOR RECTAL CANCER

SANG MIN YOON, M.D., DAE YONG KIM, M.D., TAE HYUN KIM, M.D., KYUNG HAE JUNG, M.D.,HEE JIN CHANG, M.D., WOONG SUB KOOM, M.D., SEOK-BYUNG LIM, M.D., HYO SEONG CHOI, M.D.,

SEUNG-YONG JEONG, M.D., AND JAE-GAHB PARK, M.D.

Center for Colorectal Cancer, Research Institute and Hospital, National Cancer Center, Goyang, Republic of Korea

Purpose: To identify pretreatment clinical parameters that could predict pathologic tumor response to preopera-tive chemoradiotherapy (CRT) for rectal cancer.Methods and Materials: The study involved 351 patients who underwent preoperative CRT followed by surgerybetween October 2001 and July 2006. Tumor responses to preoperative CRT were assessed in terms of tumordownstaging and tumor regression. Statistical analyses were performed to identify clinical factors associatedwith pathologic tumor response.Results: Tumor downstaging (defined as ypT2 or less) was observed in 167 patients (47.6%), whereas tumor regres-sion (defined as Dworak’s Regression Grades 3 or 4) was observed in 103 patients (29.3%) and complete regressionin 51 patients (14.5%). Multivariate analysis found that predictors of downstaging were pretreatment hemoglobinlevel (p = 0.045), cN0 classification (p < 0.001), and serum carcinoembryonic antigen (CEA) level (p < 0.001), thatpredictors of tumor regression were cN0 classification (p = 0.044) and CEA level (p < 0.001), and that the predictorof complete regression was CEA level (p = 0.004).Conclusions: The data suggest that pretreatment CEA level is the most important clinical predictor of pathologictumor response. It may be of benefit in the selection of treatment options as well as the assessment of individualprognosis. � 2007 Elsevier Inc.

Rectal cancer, Preoperative chemoradiotherapy, Downstaging, Tumor regression.

INTRODUCTION

Postoperative chemoradiotherapy (CRT) improves local

control and survival in locally advanced rectal cancer (1).

In the past, radical surgery followed by postoperative CRT

was considered standard treatment. However, preoperative

CRT for advanced rectal cancer has been increasingly used

in recent years and has now become a standard treatment,

and a recent prospective, randomized trial confirmed the su-

periority of preoperative over postoperative CRT in terms of

local control and toxicity (2).

In contrast to the postoperative setting, preoperative CRT

allows a relatively short-term evaluation because it offers

alternative endpoints based on pathologic tumor response.

The most commonly used endpoint is tumor downstaging.

Numerous studies have reported that tumor downstaging

after preoperative radiotherapy with or without chemother-

apy followed by surgical resection is associated with

11

decreased recurrence and improved survival (3–9). Another

endpoint for assessing pathologic tumor response is tumor

regression grade after preoperative CRT. A variety of defini-

tions and techniques for identifying and scoring the residual

cancer exist, and several studies have reported correlations

between tumor regression grade and long-term clinical out-

comes (8, 10–12). On the basis of these reports, the ability

to predict the pathologic tumor response before treatment

would be of clinical advantage in that it may provide addi-

tional information for permitting tailored treatment options

as well as for assessing the individual prognosis.

Using specific molecular markers or gene expression pro-

filing, many translational studies have sought to identify the

biologic properties of tumors that might predict therapeutic

response (13–19). Although those studies identified a number

of potentially important predictors, the results have not yet

had clinical impact owing to variable findings between stud-

ies and because the techniques used are complex and time

Reprint requests to: Dae Yong Kim, M.D., Center for ColorectalCancer, National Cancer Center, 809 Madu-1-dong, Ilsandong-gu,Goyang-si, Gyeonggi-do 410-769, Republic of Korea. Tel: (+82)31-920-1721; Fax: (+82) 31-920-0149; E-mail: radiopia@ncc.re.kr

Supported by the Korea Health 21 R&D Project, Ministry of

6

Health and Welfare, Republic of Korea (Grant No. 0412-CR01-0704-0001).

Conflict of interest: none.Received Feb 1, 2007, and in revised form April 24, 2007.

Accepted for publication April 26, 2007.

7

1168 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 4, 2007

consuming in terms of clinical application. Thus, despite nu-

merous studies, neither molecular marker nor gene expres-

sion profiling data have identified definitive predictors of

tumor response. In contrast, there has been no systematic

investigation of clinical parameters that may predict preoper-

ative CRT tumor response.

With this background, the present study sought to identify

pretreatment clinical parameters that may predict pathologic

tumor response to preoperative CRT.

METHODS AND MATERIALS

PatientsBetween October 2001 and July 2006, 388 patients with primary

rectal cancer underwent preoperative CRT at the National Cancer

Center, Republic of Korea. Medical records were reviewed to ana-

lyze clinical prognostic parameters and to identify the following in-

clusion criteria: (1) histologically confirmed rectal adenocarcinoma,

(2) tumor located within 8 cm of the anal verge, (3) locally advanced

(cT3-4 classification) and curatively resectable tumor evaluated with

magnetic resonance imaging (MRI) with or without transrectal ultra-

sonography, (4) adequate bone marrow, hepatic, and renal function,

and (5) no evidence of distant metastasis.

TreatmentPreoperative radiotherapy was delivered to the whole pelvis at

a dose of 45 Gy in 25 fractions, followed by a boost of 5.4 Gy in

3 fractions to the primary tumor within 6 weeks. All patients under-

went computed tomography (CT) simulation for three-dimensional

conformal radiotherapy planning, and a three-field treatment plan

was used involving a 6-MV photon posterior–anterior field and

15-MV photon opposed lateral beams. The beam weights of the

plan were optimized to minimize the maximal dose within the target

volume. The initial radiation field encompassed a volume that

included the gross tumor, mesorectum, presacral space, whole of

the sacral hollow, and regional lymphatics. The superior border

was placed at L5/S1 and the inferior border at 3 cm or more caudal

to the gross tumor. The boost field included the gross tumor volume

and mesorectum with $2-cm margins in all directions.

Preoperative chemotherapy was initiated on the first day of

pelvic radiotherapy and was delivered concurrently with radiother-

apy. Three different chemotherapy regimens were used. The 5-

fluorouracil and leucovorin group received two cycles of an i.v.

bolus injection of 5-fluorouracil (400 mg/m2/day) and leucovorin

(20 mg/m2/day) for 3 days in the first and fifth weeks of radiotherapy

(from October 2001 to July 2006). The capecitabine group received

825 mg/m2 capecitabine orally twice daily during radiotherapy with-

out weekend breaks (from April 2003 to July 2006). The capecita-

bine and irinotecan group received 825 mg/m2 capecitabine twice

daily during radiotherapy with weekend breaks and 40 mg/m2 irino-

tecan weekly for 5 weeks (from July 2004 to July 2006).

After the completion of preoperative CRT, all patients underwent

a radical proctectomy, including high ligation of the inferior mesen-

teric vessels and total mesorectal excision. Low anterior resection

was performed in 292 patients (83.2%) and abdominoperineal resec-

tion in 59 patients (16.8%). Lateral node dissection was not

performed routinely in the surgical procedure. The median interval

between CRT and surgery was 6 weeks (range, 4–8 weeks).

EvaluationBefore preoperative CRT, patients underwent preoperative diag-

nostic and staging workups, including digital rectal examination,

full blood counts and a biochemical profile, serum carcino-

embryonic antigen (CEA) level measurement, colonoscopy with

biopsy, chest radiography, abdominopelvic CT, pelvic MRI, and/

or transrectal ultrasonography. The MRI protocol has been described

in our previous report (20). The cross-sectional areas of the primary

tumors were measured using axial T2-weighted images by tracing

the lesion boundary. On T2-weighted images, the cross-sectional

lesion areas were defined as intermediate signal intensity areas

that had a different signal intensity and contour from the normal

rectal wall. The lesion volumes were displayed automatically in

a three-dimensional format and were calculated by summing each

of the cross-sectional volumes of the entire lesion. Positive lymph

node involvement was defined as a lymph node $5 mm in the small-

est diameter observed on CT or MRI (21).

After radical surgery, tumor specimens were reviewed and the

post-CRT pathologic stage (yp) was determined according to the

TNM classification system recommended by the American Joint

Committee on Cancer (22). Tumor downstaging was determined

by comparing pretreatment clinical and postoperative pathologic

T classifications, and downstaging was defined as ypT2 or lower.

Tumors were also assessed using Dworak’s tumor regression grading

system for semiquantitative evaluation of histopathologic tumor

regression (23). Regression was graded as follows: Grade 0, no re-

gression; Grade 1, dominant tumor mass with obvious fibrosis and/

or vasculopathy (minimal regression); Grade 2, dominant fibrotic

changes with few tumor cells or groups (moderate regression); Grade

3, very few tumor cells in fibrotic tissue with or without mucous sub-

stance; Grade 4, no tumor cells, only fibrotic mass or acellular mucin

pools (complete regression). The regression grade concerned both

primary tumor and regional lymph nodes. The study defined Grades

3 and 4 as tumor regression for the purpose of statistical analysis.

StatisticsThe study was designed to identify clinical variables that could

predict pathologic stage and tumor regression. Logistic regression

analysis was used to identify correlations between pathologic re-

sponses and continuous variables. Chi-square or Fisher’s exact tests

were used to determine the significance of associations between

pathologic findings and categoric variables. The multivariate analy-

sis used to identify correlations between pathologic findings and all

potential parameters involved a backward likelihood ratio using a

logistic regression model. A receiver operating characteristic curve

was used to define the cutoff point for the various continuous

variables that were relative to the predicting tumor response. Prob-

ability (p) values of <0.05 were considered to indicate a significant

difference.

RESULTS

Patient characteristicsOf the 388 patients, 20 refused surgery, 7 were treated with

transanal excision because of comorbidities or strong refusal

of anal ablation, and 10 were transferred to other hospitals

closer to their residence. Therefore, 351 patients who met

the inclusion criteria were analyzed in this study. Table 1

shows detailed patient characteristics. The study population

was mostly male (66.7%) and had a median age of 57

years (range, 31–83 years). Almost all patients had a cT3

Clinical parameters predicting response d S. M. YOON et al. 1169

classification of their primary tumor (96.9%), and the major

type was adenocarcinoma (96.9%). Elevated serum CEA

levels were observed in 33.6% of patients at diagnosis (the

upper limit of normal was defined as 5 ng/mL).

Pathologic tumor response after preoperative CRTPathologic examination of resected specimens revealed

ypT0 in 54 patients (15.4%) (including ypT0N1 in 3 pa-

tients), ypTis in 3 patients (0.9%), ypT1 in 20 patients

(5.7%), ypT2 in 90 patients (25.6%), ypT3 in 168 patients

(47.9%), and ypT4 in 16 patients (4.5%). Downstaging to

ypT2 or less was observed in 167 patients (47.6%). Dworak’s

tumor regression grades after preoperative CRT were Grade 1

in 57 patients (16.3%), Grade 2 in 191 patients (54.4%),

Grade 3 in 52 patients (14.8%), and Grade 4 in 51 patients

(14.5%). Thus, 103 patients (29.3%) showed tumor regres-

sion (i.e., Grade 3 or 4) as defined in the present study.

Clinical parameters predicting pathologic tumor responseUnivariate predictors of downstaging (ypT0-2) were

found to be a pretreatment hemoglobin level $12.5 g/dL

Table 1. Patient characteristics

GenderMale 234 (66.7)Female 117 (33.3)

Age (y)Median 57Range 31–83

Pretreatment hemoglobin level#12.5 g/dL 142 (40.5)>12.5 g/dL 209 (59.5)

Distance from anal verge (cm)Median 5Range 0–8

Clinical T classificationcT3 340 (96.9)cT4 11 (3.1)

Clinical N classificationcN0 55 (15.7)cN+ 296 (84.3)

Histologic typeAdenocarcinoma 340 (96.9)Mucinous 9 (2.6)Signet ring cell 2 (0.5)

Histologic gradeLow grade 330 (94.0)High grade 21 (6.0)

Pretreatment tumor volume (cm3)Mean � SD 19.1 � 21.5Median 13.8Range 0.8–235.8

Pretreatment CEA level#5 ng/mL 233 (66.4)>5 ng/mL 118 (33.6)

Chemotherapy regimenFL group 146 (41.6)Capecitabine group 134 (38.2)IX group 71 (20.2)

Abbreviations: SD = standard deviation; CEA = carcinoem-bryonic antigen; FL = 5-fluorouracil and leucovorin; IX = irinotecanand capecitabine.

Values are number (percentage) unless otherwise noted.

(p = 0.016), cT3 classification (p = 0.047), cN0 classification

(p < 0.001), pretreatment tumor volume #13 cm3 (p <

0.001), and CEA levels #5 ng/mL (p < 0.001). Univariate

predictors of tumor regression were found to be a cN0 class-

ification (p = 0.011), pretreatment tumor volume #13 cm3

(p = 0.018), and CEA levels #5 ng/mL (p < 0.001). Univar-

iate predictors of complete regression were found to be

a cN0 classification (p = 0.037) and CEA levels #5 ng/mL

(p < 0.001) (Table 2).

Multivariate analysis revealed that pretreatment hemoglo-

bin levels $12.5 g/dL (p = 0.045), cN0 classification (p <

0.001), and CEA levels #5 ng/mL (p < 0.001) were predic-

tors of downstaging, that cN0 classification (p = 0.044) and

CEA levels #5 ng/mL (p < 0.001) were predictors of tumor

regression, and that CEA levels #5 ng/mL (p = 0.004) was

a predictor of complete regression (Table 3).

DISCUSSION

A wide spectrum of tumor responses has been reported

after preoperative CRT for locally advanced rectal cancer, and

the clinical meaning of such responses in terms of prognosis

has been the subject of many investigations. Tumor response

prediction before surgery may be of benefit to effective man-

agement. The prognostic factors for predicting long-term

clinical outcomes have mostly been based on histopathologic

parameters after curative resection in preoperative setting

studies (3–12, 24). The potential advantage of using these

pathologic parameters includes their availability in the rela-

tively short-term and that they are objectively measurable

(25). However, a significant proportion of patients are un-

likely to respond to standard preoperative CRT. Consider-

ation of more intensive and individualized CRT regimens

may be indicated for these patients. Conversely, less aggres-

sive surgical procedures may be an alternative to sphincter-

ablating surgery in highly selected patients with distal rectal

cancers who had marked regression after preoperative CRT.

Thus, there is an obvious need for new markers to predict path-

ologic response before preoperative CRT with a perspective of

tailored treatment. Histologic tumor response is associated

with several factors, including patient-, tumor-, and treat-

ment-related factors. In the present study, treatment-related

factors, such as radiation dose schedules and concomitant

administration of chemotherapy, were relatively homogeneous

for all patients, and, thus, any differences in downstaging and

tumor regression were likely to be due to factors relating to

individual patients and particular tumor characteristics.

In the present study, multivariate analysis found that

pretreatment CEA level was the common predictor of down-

staging, tumor regression, and complete regression. Carcino-

embryonic antigen levels have long been known to provide

potential benefit for predicting outcomes. Furthermore, the

preoperative CEA level is considered a Category I prognostic

indicator in the College of American Pathologists Consensus

Statement 1999 (26). However, few studies have validated

the predictive value of the CEA level as a marker for tumor

response after preoperative CRT. In a recent study, Park

1170 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 4, 2007

Table 2. Univariate analysis to identify predictors of downstaging, tumor regression, and complete regresion

Downstaging Tumor regression grade (1–2 vs. 3–4) Tumor regression grade (1–3 vs. 4)

Variable No (n = 184) Yes (n = 167) 1–2 (n = 248) 3–4 (n = 103) 1–3 (n = 300) 4 (n = 51)

GenderMale 124 (53.0) 110 (47.0) 167 (71.4) 67 (28.6) 199 (85.0) 35 (15.0)Female 60 (51.3) 57 (48.7) 81 (69.2) 36 (30.8) 101 (86.3) 16 (13.7)p* 0.762 0.679 0.748

Age (y)#57 94 (52.5) 85 (47.5) 122 (68.2) 57 (31.8) 147 (82.1) 32 (17.9)>57 90 (52.3) 82 (47.7) 126 (73.3) 46 (26.7) 153 (89.0) 19 (11.0)p 0.972 0.294 0.069

PretreatmentHb level (g/dL)

#12.5 85 (60.3) 56 (39.7) 105 (74.5) 36 (25.5) 122 (86.5) 19 (13.5)>12.5 99 (47.1) 111 (52.9) 143 (68.1) 67 (31.9) 178 (84.8) 32 (15.2)p 0.016 0.199 0.646

Distance fromanal verge (cm)

<5 74 (50.3) 73 (49.7) 105 (71.4) 42 (28.6) 125 (85.0) 22 (15.0)$5 110 (53.9) 94 (46.1) 143 (70.1) 61 (29.9) 175 (85.8) 29 (14.2)p 0.507 0.787 0.844

T classificationcT3 175 (51.5) 165 (48.5) 239 (70.3) 101 (29.7) 291 (85.6) 49 (14.4)cT4 9 (81.8) 2 (18.2) 9 (81.8) 2 (18.2) 9 (81.8) 2 (18.2)p 0.047 0.519 0.665

N classificationcN (�) 15 (27.3) 40 (72.7) 31 (56.4) 24 (43.6) 42 (76.4) 13 (23.6)cN (+) 169 (57.1) 127 (42.9) 217 (73.3) 79 (26.7) 258 (87.2) 38 (12.8)p <0.001 0.011 0.037

Histologic gradeLow grade 171 (51.8) 159 (48.2) 232 (70.3) 98 (29.7) 282 (85.5) 48 (14.5)High grade 13 (61.9) 8 (38.1) 16 (76.2) 5 (23.8) 18 (85.7) 3 (14.3)p 0.369 0.566 0.974

Pretreatment tumorvolume (cm3)

#13 67 (41.9) 93 (58.1) 103 (64.4) 57 (35.6) 133 (83.1) 27 (16.9)>13 117 (61.3) 74 (38.7) 145 (75.9) 46 (24.1) 167 (87.4) 24 (12.6)p <0.001 0.018 0.254

Pretreatment CEAlevel (ng/mL)

#5 98 (42.1) 135 (57.9) 148 (63.5) 85 (36.5) 189 (81.1) 44 (18.9)>5 86 (72.9) 32 (27.1) 100 (84.7) 18 (15.3) 111 (94.1) 7 (5.9)p <0.001 <0.001 <0.001

CRT–Op interval (wk)#6 109 (54.5) 91 (45.5) 147 (73.5) 53 (26.5) 172 (86.0) 28 (14.0)>6 75 (49.7) 76 (50.3) 101 (66.9) 50 (33.1) 128 (84.8) 23 (15.2)p 0.370 0.178 0.746

ChemotherapyFL 82 (56.2) 64 (43.8) 110 (75.3) 36 (24.7) 131 (89.7) 15 (10.3)Capecitabine 68 (50.7) 66 (49.3) 94 (70.1) 40 (29.9) 112 (83.6) 22 (16.4)IX 34 (47.9) 37 (52.1) 44 (62.0) 27 (38.0) 57 (80.3) 14 (19.7)p 0.459 0.126 0.132

Abbreviations: Hb = hemoglobin; CEA = carcinoembryonic antigen; CRT = chemoradiotherapy; Op = operation; FL = 5-fluorouracil andleucovorin; IX = irinotecan and capecitabine.

Values are number (percentage).* Determined by chi-square or Fisher exact test.

et al. (27) reported that an elevated pre-CRT serum CEA

level predicted a poor tumor response to preoperative CRT

in rectal cancer patients, independent of other clinicopatho-

logic features in their analysis. Those data are consistent

with the present study results. Another recent study, by Das

et al. (28), demonstrated that circumferential extent of tumor,

CEA level, and distance from the anal verge predicted for

the pathologic response to preoperative CRT for patients

with rectal cancer. These findings indicate that further studies

are warranted to confirm the apparent association between

serum CEA level and tumor response after preoperative

CRT.

Table 3. Multivariate analysis to identify predictors of downstaging, tumor regression, and complete regression

Downstaging Tumor regression grade (1–2 vs. 3–4) Tumor regression grade (1–3 vs. 4)

Odds ratio 95% CI p* Odds ratio 95% CI p Odds ratio 95% CI p

Pretreatment Hb level 1.595 1.002-2.532 0.045 — — NS — — NSN classification 3.303 1.691-6.452 <0.001 1.863 1.016-3.415 0.044 — — NSPretreatment CEA level 3.444 2.097-5.654 <0.001 3.023 1.705-5.358 <0.001 3.376 1.460-7.808 0.004

Abbreviations: CI = confidence interval; Hb = hemoglobin; CEA = carcinoembryonic antigen.* Determined by logistic regression analysis.

Clinical parameters predicting response d S. M. YOON et al. 1171

Interestingly, the present study found that pretreatment

hemoglobin level was a predictor for tumor downstaging.

Anemia is present in many cancer patients at the time of

diagnosis and has been hypothesized to lead to tumor hypoxia,

angiogenesis, and resistance to chemotherapy and radiother-

apy. Many cervical cancer and head-and-neck cancer studies

have shown a correlation between low hemoglobin levels

and poorer prognosis, despite the use of different definitions

of anemia (varying from 10 to 14.5 g/dL) (29). However,

few studies have evaluated the prognostic value of the hemo-

globin level in rectal cancer, and none has investigated its

potential as a predictor of tumor response. One recent study

revealed that the hazard ratio of death was 0.35 (95% confi-

dence interval 0.19–0.65, p = 0.001) in patients without preop-

erative anemia, which indicates a threefold higher mortality

risk for anemic patients with rectal cancer treated with total

mesorectal excision (30). Although no studies report investi-

gating whether pretreatment hemoglobin levels can predict

preoperative CRT response, the current findings indicate that

a prospective trial involving rectal cancer patients is warranted.

The present study found that pretreatment nodal classifica-

tion was a predictor of pathologic downstaging and tumor

regression. Pathologic tumor or nodal classification were the

most important prognostic factors for rectal cancer in either

neoadjuvant or adjuvant settings. However, questions sur-

round pre-CRT clinical staging and its prognostic significance,

owing to the low accuracy of exact staging, especially in lymph

node evaluation (31). Previous studies found that pretreatment

nodal status did not influence levels of tumor regression by the

pretreatment characteristics analysis in the same manner (11,

24, 27). Therefore, the present correlation between cN classifi-

cations and downstaging or tumor regression should be inter-

preted cautiously until more accurate methods of lymph

node evaluation are developed.

The present study also identified a range of clinical factors

that were not associated with histologic response, and such

a lack of association with these factors has been reported in

other tumor response studies (5–7, 11, 27). Pretreatment tu-

mor volume was associated with the degree of both down-

staging and tumor regression in univariate analysis, but this

was not an independent prognostic factor in further multivar-

iate analysis. Theoretically, the response to radiotherapy is

related to the number of cells to be killed according to the ra-

diobiologic principle, and the correlation between tumor size

and treatment response was revealed in other reports (24, 32).

However, this result was not confirmed in our study. This

finding might be related to an individual tumor gene profile

being more important than the number of tumor cells in a re-

sponse to CRT, and hopefully future studies will clarify this

genetic characteristic. The duration of the interval between

CRT and surgery did not have an influence on pathologic tu-

mor response. Because the interval between CRT and surgery

ranged from 4 to 8 weeks in the present study, we could

assume that the difference of pathologic tumor response is

not significant within these intervals, although it may have

an influence on downstaging for rectal cancer (33).

This study had some limitations. First, the analysis end-

points were limited to the pathologic response and not

extended to long-term clinical outcomes. It would be very

useful to analyze the present population in terms of early re-

sponse and survival data, such as disease-free and overall sur-

vival. Second, overstaging a tumor remains a possibility in

staging analysis. Thus, to enhance the accuracy, only the T

classification was used in downstaging analysis in the present

study. Transrectal ultrasonography and MRI are more reli-

able methods for determining T classification than evaluation

of lymph node status. However, staging failures still occur,

owing to difficulties in accurate discrimination between T2

and T3 lesions, which are caused by perirectal fat desmoplas-

tic reactions on MRI (34). These inaccuracies may challenge

the validity of using downstaging parameters to predict treat-

ment response. Third, there was a discrepancy between the

factors predicting downstaging and those predicting tumor

regressions. Although CRT-induced tumor shrinkage may

lead to complete or partial pathologic regression, in many

cases the pathologic stage does not change even if the tumor

cell density decreases. The differing criteria for defining

tumor response, and the different number of patients in

each group, may contribute to such a discrepancy. However,

it should be noted that each endpoint had a clinical meaning

in previous reports, and the results should be interpreted

separately concerning clinical outcomes.

In summary, the present data suggest that pretreatment

CEA level is the most important clinical predictor of patho-

logic tumor response. It may be of benefit for selecting treat-

ment options and for assessing individual prognosis. These

findings using retrospective data indicate the need for a larger,

prospective clinical trial. Furthermore, a comprehensive

comparison between these surrogate endpoints and long-

term survival data should be undertaken to determine the

prognostic significance of the present endpoints in a random-

ized preoperative setting.

1172 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 4, 2007

REFERENCES

1. NIH consensus conference. Adjuvant therapy for patients withcolon and rectal cancer. JAMA 1990;264:1444–1450.

2. Sauer R, Becker H, Hohenberger W, et al. Preoperative versuspostoperative chemoradiotherapy for rectal cancer. N Engl JMed 2004;351:1731–1740.

3. Chan AK, Wong A, Jenken D, et al. Posttreatment TNM stagingis a prognostic indicator of survival and recurrence in tethered orfixed rectal carcinoma after preoperative chemotherapy andradiotherapy. Int J Radiat Oncol Biol Phys 2005;61:665–677.

4. Das P, Skibber JM, Rodriguez-Bigas MA, et al. Clinical andpathologic predictors of locoregional recurrence, distant metas-tasis, and overall survival in patients treated with chemoradia-tion and mesorectal excision for rectal cancer. Am J ClinOncol 2006;29:219–224.

5. Garcia-Aguilar J, Hernandez de Anda E, Sirivongs P, et al. Apathologic complete response to preoperative chemoradiationis associated with lower local recurrence and improved survivalin rectal cancer patients treated by mesorectal excision. DisColon Rectum 2003;46:298–304.

6. Kaminsky-Forrett MC, Conroy T, Luporsi E, et al. Prognosticimplications of downstaging following preoperative radiationtherapy for operable T3-T4 rectal cancer. Int J Radiat OncolBiol Phys 1998;42:935–941.

7. Theodoropoulos G, Wise WE, Padmanabhan A, et al. T-leveldownstaging and complete pathologic response after preopera-tive chemoradiation for advanced rectal cancer result indecreased recurrence and improved disease-free survival. DisColon Rectum 2002;45:895–903.

8. Losi L, Luppi G, Gavioli M, et al. Prognostic value of Dworakgrade of regression (GR) in patients with rectal carcinomatreated with preoperative radiochemotherapy. Int J ColorectalDis 2006;21:645–651.

9. Valentini V, Coco C, Picciocchi A, et al. Does downstaging pre-dict improved outcome after preoperative chemoradiation forextraperitoneal locally advanced rectal cancer? A long-termanalysis of 165 patients. Int J Radiat Oncol Biol Phys 2002;53:664–674.

10. Bouzourene H, Bosman FT, Seelentag W, et al. Importance oftumor regression assessment in predicting the outcome inpatients with locally advanced rectal carcinoma who are treatedwith preoperative radiotherapy. Cancer 2002;94:1121–1130.

11. Rodel C, Martus P, Papadoupolos T, et al. Prognostic signifi-cance of tumor regression after preoperative chemoradiotherapyfor rectal cancer. J Clin Oncol 2005;23:8688–8696.

12. Vecchio FM, Valentini V, Minsky BD, et al. The relationship ofpathologic tumor regression grade (TRG) and outcomes afterpreoperative therapy in rectal cancer. Int J Radiat Oncol BiolPhys 2005;62:752–760.

13. Chang HJ, Jung KH, Kim DY, et al. Bax, a predictive markerfor therapeutic response to preoperative chemoradiotherapy inpatients with rectal carcinoma. Hum Pathol 2005;36:364–371.

14. Giralt J, Eraso A, Armengol M, et al. Epidermal growth factorreceptor is a predictor of tumor response in locally advancedrectal cancer patients treated with preoperative radiotherapy.Int J Radiat Oncol Biol Phys 2002;54:1460–1465.

15. Rau B, Sturm I, Lage H, et al. Dynamic expression profile ofp21WAF1/CIP1 and Ki-67 predicts survival in rectal carcinomatreated with preoperative radiochemotherapy. J Clin Oncol2003;21:3391–3401.

16. Smith FM, Reynolds JV, Kay EW, et al. COX-2 overexpressionin pretreatment biopsies predicts response of rectal cancers to

neoadjuvant radiochemotherapy. Int J Radiat Oncol Biol Phys2006;64:466–472.

17. Zlobec I, Steele R, Compton CC. VEGF as a predictive markerof rectal tumor response to preoperative radiotherapy. Cancer2005;104:2517–2521.

18. Watanabe T, Komuro Y, Kiyomatsu T, et al. Prediction ofsensitivity of rectal cancer cells in response to preoperativeradiotherapy by DNA microarray analysis of gene expressionprofiles. Cancer Res 2006;66:3370–3374.

19. Ghadimi BM, Grade M, Difilippantonio MJ, et al. Effectivenessof gene expression profiling for response prediction of rectaladenocarcinomas to preoperative chemoradiotherapy. J ClinOncol 2005;23:1826–1838.

20. Kim YH, Kim DY, Kim TH, et al. Usefulness of magnetic res-onance volumetric evaluation in predicting response to preoper-ative concurrent chemoradiotherapy in patients with resectablerectal cancer. Int J Radiat Oncol Biol Phys 2005;62:761–768.

21. Kim JH, Beets GL, Kim MJ, et al. High-resolution MR imagingfor nodal staging in rectal cancer: Are there any criteria in addi-tion to the size? Eur J Radiol 2004;52:78–83.

22. Greene FL, Page DL, Fleming ID, et al., editors. American JointCommittee on Cancer staging manual. 6th ed. Philadelphia:Springer; 2002.

23. Dworak O, Keilholz L, Hoffmann A. Pathological features ofrectal cancer after preoperative radiochemotherapy. Int J Colo-rectal Dis 1997;12:19–23.

24. Janjan NA, Crane C, Feig BW, et al. Improved overall survivalamong responders to preoperative chemoradiation for locallyadvanced rectal cancer. Am J Clin Oncol 2001;24:107–112.

25. Glynne-Jones R, Mawdsley S, Pearce T, et al. Alternativeclinical end points in rectal cancer—Are we getting closer?Ann Oncol 2006;17:1239–1248.

26. Compton CC, Fielding LP, Burgart LJ, et al. Prognostic factorsin colorectal cancer. College of American Pathologists Consen-sus Statement 1999. Arch Pathol Lab Med 2000;124:979–994.

27. Park YA, Sohn SK, Seong J, et al. Serum CEA as a predictor forthe response to preoperative chemoradiation in rectal cancer.J Surg Oncol 2006;93:145–150.

28. Das P, Skibber JM, Rodriguez-Bigas MA, et al. Predictors oftumor response and downstaging in patients who receive preop-erative chemoradiation for rectal cancer. Cancer 2007;109:1750–1755.

29. Varlotto J, Stevenson MA. Anemia, tumor hypoxemia, and thecancer patient. Int J Radiat Oncol Biol Phys 2005;63:25–36.

30. van Halteren HK, Houterman S, Verheij CD, et al. Anaemiaprior to operation is related with poorer long-term survival inpatients with operable rectal cancer. Eur J Surg Oncol 2004;30:628–632.

31. Nastro P, Beral D, Hartley J, et al. Local excision of rectalcancer: Review of literature. Dig Surg 2005;22:6–15.

32. Willett CG, Warland G, Coen J, et al. Rectal cancer: Theinfluence of tumor proliferation on response to preoperativeirradiation. Int J Radiat Oncol Biol Phys 1995;32:57–61.

33. Francis Y, Nemoz CJ, Baulieux J, et al. Influence of the intervalbetween preoperative radiation therapy and surgery on down-staging and on the rate of sphincter-sparing surgery for rectalcancer: The Lyon R90-01 randomized trial. J Clin Oncol1999;17:2396–2402.

34. Beets-Tan RG, Beets GL. Rectal cancer: Review with emphasison MR imaging. Radiology 2004;232:335–346.